Document RJmR9DZMpm3EvNvkN8rrXakBa

Environmental issue report No 22 Late lessons from early warnings: the precautionary principle 1896-2000 Editorial team: Poul Harremoes (Chairman) David Gee (EEA editor) Malcolm MacGarvin (Executive editor) Andy Stirling (Editor) Jane Keys (Editor) Brian Wynne (Editor) Sofia Guedes Vaz (EEA editor) Project managers: David Gee and Sofia Guedes Vaz European Environment Agency European Environment Agency Layout: Brandenborg a/s Note The contents of this publication do not necessarily reflect the official opinions of the European Commission or other European Communities institutions. Neither the European Environment Agency nor any person or company acting on the behalf of the Agency is responsible for the use that may be made of the information contained in this report. All rights reserved No part of this publication may be reproduced in any form or by any means electronic or mechanical, including photocopying, recording or by any information storage retrieval system without the permission in writing from the copyright holder. For rights of translation or reproduction please contact EEA project manager Ove Caspersen (address information below). A great deal of additional information on the European Union is available on the Internet. It can be accessed through the Europa server (http://europa.eu.int). Cataloguing data can be found at the end of this publication. Luxembourg: Office for Official Publications of the European Communities, 2001 ISBN 92-9167-323-4 EEA, Copenhagen, 2001 European Environment Agency Kongens Nytorv 6 DK-1050 Copenhagen K Denmark Tel: (45) 33 36 71 00 Fax: (45) 33 36 71 99 E-mail: eea@eea.eu.int Internet: http://www.eea.eu.int Preface To know and not to know. To act or not to act...? The task o f the European Environment Agency (EEA) is to provide inform ation o f direct use for improving decision-making and public participation. We often provide information in situations o f scientific uncertainty, in which the precautionary principle, enshrined in the Treaty o f the E u ro p ean U nion, is increasingly relevant. The growing innovative powers o f science seem to be outstripping its ability to predict the consequences o f its applications, whilst the scale o f hum an interventions in nature increases the chances that any hazardous im pacts may be serious an d global. It is therefore im portant to take stock o f past experiences, and learn how we can adapt to these changing circumstances and improve our work, particularly in relation to the provision o f information and the identification o f early warnings. Late lessons from early warnings is about the gathering of information on the hazards of hum an econom ic activities and its use in taking action to better protect both the environment and the health o f the species an d ecosystems that are d ep en d en t on it, an d then living with the consequences. The report is based on case studies. The authors o f the case studies, all experts in their particular field o f environmental, occupational and consum er hazards, were asked to identify the dates o f early warnings, to analyse how this inform ation was used, or not used, in reducing hazards, and to describe the resulting costs, benefits and lessons for the future. The lessons they drew from their histories were then distilled into twelve `late lesson s' by the editorial team, under the guidance of the EEA Scientific Committee. In a separate EEA publication som e im plications o f the late lessons for the policy process and associated inform ation flows will be further explored. The precautionary principle is n o tju st an issue for the E u ro p ean U n ion (E U ): its potential im pact on trade m eans that its application can have global repercussions. The current dialogue between the EU and the United States on the use and application o f precaution is partly affected by confusion about the m eaning of terms used in the debate. This report should contribute to a greater and shared understanding about past decisions on hazardous technologies and therefore, we hope, to im proved transatlantic agreem ent about future decisions. It may also help the dialogue within both the EU an d the United States, where there are healthy debates about the pros and cons of applying the precautionary principle. That we have all acted too late in many areas is now well known. Over the n ext 50 years we will see som e thousands o f extra skin cancers as today's children grow up exposed to the higher levels o f ultraviolet radiation penetrating the normally protective ozone layer through the `h ole' created by chlorofluorocarbons (CFCs) and other synthetic chemicals. Over the same period many thousands o f E u ro pean s will die from one of the m ost painful and terminal of cancers, mesotheliom a, caused by the inhalation o f asbestos dust. In both cases we were taken by surprise: the hazards o f these beneficial technologies were not `known about' until it was too late to stop irreversible impacts. Both phenom ena had such long latent periods between first exposures and late effects that `pipelines' o f unstoppable consequences, decades long, were set in place before actions could have been taken to stop further exposures. The first reports of injuries from radiation were m ade as early as 1896 (hence the title o f the report). The first clear and credible early warning abo u t asbestos cam e two years later in 1898. A similar signal for action on CFCs came in 1974, though some may argue that im portant clues were missed earlier. Eleven other well-known hazards are dealt with in this report. We invite the reader to ju d g e whether, as in the cases of radiation, asbestos and CFCs, the early warnings could have led to earlier actions to reduce hazards, at a lower overall cost to society. The costs of preventive actions are usually tangible, clearly allocated and often short term, whereas the costs o f failing to act are less tangible, less clearly distributed and usually longer term, posing particular problem s o f governance. Weighing up the overall pros an d cons o f action, or inaction, is therefore very difficult, involving ethical as well as econom ic considerations, as the case studies illustrate. A key question arising from the case studies is how to acknowledge and respond not only to scientific uncertainty but also to ignorance, a state o f not knowing from which springs both scientific discoveries and unpleasant `surprises', such as ozone holes and rare cancers. Socrates had a response to this when he acknowledged ignorance as a source of wisdom. O ur report shows that this is a lesson from history that many people have forgotten. Misplaced `certainty' about the absence o f harm played a key role in delaying preventive actions in m ost of the case studies. However, there is clearly nothing scientific about the pretence o f knowledge. Such `certainty' does little to reduce ignorance, which requires m ore scientific research and long-term m onitoring in order to identify the unintended impacts o f hum an activities. Could we have known about, or anticipated, the hazards any earlier? Are there ways o f `knowing m ore' or `knowing better' that could help justify our self-awarded title o f Homo sapiens -- the `wise on e s'? R eaders o f the case studies in Late lessons may conclude that we have a long way to go. Som e possible directions are indicated in the chapter `Twelve late lesson s', derived from the case studies. A phenom enon that Socrates probably did n ot know about, but may have suspected, is that `everything connects' -- or at least, so m any things d o react with each other that the sim ple science o f linear, mechanistic proposition s n eeds to be su pplem en ted with the dynamic and em ergent properties of systems science. The potential systemic instabilities of such com plex phenom ena as climate change or brain cell behaviour may be critical yet unpredictable determinants of our fate, whether they be systems that govern the stability o f the G ulf Stream or that generate the `genom ic instabilities' of irradiated cells. Com partmentalised science, no matter how erudite, is an insufficient base for knowing enough to anticipate or mitigate the impacts o f such com plex systems: integrated and synthesised knowledge, which pools the wisdom from many natural and social sciences, is a necessary condition for being Homo sapiens. B u tju st knowing enough is n ot o f itself sufficient: acting wisely, an d in good time, is also necessary. It is p art o f the EEA's task to help expand the knowledge base through integrated assessments, thereby assisting decision-makers to foresee the possible consequences of regulatory and stakeholder actions and inactions. Knowing enough, and acting wisely enough, across the full range of environmental and related health issues seems daunting. The interconnections between issues, the pace of technological change, our limited understanding and the `time to harm and then to h eal' o f the ecological and biological systems that can be perturbed over decades by our technologies together present an unforgiving context. Some people fear or imagine that a more precautionary approach to forestalling potentially irreversible hazards will stifle innovation or com prom ise science. However, there are im m ense challenges and opportunities in understanding com plex and em ergent systems while m eeting hum an needs with lower health and ecological costs. Many o f the case studies suggest that wider use o f the precautionary principle can help stimulate both innovation and science, replacing the 19th century technologies and simple science of the first industrial revolution with the `eco-efficient' technologies and systems science o f the third. One final and obvious question arises from the case studies in Late lessons: why were n ot only the early warnings but also the `loud and late' warnings often ignored for so long? This question we largely leave to the reader, whilst noting that the absence o f political will to take action to reduce hazards, in the face o f conflicting costs and benefits, seems to be an even m ore im portant factor in these histories than is the availability o f trusted information. However, as Aristotle observed, the way we perceive the world determines in large part how we act, and information plays a critical role in how we see the world. But whose inform ation is received? Is it `true, fair and in d ep en d en t'? And is it un derstandable to the politicians and business people who are rarely experts but nevertheless have to make the difficult decisions? This report notes the im portance of trusted and shared information for effective policy making and stakeholder participation in decision-making, especially in the context o f complexity, ignorance, high stakes and the n eed for `collective learn in g'. We m ust not forget that EU product legislation defines as safe any product that does not present `unacceptable risks' under norm al or foreseeable conditions o f use. Public acceptability o f risks requires public participation in the decisions that create and m anage such risks, including the consideration o f values, attitudes and overall benefits. Sound public policy-making on issues involving science therefore requires m ore than good science: ethical as well as economic choices are at stake. Such matters concern not only the experts and the politicians but all o f us. It is therefore my hope that this report contributes to better and m ore accessible science based information and m ore effective stakeholder participation in the governance o f econom ic activity so as to help minimise environmental and health costs and maximise innovation. Decision-makers need not only to have more and better quality inform ation but also to act wisely m ore often so as to achieve a better balance between the benefits o f innovations and their hazards. Learning and applying these late lessons from the last century's early warnings could help all o f us to achieve this better balance during this century. I would like to thank the editorial team and the authors who took up the challenge of m aking this report, as well as the peer reviewers and the EEA Scientific Committee who also played an im portant role. Finally I would like to thank David Gee, who initiated this report and many EEA staff whose contributions m ade this report possible. Domingo Jim nez Beltrn Executive Director Acknowledgements A cknow ledgem ent is given to the editorial team for their invaluable help and assistance in producing this book: Poul H arrem oes (Chairm an), Malcolm MacGarvin (Executive editor), Andy Stirling (Editor), Brian Wynne (Editor), Jan e Keys (Editor), David Gee and Sofia G uedes Vaz (EEA editors). Acknowledgem ent is also given to the authors and the peer reviewers: Authors: Malcolm MacGarvin, Barrie Lam bert, Peter Infante, Morris Greenberg, David Gee, Ja n n a G. Koppe, Jan e Keys, Jo e Farm an, Dolores Ibarreta, Shanna H. Swan, Lars-Erik Edqvist, Knud B 0rge Pedersen, Arne Semb, Martin Krayer von Krauss, Poul Harrem oes, Michael Gilbertson, David Santillo, Paul Johnston, William J. Langston, Jim W. Bridges, O lga Bridges, Patrick van Zwanenberg and Erik Millstone. Peer reviewers: Jo e l Tickner, Richard Young, W olfgang Witte, Stuart Levy, H en k Aarts, Jacques Nouws, Michael Skou Andersen, Anthony Seaton, H arry Aiking, L eon ard Levy, Ron Hoogenboom , Glen Fox, Andrew Watterson, Arturo Keller, Finn Bro Rasmussen, Sandrine Dixson-Decleve, Hilkka Vahervuori, Grete 0stergaard, Erik Arvin, Eberhard M orgenroth, Guus Velders, M adhava Sarna, Tom Webster, Niklas Johansson, Ernst Boersm a , Nigel Bell, Beldrich Moldan, Jo h n Seager, Kees van Leeuwen, Mark Montforts and Robert Luttik. Finally, acknow ledgem ent is also given to m embers of the EEA Scientific Committee, Philippe Bourdeau, Eileen Buttle, Robert Kroes, Bo Jansson, and Poul H arrem oes, the Chair o f the editorial team. Contents P r e f a c e .................................................................................................................................... 3 Acknowledgem ents............................................................................................................. 6 1. Introduction.................................................................................................................... 11 1.1. Late lessons from early warnings: an approach tolearning from history 11 1.2. What is the 'precautionary principle'?.......................................................... 13 1.3. An early use of the precautionary principle: London, 1854 ...................... 14 1.4. Forestalling disasters: integrating science and public p o licy................... 15 1.5. References........................................................................................................... 16 2. Fisheries: taking stock Malcolm M acG arvin...................................................................................................... 17 2.1. Early w arning s.................................................................................................... 17 2.2. 19th century British fish eries.......................................................................... 17 2.3. Californian sardine fishery 1920s to 1942 ................................................... 19 2.4. Newfoundland cod ........................................................................................... 20 2.5. Precaution becomes e x p lic it.......................................................................... 23 2.6. The ecosystem approach................................................................................. 25 2.7. Late lessons ......................................................................................................... 26 2.8. References ........................................................................................................... 27 3. Radiation: early warnings; late effects BarrieLambert ................................................................................................................. 31 3.1. X-rays .................................................................................................................. 31 3.2. Radioactivity and radioactive m aterials........................................................ 32 3.3. Early moves towards control of e x p o su re................................................... 33 3.4. The post-war watershed: justification, optimisation, limitation.............. 34 3.5. Conclusions ......................................................................................................... 35 3.6. References ........................................................................................................... 36 4. Benzene: an historical perspective on the American and European occupational setting Peter F. Infante ............................................................................................................. 38 4.1. Early w arning s.................................................................................................... 38 4.2. Actions and inactions ...................................................................................... 39 4.3. Discussion ........................................................................................................... 42 4.4. Conclusions and lessons for the future ........................................................ 46 4.5. References ........................................................................................................... 47 5. Asbestos: from 'magic' to malevolent mineral David Gee and Morris Greenberg .............................................................................. 52 5.1. Introduction......................................................................................................... 52 5.2. The first 'early warnings' of asbestosis and someresponses .................... 53 5.3. Early warnings on asbestos cancers ............................................................ 54 5.4. Early, devastating warnings about mesothelioma cancer ....................... 55 5.5. Actions and inactions by regulatory authorities and others ................... 56 5.6. The costs and benefits of actions and inactions ........................................ 58 5.7. What are the lessons of the asbestos story? .............................................. 59 5.8. References ........................................................................................................... 61 8 Late lessons from early warnings: the precautionary principle 1896-2000 6. PCBs and the precautionary principle Janna G. Koppe and Jane K e y s ................................................................................. 6.1. Introduction......................................................................................................... 6.2. Growing evidence of persistence, presence and toxicity ....................... 6.3. Action from industry and governments in the 1970s .............................. 6.4. Scientific understanding becomes more sophisticated ............................ 6.5. Government action in the 1980s and 1990s ............................................... 6.6. Routes of environmental exposure ............................................................... 6.7. The most recent PCB accident........................................................................ 6.8. Conclusion ........................................................................................................... 6.9. References ........................................................................................................... 7. Halocarbons, the ozone layer and the precautionary principle Joe Farman .................................................................................................................... 7.1. O v e rv ie w ............................................................................................................. 7.2. Early history......................................................................................................... 7.3. The 1930s -- the CFC industry is born ........................................................ 7.4. The 1970s -- the seeds of doubt ................................................................. 7.5. The Montreal Protocol and the ozone h o le ................................................. 7.6. Late le sso n s......................................................................................................... 7.7. References ........................................................................................................... 8. The DES story: long-term consequences of prenatal exposure Dolores Ibarreta and Shanna H. Swan ...................................................................... 8.1. Introduction) ...................................................................................................... 8.2. Optimistic beginnings ...................................................................................... 8.3. Tragic consequences ........................................................................................ 8.4. DES ineffective for prevention of miscarriage ............................................ 8.5. Assessing the extent of the damage ............................................................ 8.6. Lessons from the DES story ............................................................................. 8.7. References ........................................................................................................... 9. Antimicrobials as growth promoters: resistance to common sense Lars-Erik Edqvist and Knud Borge Pedersen .......................................................... 9.1. Introduction......................................................................................................... 9.2. The first early warning ...................................................................................... 9.3. Subsequent action or inaction ........................................................................ 9.4. Advantages and disadvantages of the use of growth promoters ......... 9.5. Conclusions and lessons for the future ........................................................ 9.6. References ........................................................................................................... 10. Sulphur dioxide: from protection of human lungs to remote lake restoration Arne S e m b ...................................................................................................................... 10.1. Dead fish, dying fo rests.................................................................................... 10.2. The 1985 CLRTAP Protocol and beyond ...................................................... 10.3. Late le sso n s......................................................................................................... 10.4. References........................................................................................................... 11. MTBE in petrol as a substitute for lead Martin Krayer von Krauss and Poul H a rrem oes...................................................... 11.1. Introduction......................................................................................................... 11.2. Lead in petrol .................................................................................................... 11.3. The MTBE case .................................................................................................. 11.4. The benefits of MTBE ..................................................................................... 64 64 66 66 67 69 70 71 71 73 76 76 78 79 79 80 82 83 84 84 84 85 86 87 88 90 93 93 94 94 97 98 99 101 102 105 106 108 110 110 110 110 111 11.5. The impacts of MTBE ...................................................................................... 11.6. R esp o n ses........................................................................................................... 11.7. Present tre n d s .................................................................................................... 11.8. Discussion in relation to the precautionary principle................................. 11.9. Conclusion........................................................................................................... 11.10. References........................................................................................................... 12. The precautionary principle and early warnings of chemical contamination of the Great Lakes Michael G ilbertson......................................................................................................... 12.1. The first significant early warnings ............................................................... 12.2. Date and nature of subsequent action or inaction ................................... 12.3. Consequences of institutional responses...................................................... 12.4. Costs and benefits............................................................................................. 12.5. Conclusions and lessons for the future ........................................................ 12.6. References........................................................................................................... 13. Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex David Santillo, Paul Johnston and William J. Langston ........................................ 13.1. Introduction......................................................................................................... 13.2. The emergence of the TBT problem ............................................................ 13.3. Arcachon B a y ...................................................................................................... 13.4. UK harbours and coastal waters ................................................................... 13.5. A global p o llutant............................................................................................. 13.6. Effectiveness of controls on small v e s s e ls ................................................... 13.7. The significance of seagoing v e s s e ls ............................................................ 13.8. Progress towards a global phase-out .......................................................... 13.9. The question of alternatives .......................................................................... 13.10. Late lessons from the TBT s t o r y ................................................................... 13.11. Conclusions: precaution or retrospective action? ..................................... 13.12. References........................................................................................................... 14. Hormones as growth promoters: the precautionary principle or a political risk assessment? Jim W. Bridges and Olga Bridges ............................................................................. 14.1. Introduction......................................................................................................... 14.2. Impacts of oestrogenic compounds on w ild life .......................................... 14.3. What were the uncertainties regarding the use of oestrogenic growth promoters for human health? ........................................................................ 14.4. Has the approach adopted by the European Commission proved to be so u n d ?.................................................................................................................. 14.5. Overall conclusions........................................................................................... 14.6. References........................................................................................................... 15. 'Mad cow disease' 1980s-2000: how reassurances undermined precaution Patrick van Zwanenberg and Erik M illstone............................................................ 15.1. Introduction......................................................................................................... 15.2. A new cattle d is e a s e ........................................................................................ 15.3. Initial decisions .................................................................................................. 15.4. Expert advice and regulatory controls ........................................................ 15.5. Constructing a house of c a r d s ........................................................................ 15.6. The failures and eventual collapse of the policy edifice .......................... 15.7. Conclusions......................................................................................................... 15.8. References........................................................................................................... 111 114 115 115 120 121 126 126 127 129 130 131 132 135 135 135 136 137 138 138 139 140 141 141 142 143 149 149 151 152 153 154 155 157 157 157 158 159 161 164 164 166 Contents 9 10 Late lessons from early warnings: the precautionary principle 1896-2000 16. Twelve late lessons Editorial team ................................................................................................................ 168 16.1. Introduction......................................................................................................... 168 16.2. Twelve late lessons ........................................................................................... 169 16.3. The wider implicationsof precaution ............................................................ 182 16.4. References........................................................................................................... 189 17. Conclusions .................................................................................................................... 192 17.1. Late lessons from earlyw arnings.................................................................... 193 Author biographies................................................................................................................ 195 Index: Late lessons from early warnings: the precautionary principle 1896-2000 . 200 1. Introduction Introduction 11 1.1. Late lessons from early warnings: an approach to learning from history In 1898, Lucy Deane, a UK Factory Inspector, observed: `The evil effects of asbestos dust have also instigated a microscopic examination of the mineral dust by HM Medical Inspector. Clearly revealed was the sharp glass-like jagged nature of theparticles, and where they are allowed to rise and to remain suspended in the air of the room in any quantity, the effects have beenfound to be injurious as might have been expected. ' (D eane, 1898) O ne hundred years later, in 1998, the UK governm ent d ecided to ban `w hite' asbestos, a decision that was echoed by the European U nion (EU ) the following year. The current asbestos-induced death rate in the United K ingdom is about 3 000 deaths per year, and some 250 000-400 000 asbestos cancers are expected in western Europe over the next 35 years, due to past exposures (Peto, 1999). In trying to reduce current and future risks the lessons of history have rarely been used. T he histories o f a selection o f hazards is therefore the subject m atter o f Late lessons. Fourteen case studies (arranged chronologically according to the first date of early warning) have been chosen from a range o f well-known hazards to workers, the public and the environment, where sufficient is now known abo u t their im pacts to enable conclusions to be drawn about how well they were dealt with by governm ents and civil society. Such conclusions should be based on `the spirit of the tim es' and not on the luxury of hindsight. There are other public health effects and environmental disasters that have not been looked at, such as thalidom ide (James, 1965), lead (Millstone, 1997), and the Aral Sea (Small, 2001) These provide additional information about unintended consequences, and the conflict between econom ic and social interests, from which additional lessons from history can be drawn. The hundred years between the 1890s and 1990s is the m ain focus o f this report into the use, neglect and possible misuse o f the concept o f precaution in dealing with a selection o f occupational, public and environmental hazards. The costs and benefits o f the actions or inactions of governm ents and others in responding to `early warnings' about hazards provide us with its content. T he aim o f this repo rt is to see if something can be learnt from these histories that can help us prevent, or at least minimise, future impacts o f other agents that may turn out to be harmful, and to do so without stifling innovation or compromising science. The report is an exam ple o f the inform ation needed to help the EU and EEA m em ber countries to frame and identify sound and effective policies that protect the environment and contribute to sustainable developm ent. Providing such inform ation is the regulatory duty o f the European Environment Agency (EEA), an independent Agency of the European Community established in 1993 to provide objective information to the policy-making bodies of the EU and its M em ber States (Council Regulations, 1210/90 and 993/99). The authors o f the case studies were asked to structure their chapters around four key questions: When was the first credible scientific `early warning' of potential harm? When and what were the m ain actions or inactions on risk reduction taken by regulatory authorities and others? What were the resulting costs and benefits of the actions or inactions, including their distribution between groups and across time? What lessons can be drawn that may help future decision-making? The case studies and authors have also been chosen with a transatlantic audience in mind. Three chapters are focused either on a North American issue (pollution o f the Great Lakes) or primarily on the North American handling o f issues that are also directly relevant to Europe (benzene, and DES adm inistered in pregnancy) and authored by scientists from North Am erica (Gilbertson, Infante, and co-author Swann, respectively). Three chapters cover issues o f some conflict between North America and Europe (horm ones as growth prom oters, asbestos, and MTBE in petrol); and all other chapters 12 Late lessons from early warnings: the precautionary principle 1896-2000 Table 1.1. are as relevant to North Americans, their public health and their environments as they are to Europeans. It is som etim es said that the U nited States does not use the precautionary principle, but it is worth noting (see Table 1.1.) that the United States has helped to prom ote what could be called `precautionary prevention', without necessarily calling it `the precautionary principle'. Som e exam ples of 'precautionary prevention' in the United States Issue Food safety (carcinogenic additives) Food safety (BSE) Environmental safety (CFCs) Public health (DES) 'Precautionary prevention' The D elaney Clause in the Food , Drug and Cosm etics A c t, 19 57-96, which banned animal carcinogens from the human food chain A ban on the use of scrapie-infected sheep and goat m eat in the animal and human food chain in the early 1970s w hich may have helped the United States to avoid BSE A ban on the use of chlorofluorocarbons (C FC s) in aerosols in 1977, several years before sim ilar action in most of Europe A ban on the use of DES as a growth prom oter in beef, 19 72-79, nearly 10 years before the EU ban in 1987 The precautionary principle has become controversial, not least because o f the disputes between the EU and the United States over horm ones in beef, genetically modified organisms (GMOs), global warming and other issues in which precautionary approaches have been invoked. There is now considerable debate (not to say term inological confusion, particularly between politicians on different sides o f the Atlantic) as to what the precautionary principle means, and how it can be im plem ented. One aim o f this report is to try to im prove transatlantic understanding on the use o f precaution in policy-making. benzene-exposed workers, and has worked for many years in the U S H ealth and Safety D epartm ent to reduce workers' exposure to benzene and other pollutants. All other authors, to varying degrees, have had significant involvement in the subject o f their chapters: indeed they would not have been approached if they had not already extensively studied the case that they were asked to write about. All o f them, as respected scientists in their fields, were expected to be as objective as possible in answering the four questions put to them. This involvement o f the authors in the histories o f their case studies is therefore brought to the attention o f readers. The authors o f the case studies, who provided their services pro bono, were asked to keep their contributions brief, which obviously inhibits detailed treatm ent o f the issues. However, we wanted to elicit key conclusions from the histories and not the detailed post m ortems that others have produced: these can be accessed via the references in each chapter. It has been pointed out that the case study authors are n ot w ithout strong views, being for the m ost part active participants in the process o f making the histories that are sum m arised in each chapter. Jo e Farman, the author o f the chapter on halocarbons, for example, discovered the `hole' in the stratospheric ozone layer; Morris Greenberg helped to set up the first asbestos m esotheliom a register; Michael Gilbertson has spent m ost o f his professional life researching Great Lakes pollution and advocating its clean-up; and Peter Infante did the first cohort epidem iological study of The case studies are all about `false negatives' in the sense that they are agents or activities that were regarded at one time as harm less by governments and others, at prevailing levels of exposure and `control', until evidence about their harmful effects em erged. But are there no `false positives', where action was taken on the basis o f a precautionary approach that turned out to be unnecessary? It was felt necessary to include such examples, but despite inviting some industry representatives to submit them, and discussing these in some detail, no suitable exam ples emerged. Attention was drawn to a U S publication, Facts versus fears (Lieberm an and Kwon, 1998), which attem pted to provide som e 25 exam ples o f `false positives'. However, on closer exam ination these turned out not to be robust enough for those who recom m ended them to accept our invitation to use the strongest half dozen in this report. The challenge o f dem onstrating `false positives' remains: possible candidates that have been m entioned include the ban on Introduction 13 dum ping sewage sludge in the North Sea, an d the `Y2K m illennium b u g '. 1.2. What is the 'precautionary principle'? Albert Schweitzer (1875-1965) may have been pessim istic when he said `M an has lost the capacity to foresee an d forestall... he will end up destroying the earth'. However, being wise before it is too late is n ot easy, especially when the environmental or health impacts may be far into the future and the real, or perceived, costs of averting them are large and immediate. Forestalling disasters usually requires acting before there is strong p roo f o f harm, particularly if the harm may be delayed and irreversible, an approach to scientific evidence and policy-making which is p art o f what is now called the precautionary principle. Precautionary prevention has often been used in m edicine and public health, where the benefit o f d ou bt abo u t a diagnosis is usually given to the patient ( `better safe than sorry'). However, the precautionary principle and its application to environm ental hazards and their uncertainties only began to emerge as an explicit and coherent concept within environmental science in the 1970s, when German scientists and policy-makers were trying to deal with `forest d eath ' ( Waldsterben) and its possible causes, including air pollution. The main element of the precautionary principle they developed was a general rule o f public policy action to be used in situations o f potentially serious or irreversible threats to health or the environment, where there is a need to act to reduce potential hazards before there is strong p ro o f o f harm , taking into account the likely costs and benefits o f action and inaction. A precautionary approach, however, requires much m ore than establishing the level o f proof needed to justify action to reduce hazards (the `trigger' for action). The Vorsorgeprinzip ( `foresigh t' or `precaution ary' principle), in the Germ an Clean Air Act of 1974, as elaborated in the 1985 report on the Clean Air Act (Boehmer-Christiansen, 1994) also included elements such as: research and m onitoring for the early detection of hazards; a general reduction of environmental burdens; the prom otion of `clean production' and innovation; the proportionality principle, where the costs o f actions to prevent hazards should not be disproportionate to the likely benefits; a cooperative approach between stakeholders to solving com m on problem s via integrated policy m easures that aim to improve the environment, competitiveness and employment; action to reduce risks before full `p r o o f o f harm is available if im pacts could be serious or irreversible. Since the 1970s, the precautionary principle has risen rapidly up the political agenda, and has been incorporated into many international agreements, particularly in the marine environment, where an abundance of ecological data on pollution yielded little understanding but m uch concern: `huge amounts o f data are available, but despite these data... we have reached a sort o f plateau in our understanding o f what that inform ation is for... This is what led to the precautionary prin ciple' (Marine Pollution Bulletin, 1997). M ore generally, Principle 15 of the UN Rio Declaration on Environment and Developm ent 1992 (see Table 1.2.) extended the idea to the whole environment. The use o f different terms in these treaties and agreements such as `precautionary principle', `precautionary ap p roach ' and `precautionary m easures' can cause difficulties for communication and dialogue on how best to deal with scientific uncertainties and potential hazards. The concluding chapters o f this report attem pt to clarify som e o f these ambiguities. In Europe, the most significant support for the precautionary principle has come from the E u ro pean Com m ission's Com m unication on the Precautionary Principle (European Commission, 2000) and the Council of Ministers Nice Decision, both in 2000. They have m ade significant contributions to the practical im plem entation o f the precautionary principle, especially concerning stakeholder involvement and the avoidance o f trade disputes. Some o f the m ain issues raised by the case studies and by the E u ro pean Com m ission's Com m unication are elaborated in the concluding chapters. 14 Late lessons from early warnings: the precautionary principle 1896-2000 Table 1.2. Source: EEA The 'precautionary principle' in som e international treaties and agreem ents Montreal Protocol on Substances that Deplete the Ozone Layer, 1987 'Parties to this protocol... determined to protect the ozone layer by taking precautionary measures to control equitably total global emissions of substances that deplete it...' Third North Sea Conference, 1990 'The p a rticipants... will continue to apply the precautionary principle, that is to take action to avoid potentially damaging im pacts of substances that are persistent, to xic, and liable to bioaccum ulate even w here there is no scientific evidence to prove a causal link between em issions and effects.' The Rio Declaration on Environm ent and Developm ent, 1992 'In order to protect the environm ent the Precautionary Approach shall be w id ely applied by states according to their capabilities. W here there are threats of serious or irreversible dam age, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.' Fram ew ork Convention on C lim ate C hang e, 1992 'The Parties should take precautionary measures to anticipate, prevent or minimise the causes of climate change and mitigate its adverse effects. W here there are threats of serious or irreversible dam age, lack of full scientific certainty should not be used as a reason for postponing such m easures, taking into account that policies and m easures to deal with climate change should be cost-effective so as to ensure global benefits at the lowest possible cost.' Treaty on European Union (M aastricht Treaty), 1992 'Com m unity policy on the environm ent... shall be based on the precautionary principle and on the principles that preventive actions should be taken, that the environmental damage should as a priority be rectified at source and that the polluter should pay.' C artagena Protocol on Biosafety, 2000 'In accordance with the precautionary approach the objective of this Protocol is to contribute to ensuring an adequate level of protection in the field of the safe transfer, handling and use of living m odified organism s resulting from modern biotechnology that may have adverse effects on the conservation and sustainable use of biological diversity, taking also into account risks to human health, and specifically focusing on transboundary m o ve m e n ts.' Stockholm Convention on Persistent Organic Pollutants (POPs) 2001 Precaution, including transparency and public participation, is operationalised throughout the treaty, with exp licit references in the pream ble, objective, provisions for adding PO Ps and determ ination of best available technologies. The objective states: 'M indful of the Precautionary Approach as set forth in Principle 15 of the Rio Declaration on Environm ent and D evelopm ent, the objective of this Convention is to protect human health and the environm ent from persistent organic pollutants.' 1.3. An early use of the precautionary principle: London, 1854 The use o f precautionary approaches to hazards began well before the 1970s, particularly in the field o f public health. One early application in Europe was by Dr Jo h n Snow, who in 1854 recom m ended removing the handle from the Broad Street water pum p in an attem pt to stop the cholera epidem ic that was then ravaging central London. Some evidence for a correlation between the polluted water and cholera had been published five years earlier by Snow him self (Snow, 1849). This evidence was not `proo f beyond reasonable doubt'. However, it was p ro o f enough for Snow to recom m end the necessary public health action, where the likely costs o f inaction would have been far greater than the possible costs o f action (see Box 1.1.). The costs o f Snow being wrong in getting the pum p handle removed would essentially have been angry and inconvenienced citizens who nevertheless wanted cholera stopped. These costs were small in relation to the cost of bein g w rong in not rem oving the pum p handle, once the evidence of the link between the Broad Street pum p water and the cholera was available. His evidence that there seem ed to be a link was reliable enough to help make a public policy decision that proved correct: the cholera was being caused by sewage-contaminated water and removing exposure helped to remove the risk. The story ofJo h n Snow and cholera has sometimes been misinterpreted as an exam ple o f how very strong evidence of harm and its causes can be used in a relatively uncontroversial way. However, it was a classic case of precautionary prevention, containing several o f the key elem ents o f an approach to scientific uncertainty, ignorance and policy making. These elements include the difference between `knowing' about a hazard and its likely causes and `understanding' the chemical and biological or other processes underlying the link; a focus on the potential costs of being wrong; and the use of minority scientific opinions in public policy-making. These issues are taken up in the concluding chapters o f this report. There are many differences between cholera, asbestos (which came into use at about the Box 1.1. John Snow's 'precautionary prevention' In a 10-day period from 31 A u gust to 9 Septem ber 1854, there w ere about 500 deaths from cholera in the parish of St. Jam es, which included the Golden Square area of Central London. John Snow, a London physician, investigated the outbreak, having previously written The M o d e o f Com m unication o f Cholera, a pamphlet of 30 pages which he published at his own expense in 1849. Prior to the Golden Square outbreak, Snow was studying cholera and the water supplies from two different w ater com panies in South London: one 'clean' and the other 'polluted' with sewage. This incomplete study was already producing data that supported his theory that cholera w as caused by contam inated w ater when he w en t to investigate the Golden Square outbreak. A short investigation revealed that virtually all of the 83 people w ho had died in the Golden Square area between 31 August and 5 Septem ber had drawn w ater from the popular Broad Street water pump, rather than from the available, and cleaner yet less popular, piped water supplies. On 7 Septem ber, Snow recommended the removal of the Broad Street w ater pump on the grounds that there was 'no... Cholera... except amongst persons, w ho w ere in the habit of drinking the water of the (Broad Street) water pump'. The authorities removed the pump handle the next day, thereby helping to speed up the declining cholera outbreak and preventing further infection from that source. Snow later produced one of the first epidem iological maps of disease and possible causes at a presentation to the Epidemiological Society of London on 4 Decem ber 1854, which included a map of cholera deaths and the wells nearest to Broad Street. Snow's views on cholera causation were not shared by the majority of relevant scientists. The Royal College of Physicians inquiry into the earlier 1853 54 cholera outbreak had considered Snow's thesis and rejected it as 'untenable', as had the General Board of Health in 1854: 'w e see no reason to adopt this belief'. They believed that cholera was caused by airborne contamination. The biological mechanism underlying the link between polluted water and cholera was unknown at the time of this successful 'precautionary prevention' in 1854: that cam e 30 years later, in 1884, when Koch announced his discovery of the cholera vibrio in G erm any. E E A , based on Brody e t al., 2000 time o f Snow 's action ), an d the other harm ful agents in the case studies, not least being the time lag between exposure to the harm ful agent and the health dam age, which was hours in the case o f cholera but decades in the case of asbestos and most of the other agents studied. Yet had governm ents adopted a similar approach to precautionary prevention as Dr Snow, once the early warnings on asbestos had been published, much o f the tragedy and the huge costs of asbestos exposure could have been averted. 1.4. Forestalling disasters: integrating science and public policy Snow was working, both as a scientist and policy-maker, in the conditions o f scientific uncertainty and political stress shared by all who are charged with responsibilities for protecting the public and the environment from potentially harm ful econom ic activity. Politicians today are working in similar conditions o f scientific uncertainty and stress as Dr Snow, but now m ade m ore difficult by the higher risks and uncertainties (economic, health and ecological) of largerscale activities (Beck, 1992) and by greater pressure from the mass m edia (Smith, 2000). They also work with m ore dem ocratic institutions, and are accountable to a bettereducated and involved citizenry which can have good access to inform ation from the Internet. Globalisation and free trade issues add further complications, as does the emerging science of complexity and chaos, which can require m ore humility and less hubris in science. It is in these circumstances o f trying to prevent potentially serious and irreversible effects, without disproportionate costs, that the precautionary principle can be useful. It helps policy-makers and politicians, in circumstances in which waiting for very strong evidence of harm before taking precautionary action, may seriously com prom ise public health or the environment, or both. Achieving consensus on the history of accepted hazards, such as asbestos, chlorofluorocarbons (CFCs) and the other case studies, is n ot easy, bu t it is easier than achieving consensus on how to deal with current controversies such as climate change, mobile phones or GMOs. There are some well established criteria for helping scientists to move from `association' to `causation', in health hazard identification (Hill, 1965), but there are no generally accepted criteria for helping politicians to make sound public policy decisions in the face o f scientific uncertainty, despite several good proposals (Raffensperger and Tickner, 1999; Gee, 1997). T here is already a large literature on risk assessment and hazard reduction which can assist decision-makers in certain circumstances, but an historical perspective m ight also help. There is a rich history o f hazards covered in this report from which som ething o f value can be learnt. In chapter Introduction 15 16 Late lessons from early warnings: the precautionary principle 1896-2000 16 lessons are drawn that can help fram e and identify sound and effective public policy m easures. These may help minimise the future costs of being wrong about environm ental an d health risks. T here is particular concern to see fewer `false negatives' in the future, but the late lessons should also help reduce the smaller but commonly feared risk o f `false positives'. European Commission, 2000. Com munication from the Commission on the Precautionary Principle, COM (2000) 1, Brussels. Gee, D,1997. `A pproach es to Scientific U ncertainty', in Fletcher, T. an d M ichael, A. J . M. (eds), Health at the crossroads: Transport policy and urban health, Wiley, L on don . Public trust in the politicians and scientists who are trying to protect people and the plan et from hazards is very low, especially in Europe, where BSE in the United Kingdom and elsewhere, dioxins in Belgium, and the HIV-contaminated blood transfusion affair in France have contributed to a general sense of malaise. Governments are aware o f this and are developing responses, such as the EU White Paper on European Governance (July 2001). This includes recommendations for improving public participation in m anaging the inter-reactions between science, technologies and society. This report aims to contribute to the debate on the em erging issue of democratising scientific expertise. 1.5. References Beck, U., 1992. Risk Society, Sage, Lon don . Hill, A. B., 1996), `The environment and disease: A ssociation or causation ?' Proceedings of the Royal Society ofMedicine Vol. 58, pp. 295 300. Jam es, WH, 1965. `Teratogenetic properties o f thalidom ide' Br MedJ, 1965 Oct 30; 5469:1064 Lieberm an, A. J . and Kwon, S. C., 1998. Facts versus fears: A review of the greatest unfounded health scares of recent times, 3rd ed., revised June 1998, American Council on Science and Health, New York, at http://w w w.acsh.org. M illstone, E., 1997. Lead and public health, Earthscan, London. Marine Pollution Bulletin, 1997. Vol. 34, No 9, pp. 680-681. Boehm er-Christiansen, S., 1994. `The precautionary principle in Germany: Enabling govern m en t', in O 'R iordan, T. and C am eron, J . (eds), Interpreting theprecautionary principle', p. 3, C am eron an d May, Lon don . Brody, H. et al., 2000. `M ap-making an d myth making in Broad Street: the London cholera epidem ic, 1854', Lancet Vol. 356, pp. 64--68. Deane, Lucy (1898). Report on the health of workers in asbestos and other dusty trades', in HM Chief Inspector o f Factories and W orkshops, 1899, A nnual Reportfor 1898, pp. 171-172, HMSO, London (see also the Annual Reports for 1899 and 1900, p502). Peto, J., 1999. `The European mesotheliom a epidem ic', B. J. CancerVol. 79, February, pp. 666-672. Raffensperger, C. and Tickner, J., 1999. Protectingpublic health and the environment: Implementing the precautionary principle, Island Press, Washington, DC. Snow, J., 1849. On the mode of communication of cholera, Lon don . Smith, J . 2000. The Daily Globe: Environmental change, the public and the media, Earthscan, London. 2. Fisheries: taking stock Malcolm MacGarvin Fisheries: taking stock 17 Fisheries have had to deal with uncertainties -- with attem pts m ade to m anage them -- for centuries. T he topic is therefore doubly challenging as, while the underlying precautionary principle is the sam e as for pollutants, the practical approach to im plem entation is necessarily very different. `Late lesson s' is certainly an appropriate topic for m arine capture fisheries. Awareness of events such as those described here -- from the M iddle Ages, through to 19th century Scottish fisheries, the mid-20th century Californian sardine fishery crash and the collapse of Canadian northern cod stocks in the 1990s -- provide a sometimes gloomy awareness o f history repeating itself. But there are also positive things to learn from the past. The need for an explicitly precautionary approach has been increasingly recognised in the last decade. In areas such as the North Atlantic, the principal problem now is finding a m eans to allow fishers to turn theory into practice, against the backdrop of short-term economic pressures. 2.1. Early warnings The relationships between precautionary attitudes, cultural perspectives, technological ability, an d risks an d benefits, are com plex. In a sense there is nothing new in varying degrees of precautionary management, and warnings of the risk of overexploitation. There is som e evidence that certain Native North American communities may have died out because they overexploited marine resources. Others, however, circum scribed their catches by rights and taboo, re-enforced by knowledge of fish ecology. They maintained substantial catches for centuries, for exam ple o f salmon, in contrast to the commercial fishers who displaced them and extinguished the stocks (McEvoy, 1986). In medieval Europe people were also aware that fish could be overexploited. An early call for precautionary action comes from 1376 77 (March, 1953). A petition was presented to the English parliam ent calling for the prohibition o f a net `of so small a mesh, no m anner of fish, however small, entering within it can pass out and is com pelled to remain therein and be taken... by m eans o f which instrum ent the fisherm en aforesaid take so great an abundance of small fish aforesaid, that they know n ot what to do with them, but feed and fatten pigs with them, to the great dam age of the whole commons of the kingdom, and the destruction o f the fisheries in like places, for which they pray remedy.' The response to this assertion was to set up a comm ission `by qualified persons to inquire and certify on the truth o f this allegation, and thereon let right be done'. Between then and the late 19th century there were num erous attempts in England (as elsewhere) to regulate fishing by still familiar m eans (such as, in 1716, a minim um mesh size, a ban on circumvention by placing one net within another and m inim um fish landing sizes). 2.2. 19th century British fisheries In the 19th century British fisheries, as elsewhere, grew rapidly. Uncertainty about the consequences for the valuable herring fishery, and the expansion of trawling for other species (recently unearthed by a `lay' Wick inhabitant, historian Ian Sutherland (Sutherland, n.d.)) provoked, am ong other events, a series o f official enquiries between 1866 and 1893 (Report, 1866; Report, 1885). 2.2.1. Scottish herring fishery Many marine fish species markedly fluctuate in abundance as a result of natural causes. In this, herring and its relatives, such as sardine, can show extreme changes over tens of years. Given this context, it took a long time for it to be accepted that fisheries could exacerbate such natural cycles. In 1865 the prescient Jam es Bertram (Bertram, 1865, pp. 277-282) docum ented inshore Scottish herring catches between 1818 and 1863, when the area of drift nets carried per boat grew from 4 500 to 16 800 square yards, yet the catch fell from 125 to 82 crans (barrels). Such figures spoke too plainly to him to `expend further argu m en t'. H e wrote `I have always been slow to believe in the inexhaustibility o f the shoals, and can easily im agine that overfishing, which som e people pooh-pooh so glibly, could easily be possible... As it is, I fear the great Wick fishery m ust come som e day to an end. When (it) first began the 18 Late lessons from early warnings: the precautionary principle 1896-2000 fisherm an could carry in a creel on his back the nets he required; now he requires a cart and a good strong horse.' Subsequent action or inaction B ertram 's words proved prophetic. Yet several decades later Thom as Huxley, President o f the Royal Society and Inspector o f Fisheries (no fool, yet now infamous for his com m ent that `m arine fisheries are inexhaustible' (Huxley, 1883)), could still maintain for British herring in general `nothing to show, so far as I am aware, that, taking an average o f years, they were ever either m ore or less num erous than they are at present' (Huxley, 1881), while in 1893 a parliam entary comm ittee also saw `no indication of any falling off in the num bers of herrings to be found off our coast'. Presumably landings were equated with stock size, without allowing for increasing effort, area covered and length of season. In the 1890s the Germ an Heincke established, contrary to general belief, that herring existed as isolated races, implying m anagem ent at this level. But by then the innerm ost shoals o f Wick and the Moray Firth were gone. The fishing moved further offshore and, overall, North Sea herring landings fluctuated without trend during the first h alf o f the 20th century. This was not so m uch due to lessons learned and action taken, but to turmoil in the continental European markets that restrained the technological potential o f the new m otorised drifters. When an `industrial' fishery for fishmeal and oil subsequently developed in the second h alf o f the 20th century, both the new technology and the necessary safety m argins were underestimated. This brought the North Sea herring to the brink of collapse by the 1970s, forcing a m oratorium on the fishery. The stocks did recover, and the industrial fishery upon herring was subsequently restricted when fishing reopened; nevertheless the pressure from `human consum ption' fisheries remained so great that, by the mid-1990s, further emergency m easures had to be im posed, until once m ore the stocks recovered. Although we appear to have learned enough by the late 1990s to stave off total collapse, it was hardly a m odel o f effective m anagem ent. 2.2.2. The advent of steam trawlers Another m ajor developm ent o f the late 19th century was steam-powered trawling for `dem ersal' fish (such as cod, haddock, whiting and flatfish). This allowed access to areas too dangerous or inaccessible for sail or oar (a natural equivalent o f m odern `no-take' zones), and increased the ability to drag and haul nets, resulting in much bigger catches. Trawling divided scientists and fishers alike. The principal com plaints were that trawling destroyed spawn on the seabed; caught im m ature fish; resulted in wastage through dam aged fish; in terfered with other fishers; and that trawlermen -- mobile outsiders -- depleted the traditional grounds o f others. Interestingly, in 1883 trawlermen based in Yarmouth themselves voluntarily agreed not to work certain offshore grounds at certain seasons to avoid catching im m ature fish, although this was underm ined by other trawlermen and abandoned. Subsequent action or inaction In 1865 Bertram raised such accusations against trawling, although he also thought it was, used wisely, the best tool for certain fisheries (Bertram, 1865, p. 308). By 1883 another parliamentary enquiry had comm issioned research which concluded that the im m ature catch from trawling was insignificant (less than by traditional m ultiple hook and line `long-lining'); that m ost fish caught were undam aged for sale and that, o f the comm ercial species, only herring spawn lay on the seabed, and even here it was questionable whether injury would occur. But regarding the effects o f trawling within the m ost accessible inshore waters, M cIntosh -- the research 's author -- stated that these `could very easily be trawled out' and that the suspension o f trawling `w ould be a very valuable experim ent. It would set at rest all sorts o f notions, and it is upon safe lines, and does no harm to anyone'. This was supported by Scottish fisheries officers, other scientists, some trawlermen, and evidence from neighbouring countries and the United States. However the enquiry concluded that, although fishers tended to blame each other for any decline, natural fluctuations `largely influence the supply of fish'. They could not justify prohibition unless experim ental evidence `decisively settled' the issue, because o f the loss of valuable catches. This was controversial. In Scotland (only), despite the enquiry, a political decision was m ade to ban trawling in inshore waters, including the Moray and other firths. This continued until the 1920s when fishers, under pressure from the failing herring m arket and searching for alternatives, started dragging their `seine' nets, effectively using Fisheries: taking stock 19 these as a trawl. T he initially large landings o f codfish and flatfish suggested that the previous policy had successfully nurtured stocks. 2.2.3. Costs and benefits It is possible to give som e qualitative statements about the im m ediate costs and benefits o f actions taken in the 19th century regarding British herring fishing and trawling (MacGarvin and Jones, 2000). Scottish towns such as Wick, heavily dependent on local herring fisheries, went into decline. The centralisation that accom panied trawling contributed to an increase in the prosperity o f larger centres, such as Fraserburgh, Peterhead and Aberdeen, at the expense o f smaller communities. The same expansion was true o f the hom e ports o f widely ranging trawler fleets south o f the border, such as Grimsby, Hull, Newlyn, Fleetwood and Swansea. Over the longer term, the 19th century set in train events that can be traced through to the presen t day. Concentrating on Scotland, landings and boat tonnage (including both the herring and derm ersal fleet) stayed remarkably constant between 1898 and 1998, at ca. 333 000 tonnes and 109 000 tonnes respectively (Scottish Office, 1898 to present). But the num ber o f boats fell from 11 536 to 2 661, an d the n um ber o f fishers from 36 161 to 7 771. Sail an d oar boats fell from 11 383 to nil, an d fossil fuel use increased enormously. Prosperity fluctuated widely, bu t overall was n ot m aintained. By the late 1990s the average net profit on capital invested for the Scottish dem ersal fleet was ju st 0.1 % per annum , according to the 2001 European Commission Fisheries Green Paper (European Commission, 2001b). One can question ju st what had been gained by this investment in technology. But on the benefits side, one change has been a marked fall in deaths at sea (although fishing remains dangerous). In the good years, it may also have provided som e individuals with a m eans o f acquiring significant capital in a relatively rem ote area with few econom ic opportunities. This could then be invested in other activities, or elsewhere, resulting in higher returns. In such circumstances there is no particular interest in m aintaining a sustainable fishery. But for others (probably the majority) it was an d is a way o f life, with a deeply ingrained wish to m aintain this for future generations, and a sense of being forced down a track whose direction they do not control. The changing catch composition for Scotland was broadly similar to that for the North Sea as a whole during this period of over a century. First, overall in the North Sea, herring caught for human consumption become less important. Dem ersal catches increased and then declined in turn, being supplanted by `industrial' catches (initially for herring, sprats and mackerel, and later for species such as sand-eels) which grew explosively from the 1950s. These peaked, for the N orth Sea as a whole, in the 1970s with 2.2 o f the 3.5 m illion tonne total catch. This reflects a global pattern of `fishing down the food w eb' (Pauly et a l , 1998), proceedin g from (potentially) high-value species to industrial stocks lower in the food web, and o f lower unit value. One can surmise that the vast rem oval o f biom ass by the fisheries will have also had an im pact on other species, but comprehensive data are scarce. 2.3. Californian sardine fishery 1920s to 1942 By the 1920s both exploitation strategies and scientific and m anagerial arrangem ents were beginning to take their current form. In this the Californian sardine fishery was precocious. The fish were originally canned for hum an consumption, but similar market conditions to those faced by the British herring fishers m eant that their reduction to fishmeal and oil becam e far m ore important, driving the economics. At this time there were no catch limits. Californian state scientists, involved in overseeing the fishery, were by the mid-1920s emphasising what was clearly precautionary action: `Unnecessary drain upon the supply should be avoided until research has shown that it is possible to detect overfishing in tim e' (McEvoy, 1986, pp. 160-161). It was the function o f governm ent `not only to aid in the greatest possible use, but to ensure its continuance because it is the only agency uniting all factions and successive generations' (McEvoy, 1986, p. 159). As in Scotland decades earlier, the catch was no longer increasing in proportion to the increasing effort; the average age o f the fish was falling, and ships had to travel further and fish longer -- classic signs of overfishing. A limit was recom m ended. While there were uncertainties the State Fisheries Laboratory concluded that the growing intensity o f the 20 Late lessons from early warnings: the precautionary principle 1896-2000 fishery was `fact enough to make us sure that we are headed for destruction and a great loss'. 2.3.1. Subsequent action or inaction The federal authority, the United States B u reau o f Fisheries, took a different view. While acknowledging that the evidence of depletion was strong, there was `no clear-cut or convincing evidence that will satisfy everyone' that the sardines were overexploited. Its view was `to us conservation m eans wise use. We do n ot believe in hoarding our fisheries resources'. Rather, `We believe very firmly that restrictions that are unnecessary ham per or restrict legitimate business enterprise' (McEvoy, 1986, pp. 162-166). The fishers were divided. Those disadvantaged by the developm ents saw depletion as inevitable. Those behind the expan sion saw a plot to `im poverish one o f the few successful enterprises in the D epression'; that any changes could be attributed to environmental fluctuations, and that in any case sardines were so fecund they would soon grow back. To simply discount their testimony that the sardines were abundant was `brutal... medieval sch olasticism ' . The dispute rum bled on unresolved until, in 1939, the new Californian Governor replaced the state experts with `em ergency' appointees, and the tone o f the reports changed from `unm istakable' signs of depletion and an `im perative' need to reduce the catches (1938) to `no reason to be concerned' (1942). It was then agreed to increase the catch (to `assist the war-effort'), but in that year the sardine stock collapsed. It only began to show signs o f a recovery in the mid-1980s (San Diego Natural History Museum, 2000). 2.3.2. Costs and benefits There are similar im ponderables regarding the im pact on other species as for the British fisheries. Regarding the economics, the rate o f return on processing fishmeal and oil p rior to the collapse was extraordinary, with many plants in the 1930s recovering the entire investm ent in one season (McEvoy, 1986, p. 145). So far as the processing companies such as Starkist and Van Camp were concerned, it could be argued that they had followed an optim um econom ic strategy, effectively `m in in g' the fish in a seller's m arket as quickly as possible, and then shifting equipm ent and operations down to South A m erica, with heavy involvement in open in g the Peruvian anchoveta fishery, whose exploitation followed the Californian model, and which collapsed in the early 1970s (McEvoy, 1986, p. 155). 2.4. Newfoundland cod Soon after the Californian experience, m anagers turned to increasingly complex mathematics in the attem pt to squeeze more from the limited inform ation on stock sizes, an d the effect o f fishing intensity. T his was assisted by ever increasing computational powers. By the 1970s there was optim ism that past mistakes could be avoided. N o m ore was this so than for the New foundland `northern co d ' stock. This, historically the largest cod stock in the world, had been exploited by European fishers since the 16th century (DFO, 2000). However, fishing intensity grew dramatically in the 1960s with a peak catch o f 800 000 tonnes in 1968, after which it dropped well below those actually authorised by the international regulatory body. This, the International Commission for the Northwest Atlantic Fishery, ICNAF, was regarded by many Canadians as ineffective (O 'Reilly Hinds, 1995), a consideration also true o f its successor, the Northwest Atlantic Fisheries O rganization, NAFO (Day, 1995). In the late 1970s, using the then novel UN Convention on the Law of the Sea, Canada extended its jurisdiction from 12 to 200 nautical miles, one intention being to bring m uch o f this stock under its control. Its goal was to im pose what many would even now consider as precautionary measures, setting `deliberately conservative' restrictions on catches, aim ing to limit these to ca. 20 % of the stock, with the intention o f rebuilding. T he D epartm ent o f Fisheries an d O cean 's calculations indicated that this was happening, and offshore catches by Canadian trawlers increased. In 1988 the Departm ent of Fisheries and Oceans (DFO) claim ed a `five fold increase in N orthern Cod since 1976', and it was held in wide regard as an exam ple o f how cautious, science-driven, m anagem ent could turn around a seemingly hopeless situation. Fisheries: taking stock 21 2.4.1. Subsequent action or inaction The only dissenting group were inshore fishers, who (unlike those offshore) had generally not upgraded their technology, and who could not reconcile their falling catches with this supposed increase. Their protests disregarded, they com m issioned what becam e the Keats Report, published in 1986. Keats highlighted the D F O 's own (downplayed) retrospective analyses that indicated consistent and severe underestimation of the fishing pressure on the stock since the imposition o f Canadian control, with `the result that we have consistently taken from 1.5-3 times the (20 % o f stock) catch since 1977' (Keats et a l, 1986). The DFO dismissed this `as biased pseudoscience written to support a political agen da' (Finlayson, 1994). Keats nevertheless gained m edia attention, forcing the commissioning, by the federal fisheries minister, o f an official report. This 1988 Alverson R eport had authority, in official eyes, because it was prepared by fisheries scientists. In the body o f the report they concluded that the stock had increased since 1977, although after 1982 it increased `probably only very slowly'. But they too stated that the `fishing mortality actually exerted has been considerably in excess o f target mortality' because o f the `consistent overestimation o f the current stock size'. The prob lem was that, for any year, som e five subsequent years' data were required before the estimated fishing mortality and biomass estimates for the original year `effectively converged to the correct answer'. The shorter this period the m ore it simply reflected the assum ed level o f fishing mortality. This is a critical flaw where a stock becom es so depleted that it depends on the last few years' breeding success (now comm onplace for many stocks). Moreover, if a stock goes through a period o f sharp decline, the m ethod instils a false sense of assurance. Alverson dem onstrated how sensitive conclusions about stock size were to the wide range o f fishing mortality estimates that m ight be im plied by the data. However in the executive summary this was turned on its head, it being stating that the DFO calculations o f fishing m ortality fell `within the range o f estimates supported by the data', albeit at the lower end. The DFO had responsibility for the m anagem ent o f the release o f the Alverson R eport. Publicly Alverson n oted that `it's rather amazing that we are as close to each other as we a re '. The DFO in its response em phasised that `the difference in num bers overall was about 4 to 5 percent... the conclusions... are quite similar (with respect to stock size and cause o f the decline in inshore fisheries)... the credibility o f DFO science was not questioned'. Privately the Alverson team were less sanguine and there was an internal reappraisal o f DFO methodology. The 1989 assessment assumed that fishing mortality was higher, concluded that the stock was not growing, and recom m ended that the offshore catch be virtually halved. This reappraisal was seen as an admission that the DFO had got it wrong all along. This caused serious problem s for the administration, as they depended on the `science' touchstone as an arbiter to conflicting claims to resources. Now the offshore fishers were com plaining bitterly that there was no evidence that stocks had fallen. The fisheries minister called a new enquiry, the 1990 Harris Report, this time fully independent o f DFO control. Harris also concluded that, prior to 1989, fishing mortality was probably m ore than double that intended, and the stock little more than h alf the assum ed size, with the result that the stock had been fished at levels that pointed towards commercial extinction -- a conclusion widely reported by the m edia. The Harris Report cautiously concluded that the revised DFO 1989 assessment was a better approxim ation o f reality. But also, over many pages, it pointed to m ajor issues, not easily resolved, at every conceivable level. `We acted in substantial ignorance of the animals in which we were principally interested and in alm ost total ignorance o f the dynamics o f the ecosystems in which they existed.' `We continued for too long to wear rose tinted glasses and to interpret all data in the m anner best calculated to support and confirm the m odel o f growth upon which our hearts had been set.' What were (and often still are) `believed to be the best available m anagem ent theory, data and assessm ent m eth odologies will legitim ately support claims of stock status ranging from sustainable growth to dangerous decline', (Harris, 1990). 2.4.2. Costs and benefits Harris estimated that the total allowable catch, TAC, would have be reduced from 235 000 tonnes in 1989 to ca. 125 000 in 1990 to bring this into line with the goal o f no 22 Late lessons from early warnings: the precautionary principle 1896-2000 m ore that 20 % stock removal, but that this `would precipitate social an d econom ic repercussions of a particularly drastic nature'. Instead Harris suggested a TAC of 190 000 tonnes (ca. 30 % removal), although cautioning that `this may contribute to further decline'. Yet this was done, at a loss o f CAD 26 m illion (ca. E U R 21 m illion) o f landings, CAD 66.6 million (EUR 53 million) processed product and the equivalent of som e 1 000jo bs. Similar limits were set for 1991-92. But during the 1992 fishing season it becam e apparent that there was little left to catch. The situation was far worse than even the most pessimistic projections. An emergency m oratorium was im posed in July 1992, initially for two years. B ut the stock failed to rebound, and it was not until 1999 that an inshore fishery ofju st 9 000 tonnes was perm itted. The financial cost during the 1990s, including lost sales, unemployment benefit and financial assistance, was in excess o f several billion Canadian dollars (MacGarvin, 2001a). Somewhat depressingly, the 2000 DFO assessm ent concluded that 1999 catch was already in excess o f the 20 % reference level, som ething `unacceptable under a precautionary approach', and that the stock rem ained so weak that even an index fishery (i.e. to m onitor the state o f the stock) `may be associated with an increased risk o f the inshore (stock) declining and the offshore not recovering'. The 2000 DFO assessment also makes it clear ju st how little is understood about this most studied stock -- why it collapsed, why it has failed to recover, what proportion are being taken by predators and other fishing activities -- and o f the poor status o f capelin, small fish that are im portant prey for the northern cod. Even the linkage between inshore stocks and those offshore -- the source of the original controversy -- is now open to question. There is also growing awareness that the `stock' is m ade up o f m ore or less discrete local p opulation s (cf. H ein ck e's conclusions about herring in the 19th century), and this has considerable implications for recovery program m es (Kent Sm edbol and Wroblewski, 2000). On the positive side the assessments are freely available and the uncertainties clearly set out; attem pts are being m ade to incorporate `lay' assessments o f stock strength m ade by fishers (DFO, 2000); a constructive debate has been initiated concerning the future o f the fisheries (Atlantic Fisheries Policy Review, 2000); and m anagem ent methods, such as no-take zones, that are not so dependent on the accuracy or theory of stock assessments are bein g evaluated (G uenette et al., 2000). 2.4.3. The human dimension As described so far, the dem ise o f the northern cod is already a rem arkable event, one that challenges the likely success of precautionary approaches, still advocated, based on setting a limit on calculated fishing mortality as part o f an intensive reliance on stock m odelling and prediction (see below). Yet there is m ore. Fishingfor truth (Finlayson, 1994), a rem arkable sociological study com pleted ju st before the final collapse, contains m any illum inating interviews with key participants, detailing the hum an d im en sion . According to the views quoted there, DFO scientists warned of the uncertainties at an early stage o f Canadian m anagem ent, but in the event not loudly enough (ibid., p. 136). Pleas then that the information dem anded (for long-term forecasts) was im possible to provide were overridden, and they apparently took seriously the threat that if they did not do the jo b , then economists would do it for them (ibid., p. 135). There was a double pressure to downplay the uncertainty; on one side the politicians pleaded for constancy and certainty to aid them resolve disputes (ibid., pp. 132-133, p. 142). The scientists -- believing that, given a range o f estimates, the highest catch would always be taken -- also tended towards lower (and they believed precautious) estimates, and with a greater assertion o f precision than their internal assessments suggested (ibid., p. 141). Ultimately this public overemphasis of confidence hoist them on their own petard, when even their m ost pessim istic view turned out to be based on underestim ates o f the historic level o f fishing mortality. Moreover the scientists were aware o f wider shortcomings regarding biological and physical param eters (ibid.), fundam ental sam pling problem s (ibid., pp. 73-74), and the often dubious nature o f scientific advice ham m ered out each year for the many stocks in a few weeks o f intensive m eetings (ibid., p. 79). Harris likened fisheries science to the Ptolem aic m odel o f the solar system (which placed Earth at the centre of the universe) where, when observations did not fit the theory, an additional layer o f complexity was added, rather than questioning the basic theory (ibid., p. 69). This accorded with Fisheries: taking stock 23 strong criticism, at that time, from related but separate disciplines, such as theoretical ecology (Peters, 1991). However none o f this happened because the participants were stupid, careless or lacking in intent to restore the stock. It was a systems failure that prevented fishers, scientists and politicians from responding to existing information or extracting themselves from the situation. goal o f optim um stock size), and at the same time, the action to be taken if they are exceeded' and `limit reference points (the negative goal o f the lowest acceptable stock size), and at the same time, the action to be taken if they are exceeded; when a limit reference poin t is approach ed m easures should be taken to ensure that it will n ot be exceeded'. The explicit reference to targets, limits and predeterm ination were novel d evelop m en ts. 2.5. Precaution becomes explicit By the early 1990s precaution had becom e explicit, due both to disasters such as northern cod, and to its increasing prom inence in other fields. The most significant global developm ent was the negotiation o f two related docum ents in 1995: the UN FAO Code o f Conduct for Responsible Fisheries (FAO, 1995) and the UN Agreem ent on Straddling Fish Stocks and Highly Migratory Fish Stocks (UN, 1995). The FAO Code applies a 'precautionary approach' to fisheries. The adoption o f this terminology initially arose from a nervousness am ongst the fishing industry and m anagers that the precautionary 'principle' was a concept captured by environmental NGOs, which might unjustifiably be used as a w eapon to substantially reduce or even halt fishing. Their use o f 'approach' is therefore probably best seen as an initial statement about ownership o f the process, rather than any close analysis, or fundam ental difference o f outcom e, arising from this choice of terminology. However, as described in the opening paragraph, there are advantages in maintaining a distinction between a universally applicable principle and the detailed approach to im plem entation, which will differ from field to field, although this was not a distinction m ade by those who negotiated the Code. Precaution in the Code covers uncertainties relating to individual stocks, other affected species, and environmental and socio econom ic conditions. However, the aspect em phasised is the m echanism for taking account o f uncertainties in stock assessment models, namely to determ ine `stock-specific target reference points (i.e. the `positive' The U N agreem ent provides details, notably that `the fishing mortality rate that generates m axim um sustainable yield (MSY) should be regarded as a minimum standard for limit reference points (i.e. the `negative' go al)'. It is a m inim um because the old con cept o f MSY (yet a widely used m ethod of calculation) is known to overestimate sustainable yields. The implications are profound, because many stocks are depleted well beyond form al MSY. It continues, `For stocks which are not overfished, fisheries m anagem ent strategies shall ensure that fishing mortality does not exceed that which corresponds to m axim um sustainable yield, and that the biomass does not fall below a predefined threshold'. For overfished stocks `the biom ass which would produce m aximum sustainable yield can serve as a rebuilding target'. The possible contradiction here may well be the result o f a negotiated com prom ise -- it can be a m istake to try to interpret such documents as the product of a single consistent mind. Nevertheless, overall the UN agreem ent and, indirectly, the FAO code emphasise precaution at the level o f maximising long term yields. 2.5.1. Subsequent action or inaction U S federal guidelines (R estrepo et a l, 1998) recognise that stocks should be m anaged in a m anner consistent with the U N agreem ent and FAO code (3). T he C anadian position (Richards an d Schnute, 2000) is that stocks must be equal to or greater than, and fishing mortality less than, that resulting from MSY; or for equivalent proxies to be used where MSY is inappropriate or inapplicable. Unlike the United States and Canada, fisheries responsibilities within the seas under EU M em ber State control are split. (3) The US position and technical guidance is of interest beyond fisheries. It requires the im plem entation of protective m easures 'even in the absence of scientific certainty that stocks are being exp lo ited '. Prior to this approach there was a 'perceived... inability to implement tim ely conservation measures without scientific proof of overfishing. Thus, the precautionary approach is essentially a reversal of the 'burden of proof'.' 24 Late lessons from early warnings: the precautionary principle 1896-2000 The International Council for the E xploration o f the Sea -- IC ES -- is responsible for technical advice, but the responsibility for m anagem ent, including the setting o f targets, resides with the European Commission and M ember States. This causes complications. Currently ICES advice centres arou n d Blim (biom ass lower limit) an d a larger `precautionary ap p roach ' target stock size, Bpa, at least accordin g to interpretations arising from within the FAO (Garcia, 2000). Two equivalent levels o f fishing mortality believed to achieve these reference points are also used: Fliman d Fpa. B limis set n ot to MSY but to a level where, if depleted further, a stock `is in im m ediate dan ger o f co llapse' (Garcia, 2000, p. 22). Bpa is set inconsistently, depending on the state o f the stock. For those in a poor state, such as North Sea cod, haddock and plaice, it is set at or close to the m inim um biologically acceptable level (M BA L), below which there is ju d g e d to be unacceptable risk that it could reach Blim. Unfortunately many stocks are ju d ged by IC ES to be below Bpa, even close to B lim (cf. ICES advice on North Sea cod in 2000 (ACFM ICES, 2000a)). This emphasis and interpretation of reference points has been seen as conflicting with the FA O /U N approach, resulting in peer-group criticism from within the FAO (Garcia, 2000, p. 23), the U nited States (R estrepo et a l, 1998, p. 24) and Canada (Richards and Schnute, 2000, p. 7). In effect the earlier m anagem ent regime (a goal o f maintaining stocks above MBAL) often remained unchanged, although the language has been given a precautionary gloss. ICES may concur, noting that this use o f limit reference points `is a needlessly restricted interpretation o f a concept' and -- obliquely -- that `the adoption of precautionary referen ce points requires discussion with fisheries m anagem ent agencies' (ACFM ICES, 2000b, p. 55). This illustrates the tensions that arise from split responsibility. IC ES itself does n ot refer to Bpa as a `target' reference point, but as a `buffer' or a `precautionary reference poin t' (ACFM ICES, 2000b, p. 2), presumably because setting a `target' is seen as im pinging on the responsibilities o f the Commission and M ember States. The 1995 criticism by the Com m ission's Scientific, Technical and Economic Committee (European Commission, 1995) o f a m anagem ent system `at ease with crisis m anagem ent or an unwillingness or inability to state specifically a m ore positive goal aim ed at enhancing the productivity o f fisheries in either a biological, econom ic or social sense' still appears valid. As yet M ember States appear unwilling, or unable, to fund the short-term investment involved in restoring stocks to optim al levels. The European Commission 2000 comm unication on the application o f the precautionary principle (European Commission, 2000) argues that the reason why ICES does not refer to MSY is twofold: namely that for `a num ber o f stocks' the conditions whereby sustainable yields can be obtained are `difficult, if not im possible' to define, and that fishing pressure on many EU stocks is `well above that which would correspond to the maximising of yield'. A likely respon se is that establishing m axim um yields is no m ore or less certain than establishing the level at which stocks will collapse, while high fishing pressure is the reason why the code is necessary, n ot a reason why it should n ot be applied. The relationship between EU policy and the FAO code and UN agreem ent certainly has the potential to becom e a highly contentious issue. However the Commission 2001 Green Paper on the future o f the Com m on Fisheries Policy (European Com mission, 2001a), while general in scope, contains a frank identification o f the problem s as well providing a constructive basis for discussing the way forward. C oncentrating on the future rather than dwelling overmuch on the past may be the m ost constructive way forward. Turning back to the code and agreem ent in general, one problem remains unchanged: the quality o f the data. For Canadian northern cod the underestim ation o f a vital statistic, fishing mortality, was a key factor. Yet a retrospective analysis in 1999 o f fishing mortality rates o f m ajor North Sea stocks for which the best data were available (cod, haddock, whiting, plaice and sole) found the same fundam ental problem s (van Beek and Pastoors, 1999; ACFM ICES, 1999, p.12). For cod, h addock an d whiting m ortality was much higher than originally stated, for plaice there was no correlation, an d for sole there was a possible negative correlation. This too noted that underestimating fishing mortality is associated with an overestim ate o f stock size. M oreover this problem had been identified as early as 1977. The parallels are startling. One conclusion drawn is that Northeast Atlantic cod can survive higher levels o f mortality that the northern cod, but this seem s to be based on the circular Fisheries: taking stock 25 argum ent that they have not yet crashed. Northern cod also survived periods o f intense fishing pressure beyond that prevalent at the time of the crash. Indeed the goal that the Canadians adopted after taking over control o f northern cod m anagem ent was identical to that now advocated under the precautionary approach: to severely curtail fishing effort and to build up the stock to a level above that predicted for optim um econom ic yields. Yet still it crashed. So it is legitim ate to ask whether the change -- as yet -- is sufficient. 2.6. The ecosystem approach Another criticism o f the U N /FA O approach is that while it nominally requires precaution at the ecosystem level, the practical emphasis remains on single stock management. Moreover, o n e 's value ju d g e m e n t m ight be to target precaution not at avoiding stock collapses (cf. IC E S/E U ), or even ensuring that stocks are maintained above MSY (UN / FAO), but at preventing adverse effects on other species dependent on the fish. G reenpeace was an early advocate o f this in its 1994 precautionary approach (Earl, 1994), although since then others have also argued that catches, in order to be sustainable, should be reduced the same order of m agnitude as other predators (Fowler, 1999). There is thus no one `correct' precautionary approach; it depends on the objectives set (MacGarvin, 2001b). Even between different fisheries in the sam e area, ju d ge m en ts will necessarily differ as to the costs and benefits, and the acceptable risks, o f different elements o f a precautionary response, dependent on how these im pinge on activities and interests. Indeed, stocks cannot be treated in isolation from one another, because many o f them are m ajor predators on each other (Swain et a l, 2000). Fishing one stock has implications for the others. An attem pt is som etim es m ade to allow for this in the single species models, but it falls far short o f the evident complexity. Indeed the gulf between the disciplines o f fisheries science and theoretical or community ecology is remarkable. In the early 1990s prom inent ecologists commented that fisheries m anagem ent was, for exam ple, a field `so accustom ed to inaccuracy in its basic m odels that striking differences between m odel and observation are scarcely noted... Nevertheless fisheries biologists fit data to m odels that are clearly inaccurate and make decisions on that basis'. Nevertheless there are signs o f change, with a new em phasis by regulators on the need to adopt an `ecosystem approach '. Notably in the United States (Ecosystems Principles Advisory Panel, 1998) bu t also within IC E S's Advisory Committee on the Marine Environment (ACME ICES, 2000) and in C an ada (M urphy an d O 'Boyle, 2000) there is an increasing involvement of theoretical and community ecologists and concepts. The US report (Ecosystems Principles Advisory Panel, 1998) notes the role o f chaotic population dynamics that may make systems fundamentally unpredictable. Indeed an earlier study o f interactions between fish species on the Canadian Grand Banks suggests that the m ore realism that is incorporated into the models, the more unpredictable becom e the effects of a change in any one species -- in effect the system may work as a gigantic random num ber generator (Gomes, 1993). However, the U S report also em phasises that we do know that ecosystems have limits which, when exceeded, can result in irreversible changes; that diversity is important; that systems operate at multiple scales; and that boundaries, to the great inconvenience o f m anagers, are indistinct. `T h ere is simply n ot enough money, time or talent to develop a synthetic and completely inform ed view o f how fisheries operate in an ecosystem context. T h ere will always be unm easured entities, random effects, and substantial uncertainties, but these are not acceptable excuses to delay im plem enting an ecosystem-based m anagem ent strategy.' Similar conclusions have been drawn by European experts (Daan, 1998). For exam ple, we have been unable to fill a crucial gap identified as long ago as 1914 (Hjort, 1914): that o f being able to predict, from the num ber of eggs laid, the num ber of fish that will subsequently m ature to ap p ear in the stocks. However, according to the US approach, we know enough about ecosystem functioning to do a better jo b of m anagem ent than in the past. Whilst in 1919 a Californian fisheries regulator required that `p ro o f that seeks to change the ways o f commerce and sport must be overwhelming' (Thompson, 1919), now the burden of proof is on fisheries to take accoun t o f uncertainty regarding ecosystem effects. At an operational level stakeholder involvement is essential. What m ight be called a `second 26 Late lessons from early warnings: the precautionary principle 1896-2000 generation' precautionary approach, less centred on stock assessment models (even with their precautionary attem pts to incorporate error), and including errorresilient concepts such as no-take zones, appears to be producing results for finfish and shellfish on the US Georges Bank (Murawski et al., 2000). The Canadian approach in particular highlights the active involvement o f fishers, and the stock assessments now attem pt to incorporate their knowledge (for example, DFO, 2000), while in northern Europe considerable attention has been devoted to the policy implications o f ecosystem m anagem ent (Norwegian Ministry o f the Environment, 1997; Nordic Council of Ministers, 1998). While at an early stage, and as yet without overall consensus, taken together they provide the elements o f a new approach, incorporating precaution, which has considerable potential if implemented. 2.7. Late lessons Fisheries provide a rich seam of lessons regarding the precautionary approach, of wider interest than to fisheries alone. These include: avoiding dom inance by any one discipline or sub-discipline (the general dominance o f stock modellers); accounting for `real-world' conditions (underestim ation o f real fishing mortality and technologies); taking full account o f the pros and cons of any one approach (stock assessm ent vs. wider approaches); using lay knowledge (Native American, Scottish 19th century fishers, Canadian inshore fishers); taking account o f wider social perspectives, acknowledging the im portance of value judgem ents and evaluating all the options available (different interpretations of precautionary approach); avoiding reliance on ever more elaborate m odels to explain away predictive failures (Harris, Ptolemaic astronomers and analytical fisheries science); dealing with institutional obstacles and regulatory independence (reluctance to address fundam ental economic issues, blurred independence o f technical advisers and policy-makers from Californian sardines to present); m aintaining due humility (`biased pseudoscience' 1986 response o f DFO to criticism ). the distinction between the precautionary principle and precautionary approaches (logical distinction between a simple 'principle' and differing practical im plem entation in different fields; also a 'political' distinction as in the FAO C ode); appropriate levels o f proo f (19th century Scottish fishery, 1920s Californian state scientists, 1990s US ecosystem approach); distinguishing between uncertainty and ignorance (Harris Report, uncertainties attached to estimates, ignorance regarding ecology); unrealistic expectations (or incredible claims) as to the `soundness' of scientific conclusions (specifically northern cod, but a general feature); drawing upon historical knowledge (H eincke's im portance o f sub-stocks, Scottish natural no-take zones, success of areas protected from trawling); not brushing `blind spots' under the carpet (Harris Report, rose-tinted spectacles); Positive changes o f attitude are occurring. The question is whether they are happening fast enough to stave off further collapses. While there will be som e interest groups who see it as in their interest to pursue a short term strategy, the n eed for precaution is generally not som ething that should need to be laboured within the fishing industry. Not cutting stocks so `close to the bon e' not only takes us into areas o f greater certainty regardin g the m aintenance o f stocks, but will also greatly increase the economic returns (W hitmarsh et al., 2000). T he problem is that the natural capital has been run down so far that industry in many cases cannot absorb the short-term hit necessary to rebuild the stocks. Interest groups need to redirect their attention from arguin g with each other and instead learn the language of, and engage with, the econom ic ministries who -- not surprisingly given the history of inappropriate subsidy -- have a highly jau n d iced view o f the merits o f investing in recovery programmes. Fisheries: taking stock 27 Fisheries: early warnings and actions Early fisheries in the UK 1376-77 Setting up a com m ittee is the answer of the English parliam ent to a call for precautionary actions in fishing by controlling the mesh of the nets 1866-93 Following uncertainty about the consequences of fisheries growth official enquiries take place, but no action is taken Sardine fisheries in California Mid-1920s Californian sardine fisheries scientists call for precaution and research 1942 Continuous inaction leads to the collapse of the sardine stock (signs of recovery only in mid1980s) Northern cod fisheries in Canada Late 1970s Canada starts managing the northern cod fisheries up to 200 nautical m iles, claiming it is doing so precautiously 1986 Keats Report commissioned by inshore fishers indicates severe underestimation of the fishing pressure 1988 Alverson Report, prepared by fishery scientists, states that consistent overestim ation of the stock size leads to overfishing 1989 A new governm ent assessm ent recomm ends that the offshore catch should be halved 1990 A new independent assessm ent, the Harris Report, confirms overfishing 1992 The stock collapses and a moratorium is imposed 1999 Restart of the northern cod fisheries at low levels, but dissidents state it is not low enough General fisheries 1990s The ecosystem approach slow ly settles in fisheries m anagem ent procedures 1995 FA O Code of Conduct for Responsible Fisheries and the UN A greem ent on Straddling Fish Stocks and Highly M igratory Fish Stocks are negotiated and published 2001 European Commission Green Paper on the future of the Common Fisheries Policy 2001 Positive changes are occurring, but are they happening fast enough to avoid further collapses? 2.8. References ACFM ICES, 1999. Report of the working group on the assessment of dermersal stocks in the North Sea Part 1, IC ES CM 2000/A CFM :7, Advisory Committee on Fishery M anagement, International Council for the Exploration of the Sea, Copenhagen. ACFM ICES, 2000a. Cod in sub-area IV (North Sea), division Vlld (Eastern Channel) and division IIIa (Skagerrak), Advisory Com m ittee on Fishery M anagement, International Council for the Exploration o f the Sea, Copenhagen, at http://www.ices.dk/ com m itte/acfm /com w ork/report/2000/ O ct/cod -3 4 7 d .p d f (accessed 15 April 2001). ACME ICES, 2000. Report of the working group on ecosystem effects offishing activities, IC ES CM 2000/ACM E:2 Ref.:ACFM+E, Advisory Committee on the Marine Environment, International Council for the Exploration of the Sea, Copenhagen. Atlantic Fisheries Policy Review, 2000. The management offisheries on C an ada's Atlantic coast: A discussion document on policy direction and principles, Atlantic Fisheries Policy Review, Ottawa, at http://w w w.dfo-m po.gc.ca/afprrppa/lin ksto_discodoc_e.h tm (accessed 19 March 2001). Bertram , J . G., 1865. The harvest of the sea, Jo h n Murray, London. ACFM ICES, 2000b. Report of the CWP intersessional meeting: Working group on precautionary approach terminology and CWP sub group on the publication of integrated catch statistics for the Atlantic, Advisory Com m ittee on Fishery M anagement, International Council for the Exploration o f the Sea, Copenhagen, p. 55, at http://w w w.ices.dk/ (accessed 25 April 2000). Daan, N., 1998. `Structure and dynamics o f the N orth Sea ecosystem ' in Workshop on the ecosystem approach to the management and protection of the North Sea, Oslo, Norway 15-17 June 1998, Tem aNord 1998:579, pp. 56-59, Nordic Council of Ministers, Copenhagen. Day, D., 1995. `Tending the A chilles' heel o f NAFO: Canada acts to protect the nose and tail o f the G rand B an k s', Marine Policy Vol. 19, pp. 257-270. Table 2.1. 28 Late lessons from early warnings: the precautionary principle 1896-2000 DFO , 2000. Northern (2J3KL) cod, DFO Science Stock Status Report A2-10, Departm ent of Fisheries and Oceans, Ontario, at http://www.dfo-mpo.gc.ca/ csas/ csas/status/2001/a3-01e.pdf (accessed 24 March 2001). Earl, M., 1994. A precautionary approach to the management offisheries in the North Atlantic, Greenpeace International, Amsterdam. Ecosystems Principles Advisory Panel, 1998. Ecosystems-basedfisheries management: A report to Congress by the Ecosystems Principles Advisory Panel, Ecosystems Principles Advisory Panel, Seattle, 54 pp. European Commission, 1995. `Second report o f the Scientific, Technical and Econom ic Com m ittee for Fish eries', E C 's Staff Working Paper SEC(95)2160, European Commission, Brussels. European Commission, 2000. Com munication from the Commission to the Council and European Parliament: Application o f the precautionary principle and multiannual arrangem ents for setting TACs, 01.12.2000, COM (2000) 803 final, European Commission, Brussels. European Commission, 2001a. Green Paper on the future o f the Com m on Fisheries Policy Volume 1, 20.3.2001, C O M (2001) 135, European Commission, Brussels. European Commission, 2001b. Green Paper on the future o f the Com m on Fisheries Policy Volume 2: R ep ort on the econom ic and social situation of coastal communities, CO M (2001) 135, European Commission, Brussels. FAO, 1995. Code of conduct for responsible fisheries, Food an d Agriculture O rganization o f the United Nations, at h ttp :// www.fao.org/fi/ag reem /co d eco n d / ficonde.asp#7 (accessed 25 April 2000). Finlayson, A. C., 1994. Fishingfor truth: A sociological analysis of northern cod stock assessments from 1977 to 1990, Social and Econ om ic Studies Vol. 52, Institute o f Social and Economic Research, Memorial University o f Newfoundland, St Johns, 176 pages. Fowler, C. W., 1999. `M an agem en t o f m ulti species fisheries: From overfishing to sustainability', ICESJournal ofMarine Science Vol. 56, pp. 927-932. Garcia, S. M., 2000. `The precautionary approach to fisheries 1995-2000: Progress review and m ain issues', Report of the CWP intersessional meeting: working group on precautionary approach terminology and CWP sub group on the publication of integrated catch statistics for the Atlantic, Advisory Com m ittee on Fishery M anagement, International Council for the Exploration o f the Sea, Copenhagen, p.22, at O :\A CFM \W G REPS\Cw p\Reports\2000\CW P 2000.Doc Gom es, M do C., 1993. Predictions under uncertainty: Fish assemblages and food webs on the Grand Banks ofNewfoundland, ISER, M em orial University, St Joh n s, 205 pages. G unette, S., Pitcher, T. J . an d Walters, C. J., 2000. `The potential o f marine reserves for the m anagem ent o f northern cod in N ew foundland', Bulletin ofMarine Science Vol. 66, p p .831-852. H arris, L., 1990. Independent review of the state of the northern cod stock, C om m unications Directorate, Departm ent of Fisheries and Oceans, Ontario. Hjort, J., 1914. `Fluctuations in the great fisheries o f northern E u ro p e ', Rapports et Procs-verbaux des Runions, Conseil international pour l 'exploration da la Mer Vol. 20, pp. 1-228. Huxley, T. H., 1881. `T he h errin g', Nature Vol. 23, pp. 607-613, cited in Sinclair, M., 1988, Marine populations: An essay on population regulation and speciation, Books in Recruitment Fishery Oceanography, W ashington Sea Grant Program , University o f Washington Press, Seattle. Huxley, T. H., 1883. `In augural ad d ress', The Fisheries Exhibition Literature, International Fisheries Exhibition, London Vol. 4, pp. 1-22, cited in Ecosystems Principles Advisory Panel, 1998, Ecosystems-basedfisheries management: A report to Congress by the Ecosystems Principles Advisory Panel, Ecosystems Principles Advisory Panel, Seattle. Keats, D., Steele, D. H. an d G reen, J . M., 1986. A review of the recent status of the northern cod stock (NAFO divisions 2J, 3K, and 3L) and the declining inshorefishery, R eport to the Newfoundland Inshore Fisheries Association, Fisheries: taking stock 29 cited in Finlayson, A. C., 1994, Fishingfor truth: A sociological analysis of northern cod stock assessments from 1977 to 1990, Social and Econ om ic Studies Vol. 52, Institute o f Social and Economic Research, Memorial University o f Newfoundland, St Johns. Kent Sm edbol, R. and Wroblewski, J. S., 2000. Metapopulation theory and northern cod population structure: Interdependency of subpopulations in recovery of a groundfish population, Canadian Stock Assessm ent Secretariat Document 2000/087, at h ttp :// w w w.dfo-m po.gc.ca/csas/csas/English/ R esea rch _ Y e a rs/2 0 0 0 /2 0 0 0 _ 0 8 7 e.h tm (accessed 24 March 2001). M acGarvin, M., 2001a. A comparison of thefate of Canadian and UK whitefish fisheries, m odus vivendi / WWF-UK, Glenlivet / Godalm ing, UK. M acGarvin, M., 2001b. `Precaution, science, facts an d values' in O 'R iordan et al. (eds), Reinterpreting the precautionary principle, C am eron & May, L on don . M acGarvin, M. an d Jo n e s, S., 2000. Choose or lose: A recovery plan forfish stocks and the UK fishing industry, WWF-UK, G odalm ing, at http://www.wwf-uk.org/ orca/info.htm (accessed 15 M arch 2001). M arch, E. J., 1953. Sailing trawlers: The story of deep-seafishing with long-line and trawl, Percival M arshall & Co., London. McEvoy, A., 1986. Thefisherman's problem: Ecology and law in the Californian fisheries 1850 1980, Cam bridge University Press, C am bridge. Murawski, S. A., Brown, R., Lai, H.-L., Rago, P. J . and H endrickson, L ., 2000. `Large-scale closed areas as a fisheries-management tool in tem perate m arine systems: The Georges B ank exp erien ce', Bulletin ofMarine Science Vol. 66, pp. 775-798. Murphy, O. and O 'Boyle, R., 2000. Proceedings of the ecosystem approaches tofisheries management workshop, 31 Auggust-2 September 1999, Canadian Stock Assessment Proceedings Series No 99/38, Fisheries and Oceans Canada, Maritimes Region, Dartmouth. N ordic Council o f Ministers, 1998. Workshop on the ecosystem approach to the management and protection of the North Sea, Oslo, Norway 15-17 June 1998, Tem aNord 1998:579, Nordic Council of Ministers, Copenhagen. Norwegian Ministry o f the Environment, 1997. Statement of conclusions from the intermediate ministerial meeting on the integration offisheries and environmental issues: Intermediate ministerial meeting on the integration offisheries and environmental issues, Bergen, 13-14 March 1997, Norwegian Ministry o f the Environment, Oslo, at h ttp :// www.odin.dep.no/nsc/m eeting1997/ conclusions.shtml (accessed 2 May 2000). O 'Reilly H inds, L., 1995. `Crisis in C an ad a's Atlantic sea fish eries', Marine Policy Vol. 19, p p .271-283. Pauly, D., Christensen, V., D alsgaard, J., Froese, R. an d Torres Jr, F., 1998. `Fishing down m arine fo od webs', Science Vol. 279, pp. 860-863. Peters, R. H., 1991. A critiquefor ecology, Cam bridge University Press, Cam bridge. Report, 1866. Report of the commissioners appointed to inquire into the sea fisheries of the United Kingdom, presen ted to both H ouses o f Parliament, London. Report, 1885. Report of the commissioners appointed to inquire and report upon the complaints that have been made by line and drift netfishermen of injuries sustained by them in their calling owing to the use of the trawl net and beam trawl in the territorial waters of the United Kingdom, presen ted to both H ouses o f Parliament, London. Restrepo, V. R. et al., 1998. Technical guidance on the use of the precautionary approaches to implementingNational Standard 1 of the Magnuson-Stevens Fisheries Conservation and Management Act, NOAA Technical M em orandum NM FS-F/SPO, 17 July. Richards, L. J. an d Schnute, J . T., 2000. `Science strategic project on the precautionary approach in C an ada' in H aigh, R. an d Sinclair, C. ( eds), Proceedings of the second workshop, pp. 59-60, C anadian Stock Assessment Proceedings No 99/41, Fisheries and Oceans Canada. San Diego Natural History Museum, 2000. Ocean Oasis Field Guide, Sardinops sagax caerulea, California pilchard, Pacific sardine, Sardina Monterrey, at h t t p :// 30 Late lessons from early warnings: the precautionary principle 1896-2000 w w w.oceanoasis.org/fieldguide/sardcae.html (accessed 5 May 2001). Scottish Office, 1898 to present. Annual Report of the Fisheries Boardfor Scotlandfor 1898, Scottish Office, Edinburgh, et seq. Sutherland, I., n.d. From herring to seine net fishing on the east coast of Scotland, Cam ps Bookshop, Wick (ISBN 0 9508697 2 4), 231 pages. Swain, D. P., Sinclair, A. F., C houinard, G. A. an d Drinkwater, K. F., 2000. Ecosystem effects on pre-recruit survival of cod in the southern G u lf of St. Lawrence, C anadian Stock A ssessm ent Secretariat Document 2000/147, at h ttp :// www.dfo-mpo.gc.ca/csas/csas/E n glish / R esea rch _ Y ears/2 0 0 0 /2 0 0 0 _ 1 4 7 E .h tm (accessed 15 M arch 2001). T hom pson , W. F., 1919. `T he scientific investigation o f m arine fisheries, as related to the work o f the Fish and Game Commission in Southern C alifornia', Fisheries Bulletin (C an ada) Vol. 2, pp. 3-27. UN, 1982. UN Convention on the Law o f the Sea, United Nations, at go p h er:// gopher.un.org: 70/11/L O S/U N C L O S82 (accessed 25 April 2000). UN, 1995 Draft Agreem ent for the Implem entation o f the Provisions o f the United Nations Convention on the Law o f the Sea o f 10 D ecem ber 1982 relating to the Conservation and M anagement o f Straddling Fish Stocks and Highly Migratory Fish Stocks United Nations,accessable via h ttp :// w w w.igc.org/habitat/un-proc/ Accessed 13th Septem ber 2001. van Beek, F. A. an d Pastoors, M. A., 1999. `Evaluating ICES catch forecasts: the relationship between implied and realised fishing m ortality', Theme session: Fishing capacity, effort and mortality, IC ES C.M. 1 9 9 9 / R:04, International Council for the Exploration o f the Sea, Copenhagen. Whitmarsh, D., Jam es, C., Pickering, H. and Neiland, A., 2000. `The profitability of m arine com m ercial fisheries: A review o f econom ic inform ation needs with particular reference to the U K ', Marine Policy Vol. 24, pp. 257-263. Radiation: early warnings; late effects 31 3. Radiation: early warnings; late effects Barrie Lambert Since the discovery o f ionising radiation ju st over 100 years ago, it has been appreciated that injudicious exposure could produce harm ful effects, even death. However, the general excitem ent in the scientific community and the, often inappropriate, publicity which followed these discoveries ensured that the dam age to health, particularly in the long term, was not given any prominence. The undoubted medical diagnostic and therapeutic value o f X-rays and radioisotopes m eant that caution tended to be thrown away and it was several decades before control over exposure of the public and workers was put in place. This control has slowly evolved as m ore know ledge o f the processes of interaction between radiation and biological tissue has accrued, but it has often lagged behind clear evidence o f effect. Changes over the last 70 years to recommendations governing radiation exposure have usually been m ore restrictive; even over the last 20 years we have perceived that the risk is about four to five times larger than hitherto thought. However, the controls have n ot always m an aged to adequately balance risk and benefit. To understand and analyse the development o f radiation protection it is necessary to go back m ore than 100 years. 3.1. X-rays W ilhelm C on rad R oentgen is credited with the discovery o f X-rays at the University of W urtzburg in 1895 bu t there is reasonable evidence that a num ber o f other physicists, notably G oodspeed in 1890, had produced similar penetrating radiations without appreciating their significance. Roentgen was the first to publish an account o f the production o f X-rays (Roentgen, 1895) and immediately recognised their value to m edical diagnosis -- in fact he publicised his work by sen din g an X-ray o f his wife's h an d to prom inent scientists. There was an im m ediate worldwide interest in this tool for m edical diagnosis, and because o f the cavalier approach generally adopted by physicians it was inevitable that X-ray injuries w ould soon appear. T he scientific an d lay worlds were captivated by the new phenom enon, which could penetrate hum an tissues and reveal bony structures. Despite the occasional ominous warning to the contrary (Thom pson, 1898), the general consensus was that X-rays, used judiciously, were without adverse effects. Simplistically it was assum ed that harm could not result from an agent that could not be appreciated by the senses -- ironically, this is the very reason given, nowadays, for the unreasonable fear o f radiation. There were reports o f injuries as early as 1896 -- for instance, Thom as Edison, Tesla and Grubbe noted eye and skin injuries and the former, particularly, cautioned about excessive exposure to X-rays (Edison, 1896). Unfortunately, this was too late for E d iso n 's assistant, C larence Dally, who suffered from severe radioderm atitis resulting in the amputation of his arm and his subsequent death in 1904. By the late 1890s there were num erous reports o f radiation skin burns and loss of hair (epilation) in the scientific literature bearing testimony to the apparent cavalier attitudes and the size o f the doses which were being experienced. One o f the m ore absurd actions was that o f the wellknown American physicist Elihu Thomson, who purposely exposed the little finger o f his left hand to the direct beam o f an X-ray tube over a period o f several days. The inevitable severe dam age to his finger m ade him caution against overexposure `...or there may be cause for regret when too late' (Thomson, 1896). Ironically, because o f the increasing n um ber o f reports o f radiation injury, som e physicians recognised the possible therapeutic value of the rays and the first `treatm ent' was reported in 1896 (Stone, 1946) when a woman with advanced carcinom a o f the left breast was treated in Chicago. At about this time there was som e pressure in the m edia. Jo h n Dennis, a New Yorkjournalist, who could possibly be considered the first radiation `whistle blower', cam paigned for controls on radiologists and radiographers by licence issued by the state, and suggested that injury to a patient was a criminal act (Dennis, 1899). It was many decades before this was acted upon. 32 Late lessons from early warnings: the precautionary principle 1896-2000 Despite the reports and cautions o f adverse effects, and even the use o f X-rays for therapy, a degree o f overconfidence existed in the m edical fraternity regarding the use of X-rays. The theories were that the effects were not caused by the interactions o f the Xrays themselves but by static electricity or individual sensitivity -- there was even total denial o f the existence o f X-ray effects (Scott, 1897). Perhaps the first and m ost im portant person in the role o f radiation protection pioneer was a Boston dentist, William Rollins. Rollins, who was a Harvard graduate in dentistry and m edicine, was the first to suggest a `tolerance' dose or exposure for X-rays and also the first to recom m en d ways o f shielding and collimating X-ray tubes. His criterion or standard was to expose a photographic plate outside the tube; if the plate did not fog in 7 m inutes the shielding was adequate. During the period 1900-1904 he published more than 200 papers urging physicians to use the minimum possible exposure, and he made many suggestions on how to reduce the exposure of both radiologist and patient (Rollins, 1904). This latter point has only recently been addressed again by the United K in gd om 's N ational R adiological Protection Board (NRPB, which advises the governm ent). Rollins also recognised the potential for the induction of cataract and he carried out some animal experiments that showed, am ongst other effects, the possibility o f acute (teratological) damage to the foetus. H e was the first to warn o f the risks in the Xray o f women for the diagnosis of pregnancy (pelvimetry). However, his cautions often went unheeded -- his warnings about pelvim etry were only revisited, albeit with reference to late effects, by Alice Stewart, the British epidemiologist, about 40 years later (Stewart et al., 1958). It is noteworthy that her work was also at first rejected by the medical authorities (see below). It would be reasonable to assum e that, once it had been identified that excessive X-ray exposure produced tissue dam age, care would be exercised. This was not so -- even in 1903, Albers-Schonberg (1903) who produced a set o f rules for the use of radiologists in protecting themselves, suggested that the regularly used technique o f testing the `hardness' o f the X-ray tube by placing the hand between the tube and fluorescent screen was dangerous. It m ust be said that, in this period, the absence o f an agreed unit o f radiation exposure or dose was a problem to those who wanted to establish effective standards o f protection -- this had to wait until the adoption o f the roentgen as the unit o f exposure in 1928. However, notwithstanding the absence o f units, the first set of published rules of radiation protection was produced by the Germ an Radiological Society in 1913 (Taylor, 1979). Shortly after this, also in 1913, C oolidge invented the hot-cathode, tungsten-target Xray tube which contributed im m easurably to lower doses to both patients and radiographers (for instance, with early tubes operating at low voltages, exposures o f m ore than an hour were com m on). 3.2. Radioactivity and radioactive materials Unfortunately, another hazard became ap p aren t very soon after R oen tgen 's work because within weeks Henri Becquerel had discovered radioactivity and then, in 1898, Marie and Pierre Curie reported the discovery o f radium. The extent o f the hazard o f radioactivity was no m ore recognised than that o f X-rays and both Becquerel and Pierre Curie suffered skin erythemas from carrying sam ples of radioactive materials in their pockets. Although it was soon realised that radium could be used therapeutically, for example for killing m alignant cells, for som e reason the public becam e besotted with the idea o f radium (and radium-emanation, radon) being a general panacea. The precautionary principle seem ed to act even slower in this field because not until about 1920 was it realised that controls were necessary. This was initiated at least partly by the use o f radium in lum inous paint that was used extensively in the First World War. The radium-activated paint was applied by brush and the painters, mostly young women (in New Jersey and elsewhere), found they could work faster and earn m ore by tipping their brushes with their lips -- in this way they ingested considerable quantities of radium. Very little attention was paid to industrial hygiene and the workers were irradiated internally from the radium they had taken in, externally from accumulations of the paint which contam inated their workplaces and from the inhalation of radon. The hazard of this work was not at first recognised but in 1924 a New York dentist, Theodore Blum, published a paper that identified a new disease which he called `radium jaw ' (som etim es known as `phossy jaw ') which he Radiation: early warnings; late effects 33 had seen in his patients who were ex-dial painters. He attributed the condition to the toxicity o f phosphorus. However, a local New Jersey pathologist, Harrison Martland, recognised the bone lesions as being caused by radium and in 1925 started a study that unearthed the whole sorry story (Martland and Humphries, 1929). The first bone sarcom a was recorded in this group o f women in 1923 and there have been 55 such cancers studied in a population o f nearly 3 000 w om en (Rowland et al., 1983) -- altogether about one third have died of various m alignancies (including leukaemia and breast cancer). The data derived from the experiences o f these women did eventually set the standard for the intake o f radioactive m aterial for many years -- this was the so-called radium standard. This standard was set at the am ount o f radium in the body which apparently produced no effect. There was then an assum ption o f a threshold for the effects but this was in accord with the gen eral attitude in the years up to 1930 o f setting a `tolerance' dose. The radium standard was set at a level o f radioactivity o f 0.1 m icrocuries (3.7 kilobecquerels) o f radium which would deliver a radiation dose o f 150 millisieverts to the bone. In a som ew hat bizarre way, in the 1920s radium was looked upon as a source o f health and healing. This cumulated in many potions being sold containing radium, the most fam ous being Radiothor. Four hundred thousand bottles o f this quack nostrum, which was said to cure a range o f maladies from stom ach ulcers to im potence, were sold between 1925 and 1930. The more dangerous aspects o f this were highlighted when a fam ous US golfer, industrialist and millionaire, Eben Byers, died from radiationinduced disease after consuming about 1 000 bottles over a long period o f time (Macklis, 1993). This case did quite a lot to encourage a m ore restrictive approach to the use of radium , as did the death o f Marie Curie in 1934 from (probably) aplastic anaem ia (which was at the time attributed to the effects of radium ). Nevertheless, such use of radium and radon has lasted to this day in the form of, for instance, `em antoria' where radon is breathed for (presum ed) beneficial purposes, for exam ple in Salzburg. 3.3. Early moves towards control of exposure During the 1920s the concept o f radiation dose was not defined but there were a num ber of reports aim ed at restricting exposure. These often quoted a level which could be `tolerated'. One o f these was a fraction o f the skin erythema dose (SED) which was suggested by an Am erican physicist, Arthur M utscheller (Mutscheller, 1925). His proposal was one hundredth o f the SED in a month. This would be very roughly equivalent to an annual dose limit of about 700 millisieverts (the contem porary dose limit for workers is now 20 millisieverts per year). It is noteworthy that during this period the em phasis was on limits driven by a desire to control the im m ediate effects of radiation. There seem ed to be no realisation that cancer would follow after a long lag, or latent, period. Clearly there was som e pressure from within part o f the scientific community for control o f the use o f radiation an d with the establishm ent o f the International X-ray and Radium Protection Committee (IXRPC) at the second International Congress of Radiology in 1928 the setting o f standards becam e m ore regularised. However, comments on the misuse o f radiation were often missing. For som e reason in the early pontifications o f the IXRPC em phasis was placed on leisure activities for radiographers (Desjardins, 1923). For example, `The cultivation o f an outdoor hobby is o f special im portance to all persons exposed to radiation.' The IXRPC eventually m etam orphosed into the International Committee on Radiological Protection (ICRP) but again it was som e time before this committee began to recom m end dose limits without connotation o f a dose threshold. All this was against a background o f reports of more than 200 radiologists who had died of what were thought to be radiation-induced m alignant diseases (Colwell and Russ, 1934), particularly the pioneering British radiologist Ironside-Bruce in March 1921. As a result of his death there were several press articles commenting on the adequacy of shielding of X-ray tubes which prom pted the Roentgen Society ( `Editorial', 1921) to state `the scientific com petency o f the press is less than its ability to write lurid jo u rn alise'. 34 Late lessons from early warnings: the precautionary principle 1896-2000 3.4. The post-war watershed: justification, optimisation, limitation The essential change in radiation protection philosophy came at a m eeting in Canada in 1949 (NBS, 1954) when it was concluded that, `there may be som e degree o f risk at any level o f exposure' and `the risk to the individual is n ot precisely determ inable but however small is believed not to be zero'. The additional philosophy from this m eeting, which is o f great im portance, was that `radiation exposures from whatever sources should be as low as practicable'. This is what is now known as the optim isation principle. T he risk vs. benefit (justification) principle, which is probably unique to radiation as a pollutant, was also introduced. The ICRP was set up to do no m ore than make `recom m endations' which, presumably, could have been accepted or rejected by national governm ents, but from its inception its role has been criticised. For exam ple, it took no stand about the testing o f nuclear weapons in the atm osphere which produced worldwide fallout. In addition it has generally been the work o f individuals, rather than the ICRP, which has prevented the misuse o f radiation. There are num erous examples of the ill-conceived use o f radiation, including: The widespread use o f Pedascopes for fitting children's shoes. These X-ray fluoroscopy devices were in nearly every shoe store in the 1940s and 1950s and could produce reported dose rates of 1 roentgen per minute. They did no more than keep children am used whilst their parents selected shoes and thus the radiation doses received by children and shop staff were totally unnecessary. Children who had ringworm were treated with X-rays to p rodu ce epilation bu t many subsequently developed cancer (see, for exam ple, R on et al., 1989). M ental patients were `treated' with radium in the 1930s. X-rays were used for the removal of unwanted hair in beauty shops in the 1930s and 1940s. These misuses o f radiation were largely uncontrolled because there were, at that time, no specific legal regulations governing radiation safety, only recom m endations. In the United Kingdom legal regulation was first encapsulated in the Ionising Radiations Regulations (1961); and later separately for m edical radiations (POPUMET, 1988). Following the Second World War there was a rush to develop both nuclear power and nuclear weapons. The radiation protection community was faced with the problem o f setting dose limits that did not appear to restrict the expansion of these industries -- politics entered the scene. At first the public were beguiled by the promise of endless cheap nuclear power but demonstrations of nuclear weapons produced a different reaction. Gradually people began to have less and less confidence in, and m ore suspicion of, the motives o f governments, particularly with regard to bland reassurances about the effects o f radioactive contamination o f the environment. This apprehension was fuelled by the rise o f the `green m ovem ent' and was to a certain extentjustified -- it is only surprising that it took so long to develop. This was maybe because the early uses o f radiation had to a large extent been in medicine and the public had a trust of doctors. However, the motives o f the nuclear industry were seen as less likely to be for the good o f the individual. Even confidence in m edical radiology received a jo lt in the late 1950s when a well-known public health epidem iologist, Alice Stewart, carried out som e studies which linked radiology during pregnancy (pelvimetry) with leukaem ia in children (Stewart et al., 1958). This finding was at first controversial and disbelieved but, after bein g repeated by others, it is now accepted that there is a significant risk of leukaemia from even small radiation doses received by the embryo or foetus. Nowadays it is recom m ended (RCR, 1993) that no obstetrician should consider the use o f X-rays if som e other diagnostic tool is available. It has been estimated (Doll, 1989) that about 5 % o f all cases o f childhood cancer were caused by pelvimetry, which in the United Kingdom was about 75 cases each year and in the United States about 300. In could be claimed that these numbers o f leukaemias would have been saved had the work of Stewart et al. been acted on earlier. A sim ilar and contem porary story may be unfolding in relation to the childhood leukaem ia risk in proximity to overhead power lines in the United States. Contem porary risk estimates for radiation are probably more quantified and more soundly based than risks from any other environm ental pollutant. However, there are problem s with even these estimates because Radiation: early warnings; late effects 35 they are derived alm ost exclusively from the health records o f the survivors o f the atomic bom bings in Jap an in 1945 i.e. at high dose and dose rate. A conservative linear dose effect relationship is assum ed an d it is therefore appreciated that there is a risk at all doses. Exposures to radiation are therefore associated with a certain acceptance o f risk. For this reason the ICRP have based their philosophy (ICRP, 1977) around three tenets: Justification -- all uses o f radiation have to beju stified so that the detrim ent is offset by some net benefit. Optimisation -- all exposures m ust be kept as low as reasonably achievable, social and economic factors being taken into account. Limitation -- all exposures m ust be below the appropriate dose limit. It is o f interest to exam ine the first and second o f these tenets to see what progress has been m ade in the 100 years o f radiation use, using medical radiology as an example. In m edical radiation exposure it is intended that there should be som e benefit to the patient. A lthough this is usually true there are an increasing num ber o f occasions when it is doubtful, for exam ple the use o f X-rays as part o f health m onitoring for jo b selection, or some screening procedures. The NRPB has estimated (NRPB, 1990) that about 20 % o f all X-rays carried out in the United Kingdom are clinically unhelpful. Thus it has been stated in guidance principles given to radiographers that `there should be a valid clinical indication for all exam inations of patients where ionising radiation is u se d .' This constitutes `ju stificatio n ' an d is a m ajor step forward com pared to the radiology of, say, even 40 years ago, particularly as it refers to the patient and notju st the radiologist. In addition to this criterion, the dose received by the patient should be optimised, and here progress is not so good. The NRPB has estim ated (NRPB, 1990) that the total annual collective dose from m edical exam inations in the United Kingdom is about 16 000 mansieverts. It also suggests a num ber o f patient dose reduction m ethods which should result in a reduction in this collective dose o f about 7 500 man-sieverts, i.e. nearly 50 %. It has also been shown, m ore recently, that the variation in the dose received at different hospitals for the same exam ination can be m ore than an order o f m agnitude (Wall and Hart, 1997). Thus, although the individual radiation dose is now m aybe two orders o f m agnitude lower than 60 years ago, the problem o f optim isation o f that dose still exists. 3.5. Conclusions Overall it can be concluded that radiation protection standards have slowly evolved as the perception of radiation effects has developed. However, there have also been som e people, maybe ahead o f their time, who warned of im pending doom. Thus, there have always been periods when changes in limits have lagged some years behind clear evidence o f harm to hum an health. There are now substantial lobbies for changes which include both re-introducing the concept o f thresholds and considerations of hormesis (small doses which are thought to do some good) -- these have been resisted by the ICRP. Radiation protection is now firmly established in legislation both in the European Union (by directives) and, internationally, in the Basic Safety Standards of the International Atomic Energy Agency (IAEA). All o f these use the recom m endations o f the ICRP as their basis. In the United Kingdom the most recent legislation which covers workers and m em bers o f the public is the Ionising Radiations Regulations o f 1999. These are intended to im plem ent EU Directive 96/29 (laying down basic safety standards for the protection o f the health o f workers and the general public against the dangers arising from ionising radiation ). This directive will (eventually) be im plem ented throughout Europe, and similar regulations from the IAEA should apply in other countries. There are now, also, regulations which cover the use o f radiation in medicine and the limitation of doses to patients. However, it has been found difficult to ensure that radiation protection legislation is im plem ented uniform ly and there continue to be exam ples o f careless or irresponsible attitudes towards radiation sources and waste which have resulted in horrendous injuries and death, such as the caesium-137 incident at Goiania (Rosenthal et al., 1991). Historically, less stringent dose limits have also given rise to claims from workers for com pensation for cancer allegedly caused by radiation exposure. In this context, the issue o f liability for radiation injury has some lessons for other `long latent period' hazardous agents. In the United Kingdom 36 Late lessons from early warnings: the precautionary principle 1896-2000 Table 3.1. liability in the nuclear industry was originally state funded (Nuclear Installations Act 1965) but the Radiation Workers Com pensation Schem e, which is run jointly by trades unions and the nuclear industry, has been extremely successful in providing an alternative to litigation. Thus, although we have learnt m uch about the risks o f radiation exposure in the last 100 years (probably m ore than about any other environm ental pollutant) we are still constantly having to react to new knowledge. For instance, the risk rate for radiationinduced cancer was perceived (by ICRP) as four to five times h igher in 1990 as com pared to 1977. This resulted in changes in dose limits but was a belated response to m ounting incontrovertible evidence, a situation which has been a recurring theme in the history o f radiation protection, where precaution has sometimes been lacking despite the clear warnings given from the discovery o f radiation to the presen t day. Thus it m ust be concluded that the precautionary principle suggests that epidem iological databases of long-term effects must be funded and m aintained for the future even when an im m ediate n eed is not perceived. Radiation: early warnings and actions 1896 1899 1904 1904 1913 1924 1925-29 1928 1934 1949 1958 1961 1977 1988 1990-97 1990 1996 Injuries from exposure to X-rays noted by Edison, Tesla and Grubbe. John Dennis, New York journalist, campaigns for licensing of radiologists and warns of harm from X-rays. Death of Edison's assistant from complications arising from severe X-ray radiodermatitis. William Rollins, Harvard dentist/doctor, publishes many warnings on X-ray hazards, and recommendations on prevention for radiologists and patients, including pregnant women. First published rules of voluntary radiological protection by German Radiological Society. New York dentist, Theodore Blum, identifies 'radium jaw' in radium dial painters: but wrongly attributes this to phosphorous. Harrison Martland, New Jersey pathologist, identifies radium as the cause of the jawbone cancers in the dial painters studied. Establishment of the International X-ray and Radium Protection Committee: which later became the International Committee on Radiological Protection (ICRP). Reports by Colwell and Russ, on the death of more than 200 radiologists from radiation-induced cancers. ICRP concludes that there is no dose threshold for radiation-induced cancer and optimisation of all exposures is crucial. Alice Stewart reports that 'low dose' X-rays to pregnant women can cause leukaemia in their children. Not generally accepted until the 1970s. UK publishes regulations covering the use of radioactive substances. ICRP updates its radiation protection recommendations and links dose limits to risk. Regulations covering radiation doses to patients produced in the UK NRPB reports 20 % of medical X-rays are probably clinically unhelpful; that 50 % of the collective dose to patients could be avoided; and that individual doses for the same X-ray vary by 100x between hospitals. ICRP concludes in Publication 60 that the risk of radiation-induced cancer is 4-5 times greater than estimated in 1977 -- reduces the occupational dose limit to 20mSv per year. EU Directive on Ionising Radiations based on ICRP 60 which will be mandatory on member states. 3.6. References Albers-Schnberg, H. E., 1903. `U ber eine bisher unbekannte Wirkung der Rntgenstrahlen aurden Organismus der T ie re ', Mnchen med. Wchnschr. Vol. 50, pp. 1859-1863. Colwell, H. A. an d Russ, S., 1934. X-ray and radium injuries. Prevention and treatment, O xford University Press, London. Dennis, J., 1899. `The roentgen energy today', Dental Cosmos Vol. 41, p. 853. Desjardins, A. U., 1923. `Protection against radiation ', Radiology Vol. 1, p. 221. Doll, R., 1989. `The epidem iology o f childhood leu kaem ia', J. Royal Statistic. Soc. (Series A) Vol. 152, pp. 341-351. Edison, T. A., 1896. `Effect o f X-rays u p o n the eye', Nature Vol. 53, p. 421. Radiation: early warnings; late effects 37 `E d itorial', 1921. Archives of the Roentgen Society Vol. XVII, pp. 50-51. ICRP, 1977. `R ecom m endations o f the International Commission on Radiological Protection (Publication 2 6 )', Ann. IC R P Vol. 1, No 3. Ionising Radiations Regulations (Sealed Sources), 1961. HMSO, London. Macklis, R. M., 1993. `T he great radium scan d al', Scientific American Vol. 269, pp. 94 99. Martland, H. S. and Humphries, R. E., 1929. `Osteogenic sarcom a in dial painters using lum inous p a in t', Arch. Pathol. Vol. 7, pp. 406 417. Mutscheller, A., 1925. `Physical standards o f protection against roentgen ray dangers', Amer. J . Roentgen Vol. 13, p. 65. National Bureau of Standards H andbook No. 59 (1954). NRPB, 1990. `Patient dose reduction in diagnostic radiology', Docs of National Radiological Protection Board Vol. 1, No 3. POPUMET, 1988. Ionising Radiation (Protection o f Persons Undergoing Medical Examination or Treatment) Regulations, HMSO, London. RCR, 1993. M aking the best use of a department of clinical radiology, Royal College o f Radiologists, London. R oentgen, W. C., 1895. `U b er eine neue Art von Strah len ', Sitzungs-Berichte der Physikalisch medicinischen Gesellschaft zu Wurtzburg Vol. 9, p p .132-141. Rollins, W. H., 1904. `O n the tyranny o f old ideas as illustrated by the X-light used in th erapeutics', Elect. Rev. Vol. 43, p. 120. Ron, E. B., M odan, D. and Preston, D. L., 1989. `Thyroid neoplasia following low dose irradiation in ch ild h ood ', Radiat. Res. Vol. 120, p p .516-531. Rosenthal, J. J., de Almeida, C. E. and M endonca, A. H., 1991. `The radiological accident in Goiania: the initial rem edial action s', Health Physics Vol. 60, pp. 7-15. Rowland, R. E., Stehney, A. F. an d Lucas, H. F., 1983. `Dose-response relationships for radium in duced bon e sarco m as', Health Physics Vol. 44, Suppl. 1, pp. 15-31. Scott, N. S., 1897. `X-ray in juries', Am. X-ray Vol. 1, pp. 57-76. Stewart, A. M., Webb, J . W. an d Hewitt, D., 1958. `A survey o f childhood m align ancies', B. Med. J. Vol. 1, p. 1495. Stone, R. S., 1946. `Fifty years o f radiology: From Roentgen to the era of atomic power', WesternJ . Surg. Vol. 54, p. 153. Taylor, L. S., 1979. Organisation for Radiation Protection. The operations of the ICRP and NCRP 1928-1974, Report D O E/TIC-10124, US Dept o f Energy, Washington, DC. Thom son, E., 1896. `R oentgen rays act strongly on the tissues', Elec. Engr. Vol. 22, p. 534. T hom pson, S. P., 1898. `Presidential ad d ress', Arch. Roentgen Ray Vol. II, p. 28. Wall, B. F. an d Hart, D., 1997. `Revised radiation doses for typical X-ray exam in ation s', Brit. J. Radiol. Vol. 70, pp. 437 439. 38 Late lessons from early warnings: the precautionary principle 1896-2000 4. Benzene: an historical perspective on the American and European occupational setting Peter F. Infante 4.1. Early warnings Since the 1897 report of Santessen, who observed aplastic anaem ia am ong young women engaged in the m anufacture o f bicycle tyres in Sweden, and the report in the sam e year by LeN oir and Claude, who observed haem orrhaging in a young man engaged in a dry-cleaning operation in France, benzene has been known to be a powerful bone marrow poison. Similar reports o f workers developing benzenerelated diseases of the bone marrow increased dramatically through the first half o f the 20th century. associated with this extremely high prevalence o f abnorm al white blood cell counts were 90 ppm and higher. Greenburg (1926) recommended medical removal if an employee developed clinical symptoms of poisoning, or if blood values for individual workers dropped 25 % or more. Glossary Leukaem ia: a progressive cancer of the blood. It can be classified clinically by duration of the disease (acute or chronic), and by character of the disease, including m yelogenous (myeloid cells in excess; lym phatic (lymphoid cells in excess); m onocytic (m onocytes in excess) Between 1910 and 1914, the first major use of benzene as a solvent in the rubber industry took place. Production o f benzene was also stimulated greatly by the dem and for toluene in the m anufacture o f explosives during the First World War. E xp an d ed use o f benzene in industry after the war led to an increased use o f benzene as a solvent in the artificial leather industry, rubber goods, glue manufacturing, hat manufacturing, rotogravure printing, paint, adhesives, coatings, dry cleaning, automobile m anufacturing, tin-can assembly, as a starting m aterial in organic synthesis, in petroleum products and in the blending o f m otor fuels. Aplastic anaemia: a condition whereby the bone marrow fails to produce enough red blood cells, white blood cells and platelets. Therefore the individual is usually tired because there are not enough red blood cells available, is susceptible to infection because there are not enough white blood cells present, and bleeds easily because there are not enough blood platelets available Benzene poisoning: usually refers to aplastic anaemia Hypersusceptibility: a condition of increased susceptibility to toxic exp osures in relation to the average person Haem opathy: an abnorm ality of the blood Multiple myeloma: a cancer of the lymphatic system This expansion in the industrial uses of benzene was accom panied by a vast increase in the num bers o f reported cases o f aplastic anaem ia, generally referred to as `benzene poisoning'. Som e individuals were diagnosed with benzene p oison in g within a few weeks o f initial employm ent and som e died within a few m onths o f beginning their jo b s (H ogan and Schrader, 1923). These poisonings were associated with benzene levels ranging mostly between 200 ppm (parts per million) and 1 000 ppm. Greenburg and colleagues (1926) m ade a survey o f 12 plants in the United States that used benzene and observed that 32 % of the workers had abnormally low white blood cell counts (below 5 500 per cc), and 12 % had below 4 000 per cc. The benzene exposure levels 4.1.1. First report o f benzene-induced leu kaem ia In 1928, Dolore and Borgom ano published the first case o f benzene-induced leukaemia. This case o f acute lymphatic leukaem ia was identified in a pharm aceutical worker who had a jo b considered dangerous because of high benzene exposure levels. Another worker at the same plant died of aplastic anaem ia and the authors were o f the opinion that some previous cases of aplastic anaem ia reported in association with benzene exposure may also have been leukaemia. The com pany's m eans o f dealing with the high exposure levels that caused these blood diseases was to rotate the workers out o f the specific jo b every month. Benzene: an historical perspective on the American and European occupational setting 39 4.2. Actions and inactions 4.2.1. Exposure recommendations By 1939, the vast num ber o f benzene poisonings am ong workers in all parts o f the world led to the recom m endation for the substitution o f benzene with other solvents by a num ber of investigators (Greenburg, 1926; E r f an d R hoads, 1939; M allory et al., 1939). In 1939 Hunter, an d M allory et al., reported 89 cases o f `poison in g' and three cases o f leukaem ia am ong workers exposed to benzene in a variety o f occupations. Two o f the `p oiso n in gs' were associated with benzene levels o f less than 25 pp m an d 10 ppm , respectively. Yet, in 1946, the A m erican C onference o f Governmental Industrial Hygienists (ACGIH) recom m ended a limit o f 100 ppm for benzene exposure in the workplace (ACGIH, 1946). Subsequently, the recom m ended value was reduced to 50 ppm in 1947 and to 35 ppm in 1948 (ACGIH, 1948). Because o f evidence of `hypersusceptibility' to the bone marrow suppressing effects of benzene, a docum ent published in 1948 by the American Petroleum Institute (API) concluded that the only absolutely safe level from exposure to benzene was zero (API, 1948). This opinion was based on observations o f som e workers developing various blood diseases indicative of bone marrow depression who had worked alongside other workers whose blood counts were in the norm al range. The API then proceeded to recom m end a limit o f `50 ppm or less'. In 1957, the A CG IH lowered its recom m ended eight-hour time-weighted average exposure limit to 25 ppm for benzene (ACGIH, 1957). 4.2.2. Disregard for recommendations In spite o f the recom m endations m entioned above, cases of aplastic anaemia, and central nervous system toxicity m anifested by headache, nausea, giddiness, staggering gait, paralysis and unconsciousness (leading to death in 13 cases in the U nited Kingdom ) continued to be reported in the 1940s and 1950s (Browning, 1965). These symptoms of the central nervous system are thought to be associated with benzene exposures ranging from 3 000 ppm to 20 000 ppm (Flury, 1928 ) -- levels 200 to 800 times higher than the limits recom m ended in the 1940s and 1950s and 2 000 times higher than the 10 ppm level already associated with aplastic anaem ia by H unter and by M allory et al. (1939). In the 1950s and 1960s, an obvious lack of precaution for workers exposed to benzene was taking place in many parts o f the world, including France, Italy, Russia, Turkey, the United Kingdom, the United States and other countries, as docum ented by the publication of case reports o f blood diseases as a result o f benzene exposures well above levels known to be toxic to the bone marrow and central nervous system o f workers. For example, Vigliani and Saita, in 1964, reported that the risk o f acute leukaemia am ong workers `heavily exposed to ben zen e' in the rotogravure and shoe industries in the Italian provinces o f M ilano and Pavia was at least 20 times higher than for the general population. Vigliani (1976) reported that over 200 cases of benzene haemopathy, including 34 cases o f acute leukaemia, had been treated at the Institutes of Milano and Pavia between 1942 and 1965. These workers were exposed to benzene levels ranging mostly from 200 to 500 ppm , with occasional peaks above these levels. In 1967, G oguel et al. reported 44 cases o f benzene-induced leukaemia, mostly chronic forms, having occurred in the Paris region of France between 1950 and 1965. Blood poisonings am ong workers exposed to benzene were being docum ented in other parts o f Europe as well. As stated by Aksoy (1977), because `benzene containing glue adhesives were extremely practical and cheaper in the market, they replaced their customary petroleum containing adhesives with the new p ro d u ct' in the shoe an d slipper industry in Turkey in 1961. Benzene exposures experienced by these workers were reported to have been between 150 ppm and 650 ppm (Aksoy, 1978). By the mid-1970s, an epidemic of aplastic anaem ia and leukaemia am ong Turkish shoe workers began to unfold as reported by Aksoy (Aksoy et. al., 1971; Aksoy 1977 and 1978). The majority of the individuals who were identified in the reports from these European countries as dying from leukaem ia as well as other benzene-related blood diseases were exposed to levels shown decades earlier to cause benzene poisoning. (For a review o f case reports o f benzenerelated blood diseases associated with the countries m entioned above, see IARC, 1974.) 4.2.3. Epidem iological evidence for leu kaem ia Beginning in the early 1970s, the University o f North Carolina in the U nited States published a series o f epidemiological studies demonstrating significant excesses of 40 Late lessons from early warnings: the precautionary principle 1896-2000 leukaemia, mostly chronic forms, am ong workers exposed to presum ably low atm ospheric levels o f benzene (M cM ichael et al., 1975). The benzene exposure resulted primarily from the use o f rubber solvents such as petroleum naphtha, toluene, mineral spirits, etc., that were contam inated with benzene ran gin g in volum e from abo u t 1 5 % in the 1940s to about 0.5 % for petroleum naphtha in the 1970s. In 1977, Infante et al. published the results o f the first cohort study o f workers exposed specifically to benzene. The workers were engaged in the manufacture o f a rubberised food wrap called Pliofilm. The study dem onstrated a 5- to 10-fold risk o f leukaem ia am ong workers exposed to technical-grade benzene at levels that were generally considered within the various limits recom m ended over the time period 1940 1971, that is a 10 ppm time-weighted average to a m axim um lim it o f 100 ppm (Infante et al., 1977a). Until this time, benzene was considered a cause of leukaemia based not upon epidemiological studies, but rather upon case reports o f leukaem ia and the clinical observation that individuals with benzene-induced aplastic anaem ia and other blood diseases transform ed into acute leu kaem ia. 4.2.4. U S attempt to control occupational exposure In 1977, on the basis o f the Infante et al. (1977a) study results supplem ented by the world literature on benzene and leukaemia, the US Departm ent o f Labor (OSHA, 1977a) issued an emergency temporary standard (ETS) that would have reduced the occupational benzene exposure limit in workplace air to 1 ppm as an eight-hour timeweighted average. The 1 ppm exposure limit was based on the Occupational Health and Safety A dm inistration's (O SH A 's) policy at the time that exposure to carcinogens in the workplace should be lowered to the lowest feasible limit. Such feasibility analyses take into consideration both technological and econom ic feasibility. The new OSHA ETS was stayed in 1977, however, in response to a challenge in the U S Court o f Appeals by the API, who in essence argued that there was no increased risk o f developing leukaemia as a result of benzene exposures below the old lim it o f 10 ppm . Then OSHA proposed a permanent standard, requested comments and held a public hearing (OSHA, 1977b). In 1978, OSHA issued a final standard (OSHA, 1978) that included a 1 ppm atm ospheric exposure limit. This standard was challenged by the API on the same grounds. The US Court of Appeals again vacated the final standard, and that decision was appealed to the Suprem e Court. In an unrelated development, benzene was voluntarily withdrawn from consum er products in the United States after it was shown that use o f a paint stripper in the hom e could generate atm ospheric levels up to 200 ppm in a short period o f time (Young et al., 1978). 4.2.5. The US Benzene Decision and doseresponse analyses In July 1980, the US Suprem e Court (IUD, 1980) issued what has becom e known as the Benzene Decision. This decision has had a m ajor im pact on O SH A 's ability to control exposures to benzene and other toxic substances in the workplace. The court stated that before OSHA can prom ulgate any perm anent health standard, the Secretary of Labor is required to m ake a threshold finding that a place o f employm ent is unsafe in the sense that significant risks are present and can be eliminated, or lessened, by a change in practices. Although the Benzene Decision recognised the uncertainties involved, it indicated that the determ ination o f `significant risk' should, if at all possible, be established on the basis o f an analysis o f the best available evidence through such m eans as quantitative risk assessments. The Suprem e Court in its general guidance for future OSHA rule-making noted that the requirem ent that a significant risk be identified is not a m athem atical straitjacket an d that it is the O SH A 's responsibility to determine what it considers to be a significant risk based largely on policy considerations. In the only concrete example o f significant risk, the court stated that if, for exam ple, the odds are one in a million that a person will die from cancer by taking a drink o f chlorinated water, the risk clearly could not be considered significant. On the other hand, if the odds are one in 1 000 that regular inhalation of gasoline vapours that are 2 % benzene will be fatal, a reasonable person m ight well consider the risk significant and take appropriate steps to decrease or elim inate it. Since the Benzene Decision, OSHA has considered an occupational lifetime risk (over a 45-year period) o f one extra case of cancer, or other material im pairm ent of health consequence, per 1 000 workers to be Benzene: an historical perspective on the American and European occupational setting 41 `significant'. It has not yet considered the other end o f the range -- what it considers to be a non-significant risk -- because all the health standards that OSHA has prom ulgated since the Benzene Decision, with the exception o f perhaps formaldehyde, have resulted in estimates o f excess risk that are greater than one per 1 000 over an occupational lifetime. (For quantitative estimates of risk for health standards prom ulgated by OSHA since the Benzene Decision, see Infante, 1995b.) A straightjacket, however, appropriately describes the risk analyses that OSHA currently engages in prior to proposing any regulatory action. The long delay recognised in OSHA prom ulgating standards before 1980 has now been superseded by additional detailed analyses o f risk related to exposure. While on the surface such analyses may seem appropriate, they have becom e encum bered by additional analyses that take into account the m echanism by which the substances being regulated may cause cancer. Since the exact m echanism by which any substance causes cancer (including benzene which has been studied for decades) has not been identified, speculation and argum ent about various unproven hypotheses for cancer causation are time consuming. Many other issues also have been added to the debate on risk assessment procedures such as whether the m ouse, the hamster, or the rat is the most appropriate species to use when hum an data are not available. Since the OSHA is required to review and com m ent on all possible cancer m echanism s, appropriateness o f species, etc., the entire `risk assessm ent' process has created additional years o f delay in standard setting. Instead of reasonable precautions being prom ulgated by governm ent and employers, years pass as analyses are perform ed to determine the dose response between exposure and risk of disease. These analyses can also include incorporation of speculative mechanistic data that have not been validated. 4.2.6. Cost in lives o f the prolonged regulatory process Eleven years after the OSHA prom ulgation o f an ETS for benzene, a new benzene standard that included a 1 ppm eight-hour timeweighted average exposure limit was finally issued (OSHA, 1987). The new limit was based on `economic feasibility', not elimination of `significant risk', as an occupational lifetime risk o f 10 extra leukaem ia deaths per 1 000 workers was associated with the 1 ppm limit. Other estimates o f leukaem ia risk, limited to the cohort data from the National Institute for Occupational Safety and Health (NIOSH) (as updated by Paxton et al., 1994) and to deaths from acute myelogenous and m onocytic leukaem ias only, indicate a range o f 0.02 to 5.1 per 1 000 excess deaths depending upon the estimates o f benzene exposure and the m odel chosen (Crump, 1994). However, these later analyses are based on a follow-up period o f the NIOSH cohort that results in selection bias for the reasons described below. B ased on O SH A 's final quantitative risk assessment for benzene and estimates of extra benzene exposure am ong the United States workforce during the 10 years that it took to com plete the benzene standard, it has been estimated that an extra 198 deaths from leukaem ia and 77 extra deaths from m ultiple m yelom a will eventually develop am ong U S workers as a result o f the 10-year delay -- deaths that could have been prevented (Infante and DiStasio, 1988). This estimate o f preventable deaths from benzene exposure did not include blood diseases other than leukaemia, which were known at the time to be caused by benzene exposure, the reason being that there were no dose-response data available for the other blood diseases. 4.2.7. Expansion o f lymphohaematopoietic diseases The quantitative estimate of extra deaths from benzene exposure indicated above did n ot include n on-H odgkin's lym phoma, which has been shown m ore recently to be associated with occupational exposure to very low levels o f benzene (Hayes et al., 1997). Quantitative risk assessment (Infante, 1997) based on the 1996 Hayes et al. study results suggests an extra risk o f 54 deaths from leukaem ia/lym phom a per 1 000 workers exposed over an occupational lifetime (45 years). This risk is 54 times greater than a level considered significant by OSHA. The 1997 Hayes et al. study results an d doseresponse analyses based upon those data clearly dem onstrate the inadequacy o f the 1 ppm exposure limit for benzene based on cancer risk alone. Fortunately, most occupational settings in the United States are able to achieve benzene exposure levels in the range o f 0.2-0.3 ppm , or lower. Additionally, the U S standard includes provisions ancillary to the exposure limit such as exposure monitoring, medical surveillance and hazard training -- all of which should presum ably further reduce the 42 Late lessons from early warnings: the precautionary principle 1896-2000 risk of benzene exposure and related diseases. In addition to quantitative risk assessment, direct observation of data from the 1997 Hayes et al. study dem onstrates significantly elevated relative risks for all lymphohaematopoietic cancers combined, and for acute non-lymphocytic leukaem ia and myelodysplastic syndrome combined, am ong the group of workers exposed to a constant average benzene level o f only 1.2 ppm for 5.5 years for a total cumulative dose o f 6.7 ppm-years (Hayes, pers. comm., 1999), which is a m uch lower cumulative dose than that allowed by a limit o f 1 ppm over an occupational lifetime of 45 years (45 ppmyears o f cumulative dose). Som e individuals within the benzene cohorts who have died from leukaemia or lymphoma experienced estimated benzene exposures of 0.5-2 ppm for only 1-2 years or less (Infante, 1992). By the 1990s, toxicologic research dem onstrated the multiple-site carcinogenicity of benzene in experimental animals, and additional epidemiological studies and case reports o f exposed workers expanded the carcinogenicity o f benzene to all m ajor forms o f leukaem ia in the aggregate (Savitz and Andrews, 1997; Infante, 1995a; Wong, 1987b) and specifically acute m yelogenous leukaem ia and its variants (Hayes et al., 1997; Browning, 1965; Rinsky et al., 1987; Bon d et al., 1986; D eCoufle et al., 1983), acute lymphatic leukaemia (H ernberg et a l, 1966; Shu et a l , 1988), chronic lym phatic leukaem ia (M cM ichael et al., 1975;), chronic myelogenous leukaemia (Browning, 1965; G oguel et al., 1967; T areeff et al., 1963; Infante, 1995a; Wong, 1987b;) and some m inor forms such as hairy cell leukaem ia (Aksoy, 1987; Flandrin and Collado, 1987), myelodysplastic syndrome (Hayes et al., 1997), m yeloproliferative disorders (Rawson et al., 1941; Tondel et al., 1995) as well as non-H odgkin's lym phom a (Hayes et al., 1997), including m ultiple m yelom a (D eC oufle et al., 1983; Rinsky et al., 1987; Ireland et al., 1997; Goldstein, 1990). An `u p d ated ' analysis o f the N IO SH benzene cohort (Infante et al., 1977a; Rinsky et al., 1987) by Wong (1995) concluded that the excess o f multiple myeloma in the NIOSH cohort was no longer statistically significant. Wong (1995), however, changed the beginning date o f follow-up from 1950 to 1940. In doing so, he introduced selection bias into his analysis because the company removed records for individuals who died at one o f the study locations for several o f the years prior to 1950. Thus, analyses o f the NIO SH cohort by Wong (1995) that begin follow-up prior to 1950 cannot be relied upon for estimating risk of death from multiple myeloma, or any other cause of death. 4.3. Discussion The response to inform ation on the toxicity o f benzene has at times demonstrated concern by som e in the field o f occupational health, particularly in the earlier years when large numbers of workers in various sectors and jobs were being surveyed to determine the extent of their blood diseases. Nevertheless, even during this period, benzene exposure levels were not reduced to levels com m ensurate with the toxicity data available at the time, and the epidem ic o f poisonings and leukaemia among benzeneexposed workers continued through the first six decades o f the 20th century. Various reasons for the limited public health response and subsequent overexposure and disease in light of the knowledge of benzene toxicity are apparent. 4.3.1. Som e reasons for lack o f precaution Lack of knowledge The lack o f precaution about exposure to benzene has been attributed in part to lack o f knowledge o f its toxicity during the first four decades o f the 20th century. For exam ple, even though Santessen (1897) reported four cases of aplastic anaem ia am ong women manufacturing raincoats in Sweden in 1897, exposures were not lowered enough and Helm er (1944) reported 60 cases of benzene poisoning (58 were women) at a single raincoat m anufacturing facility in the same country in 1940 and 1941. The epidem ic of benzene poisoning in Sweden was ascribed in part to a lack o f knowledge o f the toxicity o f benzene by plant m anagem ent and workers (Helmer, 1944). Cost of solvents Several investigators who surveyed w orkforces an d identified workers with various benzene-induced blood diseases, namely G reenburg and colleagues in 1926, E r f and R hoads in 1939 an d M allory et al. in 1939, recom m ended the substitution of benzene with other solvents. Yet, worldwide consum ption o f benzene in the m arketplace continued to expand after the Second World War. O ne o f the reasons for expan ded use o f benzene in the synthetic rubber industry is that it was such a good rubber solvent. Another reason, as expressed by Aksoy Benzene: an historical perspective on the American and European occupational setting 43 (1977), is that benzene was ch eaper than other solvents used in the shoe and slipper industry in Turkey. Thus, econom ic consequences led to other solvents being replaced with benzene in the Turkish shoe m anufacturing industry as late as 1961. This substitution with benzene led to high-level workplace atmospheric benzene concentrations and to the epidem ic of leukaemia, preleukaemia, pancytopenia and other blood diseases as reported by Aksoy (Aksoy et al., 1971; Aksoy, 1977 an d 1978). Thus, econom ic considerations in the 1960s (Aksoy, 1977) further contributed to the overexposure and subsequent benzenerelated diseases. Consensus recommendations and corporate influence T he 1939 reports by H unter an d M allory et al. indicated that some cases of benzene poisoning were associated with levels o f 25 pp m and 10 ppm . Yet, in 1946, the A CGIH recom m ended a limit o f 100 ppm . Although the ACGIH recom m endation for benzene was lowered to 35 ppm in 1948, this level was still higher than the levels reported in association with benzene poisonings. From my p erson al experien ce over the years, it is apparent that one o f the problem s being faced in the occupational health community for benzene (as well as for other toxic substances) is that consensus organisations usually base their recom m ended exposure levels on what is easily achievable in the workplace. Data related to level o f exposure and toxicity are reviewed, but are not translated into health-based exposure-limit recom m endations. Castleman and Ziem (1988) investigated this behaviour by the ACGIH and concluded that the ACGIH threshold limit values (TLVs) were based heavily on corporate influence. Thus, consensus recom m endations were inadequate and corporate influence may have played a role in these recom m endations and in the resultant proliferation o f benzeneinduced diseases in the workplace. Castleman and Ziem (1988) were of the opinion that an international effort was needed `to develop scientifically based guidelines to replace TLVs in a climate o f openness and without m anipulation by invested interests'. To date, this goal has not been achieved (Castleman and Ziem, 1994). Anti public health attitude (call for scientific certainty) In the 1970s, benzene m anufacturers and users began a new approach for conveying knowledge about the toxicity o f benzene to the public in general, and to workers and plant m anagers specifically, which contributed to a continuation of overexposure to benzene. This is the period when m anufacturers began to hire consultants to downplay the importance of the scientific observations related to the toxicity o f benzene and to introduce unresolvable arguments about dose-response analyses, which had an im pact in delaying much needed government regulations that sought to reduce benzene exposure in the workplace. Econom ic considerations were again being given a higher priority than concerns for public health, but this time the econom ic concerns appeared to be based on the cost of lowering exposures (OSHA, 1987), and perhaps on the increasing cost of litigation and liability related to workers contracting benzene-related diseases on the jo b . This era has fostered the development of arguments that seek to minimise or m isrepresent study results. In my opinion, it is part o f a new anti public health approach that calls for scientific certainty in terms of causality for every specific lym phohaem atopoietic disease related to benzene by exposure level. As a result, workers employed worldwide in benzene exposure operations today may not be afforded the appropriate protection to the extent feasible in order to reduce their risk of haem atopoietic diseases and many of those who develop benzene-related diseases may receive little or no compensation. For exam ple, during the OSHA rule-making hearings held in 1977, it was argued that the benzene cohort study conducted by NIOSH, which dem onstrated a 5- to 10-fold elevated risk o f leukaem ia (Infante et al., 1977a), was m eaningless in terms o f public health intervention to reduce exposures in the workplace. Although not persuasive, one of the argum ents raised by consultants to the industry was that the study had simply identified a random leukaem ia cluster, and since clusters o f leukaem ia are known to occur in time and space, this cluster of leukaem ia ju st happened to be identified am ong a cohort of benzene-exposed workers (Tabershaw and Lam m , 1977). (See Tabershaw an d Lam m , 1977, an d Infante et al., 1977b, for a series o f argum ents and rebuttal o f the study findings.) It was further argued that benzene could not cause leukaem ia in workers because there was no evidence that benzene caused cancer in experimental animals (Olson, 1977). This 44 Late lessons from early warnings: the precautionary principle 1896-2000 argum ent was clearly fictitious given the overwhelming evidence of carcinogenicity provided by the study o f humans. In any case, shortly thereafter the evidence of carcinogenicity of benzene in experimental animals becam e available (Maltoni and Scarnato, 1979; NTP, 1986). In the 1980s, when OSHA again published a new benzene proposal taking into consideration the guidance offered by the Suprem e C ourt's Benzene D ecision on determination o f the significance of health risk, attention during the rule-making focused on the dose-response analyses prepared by O SHA and its consultants in support o f its new standard (OSHA, 1987). M ost o f the argum ent addressed the benzene exposure assessm ent portion o f the risk assessment, with new estimates o f exposure am ong the benzene cohort m embers being provided for time periods in which there were no exposure data available. These `educated' guesses by the various parties involved in the rule-making could not be confirmed. Proposed government regulation o f benzene, however, resulted in a num ber o f new dose-response analyses and a protracted debate about which exposure assessm ent was the m ost appropriate for use in the quantitative risk assessment -- the resolution o f which can never be determ ined with scientific certainty. A rgum ent abo u t the type o f dose-response m odel that was most appropriate for benzene exposure and risk of leukaem ia was also raised. While nobody would object to debate on these issues, the continuation o f the arguments for protracted periods o f time resulted in workers and the public in general being unnecessarily exposed to benzene levels that could have otherwise been reduced through a shorter regulatory process. Studying a subject to death often results in the death o f those we are trying to protect (Infante, 1987). In the 1990s, the US National Cancer Institute (N CI), in collaboration with the Chinese Academy o f Preventive M edicine (CAPM), published a series of ongoing studies o f Chinese workers exposed to benzene. T hese studies (Hayes et al., 1997 an d 1996; D osem eci et al., 1996) dem onstrate a dose response for exposure to benzene and leukaemia, lymphoma, myelodysplastic syndrome and aplastic anaemia. They also demonstrate, through direct observation, high relative risks for leukaemia, myelodysplastic syndrome and nonH o d gk in 's lym phom a as a result o f very low average benzene exposures, such as around 1 ppm . Since the results o f these wellconducted studies may be used in the future by governments in Europe, the United States and other countries for estimating benzenerelated diseases from low-level exposure to the general population, the findings have broad implications for public health intervention. In response to the publication o f these studies, consultants to the chemical industry have published critiques o f both the health findings and the benzene exposure estimates related to those findings (Wong, 1998 an d 1999; Budinsky et al., 1999), which in the opinion of som e m isrepresent the data on health effects as well as the benzene exposure estimates m ade by the NCI/CAPM investigators (Hayes et al., 1998; Hayes, pers. comm.). Some of these same consultants have also expressed surprising views about the m ore general findings related to benzene exposure and disease (Wong, 1995 and 1996; B ergsagel et al., 1999). For exam ple, based on his analysis o f data from the NIOSH benzene cohort study, one author concluded that benzene can only cause acute myelogenous leukaem ia, in contrast to other types o f leukaemia, and that the threshold is between 370 and 530 ppm-years o f exposure (Wong, 1995). In this publication, he failed to include data from his own benzene study whereby he reported a statistically significant dose response for benzene exposure and leukaem ia am ong workers whose cumulative benzene exposures ran ged from less than 15 ppm-years to m ore than 60 ppm-years (Wong, 1987a and b). Furthermore, in the latter study, none o f the leukaem ia deaths were from acute myelogenous leukaemia. Thus, the findings and conclusions he drew from his own study contradict his opinion that benzene causes only acute myelogenous leukaem ia and that the cumulative exposure threshold is between 370 and 530 ppm-years. Wong (1995) further concluded that there is no evidence that benzene exposure was associated with m ultiple m yeloma in the NIOSH study because he could not identify a dose response based on four cases of multiple myeloma in the study population. Lack o f ability to observe a dose-response relationship based on four deaths from multiple myeloma is essentially m eaningless, because four cases are too few to allow enough statistical power to observe a dose response if in fact one were present. In any case, Rinsky et al. (1987) o f N IO SH had previously dem onstrated a significant excess o f death from multiple myeloma am ong the Benzene: an historical perspective on the American and European occupational setting 45 benzene cohort members and concluded that low-level benzene exposure was related to multiple myeloma. Wong (1995) has also argu ed that benzene in gen eral is not associated with m ultiple m yeloma and, contrary to the findings in the NCI/CAPM study, he has argu ed that there is no evidence that benzene is associated with nonH o d gk in 's lym phom a (Wong, 1998) -- conclusions that ap p ear to be at odds with the views of others on benzene toxicity (Goldstein and Shalat, 2000; Goldstein, 1990; Rinsky et al., 1987; D eCoufle et al., 1983; Infante, 1995a; Savitz and Andrews, 1996 and 1997; Hayes et al., 1998 and 2001). The argum ents about the NCI/CAPM study seem reminiscent of the protracted debate and delay in necessary governm ent action that took place after release of the NIOSH benzene study in 1977 (Infante et al., 1977a). It will be unfortunate if m ore precaution is taken with the use o f data from the N C I/ CAPM study than with the protection o f populations exposed to levels o f benzene that can be reduced with technology currently available. Scientific certainty is difficult to achieve, but stressing the uncertainty does not do justice to the data on benzene exposure and related diseases. In the case of benzene historically, taking precaution to maintain exposure levels in the workplace in accordance with the scientific data available at the time would have eliminated much needless suffering and death am ong the workers. In my opinion, the protracted argum ent about the dose response for benzene-related diseases am ong exposed workers and denial about the lym phohaem atopoietic diseases m ost likely associated with benzene exposure is counterproductive to efforts to provide a workplace relatively free o f harm . While the continuation o f this debate may be interesting from an academ ic viewpoint, it also raises the question o f whether it may be more a reflection o f economic concerns and potential liability on the part o f com panies than a concern for the p u b lic's health based on a reasonable interpretation o f available scientific data. 4.3.2. Benzene in gasoline a continuing hazard Most consumers and many medical personnel are not aware o f the fact that gasoline contains benzene. In the United States, gasoline has contained an average of abo u t 1.5 % benzene for the past two decades, but may reach 5 % by volume (Infante et al., 1990). Historically, petrol in most European countries has contained more benzene than US varieties and the trend apparently still existed through 1994 (Deschamps, 1995), but the benzene content has supposedly been reduced m ore recently. Not surprisingly, epidem iological studies, analyses and case reports have demonstrated an association between gasoline exposure and leukaem ia (Schwartz, 1987; Jak o b sso n et al., 1993; Infante et al., 1990), other blood diseases (Infante et al., 1990; Lum ley et a l, 1990; Naizi and Fleming, 1989), chrom osom al defects (Lum ley et al., 1990; H ogstedt et a l, 1991) and other m anifestations o f genetic dam age (N ilsson et al., 1996). Yet, gasoline station pum ps do not provide adequate information on the cancers known to be associated with benzene exposure. Nor do the material safety data sheets for gasoline provide the available evidence on chrom osom al or genetic damage. Because o f this lack o f candour about the hazards of benzene in gasoline, garage m echanics and highway m aintenance workers take unnecessary risks by using gasoline as a solvent in cleaning auto parts (Infante, 1993), and consum ers take unnecessary risks by using gasoline as a solvent and fail to take the necessary precautions when using gasoline in various hom e appliances such as lawn mowers, weed trimming devices, power saws, etc. In addition, a study of roadside vendors who sold re-packaged gasoline in Nigeria has demonstrated that 26 % had neutropenia as com pared to 2-10 % in controls, a significant difference (Naizi and Fleming, 1989). The hazards of benzene in gasoline have been recognised since at least 1928, when Askey (1928) reported a case o f aplastic anaem ia in a U S worker exposed to gasoline, and more recently by the report o f a case o f myelofibrosis in a Swedish petrol station attendant (Tondel et al., 1995) and a case o f aplastic anaem ia in a US roofer who used petrol to clean seam s before fitting rubberised roofin g m aterial (Infante et al., 1990). Despite the overwhelming literature on the hazards o f benzene in petrol, the public health community and safety officials, including those employed by industry, have yet to com e to grips with the task of adequately informing workers and consum ers about this hazard. 46 Late lessons from early warnings: the precautionary principle 1896-2000 4.4. Conclusions and lessons for the future The available knowledge on the toxicity o f benzene and the failure to take precautions to protect workers (and the public in general) in light o f this knowledge over the past century is cause for concern. The inaction or inadequate actions by consensus organisations and governm ents alike throw into question the ability o f these organisations to protect the health o f the public. In the case o f benzene exposure in the workplace, the precautionary principle is not relevant. Recom m endations m ade in the United States and the United Kingdom in the 1920s for the substitution o f benzene with other solvents known to be less toxic to bone marrow went unheeded for decades even though high percentages of workers being surveyed demonstrated blood disorders. Furtherm ore, benzene was not withdrawn from consum er products in the United States until 1978 and this was done by manufacturers on a voluntary basis, and it has never been adequately validated. It is also difficult to accep t the claim ed ignorance o f the toxicity o f benzene in the raincoat m anufacturing industry in Sweden in the 1940s when 60 cases of blood poisoning in a single factory were reported. Forty years earlier (in 1897) published case reports of aplastic anaem ia am ong women employed in this same industry in Sweden appeared in the literature. The claim that m anagem ent was unaware o f the hazards o f benzene exposure in such a small industry in a country known for its hum anitarian concerns is incom prehensible. In the 1940s and 1950s, 13 deaths from the neurotoxic effects o f benzene were reported to have occurred in the United Kingdom. The benzene exposure levels associated with these acute deaths were m ost likely m ore than 200 to 800 times the occupational exposure levels recom m ended at the time. Clearly, this situation could, and should, have been avoided had there been any serious concern for worker health. With the knowledge available at the time, it is also difficult to understand how benzene was substituted for other petroleum solvents in the shoe industry in Turkey in 1961. Aksoy states that the substitution was m ade because benzene was cheaper than the other solvents. Data on the costs o f benzene and the other solvents in Turkey in the 1960s are not available, but it is unlikely that the difference could have been m ore than a few cents a gallon. Yet, the epidem ic o f leukaem ia and other fatal blood diseases that followed this substitution had to have been very costly in terms of the workers' diseases, the associated expenses for health care and the loss of wages, etc. This is simply a case, as in other instances, of the cost of production being more im portant to the m anufacturers than the cost o f hum an life. Even though num erous case reports (in the thousands) o f benzene-related blood diseases, including leukaemia, were reported in the literature, precautionary m easures to reduce exposure levels below those known, or reasonably anticipated, to cause blood diseases were not taken, and recom m ended exposure limits by consensus organisations like the ACGIH were based on those that were easily achievable in the workplace. According to Castleman and Zeim, such recom m endations stemm ed from the participation o f scientists employed by various corporations on the Threshold Limit Value Committee that m ade the exposure recom m endations. Thus, one o f the lessons to be learned, if not already obvious, is that consensus organisations in the process of developing exposure limits for chemicals should maintain distance from the producers of the chemicals and their `consultants' when evaluating evidence for the diseases of concern. Finally, affixing a w arning label on gasoline pum ps that includes the cancers and other diseases known, or likely to be caused by benzene exposure, may serve to reduce unnecessary benzene exposure to garage mechanics, petrol service station attendants, highway m aintenance workers and consum ers who fill their own gas tanks, but who m ore unknowingly use the gasoline in consumer products at home, and not infrequently use gasoline as a solvent at home without full knowledge o f its cancer and non cancer disease risks. Failure adequately to inform the public o f the cancers, bone marrow proliferative diseases and genetic hazards associated with benzene in gasoline is to repeat our failures o f the 20th century in the 21st century and to make a mockery of public health education. Benzene: an historical perspective on the American and European occupational setting 47 Benzene: early warnings and actions 1897 1926 1928 1939 1946 1947 1948 1948 1957 1950s-60s 1977 1977 1978 1980 1987 1996 2001 Santessen report on observed aplastic anaem ia In Sw eden and other reports show that benzene Is a powerful bone m arrow poison G reenburg and colleagues observe abnorm ally low white blood cell counts In benzene w orkers Dolore and Borgomano publish the first case of benzene-induced leukaemia A number of investigators recommend the substitution of benzene with other solvents, but this was not implemented Am erican Conference of Governm ental Industrial Hygienists (ACGIH) recomm ends a limit of 100 ppm for benzene exposure, even though some cases of benzene poisoning were associated with levels of 25 ppm and 10 ppm Recommended value reduced to 50 ppm Further reduced to 35 ppm Am erican Petroleum Institute (API) concludes that the only absolutely safe level is zero, but recom m ends 50 ppm or less AC G IH lowers recommended exposure to 25 ppm O bvious lack of precaution for w orkers exposed to benzene in many parts of the world with fatal consequences Infante et al. publish the first cohort study of w orkers linking benzene exposure directly to leukaem ia Based on these results, the US D epartm ent of Labor w ants to reduce exposure to 1 ppm , but is challenged in the courts by API Benzene w as voluntarily w ithdraw n from consum er products in the United States US Supreme Court issues the Benzene Decision severely limiting regulatory actions N ew benzene standard of 1 ppm . Th is 10-year delay caused more than 200 deaths in the United States Studies showing benzene-related diseases from 1 ppm level of exposure Petrol contains benzene, giving public exposure risk 4.5. References ACGIH , 1946. Transactions of the Eighth A nnual Meeting, American Conference of Governmental Industrial Hygienists, p. 40, Cincinnati, Ohio. ACGIH , 1948. Transactions of the Tenth Annual Meeting, American Conference of Governmental Industrial Hygienists, p. 31, Cincinnati, Ohio. A CG IH , 1957. Transactions of the Nineteenth A nnual Meeting, American Conference of Governmental Industrial Hygienists, p. 47, Cincinnati, Ohio. Aksoy, M., 1977. `Leukem ia in workers due to occupational exposure to ben zen e', New Istanbul Contrib. Clin. Sci. Vol. 12, pp. 3-14. Aksoy, M., 1978. `Benzene and leukaem ia', Lancet Vol. I, p. 441. Aksoy, M., 1987. `Chronic lymphoid leukaem ia and hairy cell leukaem ia due to chronic exposure to benzene: report o f three cases', Br. J. Haematol. Vol. 66, pp. 209-211. of chronic benzene poisoning in 217 w orkers', Br. J . Industrl. Med. Vol. 28, pp. 296 302. API, 1948. Toxicological Review: Benzene, Departm ent o f Safety, Am erican Petroleum Institute, New York, September. Askey, J . M., 1928. `Aplastic an em ia due to benzol poiso n in g', Calif. West. Med. Vol. 29, pp. 262-263. Bergsagel, D. E., Wong, O., Bergsagel, P. L., Alexanian, R., A nderson, K., Kyle, R. A. and Raabe, G. K., 1999. `Benzene and multiple myeloma: Appraisal o f the scientific evidence', Blood Vol. 94, pp. 1174-1182. Bond, G. G., M cLaren, E. A., Baldwin, C. L. and Cook, R. R., 1986. `An update o f mortality am ong workers exposed to ben zen e', Br. J. Indust. Med. Vol. 43, pp. 685 691. Browning E., 1965. Toxicity and metabolism of industrial solvents, pp. 3-65, Elsevier Publishing Co., New York. Aksoy, M., Dincol, K., Akgun, T., Erdem , S. Budinsky, R. A., DeM ott, R. P., Wernke, M. J. an d Dincol, G., 1971. `H aem atological effects and Schell, J . D., 1999. `An evaluation o f Table 4.1. 48 Late lessons from early warnings: the precautionary principle 1896-2000 modeled benzene exposure and dose estimates published in the Chinese-National Cancer Institute collaborative epidemiology stu d ies', Reg. Toxicol. Pharm. Vol. 30, pp. 244-- 258. Castlem an, B. I. an d Ziem, G. E., 1988. `Corporate influence on threshold limit values', Am. J . Ind. Med. Vol. 13, pp. 531-559. Castlem an, B. I. an d Ziem, G. E., 1994. `Am erican Conference o f Governm ental Industrial Hygienists: Low threshold of credibility', Am. J. Ind. Med. Vol. 26, pp. 133 143. Crum p, K. S., 1994. `Risk o f benzene-induced leukem ia: A sensitivity analysis o f the Pliofilm cohort with additional follow-up and new exposure estim ates', J. Tox. Environ. Health. Vol. 42, pp. 219-242. D eCoufle, P., Blattner, W. A. an d Blair, A., 1983. `Mortality am ong chemical workers exposed to benzene and other agents', Environ. Res. Vol. 30, pp. 16-25. Descham ps, F., 1995. `U n leaded gasoline: a continuing h azard ', Occup. Med. Vol. 45, pp. 55. D olore, P. an d Borgom an o, C., 1928. `Acute leukem ia in the course o f benzene poisoning: The toxic origin of certain acute leukemias and their relationship to severe an em ia', J. Med. Lyon. Vol. 9, pp. 227-233. D osem eci, M., Yin, S-N., Linet, M. et al., 1996. `Indirect validation o f benzene exposure assessm ent by association with benzene p oiso n in g', Environ. Health. Perspect. Vol. 104, No 6, pp. 1343-1347. Erf, L. A. an d R hoads, C. P., 1939. `The hematological effects of benzene (benzol) p oiso n in g', J . Industrl. Hyg. Toxicol. Vol. 21, pp. 421-435. Flandrin, G. and Collado, S., 1987. `Is male predom inance (4 /1 ) in hairy cell leukaem ia related to occupational exposure to ionizing radiation, benzene and other solvents?', Br J. Haematol. Vol. 67, pp. 119-120. Flury, F., 1928. `M odern occupational intoxications from the aspect of pharm acology and toxicology II: M odern occupational intoxications', Arch. Exp. Path. Pharmakol. Vol. 138, pp. 65-82 (G er), cited in US Departm ent o f Health, Education and Welfare, 1974, Occupational exposure to benzene, criteria for a recommended standard, pp. 74-137, PHS, CDC, NIOSH. Goguel, A., Cavigneaux, A. and Bernard, J., 1967. `Benzene leukemia in the Paris region between 1950 and 1965: A study o f 50 cases', Nouv. Rev. Fr. Hemato. Vol. 7, pp. 465-480. Goldstein, B. D., 1990. `Is exposure to benzene a cause of hum an multiple m yelom a?', Ann. N Y Acad. Sci. Vol. 609, pp. 225-230. Goldstein, B. D. and Shalat, S. L., 2000. `The causal relation between benzene exposure and multiple m yelom a', Letter to the Editor, Blood Vol. 95, pp. 1512-1513. G reenburg, L., 1926. `Benzol poisoning as an industrial h azard', Pub. Health. Rep. Vol. 41, pp.1516-1539. Hayes, R. B. et al., 1996. `Mortality am on g benzene-exposed workers in C hina', Env. Health. Perspect. Vol. 104, N o 6, pp. 1349 1352. Hayes, R. B., Yin, S-N, D osem eci, M, Li, G-L, Wacholder, S, Travis, LB, Li, C-Y Rothm an, N, Hoover, RN, Linet, MS., 1997. `Benzene and the dose-related incidence of hematologic neoplasms in China', J . Natl. C ancer Inst. Vol. 89, pp. 1065 1071. Hayes, R. B. et al., 1998. `Benzene an d the dose-related incidence o f hematologic neoplasm s in China -- R esponse', Letter to the Editor, J. Natl. Cancer Inst. Vol. 90, pp. 470-471. Hayes, R.B., Yin, S, Dosemeci, M and Linet, M. Com m entary: Benzene and lymphohematopoietic malignancies in China. Am. J. Industrl. Med. 2001 (in press). Helmer, K. J., 1944. `A ccum ulated cases o f chronic benzene poisoning in the rubber industry', Acta Medica Scand. Vol. 118, pp. 254-275. H ernberg, S., Savilahti, M., Ahlman, K. and Asp, S., 1966. `Prognostic aspects o f benzene p oiso n in g', Br. J. Indust. Med. Vol. 23, pp. 204-209. H ogan , J . F. an d Shrader, J . H ., 1923. `Benzol p oiso n in g', Am. J . Pub. Health. Vol. 13, pp. 279-282. Benzene: an historical perspective on the American and European occupational setting 49 H ogstedt, B., H olm en, A., Karlsson, A., Raihle, G., Nillius, K. and Vestlund, K., 1991. `Gasoline pum p mechanics had increased frequencies and sizes o f m icronuclei in lymphocytes stim ulated by pokeweed m ito gen ', Mutat. Res. Vol. 263, pp. 51-55. Hunter, F. T., 1939. `Chronic exposure to benzene (benzol) II: The clinical effects', J. Industrl. Hyg. Toxicol. Vol. 21, pp. 331-354. IARC, 1974. IARC Monographs on the evaluation of carcinogenic risk of chemicals to man: Some anti-thyroid and related substances, nitrofurans and industrial chemicals, Vol. 7, pp. 203-221, International Agency for Research on Cancer, World H ealth Organization, Lyon, France. Infante, P. F., 1987. `Benzene toxicity: Studying a subject to d eath ', Am. J. Ind. Med. Vol. 11, pp. 599-604. Infante, P. F., 1992. `Benzene an d leukem ia: The 0.1 p p m A CG IH p rop o sed TLV for ben zen e', Appl. Occup. Environ. Hyg. Vol. 7, pp. 253-262. Infante, P. F., 1993. `State o f the science on the carcinogenicity o f gasoline with particular reference to cohort mortality study results', Environ. Health Perspect. Vol. 101, No 6, pp. 105-105. Infante, P. F., 1995a. `Benzene an d leukem ia: Cell types, latency an d am oun t o f exposure associated with leukem ia', in Imbriani, M., Ghittori, S., Pezzagno, G. and Capodaglio, E. (eds), Update on benzene, advances in occupational medicine and rehabilitation, pp. 107-120, Fondazione Salvatore Maugeri Edizioni, Pavia, Italy. Infante, P. F., 1995b. `C ancer in blue-collar workers: Who cares?' New Solutions Vol. 5, pp. 52-57. Infante, P. F., Rinsky, R. A., Wagoner, J . K. an d Young, R. J . , 1977a. `Leukaem ia in benzene w orkers', Lancet Vol. II, pp. 76-78. Infante, P. F., Rinsky, R. A., Wagoner, J . K. an d Young, R. J ., 1977b. `Benzene and leu kaem ia', Lancet Vol. II, pp. 867-868. Infante, P F., Schwartz, E. an d Cahill, R., 1990. `Benzene in petrol: a continuing h azard', Lancet Vol. 335, pp. 814-815. Ireland, B., Collins, J . J., Buckley, C. F. and Riordan, S. G., 1997. `Cancer mortality am ong workers with benzene exposu re', Epid. Vol. 8, pp. 318-320. IUD, 1980. AFL-CIO vs. A m erican Petroleum Institute, 1980, 448 U S 601, Industrial Union Department. Jakobsson , R., Ahlbom , A., Bellander, T. and Lun dberg, I., 1993. `Acute m yeloid leukem ia am on g petrol station atten dan ts', Arch. Environ. Health. Vol. 48, pp. 255-259. LeNoir and Claude, 1897. `On a case of purpura attributed to benzene intoxication', Bui. Mem. Soc. Med. Hop. Vol. 3, pp. 1251 1261. Lumley, M., Barker, H. and Murray, J. A., 1990. `Benzene in p etrol', Lancet Vol. 336, pp. 1318-1319. Mallory, T. B., Gall, E. A. an d Brickley, W. J . , 1939. `Chronic exposure to benzene (benzol) III: The pathologic results',J. Industrl. Hyg. Toxicol. Vol. 21, pp. 355-377. Maltoni, C. and Scarnato, C., 1979. `First experim ental demonstration of the carcinogenic effects of benzene: Long-term bioassays on Sprague-Dawley rats by oral adm inistration', Med. Lav. Vol. 70, pp. 352 357. Infante, P. F., 1997. `Quantitative risk from leukem ia/lym phom a from occupational exposure to b en zen e', The Toxicology Forum 22nd A nnual Winter Meeting, pp. 485-488, Toxicology Forum, Inc., Washington, DC. McMichael, A. J., Spirtas, R., Kupper, L. L. and Gam ble, J . F., 1975. `Solvent exposure and leukemia am ong rubber workers: An epidem iologic study', J. Occup. Med. Vol. 17, pp. 234-239. Infante, P. F. an d DiStasio, M. V., 1988. `Occupational benzene exposure: Preventable d eath s', Lancet Vol. I, pp. 1399 1400. Naizi, G. A. an d Flem ing, A. F., 1989. `Blood dyscrasia in unofficial vendors of petrol and heavy oil and motor mechanics in N igeria', Tropical Doctor Vol. 19, pp. 55-58. Nilsson, R. I., Nordlinder, R. G., Tagesson, C., Walles, S. and Jarvholm , B. G., 1996. 50 Late lessons from early warnings: the precautionary principle 1896-2000 `G enotoxic effects in workers exposed to low levels o f benzene from gasolin e', Am. J. Ind. Med. Vol. 30. myeloma in workers exposed to benzene', Letter to the Editor, Occup. Environ. Med. Vol. 53, pp. 357. NTP, 1986. Toxicology and carcinogenisis studies of benzene (CAS No 71-43-2) in F 3 4 4 /N rats and B6C3F1 mice (Gavage studies), N IH Publ. No 86-2545, National Toxicology Program Technical Report 289, Research Triangle Park, NC. Olson, R. E., 1977. Testimony before the US Departm ent of Labor, Occupational Safety and Health Administration, Washington, DC, OSHA Docket No H-059, Exhibit No 149E. OSHA, 1977a. `O ccupational exposure to benzene: Emergency temporary standards h earin g', May 3, Part IV, Fed. R egister 42, pp. 22516-22529. Savitz, D. A. an d Andrews, K. W., 1997. `Review o f epidem iologic evidence on benzene and lymphatic and hem atopoietic can cers', Am. J . Ind. Med. Vol. 31, pp. 287-95. Schwartz, E., 1987. `Proportionate mortality ratio analysis o f autom obile m echanics and gasoline service station workers in New H am psh ire', Am. J . Ind. Med. Vol. 12, pp. 91 99. Shu, X . O., Brinton, L. A., Linet, M. S., Tu, J. T., Zheng, W. an d Fraum eni, J . F., 1988. `A population-based case-control study of childhood leukem ia in S h an gh ai', CancerVol. 62, pp. 635-643. OSHA, 1977b. `O ccupational exposure to benzene: Emergency temporary standards hearing', May 27, Part VI, Fed. Register 42, pp. 27452-27478. OSHA, 1978. `O ccupational exposure to benzene: Occupational safety and health stan d ard s', February 10, Part II, Fed. Register 43, p p .5918-5970. OSHA, 1987. `O ccupational exposure to benzene: Final ru le ', Septem ber 11, Part II, Fed. Register 52, pp. 34460-34578. Paxton, M. B., Chinchilli, V. M., Brett, S. M. and Rodricks, J . V., 1994. `Leukem ia risk associated with benzene exposure in the Pliofilm cohort: Mortality update and exposure distribution', Risk Analysis Vol. 14, p p .147-154. Tabershaw, I. R. an d Lam m , S. H ., 1977. `Benzene and leukaem ia', Letter to the Editor, Lancet Vol. II, pp. 867-868. Tareeff, E. M., Kontchalovskaya, N. M. and Zorina, L. A., 1963. `Benzene leukem ias', Acta UnioInternat. Contra Cancrum Vol. 19, pp. 751-755. Tondel, M., Persson, B. and Carstensen, J., 1995. `Myelofibrosis and benzene exposure', Occup. Med. Vol. 45, pp. 51-52. Vigliani, E. C., 1976. `Leukem ia associated with benzene e xp o su re', Ann. NY Acad. Sci. Vol. 271, pp. 143-151. Vigliani, E. C. and Saita, G., 1964. `Benzene an d leukem ia', N. Engl. J. Med. Vol. 271, pp. 872-876. Rawson, R., Parker Jr., F. an d Jack so n Jr., H., 1941. `Industrial solvents as possible etiologic agents in m yeloid m etaplasia', Science Vol. 93, p p .541-542. Wong, O ., 1987a. `An industry wide mortality study o f chemical workers occupationally exposed to benzene I: G eneral results', Br. J. Ind. Med. Vol. 44, pp. 365-381. Rinsky, R. A., Smith, A. B., H ornung, R., Filloon, T. G., Young, R. J., Okun, A. H. and Lan drigan, P. J ., 1987. `Benzene and leukemia: An epidem iologic risk assessm ent', N Eng. J. Med. Vol. 316, pp. 1044-1050. Santessen, C. G., 1897. `Chronische vergiftungen mit steinkohlentheerbenzin: Vier T odesfalle' Arch. Hyg. Bakteriol. Vol. 31, pp. 336-376. Savitz, D. A. an d Andrews, K. W., 1996. `Risk of myelogenous leukaem ia and multiple Wong, O ., 1987b. `An industry wide mortality study o f chemical workers occupationally exposed to benzene II: Dose response analyses', Br. J. Ind. Med. Vol. 44, pp. 382-395. Wong, O ., 1995. `Risk o f acute myeloid leukaem ia and multiple myeloma in workers exposed to ben zen e', Occup. Environ. Med. Vol. 52, pp. 380-384. Wong, O ., 1996. `Risk o f myeloid leukaem ia and multiple myeloma in workers exposed to Benzene: an historical perspective on the American and European occupational setting 51 benzene-response', Letter to the Editor, Occup. Environ. Med. Vol. 53, pp. 357-358. Wong, O ., 1998. `Re: Benzene and the doserelated incidence o f hematologic neoplasms in C hina', Letter to the Editor, J. Natl. Cancer Inst. Vol. 90, pp. 469-470. Wong, O ., 1999. `A critique o f the exposure assessment in the epidemiologic study of benzene-exposed workers in China conducted by the Chinese Academy o f Preventive Medicine and the U S National Cancer Institute', Reg. Toxicol. Pharm. Vol. 30, p p .259-267. Young, R. J., Rinsky, R. A., Infante, P. F. and Wagoner, J. K., 1978. `Benzene in consumer prod u cts', Science Vol. 199, pp. 248. 52 Late lessons from early warnings: the precautionary principle 1896-2000 5. Asbestos: from 'magic' to malevolent mineral David Gee and Morris Greenberg `Looking back in the light ofpresent knowledge, it is impossible not tofeel that opportunities for discovery and prevention of asbestos disease were badly missed.' T hom as Legge, ex C h ief M edical Inspector o f Factories, in Industrial maladies, 1934. 5.1. Introduction On 20 May 2000, the family o f a senior U K hospital surgeon was awarded GBP 1.15 million in compensation for his death, at 47, from the asbestos cancer, mesothelioma. The disease was caused by exposure to `blue' asbestos dust from dam aged pipe insulation which was present in the com m unication tunnels under Middlesex Hospital, London, where the surgeon worked for four years as a student and trainee, during the period 1966-73 (British MedicalJournal, 2000). The m ain cause o f m esoth eliom a is asbestos. It is now estimated that some 250 000 cases of m esotheliom a, which is normally fatal within one year, will occur in the E u ro pean U nion (EU) over the next 35 years (Peto, 1999). As asbestos is also a cause o f lung cancer, the total disease burden could be around 250 000-400 000 deaths, including cases of the lung disease, asbestosis, which was the first disease to be associated with asbestos exposure. Figure 5.1. shows the peak o f asbestos imports into the United Kingdom being followed som e 50-60 years later by the estimated peak of mesotheliomas. Ninety years before this environmental exposure in the London hospital occurred, a new global public health hazard was born when m ining for chrysotile ( `white') asbestos began in Thetford, Canada, in 1879. Some years later, two other types o f asbestos, `b lu e ' (crocidolite) and `brown' (am osite) cam e to be m ined in Australia, Russia, South Africa and other countries, and the annual production o f all types o f asbestos worldwide grew to 2 million tonnes in 1998. Imports into the EU peaked in the mid-1970s and rem ained above 800 000 tonnes a year until 1980, falling to 100 000 tonnes in 1993. Today, a substantial legacy o f health and contamination costs has been left for both mining and user countries, and asbestos use is continuing, now largely in developing countries. Asbestos: from 'magic' to malevolent mineral 53 T he focus o f this chapter is prim arily on the United Kingdom but the histories o f asbestos have been sim ilar in France, Germany, Italy, Scandinavia and the United States (Castlem an, 1996), as well as in the main m ining countries of Australia, Canada, Russia and South Africa. These histories are now bein g repeated, albeit with som e differences, in Asia, Africa and South America. 5.2. The first 'early warnings' of asbestosis and some responses Within 20 years o f the start-up o f asbestos mining, over 100 products m ade from the `magic m ineral' had been developed, but reports o f serious disease had also begun to appear. The earliest account o f the health hazard of w orking with asbestos was provided by Lucy Deane, one of the first Women Inspectors of Factories in the UK. Writing in 1898, Deane included asbestos work as one of the four dusty occupations which came under observation that year, `on account o f their easily dem onstrated danger to the health of workers and because of ascertained cases of injury to bronchial tubes and lungs medically attributed to the employment o f the sufferer'. She went on to observe that: `the evil effects of asbestos dust have also instigated a m icroscopic exam ination o f the m ineral dust by HM Medical Inspector. Clearly revealed was the sharp glass-like ja g g e d nature o f the particles, and where they are allowed to rise and to rem ain suspended in the air o f the room in any quantity, the effects have been found to be injurious as might have been expected.' (Deane, 1898) Two similar observations by W omen Inspectors followed in 1909 and 1910. They appeared in the annual reports o f HM Chief Inspector o f Factories, which were widely circulated am ongst policy-makers and politicians. The observations o f these laywomen might not have been categorised as `expert opinion' but they were com petent observers whose discussion of occupational disease would have done credit to a m edical scientist. Their reports were not refuted but simply ignored. London, saw the first reported case o f lung disease attributed to inhaled asbestos dust in a 33-year-old m an. In M urray's words: `H e had been at work som e fourteen years, the first ten of which he was in what was called the carding room , which he said was the m ost risky part o f the work. H e volunteered the statement that of the ten people who were working in the room when he went into it, he was the only survivor. I have no evidence except his word for that. He said they all died somewhere about thirty years o f age.' (Murray, 1906) This observation was brought to the attention of the UK government inquiry into compensation for industrial diseases in 1906. In the sam e year, a French Factory Inspector reported some 50 deaths am ongst female asbestos textile workers (Auribault, 1906). This repo rt dealt with the nature o f asbestos, its processin g an d uses, safety an d health hazards in the spinning and weaving processes, and designs for apparatus to capture dust at source. It too was largely ignored, but it was the French ban on asbestos, som e 90 years later, which led to the high-profile case at the World Trade Organization (WTO) in 1999, discussed below. The French report provided confirmation of the earlier observations o f the British Women Inspectors. However, the 1906 British government inquiry did not include asbestos as a cause of industrial disease. Dr Murray had stated in evidence: `one hears, generally speaking, that considerable trouble is now taken to prevent the inhalation of the dust, so that the disease is not so likely to occur as heretofore.' (Murray, 1906) This may have influenced the committee. However, no attem pt was m ade to check on the truth o f Dr M urray's patien t's claim about the deaths o f nine fellow workers. Nor were the surviving workers at that factory investigated, despite the proposals from Lucy Deane about the kinds o f mortality statistics that would be helpful. Dr M urray's view that `no evidence o f h arm ' is the sam e as, `there is evidence o f no h arm ', is an early exam ple o f a com m on fallacy that has inhibited the identification of many dangerous substances which were initially considered to be harmless ( `false negatives'). O ne year after Lucy D e an e 's report, Dr Other evidence about the hazards of asbestos M ontague Murray o f Charing Cross Hospital, was noted in workers in 1910 (Collis, 1911) 54 Late lessons from early warnings: the precautionary principle 1896-2000 and in pioneering dust experim ents with rats in 1911 (Merewether and Price, 1930), and this was later considered to have been `reasonable grounds for suspicion that the inhalation o f m uch asbestos dust was to some extent harm ful', such that the Factory D epartm ent pressed for the installation of exhaust ventilation in the dusty processes (Merewether, 1933). However, subsequent Factory Departm ent inquiries in 1912 and 1917 found insufficient evidence to justify further action. Meanwhile in the United States and Canada, insurance companies had seen enough p roo f o f asbestos disease by 1918 to decline insurance cover for asbestos workers `due to the assum ed injurious conditions in the industry' (Hoffm an, 1918). Unfortunately, this early precautionary action was later forgotten, such that asbestos costs to U S insurers becam e hugely dam aging in the 1990s. In 1924 in Rochdale, hom e o f the Turner Brothers asbestos factory since 1880, was the first inquest and pathological examination of an asbestos worker. Nellie Kershaw was diagnosed as having died o f asbestos poisoning by her local doctor, Dr Joss, who observed that he saw 10-12 such cases a year. H is view was corroborated by pathologist Dr W. Cooke, who wrote the case up in the medical literature (Cooke, 1924 and 1927). In Leeds, where another Turner Brothers factory was situated, a local doctor had found enough asbestos cases to produce a doctoral thesis (Grieve, 1927). By 1930 there had been at least 12 deaths am ongst workers from these two factories with asbestosis cited as the cause or partial cause (Tweedale, 2000). In some cases tuberculosis, heart failure and pneum onia com plicated the diagnosis, as they did for the next few decades. However, the com bination o f at least som e of this evidence with two other reports in the medical literature in 1928 (Sim pson, 1928; Seiler, 1928), including four cases from South Africa, was sufficient to prom pt a m ajor government inquiry into the effects of asbestos dust by Dr Merewether, Medical Inspector o f Factories, an d C. W. Price, a Factory Inspector and pioneer of dust m onitoring and control. It included the first health study of asbestos workers and found that 66 % of those employed for 20 years or m ore suffered from asbestosis, com pared to none of those employed for less than four years, with an average o f 25 % for the 363 workers studied (Merewether and Price, 1930). This was probably an underestimate, as only current workers were examined, excluding those who had left employment through ill health. However, these results led, in 1931, to the first asbestos dust control regulations, medical surveillance and com pensation arrangem ents in the world. These rem ained largely unaltered (but also unenforced) until 1969, when new asbestos regulations were introduced in the United K in gd om . 5.3. Early warnings on asbestos cancers In 1932 in a report to the Trades Union Congress (TU C), a freelance investigator, Ronald Tage, drew attention to three asbestosis cases from the Cape Asbestos Company in Barking, London, that were complicated by cancer (Greenberg, 1993). R eports o f lung cancers being associated with asbestos appeared in the US, Germ an and U K m edical literature in the 1930s and 1940s (Lynch and Smith, 1935; Gloyne, 1935; Wedler, 1943; Heuper, 1942) including the 1938 Report o f the Chief Inspector of Factories. In 1938, when lung cancer was generally much less prevalent, the German authorities were persuaded that the association was causal, and asbestos lung cancer was m ade a com pensatable industrial disease in 1943. (D ecades later the complication of smoking-induced lung cancer m ade the link with asbestos that m uch harder to prove.) A high rate o f lung cancer found at autopsy in asbestosis cases was reported in the Chief Inspector o f Factories' Annual Report in 1949 an d industry h ad two un published U S reports o f an excess o f respiratory cancers in mice (Scheper, 1995). Three in-house investigations o f cancer mortality in the Rochdale district had failed to find evidence o f lung cancer in asbestos workers (Knox, 1952 and 1964), but the company doctor adm itted that his knowledge o f statistics was `nil' (Tweedale, 2000, p. 148). In 1953, Turner Brothers asked Richard Doll, an independent epidemiologist, to study the mortality of Rochdale asbestos workers. He found a lung cancer risk in those who had been exposed to asbestos for 20 years or m ore which was 10 times that expected in the general population. Despite attempts by the Turner directors to suppress these findings they were published in the m edical literature (Doll, 1955). However, it was to be another 30 years before the government accepted lung cancer from asbestos as a Asbestos: from 'magic' to malevolent mineral 55 compensatable industrial disease, and then only if it was accom panied by asbestosis. This was partly because future studies were to be increasingly com plicated by the rising trends o f lung cancer caused by smoking, which Doll had also discovered in a study o f British doctors in 1955. Later studies of asbestos workers showed that the com bination o f the two carcinogens, cigarette smoke and asbestos, multiplied the risks o f lung cancer. Asbestos alone increased the lung cancer risk 5-fold, and smoking alone in creased the risk 10-fold, but the two together p rod u ced n ot 15 times the risk (an additive effect) but over 50 times the risk, a multiplicative or synergistic effect. (Hammond, 1979). Smoking and radiation from uranium and other mining has a similarly synergistic effect in radiation exposed workers (Archer, 1973). This was devastating news because the exposure needed to cause m esotheliom a seem ed to be a m atter o f m onths only. In contrast, m ost lung cancer and asbestosis cases seem to n eed 10 or m ore years o f exposure to asbestos dust. The average latent period between first exposure and the m esotheliom a cancer appearing was about 40 years, in contrast to the 20-25 years for lung cancer. W agner's p ap er provided evidence o f a very strong association between asbestos and mesotheliom a, but by 1964 m ost experts accepted that the relationship was causal, based mainly on the studies o f Dr Selikoff in the United States and Dr Newhouse in the United Kingdom. Both worked independently o f the industry, using case data from unions and hospital records respectively. As with all other hum an studies o f asbestos exposure, there was only a relatively small num ber o f workers who had been working in the `new conditions' o f im proved dust control for the 20-25 years before lung cancer could appear, so it was not possible to say what the risks were in 1955 until many m ore years had passed, when again, in conditions o f improving dust control, it was impossible to know what the risks could be. This problem , which m ight be called the `latency lacu n a', characteristic o f all longlatent-period hazards under conditions of technological change, is a m ajor reason why preventative action is often too late. 5.4. Early, devastating warnings about mesothelioma cancer Cases o f mesothelioma, a normally very rare cancer o f the lining o f the chest or abdomen, h ad been observed in association with asbestos exposure in the 1940s and 1950s, but it was not until 1955 that Dr Sleggs, a local doctor in South Africa, noted a num ber of these unusual cancers at the centre o f the asbestos m ining areas and sent som e to a pathologist, Dr Wagner. T he association with asbestos was m ade and they toured the m ining areas trying to reconstruct the history of asbestos exposure o f those who had died by talking to colleagues and families. Out of 47 cases o f mesotheliom a, they found earlier asbestos exposure in all but two, an d m any o f them were environmental cases including children exposed when playing on waste dumps. They published their findings in 1960 (W agner et al., 1960). Selikoff had observed that 15 ou t o f 17 patients from the sam e asbestos manufacturing plant had asbestos diseases, but as he was refused access to company records he used trade union records to show that the users of asbestos, such as insulation workers, were at even greater risk than m anufacturing workers: o f the 392 workers exam ined with 20 years or m ore asbestos exposure, 339 had asbestosis. The lung cancer rate was seven times norm al, and a num ber h ad m esoth eliom a (Selikoff et al., 1964). The excess o f lung cancer only becam e statistically clear after 25 years of follow-up o f workers, illustrating one o f the serious limitations of so called `negative' cancer studies, which is that the power of such studies to detect long-latent-period cancers can be very low unless som e 20-30 years o f follow-up has been possible. Selikoff was to be described as a `disturbing sore thum b' by an industry representative from the Asbestos Textile Institute (Tweedale, 2000, p. 183, footnote 17). This was a sim ilar sentim ent to that expressed by the ex Chief Medical Inspector of Factories, Dr Legge, writing in his then capacity as M edical Adviser to the TUC in 1932 about Ronald Tage, whom he said the TUC could be `quit o f' by paying him a small fee (Greenberg, 1993). The practice o f attacking the purveyors of news about hazards had been well illustrated by Ibsen in his play, An enemy of the people (1882), in which the local doctor notices a health hazard which, if fully recognised, 56 Late lessons from early warnings: the precautionary principle 1896-2000 would threaten the economy o f the local town. H e descends from public hero to public enemy as the economic implications o f his observations come to be realised by the mayor, the m edia and m ost o f the citizens. and, for that matter, the product' (Sells, 1994). 5.5. Actions and inactions by regulatory authorities and others Newhouse used the long-term pathology records collected by the London Hospital between 1917 and 1964 to show that, o f 76 mesothelioma cases, over 50 % had occupational or domestic exposure (lived in the house o f an asbestos worker), whilst of the others, one third lived within half a mile o f the Cape asbestos factory (Newhouse and Thom pson, 1965). It was to be 30 years later that children exposed to asbestos in the neighbourhood of factories and who later developed m esotheliom a becam e the first successful environmental exposure cases against Turner Brothers in the United Kingdom (Tweedale, 2000, p. 272). Both Newhouse and Selikoff presented their findings at a conference in New York in O ctober 1964, organised by the New York A cadem y o f Sciences. A study o f D o ll's group o f workers from the regulated areas o f the Rochdale factory was presented as supportin g the view that `it is possible that the specific occupational hazards to life have been completely elim inated', perhaps another example of the `latency lacuna' (K nox et al., 1965) But neither Selikoff nor the UK Factory Departm ent found such evidence o f falling disease rates, mainly because they were including severe cases o f dust exposure am ongst the users, n otju st the manufacturers of asbestos, for whom dust conditions were relatively better, at least in the regulated factory production areas. This failure to appreciate the `worst case' asbestos exposure scenarios was part o f the reason for the delayed and inadequate responses to asbestos. Ju lian Peto, asbestos cancer researcher, has described the focus of asbestos cancer studies on factories, rather than users, as a `stupid mistake' (Peto, 1998). This view was shared by a form er director o f the w orld's biggest asbestos company, Jo h n s Manville, when reviewing why, although still profitable, it filed for bankruptcy in 1982 as a m eans o f dealing with asbestos pollution claims. He argued that medical research, assiduous communication, insistent warnings and a rigorous dust reduction programm e `could have saved lives and would probably have saved the stockholders, the industry The asbestos regulations o f 1931 were only partially enforced, there being only two prosecutions between 1931 and 1968 (Dalton, 1979). Their focus on ju st parts of the m anufacturing process m eant that the riskier user activities were neglected. However, the issue o f dangerous asbestos was not neglected. From 1964 to 1975 the m edia in both the United States and the United Kingdom kept asbestos high on the political agenda (Sunday Times, 1965). T he ITV program m e The World in Action in 1971, and the BBC Horizon in 1975, abo u t conditions at C a p e 's Acre Mill asbestos plant in Yorkshire, United Kingdom, helped to initiate action by authorities, such as a Parliamentary Om budsm an Report into asbestos regulation enforcem ent at the factory. This report was initiated by local MP M ax Madden, who lodged an official com plaint against non-enforcem ent o f the asbestos regulation o f 1931. The report was very critical o f the Factory Inspectors, and the governm ent responded by appointing a governm ent inquiry, the Simpson Com mittee, in 1976. Meanwhile, the 1931 asbestos regulations had been updated in 1969, and a limit for factory asbestos dust exposure of 2 million fibres per cubic metre (m 3) o f air was to be gradually introduced. Unfortunately this `hygiene standard' did not include consideration o f the lung or m esotheliom a cancer hazards. It was later to be strongly criticised an d associated with high asbestosis levels (one worker in 10 would get the disease) byJu lian Peto in evidence to the Sim pson inquiry (Peto, 1978). T he Sim pson R eport cam e ou t in 1979 with the following recom m endations: a ban on `blue' asbestos, which had already been withdrawn by the industry; a ban on insulation spraying, which was also largely defunct by then; contractors were to be licensed for the removal o f asbestos; and the asbestos limit was to be reduced to 1 million fibres/m 3 (or 1 fibre per millilitre) by 1980 for `w hite' asbestos, with a target o f 0.5 m illion fib res/m 3 (0.5 f/m l) for `brown' asbestos, which was considered to be m ore hazardous than white. An asbestos fibre that Asbestos: from 'magic' to malevolent mineral 57 is visible is abo u t the diam eter o f a hum an hair, 40 microns, but it is com posed o f a bundle o f about 2 million fibrils that can be released by abrasion or physiological processes in the body (Selikoff and Lee, 1978). Electron m icroscopy is n eed ed to accurately monitor the presence of such fibrils in air or tissues. T here was, and still is, scientific controversy about the relative cancer and asbestosis potencies o f the three types o f asbestos, with white often being regarded as less hazardous than blue or brown. By 1986 the World H ealth O rganization's International Agency for Research on Cancer (IARC) had concluded that all three types were carcinogenic and, as with other carcinogens, there was no known safe level o f exposure to any o f them. There was no similar hygiene standard for the pub lic's protection from airborne asbestos dust until the late 1980s, when the lowest limit detectable by the prevailing dust m onitoring m ethod, optical microscopy, was recom m ended by the U K Health and Safety Executive (100 000 fib re s/m 3 or 0.1 f/m l). In 1982, Yorkshire TV screened a two-hour docum entary at prim e viewing time, featuring Alice Jefferson, a 47-year-old who had contracted mesotheliom a when working for a few m onths at C a p e 's Acre Mill asbestos plant. Alice, a Fightfor Life had an im m ediate impact, even though some, like Sir Richard Doll, criticised the program m e for being unscientific and emotional. The government responded to the program m e by im plem enting the Sim pson inquiry recom m endations and, in 1984, by introducing the asbestos licensing regulations and a further lowering o f the exposure limits to 0.5 m illion fibres/m 3 (0.5 f/m l) for white and 0.2 m illion fibres/ m3 (0.2 f/m l) for brown asbestos. A voluntary labelling schem e was introduced for some uses. Pressure for further improvements continued to com e from local MPs, som e trade unions and people representing victims, such as asbestos widow, Nancy Tait. She helped reveal that the Turner Brothers' asbestos compensation arrangements, such as GBP 1 a week for the widows o f workers with asbestosis, had been largely unchanged since the 1930s. H er work helped to force Turner Brothers to improve their com pensation awards. New regulations were introduced in 1987 and tightened further in 1989. In 1998 the government adopted a ban on all forms of asbestos, which was im plem ented the following year, along with an EU ban, which is to be im plem ented by M em ber States by 2005. Canada filed a trade barriers complaint against the French and EU ban at WTO, but this was rejected by the WTO Disputes Panel. Canada appealed against this ruling to the WTO Appellate Body, which found in favour o f France and the EU (see Box 5.1.) Box 5.1. W TO upholds French and EU ban on asbestos In 1997 France banned all form s of asbestos fibres and products in order to protect the health of workers and consumers. Existing 'white' asbestos products could be exem p t on an excep tional, tem porary and annually review ed basis, if no effective substitute materials were available that posed a lower health risk to w orkers handling them . Canada objected to this ban at the W TO but the W T O found in favour of France in Septem ber 2000 (W TO , 2000). Canada appealed to the Appellate Body the W TO and the EU cross-appealed to uphold the main findings of the panel and to seek correction of some 'errors' of the panel's interpretations and conclusions. The US cross-appealed against the panel's judgem ent that glass fibres were as carcinogenic as asbestos. The Apellate Body issued a report in early 2001 (W TO , 2001), out of which a number of main points arose, which also have implications for other hazardous agents: all form s of asbestos ('w hite', 'brow n' and 'blue') are carcinogenic; there is no known threshold of safety for this carcinogen; the risk from 'w hite' asbestos in products is based on evidence which 'tends to show' a risk rather than not; w orkers handling asbestos products, such as building and brake lining w orkers, are at risk from asbestos exposure; there is no W T O requirem ent for countries to provide quantitative risk assessm ent data: qualitative evidence is sufficient; countries can base their health/ environment/ animal welfare measures on qualified and respected scientific opinions held by only a m inority of scientists: 'a M em ber is not oblig ed, in setting health policy, autom atically to follow w hat, at a given tim e, may constitute a majority scientific opinion' (p. 64). This means that a W TO Panel need not necessarily reach its decision, on the scientific evidence, based on a 'preponderant weight of the evidence', but on a lower level of proof; the efficiency of the 'controlled use' of asbestos products was not demonstrated and the residual risk to the workers would still be significant; this risk m anagem ent option could not be relied on to protect workers' health, and therefore was not a reasonable 'alternative' measure to the asbestos ban; in determ ining w hether asbestos substitutes such as glass fibre were 'like' products, four criteria have been developed by W TO , including the properties and end uses of a substance, and the tastes and habits of consum ers. Based on these criteria, the Appellate Body found that the panel had erred in finding that glass fibre products were 'like' products: they were not, principally because they were not as carcinogenic. 58 Late lessons from early warnings: the precautionary principle 1896-2000 The WTO procedures for dealing with the kind of scientific and technological complexity involved in asbestos and other health and environmental hazards has been criticised by one o f the scientific advisers involved in the WTO asbestos case (Castleman, 2001). Meanwhile the annual U K cancer death rate from mesothelioma and lung cancer from asbestos is estim ated by the H ealth an d Safety Commission (Health and Safety Commission, 1994--95) to be aroun d 3 000 deaths per year and rising (see Figure 5.1.). Despite huge amounts of research, many issues of biological mechanisms and dose-response relationships rem ain unclear, illustrating the limited relevance o f m ore research to disease. 5.6. The costs and benefits of actions and inactions It is beyond the scope o f this case study to provide a detailed evaluation o f the full costs and benefits o f the asbestos story (see Castlem an, 1996, p. 8 -9 ). However, a few illustrative figures will indicate the dim ensions o f such an evaluation. At a company level, Turner Brothers m ade arrangem ents in 1994 to pay up to GBP 1 billion in asbestos claims. The insurance underw riters Lloyd's o f L o n d o n faced near collapse in the early 1990s from U S pollution claims, many o f which were for asbestos health compensation and clean-up costs. If lives are valued at EU R 1 m illion each, which is com m on in transport studies, then the costs o f the estim ated 400 000 European asbestos cancer deaths expected over the next few decades is EU R 400 billion. The hum an costs in terms of suffering are not calculable. Removing asbestos from buildings safely at the end o f their life will cost further billions. Earlier actions to reduce asbestos exposure could have saved many o f these costs. A Dutch illustration of some of the potential savings from earlier risk reduction actions has estimated that a ban in 1965, after the m esotheliom a evidence had been widely accepted, instead o f in 1993, would have saved the country some 34 000 victims and N LG 41 billion in building and com pensation costs. This is com pared to the 52 600 victims and NLG 67 billion guilders in costs expected over the period 1969- 2030, estimated by the Dutch Ministry o f Health and Social Security (Heerings, 1999). In the United States, asbestos compensation settlem ents reach ed USD 2 billion, with Lloyds syndicates paying around h alf o f that. On the other hand, asbestos has brought some benefits, including employment. In 1919 it was estim ated that fires in the w orld's theatres in the 1870s and 1880s caused 2 216 deaths, 95 % o f which could have been saved by asbestos fire insulation, it was claim ed (Summers, 1919). Asbestos boiler insulation saved energy, and asbestos brake linings saved lives, though the extra vehicle speeds it allowed com plicates the picture. T he Lancet, a U K m edical jo u rn al, argu ed in 1967 that `it would be ludicrous to outlaw this valuable and often irreplaceable material in all circumstances (as) asbestos can save m ore lives than it can possibly endan ger' (Lancet, 1967). Apart from their gross underestimate o f the health im pacts o f asbestos, for which their expertise was at least relevant, the replaceability o f asbestos is a technological and econom ic question which doctors are not well qualified to ju d ge. Little evidence was presented to substantiate their argum ent that asbestos was `irreplaceable'. Substitutes for m ost uses o f asbestos were available by the 1970s and, in some cases, much earlier - many US oil refineries were insulated with m ineral wool in the 1940s and 1950s (Castleman, 1996, pp. 456-457). The slow spread o f asbestos substitutes was partly because asbestos industry cartels worked to inhibit their spread (Castleman, 1996, pp. 34- 38), and partly because the m arket price o f asbestos was very low com pared to its full production, health and environmental costs. This failure o f m arket prices to reflect full environmental and health costs is the comm on cause o f delay in replacing hazardous materials. Manyjobs, much profit and high dividends were generated by asbestos. Turner Brothers' profits rose strongly after 1947 and peaked at almost GBP 9 million a year in 1965 (Tweedale, 2000, p. 9). These profits suffered little from the ill health and contam ination costs o f asbestos, which were `externalised' onto workers with disease, their families, the health service, insurance carriers and building owners. An often ignored but significant non financial benefit of asbestos compensation trials is that they frequently uncover m any o f the contradictions between company words Asbestos: from 'magic' to malevolent mineral 59 on asbestos and their actions to reduce hazards (Castleman, 1996). 5.7. What are the lessons of the asbestos story? Asbestos offers many lessons that are relevant to num erous other agents or activities that have long-term hazardous impacts. 1. T he experiences o f victims, lay people and `competent observers', such as factory inspectors and family doctors, should be taken seriously by governmental and other authorities, and followed up by appropriate investigations. They can anticipate the views o f scientific experts, som etim es by many years. 2. The early warnings o f 1898-1906 in the United Kingdom and France were not followed up by the kind o f long-term m edical and dust exposure surveys o f workers that would have been possible at the time, and which would have helped strengthen the case for tighter controls on dust levels. Even now, leadin g asbestos epidem iologists can conclude: `It is unfortunate that the evolution o f the epidemic o f asbestos-induced m esotheliom a, which far exceeds the combined effects o f all other known occupational industrial carcinogens, cannot be adequately m onitored.' (Peto, 1999) Long-term environmental and health m onitoring rarely meets the short-term needs o f anyone, thus requiring particular institutional arrangem ents if it is to m eet society's long-term needs. 3. T he laws on prevention and com pensation introduced in the United Kingdom in 1931-32 were not well im plem ented, and the sanctions were trivial, a pattern that was repeated down the lon g history o f asbestos.4* 4. If early warnings had been heeded, and better control m easures adopted, either before 1930 as Dr Legge, Chief Medical Inspector of Factories, and others (Greenberg, 1994; Bartrip, 1931) have noted, or in the 1950s and 1960s, when new cancer hazards em erged and economic circumstances were good, then m uch tragic loss would have been avoided. Action to curb asbestosis prior to the discovery o f the cancers would at least have minimised the im pact of these later `surprises'. M ore strategically, tighter regulation o f asbestos would have raised its m arket price to capture m ore o f its costs o f production and use, thereby stimulating the innovation that belatedly led to better and often cheaper substitutes, as well as to improved engine and building designs that generate, at source, less waste heat. 5. Econom ic factors played a key role as in other cases o f worker, public and environmental hazards. These include the em ployers' need for profits and the workers' need for jobs, which can together produce an alliance which may not be in the long-term interests o f workers or society. The greater the size of the `external' cost of harm (dam age costs not borne by the com panies), the greater the chance that these diverging private an d social costs will inhibit preventative action. Only when full dam age costs, including health, building maintenance and site contamination costs, are borne by the polluters via the `polluter pays' principle, and through liability provisions, regulations, taxes, etc., can the private and social costs o f economic activity be brought closer together, thus allowing the m arket place to operate m ore efficiently. Penalties on wrongdoing by employers also need to be com m ensurate with the costs they inflict on others, if private and social costs and benefits are to be m ore closely aligned. B ut this is n ot easy. It is very difficult for governments to overcome powerful economic interests that usually operate on the same short timescales as most politicians, and to im plem ent decisions that are in the best, longer-term interests o f society, if they are perceived as imposing short-term costs on powerful groups. Again, appropriate institutional arrangem ents are needed to help m eet society's long-term interests: a `governance' issue that is taken up in the final chapters o f this report. 6. O ne o f the m ain reasons for the failure to im plem ent control m easures was the view that `current exposures to asbestos dust are so much lower than past exposures an d should therefore be safe ', a view offered to the UK committee o f inquiry into compensation for industrial diseases 60 Late lessons from early warnings: the precautionary principle 1896-2000 by Dr Murray in 1906, and repeated by many others ever since. As there is a 10 40-year latent period between asbestos exposure and the diseases it causes, by the time that evidence o f `today's ' exposure risks becom es available, many years o f generally reducing dust levels have ensued, making it once again possible to say that risks in the new `today' are much less than in the past, or non-existent. The point cannot be proved decisively one way or the other until another 20-40 years have passed. This `latency lacuna', which is com m on to all long-latent-period hazards, is an illustration o f the com m on error o f assum ing that `absence o f evidence o f harm ' means `evidence of absence of harm '. It does not. In the absence o f good evidence that today's exposures to carcinogens are safe, it is wiser to apply the precautionary principle, and assum e they are unsafe, especially if the disease (or ecological impacts) from higher exposures have no known threshold o f exposure below which there are no effects. This is a key lesson that is relevant to all long-latent-period hazards. The particular preventative m easures that would then be required would depend on the proportionality principle -- the expected benefits from prevention, including any `secondary benefits', would need to be significant in relation to the costs o f achieving such prevention. This m ore precautionary approach to uncertainty and ignorance would also involve switching the current bias within norm al scientific m ethods away from avoiding `false positives' (with its associated bias o f producing `false negatives' such as asbestos) towards producing a better balance between false positives and false negatives. This would increase the chances o f generating the costs o f restricting a substance or activity that might later turn out to be safe. However, the asbestos case strongly suggests that society would gain overall from a m ore ethically acceptable and economically efficient balance between generating false positives and false negatives.7 7. Im plem entation o f preventative m easures was also inhibited by the healthy survivors fallacy. This n eeds to be widely com m unicated an d avoided as it gives rise to a general but false reassurance o f safety, as it does with the gen eral hazard o f sm oking. It was first described in relation to asbestos by Lucy Deane in 1898: `Even when the evil reaches such grave proportions as to be capable o f easy and tragic proof... there is always a certain proportion o f `old workers' -- the survivors o f their mates -- who are found in every unhealthy industry an d who... appear to thrive on their unhealthy calling. In less obvious unhealthy conditions the only convincing p roof o f actual injury, viz., reliable com parative statistics o f mortality, or o f health standards, is practically unattainable in the case o f any given factory, or at any rate with the time an d opportunity at present at our disposal.' (Deane, 1898). This argum ent has been used throughout the history of asbestos. For example, Dr Knox, the Turner Brothers' U K company doctor, on visiting the Canadian asbestos mines in 1952, said: `I am assured that many workers over 70 years o f age are still employed and are active and vigorous.' (G reenberg, 2000) This view was also presented to one o f the authors o f this case study (DG) when he visited UK asbestos plants in the 1980s as the union health and safety adviser. Workers pointed to the retired workers who had worked m ore than 20-30 years or so in the factory without much harm , and who were able to turn up to the annual pensioners' party. Such pensioners were cited as p roo f o f the low or absent risks of asbestos. This could be called the `pensioners' party fallacy', as it was the workers who did not make it to the party who provided the p ro o f o f harm, and their deaths, or illnesses, m ade them relatively invisible to current workers. As Deane observed, healthy survivors needed to be related to non-survivors via appropriately analysed mortality statistics. 8. It seem s necessary to establish speedy, affordable and transparent compensation arrangements, based on agreed liabilities, as soon as any harmful effects becom e known, so as to both increase the incentives to prevent further harm and to improve the chances o f recording accurate exposure histories. Asbestos: from 'magic' to malevolent mineral 61 Elements of such anticipatory com pensation arrangem ents were established in the early days o f the nuclear industry, when the state in many countries took on future liabilities for nuclear accidents, at least up to certain limits (for exam ple the UK Nuclear Installations Act, 1965). A unique example seems to be the radiationinduced cancer compensation scheme for workers at British Nuclear Fuels (see the chapter on Radiation). 9. Views should be taken from a wide range o f all relevant disciplines and the `ignorant expert' should be curbed. Specialists in one discipline, for example medicine, provided `expert' opinions about other disciplines such as dust monitoring and control (occupational hygiene and ventilation engineering) or asbestos substitutes availability. These opinions were often mistaken but went largely unchallenged, and this contributed to misplaced complacency (Greenberg, 2000). 10. It is necessary to anticipate `surprises' and take care with substitutes. If asbestos substitutes reproduce the sam e physical form as asbestos - long, respirable (< 3 m icrons in diam eter) and durable fibres -- it is likely that they too will be carcinogenic (Roller and Pott, 1998), as was predicted by the U K H ealth and Safety Executive in 1979, and later confirmed by the IARC for some forms of synthetic m ineral fibres. However, mineral wool and glass fibre appear to be m uch less hazardous than asbestos, and they can be m anufactured to be good enough for insulation but not as thin, or durable enough in hum an tissue, to be carcinogenic. `Clean' production and user techniques that minimise exposures to atmosphere, whether occupational or environmental, via `closed loop' and eco efficient systems, are therefore essential with whatever m aterials are being used. This then minimises the size o f any future `surprise' impacts from substitutes, which is an im portant benefit o f applying the precautionary principle. Asbestos: early warnings and actions 1898 1906 1911 1911 and 1917 1918 1930 1931 1935-49 1955 1959-60 1962/64 1969 1982-9 1998-99 2000-01 UK Factory Inspector Lucy Deane warns of harmful and 'evil' effects of asbestos dust French factory report of 50 deaths in fem ale asbestos textile w orkers and recom m endation of controls 'Reasonable grounds' for suspicion, from exp erim ents with rats, that asbestos dust is harmful UK Factory Departm ent finds insufficient evidence to justify further actions US insurers refuse cover to asbestos workers due to assumptions about injurious conditions in the industry UK M erew ether Report finds 66 % of long-term w orkers in Rochdale factory with asbestosis UK A sb esto s Regulations sp ecify dust control in m anufacturing only and com pensation for asbestosis, but this is poorly im plem ented Lung cancer cases reported in asbestos m anufacturing w orkers Doll establishes high lung cancer risk in Rochdale asbestos w orkers M esotheliom a cancer in w orkers and public identified in South A frica M esotheliom a cancer identified in asbestos w orkers, in neighbourhood 'bystanders' and in relatives, in the United Kingdom and the United States, am ongst others UK Asbestos Regulations improve controls, but ignore users and cancers UK m edia, trade union and other pressure provokes tightening of asbestos controls on users and producers, and stimulates substitutes. EU and France ban all form s of asbestos W TO upholds EU/French bans against Canadian appeal 5.8. References Acheson, E.D., and Gardner, M.J., 1983 `Asbestos: T he C ontrol Lim it for A sbestos', Health & Safety Executive, HMSO, London. Archer, V.E., et al, 1973, `U ranium M ining and Cigarette Smoking Effects in M an', J. Occ.Med., 15, 204. Auribault, M., 1906. `Sur l'hygiene et la securite des ouvriers dans la filature et tissage d 'am ian te', in A nnual report of the French Labour Inspectoratefor 1906. Table 5.1. 62 Late lessons from early warnings: the precautionary principle 1896-2000 Bartrip, P., 1931. `Too little, too late? T he Hom e Office and the asbestos industry regulations 1931', Medical History Vol. 42, October, pp. 421-438. British MedicalJournal, 2000. Vol. 320, 20 May, p. 1358, at h ttp ://b m j.co m /cg i/fu ll/3 2 0 / 7246/1358/a Castleman, B., 2001. Draft paper to asbestos conference at London School o f Hygiene and Tropical Medicine, 5 June. Castlem an, B. I., 1996. Asbestos: Medical and legal aspects, 4th ed., A spen Law & Business, Englewood Cliffs, NJ. Grieve, I. M. D., 1927. `A sbestosis', MD thesis, University o f Edinburgh. Ham m ond, E.C., Selikoff, I.J., Seidman, H .,'A sbestos Exposure,C igarrette sm oking and Death R ates,' Annals o f New York Academy of Sciences, p 473-490. Health and Safety Commission, 1994-95, Health and Safety Statistics Vol. 55, pp. 148 151. H eerings, H., 1999. `A sbestos -- deep in the very fibres o f society', Contrast Advise study for Greenpeace Netherlands, September, A m ersfoort. Collis, E., 1911. Annual Report ofH M Chief Inspector ofFactories for 1910, H M SO, L on don . Cooke, W. E., 1924. `Fibrosis o f the lungs due to the inhalation o f asbestos dust', British MedicalJournal Vol. 2, 26 July, p. 147. Cooke, W. E., 1927. `Pulm onary asbestosis', British MedicalJournal Vol. 2, 3 Decem ber, pp. 1024-1025. D alton, A., 1979. Asbestos: Killer dust, British Society for Social Responsibility in Science, London. Deane, Lucy, 1898. `Report on the health of workers in asbestos and other dusty trades', in HM Chief Inspector o f Factories and W orkshops, 1899, A nnual Reportfor 1898, pp. 171-172, HM SO London (see also the Annual Reports for 1899 and 1900, p502). H euper, W. C., 1942 , `O ccupational Tum ours and Allied D iseases', Charles C. Thomas, Springfield, Illinois. H offm an, F. L., 1918. `M ortality from respiratory diseases in dusty trades', Bulletin of the US Bureau ofLabor Statistics Vol. 231, pp. 176-180. Knox, J . F., 1952. `Visits to the Thetford Mines, Asbestos, Atlas Works, Keasbey & M attison Works, Raybestos-Manhattan Works', Report to the m anagem ent of Turner Brothers Asbestos, Frames 0000 0070 1950 54 in the Chase M anhattan microfilms. Knox, J . F., 1964. `R eport o f a visit to the Thetford Mines, Asbestos and M ontreal', Report to the m anagem ent at Turner Brothers Asbestos, Discovered documents marked 015039-015041. Doll, R., 1955. `Mortality from lung cancer in asbestos w orkers', Brit. J . Industr. Med. Vol. 12, pp. 81-86. Gloyne, S. R., 1935. `Two cases o f squam os carcinom a o f the lung occurring in asbestosis', Tubercle Vol. 17, pp. 5-10. Knox J . F. et al., (1965) `C oh ort analysis o f changes in incidence o f bronchial carcinom a in a textile asbestos factory', Annals of the NY Acad. of Sciences Vol. 132, Decem ber, pp. 527 35. Lancet, 1967. 17 Ju n e , pp. 1311-1312. G reenberg, M., 1993. `R eginald Tage -- a U K prophet: A postscript', Am. J. Ind. Med. Vol. 24, pp. 521-524. G reenberg, M., 1994. `Knowledge o f the health hazard o f asbestos prior to the Merewether and Price R eport o f 1930', Social History ofMedicine, 0 7 /0 3 /, pp. 493-516. Greenberg, M., 2000. `Re call for an international ban on asbestos: Trust me, I 'm a d o cto r', Letter to the editor, Am. J . Ind. Med. Vol. 37, pp. 232-234. Legge, T., 1934, Industrial Maladies, O xford University press, Oxford. Lynch, K. M. an d Smith, W. A., 1935. `Pulm onary asbestosis 111: Carcinom a o f lung in asbestosis-silicosis', Am. J. Cancer Vol. 24, pp. 56-64. Merewether, E. R. A., 1933. `A m em oran dum on asbestosis', Tubercle Vol. 15, pp. 69-81. Merewether, E. R. A. an d Price, C. W., 1930. Report on effects of asbestos dust on the lungs and Asbestos: from 'magic' to malevolent mineral 63 dust suppression in the asbestos industry, HM SO, London. Murray, H. M., 1906. In D epartm ental Committee on Compensation for Industrial Diseases, 1907, Minutes of evidence, p. 127, paras 4076-4104, Cd 3496, HMSO, London. Newhouse, M. and Thom pson, H., 1965. `M esothelioma of pleara and pertitoneum following exposure to asbestos in the London a re a ', Brit. J . Industr. Med. pp. 261-269. Peto, J., 1978. `The hygiene standard for chrysotile asbestos', Lancet 4 March, pp. 484 489. Peto, J., 1998. `Too little, too late', Interview with Jo h n Waite, BBC R adio 4, 15 October, London. Peto, J., 1999. `The European m esotheliom a epidem ic', B. J. CancerVol. 79, February, pp. 666-672. Selikoff, I. J . et al., 1964. `A sbestos exposure an d n eo p lasia',J . Am. Med. Ass. Vol. 188, pp. 22-26. Selikoff, I. an d Lee, D. H. K., 1978. Asbestos and disease, A cadem ic Press, New York. Sells, B., 1994. `W hat asbestos taught m e abo u t m an agin g risk', H arvard Business Review M arch/April, pp. 76-89. Sim pson, F. W., 1928. `Pulm onary asbestosis in South A frica', British MedicalJournal 1 May, pp. 885-887. Sum m ers, A. L., 1919. Asbestos and the asbestos industry. Cited in Tweedale, P5, fn 10. Tweedale, G., 2000. Magic mineral to killer dust: Turner and Newall and the asbestos hazard, O xford University Press, Oxford. Sunday Times, 1965. `U rgen t probe into `new' killer dust d isease', 31 October, L on don . Roller, M. and Pott, F., 1998. `Carcinogenicity o f man-made fibres in experim ental animals and its relevance for classification of insulation w ools', Eur. J . Oncol. Vol. 3, N o 3, p p .231-239. Wagner, J . C., Sleggs, C. A. an d M archand, P., 1960. `Diffuse pleural m esotheliom a and asbestos exposure in the North Western Cape Province', Brit. J. Indust. Med. Vol. 17, pp. 260-271. Scheper, G. W. H., 1995. `Chronology o f asbestos cancer discoveries: experim ental studies at the Saranac Lab oratory', Am. J . Ind. Med. Vol. 27, pp. 593-606. Seiler, H. E., 1928. `A case o f p n eu m oco n iosis', British MedicalJournal Vol. 2, p. 982. Wedler, H. W., 1943. `U ber den Lungenkrebs bei A sbesttos', Dtsch. Arch. Klin. Med. Vol. 191, pp. 189-209. WTO, 2000. W T /D ST 35/R , 18 September. W TO, 2001. W T /D S 1 3 5 /A B /R , 12 March. 64 Late lessons from early warnings: the precautionary principle 1896-2000 6. PCBs and the precautionary principle Janna G. Koppe and Jane Keys 6.1. Introduction Polychlorinated biphenyls (PCBs) are chlorinated organic com pounds that were first synthesised in the laboratory in 1881. By 1899 a pathological condition named chloracne had been identified, a painful disfiguring skin disease that affected people employed in the chlorinated organic industry. Mass production o f PCBs for comm ercial use started in 1929. Thirty-seven years elapsed before PCBs becam e a major public issue, recognised as environmental pollutants, and a danger to animals and humans. Large-scale production worldwide, and in particular in som e eastern European countries, continued until the mid-1980s. PCBs are the first obvious exam ple of a substance that was not intentionally spread into the environment, but nevertheless becam e widespread and bioaccum ulated to high concentrations. PCBs are mixtures o f synthetic organic chem icals with the sam e basic chem ical structure, and similar physical properties, that range in nature from oily liquids to waxy solids. PCBs were well received in the m arketplace as they replaced products that were m ore flam m able, less stable and bulkier. This new group o f chemicals facilitated the production o f smaller, lighter and what were thought to be safer electrical equipment. In the United States the use o f PCBs was very im portan t in the Secon d World War. PCBs were primarily used in electrical equipment, such as capacitors and transformers, because of their insulating properties and resistance to high tem peratures. Over the years the num ber o f uses o f PCBs increased to include use as heat-transfer fluids in heat exchangers, as hydraulic fluids, as ingredients in PVC plastics, paints, adhesives, lubricants, carbonless copy paper and as im m ersion oil for microscopes. Sealants containing PCBs were widely used for constructing and renovating buildings. Between 1929 and 1988 the total world production o f PCBs (excluding the U SSR an d China) was 1.5 million tonnes. By the late 1930s M onsanto, the US producer o f PCBs, was certainly aware o f adverse health effects in workers exposed to PCBs. For example, in 1936 several workers at the Halowax C orporation in New York City exposed to PCBs (then called chlorinated diphenyls), and related chemicals called chlorinated naphthalenes, were affected by chloracne. Three workers died and autopsies o f two revealed severe liver dam age. Halowax asked H arvard University researcher Cecil K. Drinker to investigate. Drinker presented his results at a 1937 m eeting attended by Monsanto, General Electric, Halowax, the US Public Health Service and state health officials from Massachusetts and Connecticut. Like the Halowax workers, D rinker's test rats h ad suffered severe liver damage. Sanford Brown, the president of Halowax, concluded the m eeting by stressing the `necessity o f not creating m ob hysteria on the part of workmen in the plants' (Francis, 1998). The results were published but did not gain the wider attention o f policy-makers (D rinker et al., 1937). D rinker's article did however put the occupational medicine community, labour regulators and manufacturers on notice as to the concerns surrounding PCBs. The first warning that PCBs were becom ing ubiquitous in the environment came from S 0ren Je n se n . In 1966 Je n se n , while working on DD T (dichlorodiphenyltrichloroethane), fortuitously detected unknown m olecules in the muscle of white-tailed sea eagles in Sweden. The levels were appreciably higher in the fish-eating sea eagles than in fish collected from the same areas. So he concluded that the m olecules m ust be persistent in living tissues and not easily broken down. The mystery chemicals were extremely resistant to degradation, being unaffected even when boiled in concentrated sulphuric acid. It took two further years of study for Jen sen to be able to demonstrate that they were PCBs. In 1969 S0ren Je n se n published his findings (Jensen et al., 1969) which showed remarkably high PCB concentrations in a large proportion o f the Baltic Sea fauna (See Figure 6.1.). PCBs had entered the environment in large quantities PCBs and the precautionary principle 65 for m ore than 37 years and were bioaccum ulating along the food chain. In the 1960s it becam e apparent that the fertility o f all three seal species occurring in the Baltic Sea was in decline. By the 1970s nearly 80 % o f the females were infertile. Som e studies drew a link with the presence o f persistent pollutants -- high levels o f DDT and PCBs had been recorded in all three species. A clear correlation was found between the pathological uterine changes and elevated concentrations of contaminants, particularly PCBs. Further studies appeared to link high PCB levels not only with reproductive disorders in seals but with other symptoms such as dam age to skin an d claws, intestines, kidneys, adren al glands and skeleton (Swedish Environmental Protection Agency, 1998). DDT, PCBs and effects on Baltic Sea fauna World production and bans on use in Sw eden tonnes/yr 80 000 60 000 40 000 20 000 DDT 1 D D T found in B a ltic fauna 1 D D T banned in agriculture 1 D D T banned in fo restry \ 1 P C B s found in B a ltic fauna Known total world 1 `Open' uses of P C B s banned production ^V ----- PCBs `Enclosed' uses banned All u se s of P C B s banned 1 1960 1970 1980 1990 2000 Figure 6.1. Source: Bernes, C. (2001) 'W ill tim e heal every wound' The first well-publicised warning that PCBs could be harm ful to hum ans came from an accident in 1968 in Ja p a n with a mass poisoning am ong 1 800 people who ingested contam inated rice oil. The rice oil was found to contain a large am ount of Kanechlor 400, a brand o f PCB, that was believed to have leaked from a heating pipe in the factory (K im burgh et al., 1987). This resulted in serious health problems for those who had consum ed the contaminated oil (see Box 6.1.). The im pact o f this incident was such as to give rise to a new word in the Japanese language -- Yusho -- or Japanese rice oil disease. Many debates ensued as to whether it was the PCBs or their breakdown products that had caused these effects. However, it was agreed that PCBs that had been subjected to heating did have harmful effects on humans and that if PCBs were present in places where food was processed accidents like this one could happen. Percentage of white-tailed eagle pairs breeding su cce ssfu lly 100 80 a ve rag e level up to 1950s 1970 1980 1990 2000 In the late 1960s press reports about the presence of PCBs in the environment began to appear and in response Monsanto launched its public defence, denying that the chemicals were PCBs. `The Swedish and Am erican scientists... imply that polychlorinated biphenyls are `highly toxic' chem icals,' M onsanto said in a widely distributed statem ent. It continued: `This is simply not true. The source o f marine-life residue identified as PCB is n ot yet known. It will take extensive research, on a worldwide basis, to confirm or deny the initial scientific conclusions.' (Francis, 1998) Nevertheless in 1969 M onsanto privately took a different view in its internal `Pollution abatem ent plan ', which admitted that `the problem involves the entire United States, Canada and sections o f Europe, especially the United Kingdom and Sweden.... other areas o f E urope, Asia an d Latin A m erica will surely becom e involved. Evidence of contamination (has) been shown in some of the very remote parts of the world.' The plan also stated that stopping the production of PCBs was not an option as it would cause `profits to cease and liability to soar because we would be adm itting guilt by our actions' (Francis, 1998). 66 Late lessons from early warnings: the precautionary principle 1896-2000 Box 6.1. The Yusho accident The first clinical sign was a strange skin disease with conjunctivitis, swelling of eyelids and chloracne. The victim s' whole bodies, including the extrem ities, were covered with acne-like pustules. The most common symptoms were pigmentation of nails, skin and mucous m em branes; increased sweating of the palms; severe headache; swollen joints and feelings of weakness. About half of the victims coughed persistently with expectoration resulting in a chronic bronchitis. Th ese respiratory sym ptom s correlated with PCB levels in the blood. It took more than 10 years for these sym ptom s to decline. Some of the victim s were pregnant when they consum ed the PCB-contam inated oil. O f 11 babies born to these mothers, two were stillborn. All were 'C oca-C o la' coloured, grow th-retarded, had increased eye discharges and nail pigm entation. Follow-up studies of the children showed their growth to be ham pered, their IQs to be low, and their dem eanour generally apathetic and dull. There w as an increased m ortality of the whole group of Yusho patients due to malignancies of the liver and respiratory system . Their body burden of dioxins was about 450 micrograms dioxin toxic equivalent level per kilogram, not including the amounts of phenobarbital-like PCBs (Masuda, 1994). 6.2. Growing evidence of persistence, presence and toxicity During the 1970s the evidence that PCBs were ubiquitous environmental pollutants continued to build. PCBs were found in rem ote areas such as the Arctic. In the Netherlands, large amounts of PCBs were found to be entering the environment from the River Rhine. At Lobith, on the Germ an border, inputs m easured between 1976 and 1981 ranged from 14 300 kg to 24 000 kg. M ost o f the PCBs were attached to fine silt in suspension; consequently the levels in the sedim ents were highest where the flow rates o f the rivers were low. T he sedim ents in Rotterdam harbour had the highest concentrations -- 12-24 m g/kg. As these sedim ents were used for reclaim ed land, this resulted in the Netherlands being contaminated at the rate o f 5 000 kg of PCBs a year. Fatty fish caugh t o ff the D utch coast were highly polluted. Eels from the rivers and lakes, tested in 1977 and 1988, contained 3.0-131 m g/k g o f PCBs (CBS, 1980; Greve and Wegman, 1983). PCBs were found in fish, mink, seabirds and in hum ans; these studies provided further evidence that PCBs bioaccumulate. Evidence o f actual or suspected harm caused by their bioaccum ulation was also being docum ented (see the case study on the Great Lakes). What also becam e clearer throughout the 1970s was an understanding o f the m ajor reason for the disagreem ent between those who said PCBs were harm ful in low quantities and those who said they were not. It was found that different forms -- `congeners' -- of PCBs have different num bers and positions of chlorine atoms that determine the m o lecu le's physical an d chem ical properties. Studies in the late 1970s recognised the significance of these different congeners, although at first the differences were wrongly attributed solely to the level of chlorination. This proved too simplistic, and after confusing the debate for a while it becam e evident that both the position and num bers o f the chlorine atoms had an effect on toxicity -- and that different congeners have different effects (see Box 6.2.). 6.3. Action from industry and governments in the 1970s In 1971 M onsanto realised that its public position was untenable and voluntarily limited the types o f PCB mixtures that had the overall group trade nam e Aroclor to those containing less than 60 % chlorine substitution (by weight). They also reform ulated one Aroclor to reduce the percentage of higher-chlorinated PCBs. This move was based on the then current scientific opinion that m olecules with fewer chlorine atom s were less toxic. Unfortunately, as described above, this opinion was soon found to be an oversimplification. In 1972, as a result of the reports concerning the presence and effects o f PCBs in the environment, Sweden banned these substances for `op en ' uses, such as in sealants, paints and plastics, which resulted in uncontrolled losses to the environment. The first international governm ental action was in February 1973, when the Organisation for Econom ic Co-operation and Development (OECD) m ade Council Decision C(73) 1 (Final) on protection o f the environm ent by control o f polychlorinated biphenyls. The O E C D 's decision was taken as a result o f `concerns about environmental contamination by PCBs, and their health and environmental effects' (OECD, 1973). The decision required use in new open products to be banned in OECD m em ber states. However, large am ounts continued to be used in supposedly `closed systems' such as transformers, probably as the technical problem s and costs o f m easures needed to replace them were considered prohibitive. PCBs and the precautionary principle 67 In the United States, Congress also responded to the dangers associated with the use of PCBs and other toxic chemicals. In 1976 the Toxic Substances Control Act was passed. Although the act covered the regulation o f all chemical substances, the severity o f the threat posed by PCBs resulted in a section (6 (e)) devoted solely to PCBs. No other chemicals received such singular attention. During the debate over the Senate version o f the act, Senator Nelson, the author o f Section 6(e), noted that PCBs were w idespread in the environment and that they posed significant potential dangers to hum an health and to wildlife (United States Court of Appeal, 1980). The section required that, one year after the act came into effect, PCBs could only be m anufactured, processed, distributed and used in a `totally enclosed m anner'. Eighteen months later, all manufacture, processing and distribution of PCBs was prohibited. Production in the United Kingdom ended in 1978 and in the U nited States in 1979. However, elsewhere, and in particular in some eastern European countries, large-scale production continued until the mid-1980s (B oersm a et al., 1994). 6.4. Scientific understanding becomes more sophisticated In 1979 cam e another costly rem inder o f how PCBs can cause serious harm to hum an health. In Taiwan 2 000 people were poisoned by polluted rice oil. The accident, nam ed Yucheng, received m ore publicity and follow-up than the Yusho accident o f the 1960s, demonstrating how awareness of PCBs as an environmental pollutant was changing. Cases with the typical clinical features as seen in the Yusho illness were reported. The children born to poisoned mothers became the subjects o f a long-term health study. A quarter o f the children died before the age o f four years as a result o f respiratory infections. At the age o f eight years the children still had nail deform ities and chronic otitis m edia (m iddle ear inflam m ation) together with bronchitis. Follow up studies o f the adult victims o f Yucheng showed an increase in skin allergies, chloracne, headache, spine and join t diseases and goitre (Guo, 1999). A further developm ent in the understanding o f the phenom ena occurred in the 1980s, when it was realised that PCBs change during bioaccum ulation and biodegradation in the environment. This allowed many o f the earlier inconclusive studies to be reinterpreted and the apparently conflicting evidence to be resolved. Bioaccumulation through the food chain tends to concentrate congeners of higher chlorine content, producing residues that are considerably different from the original Aroclor mixtures (Schwartz et al., 1987; Oliver and Niimi, 1988). Bioaccum ulated PCBs appear to be more toxic than commercial PCBs because some toxic congeners can be preferentially retained (Aulerich et al., 1986; Hornshaw et al., 1983). M ink that were fed G reat Lakes fish contam inated with PCBs showed liver and reproductive toxicity com parable to mink fed three times the quantity of Aroclor 1254 (H ornshaw et al., 1983). Evidence o f PCB contamination of human breast milk also becam e apparent in the 1980s. Due to the high level o f PCBs in the Dutch environment, the levels of these chemicals in the Dutch population remain am on g the highest in the world, with average PCB content in adipose tissue ranging from 1.6-2.5 m g/k g fat (Greve and Wegman, 1983). In the 1980s came the first published studies on possible developmental effects in children. Examples of endocrine disruption becam e apparent. The penises of boys affected by the Yucheng accident were underdeveloped and -- another sign of endocrine disruption in the prenatal period -- their understanding of spatial relationships was dim inished, an abnorm ality also seen in boys of mothers using ph en obarbital in pregnancy (D essens et al., 1998). Children are m ore vulnerable to the toxic effects o f PCBs than adults because they have developmental periods, so called `windows o f tim e', when m etabolic systems develop (see also the case study on DES). During the intrauterine development o f a child and in the postnatal period, many processes take place that fix homeostatic systems to states that then persist throughout life. Exam ples include a steady body tem perature o f 37 degrees Celsius; the setpoints for circulating horm one levels; and the functioning o f the im m une system. This `fetotoxicity' represented a new paradigm for toxicology, one where both dose and timing is im portant. A natural exam ple is the exposure to viral rubella infection in the first three m onths of gestation when the different organs are formed. D epending on the day o f development of the organs the heart or a leg or an eye or the brain is m alform ed. 68 Late lessons from early warnings: the precautionary principle 1896-2000 Figure 6.2. Source: Soren Jensen Figure 6.3. Source: Soren Jensen Figure 6.4. Source: Soren Jensen Biphenyl skeleton structure 4 Chlorine substitutions in positions 2 and 6 in the biphenyl are known as ortho-position substitutions, in 2 and 4 as metaposition substitutions, and at the four locations as a paraposition substitution. Dibenzofuran skeleton structure 1 9 Dibenzo-p-dioxin skeleton structure Box 6.2. Explanation of the toxicity of PCBs by Soren Jensen The basic biphenyl chemical form is shown in Figure 2. Theoretically 209 chlorinated biphenyls are possible, but the greater the number of chlorine atoms already present in the m olecule, the more difficult it becom es for further substitution of hydrogen with chlorine. Only 135 congeners have been found in technical products or in biota. In most PC B congeners chlorine atom s are quite evenly distributed between the two phenyl rings. From a toxicological viewpoint positions 2 and 6 are the most important (see Figure 6.2.) -- these are called the ortho positions. The chlorine atom is rather bulky, so many chlorine substitutions in position 2 and 6 will force the biphenyl to twist around the axis. Most m olecules have two or more chlorine m olecules in the ortho position. Less frequent are m olecules with only one chlorine atom in an ortho position, and in this case the m olecule is more planar (flat). Finally the non-ortho congeners, called the co-planar PC Bs, are only present in trace am ounts in technical products. Polychlorinated dibenzofurans (see Figure 6.3.) are created as accidental by-products during the ironcatalysed synthesis of PC B. They are also formed during fires in oxygen-poor environm ents such as fires in PCB-containing transform ers; in products used as heat-transfer agents; and in other types of PCB wastes exposed to heat or fire. Polychlorinated dibenzo-p-dioxins (see Figure 6.4.) are closely related to the dibenzofurans. The 2,3,7,8-tetra chloro congener was first found as a by-product in the 2,4,5-trichloro-phenoxy-acid herbicides. TC D D (tetrachlorodibenzoparadioxin) is also form ed in fires w here chlorine is present. T C D D is the m ost toxic substance of which we know, this being the result of its interaction with a vital cell cytoplasm protein known as the Ahreceptor, which in turn affects enzym e production. Th is effect, via the A h -recep tor, is often sim ply referred to as a 'dioxin-like effect'. T C D D is a liver and nerve, and bone-marrow toxin and carcinogenic in human beings. Different chlorinated dioxins have been given a TEF-value (toxic equivalent factor), where the ability of TC D D to induce aryl-hydrocarbonhydroxylase activity via the Ah- receptor was given the value 1. The different chlorinated dibenzofurans could be given a TEF-value as they have been proven to possess a dioxin-like effect. G enerally the effect is som ew hat lower than the dioxin with the same substitution pattern. The ability to possess a dioxin-like effect is linked both to the planarity of the m olecule and the specific chlorine substitution pattern. Some PCB congeners are planar enough to have a small dioxin-like effect -- especially the non-ortho congeners. The sam e is true, at a still lower TE F value, for those PCB molecules with only one ortho chlorine. The total dioxin-like effect is, however, the TEF-value multiplied by the total amount present. As a result the net dioxin-like effect of the one ortho congeners can be substantial, as they are present in organism s at a relatively high level. The m ajority of the PC B-congeners in biological sam ples as well as in technical products have two or more ortho substituents; they are too twisted to bind to the A h -recep tor and their TEF-value is zero. They possess however a phenobarbital-like effect, affecting various enzym es in a different manner from dioxins. PCBs and the precautionary principle 69 Prenatal exposure to PCBs and furans (polychlorinated dibenzofurans, or PCDFs, see B ox 6.2.) was associated with negative behavioural effects am ongst the Yusho offspring. The children were apathetic and uninterested in their surroundings (Harada, 1976). Monkeys prenatally exposed to PCBs exhibited hyperactive behaviour during infancy, followed by inactivity at four years of age (Bowman and Heironimus, 1981). The four-year-old children o f m others whose diets had included significant am ounts o f fish from Lake Michigan also displayed reduced activity in relation to their body burdens o f PCBs (Jacob son et al., 1990). Furtherm ore, at the age o f 11 years, children m ore highly exp osed in the pren atal p eriod h ad lower IQtest scores, and showed difficulties in verbal com prehension and reduced ability to concentrate. They were also m ore than twice as likely to be two years beh in d in reading skills and word com prehension (Jacobson and Jacobson, 1996). Further research has demonstrated negative effects as a result of prenatal exposure to background levels o f PCBs in the United States, Canada and western Europe (see Box 6.3.). 6.5. Government action in the 1980s and 1990s The North Sea ministerial conferences, in conjunction with the Oslo and Paris conventions on the protection o f the N ortheast Atlantic, have been im portant international fora that have stimulated action on hazardous substances. At the first North Sea conference in 1984, it was agreed that the phasing out o f the use and discharge o f PCBs should be intensified. At the second North Sea conference in 1987 targets were agreed to reduce discharges of `substances that are toxic, persistent and liable to bioaccum ulate' by the order o f 50 % by 1995 (DoE, 1987). fires.'(OECD, 1987). The decision recom m ended that m em ber countries should cease the manufacture, import, export and sale of PCBs by 1 January 1989. The decision also called for the acceleration o f the withdrawal o f PCBs from use. During the 1980s som e national governments issued health advice suggesting the reduction o f consum ption o f PCB-contaminated fish and time limits on breast-feeding. The advice was controversial and heavily debated because o f the positive benefits to health from both breast-feeding and the consum ption o f fish (Fuerst et al., 1992). In 1990 the third North Sea conference agreed a specific plan to phase out the use o f PCBs by 1999 and safely dispose o f them by the same date (DoE, 1990). In 1995 the Barcelona Convention for the Protection of the M editerranean Sea against Pollution agreed to `reduce, by the year 2005, discharges and emissions which could reach the m arine environment, o f substances which are toxic, persistent and liable to bioaccum ulate, in particular organohalogens, to levels that are not harm ful to m an or nature with a view to their gradual elim ination'. In the same year Sweden prohibited the use of old equipm ent which contained PCBs. At the May 1995 m eeting o f the United Nations Environment Program me (UNEP) Governing Council, Decision 18/32 was adopted on persistent organic pollutants (POPs). This decision resulted in a num ber o f studies and m eetings culminating in N ovem ber 1995 with the W ashington Declaration, an agreem ent to a global program m e o f action to phase out POPs, including PCBs. The declaration was signed by 100 national governments. In 1987, the OECD m ade a further decision on PCBs, on the basis that `current controls o f polychlorinated biphenyl have not led to a clear and consistent downward trend of environmental levels o f PCBs. Previous concerns about environmental contamination by PCBs, and their health and environmental effects rem ain unabated. New concerns have arisen over the use o f PCBs particularly in situations where highly toxic products such as chlorinated dioxins or chlorinated dibenzofurans might be produced by their decom position in In 1996 Directive E C 96/59 o f the European Union called for the elimination of PCBs and PCTs (polychlorinated triphenyls) and their phase-out by 2010. However, transformers filled with PCB oil are still in use. As they become older and rust, the chances of leakage into the environment increase. Some 100 years after the first serious adverse effects had been docum ented, closure had finally been reached on the seriousness of the threat. However, the toxic legacy will remain for many decades. 70 Late lessons from early warnings: the precautionary principle 1896-2000 Box 6.3. Further research on fetotoxicity In Rotterdam and G roningen in the N etherlands, starting in 1990 and 1991, a study w as performed to investigate the effects of prenatal exposure to background levels of PCBs and dioxins on growth and developm ent of the child (Huisman, 1996; Koopman-Esseboom, 1995; Patandin, 1999; Lanting, 1999). The total study group consisted of 400 healthy mother-infant pairs, of which half the infants w ere breast-fed and half bottle-fed. Prenatal PCB exposure was estimated by the PCB-sum (PCB congeners 118, 138, 153 and 180) in maternal blood and cord blood and the total dioxin toxic eq uivalent level (TEQ ) in the m other's breast milk (17 dioxin and a total of 8 dioxin-like PCB congeners -- 3 planar, 3 mono-ortho and 2 di-ortho PCBs). Postnatal dioxin exposure w as calculated as a product of the total dioxin T E Q level in breast milk multiplied by the weeks of breast-feeding. O f the measured PCB congeners 118, 138, 153 and 180, the first is dioxin-like, but the last three are phenobarbital-like. In general 63 % of the total am ount of PC B s in human breast milk is orthosubstituted non-planar (PCB-22, -52, -138, -153 and -180) -- that is phenobarbital-like P C B s. C urrent body exposure to the PCB congeners 118, 138, 153 and 180 w as m easured in plasma at 42 months of age (Patandin, 1999). The study detected hyperactivity and slower mean reaction tim es in relation to the current PCB levels in the children at 42 months of age. Irritability and hyperactivity are well-known side-effects of the use of phenobarbital in childhood. A t the age of 42 months attention during free play behaviour was reduced relative to umbilical cord PCB concentrations and maternal PCB exposure. This persistent effect on behaviour from damage that happened prenatally is sim ilar to that found by Jacobson (Jacobson e t al., 1990; Patandin, 1999). Effects of prenatal PCB exposure w ere also adversely associated with neurological outcome at 18 months of age (Huisman e t al., 1995), but this was no longer seen at 42 months of age (Lanting, 1998b). In contrast, prenatal and lactational exposure to dioxin-like PCBs and dioxins was not shown to affect attention and activity at 42 months of age (Patandin, 1999). Thus, negative effects on cognitive and behavioural developm ent, as demonstrated by Jacobson and the Dutch study, are related to the prenatal or current accumulated exposure to phenobarbitallike PC Bs and not to dioxin-like PC Bs. A study by Seegal and Schantz (Seegal and Schantz, 1994) also demonstrated the different effects of phenobarbital-like PCB congeners and dioxin-like PCB congeners. Monkeys exposed to di-ortho PCBs were impaired on simple spatial discrimination and reversal problem s. However, those monkeys exposed to TC D D (tetrachlorodibenzoparadioxin) performed better than the control group. The di-ortho-substituted PC B-congeners tested in adult m onkeys are dopamine neurotoxicants. They reduce the amount of dopamine by inhibiting tyrosine hydroxylase -- the enzyme which controls the rate at which dopam ine is synthesised. It is probable that this is a long-term or even permanent effect (Seegal and Schantz, 1994). A particular concern is this may have im plications relating to an increased incidence of Parkinson's disease. D essens found im paired spatial ability in human adults w ho had been prenatally exposed to anticonvulsants (mostly phenobarbital) (Dessens et al., 1998). The new disease entity 'late haemorrhagic disease of the new-born', first detected in the late 1970s in Jap an and w estern Europe and originally attributed to vitamin K deficiency, might also be related to the effects of phenobarbital-like PCBs (Koppe e t al., 1989; Bouwman, 1994). Prenatal exposure to background levels of dioxin-like PCB congeners has also been shown to affect thyroid hormone m etabolism (Pluim et al., 1992; Koopman-Esseboom et a l, 1994). 6.6. Routes of environmental exposure When products containing PCBs decompose, or are destroyed by fire, significant amounts o f the PCBs survive, because o f their stability. During fires in oxygen-poor environments, and where PCBs are otherwise broken down, highly toxic polychlorinated dibenzofurans can be form ed and released (see Box 6.2.). There have also been accidental releases of PCBs and associated contaminants into the environm ent via leakage from `sealed ' PCB fluid compartments during commercial use o f transformers and capacitors, and from the im proper disposal of equipm ent or products. By the late 1990s, even organisations generally critical o f `over-regulation' o f industry had accepted that: `In the past, discharges o f PCB-laden wastes into rivers, streams, and open landfills were considered acceptable, legal, and hazard-free practices. PCBs were also som etim es intentionally released into the environment -- for exam ple, to reduce dust emissions from dirt roads, or as extenders in some agricultural pesticide formulations. In retrospect, these practices were inappropriate and potentially harm ful.' (ASCH, 1997) Once released, PCBs can volatise or disperse as aerosols, providing an effective means of transport in the environment. Strong south to north air flows, especially over west Eurasia, selectively favour the accumulation o f PCBs and certain pesticides in the Arctic, once regarded as a pristine environment (AMAP, 1997). PCBs, together with the pesticides DDT, HCH and HCB, are detectable at low concentrations (0.01-40 nanogram s (one thousand-millionth of a gram ) per gram dry weight) in all sam ples of freshwater surface sediments from Alaska, northern Canada, Greenland, Norway, Finland and Russia (1995 and 1996). Freshwater and m arine ecosystems tend to contain higher levels o f PCBs than terrestrial ecosystems. Biom agnification o f PCBs is especially significant in food webs dominated by organism s with a high fat content. Additionally, species overwintering at lower attitudes deliver POPs and m etal burdens to the Arctic during the sum m er (Holden, 1970; AMAP, 1998). H um ans absorb PCBs from ingestion, inhalation and through the skin, although over 90 % o f exposure is via food (Theelen and Lie, 1997). Animal fat is the m ajor PCBs and the precautionary principle 71 source of PCBs. Once absorbed, PCBs equilibrate am ong lipid compartm ents in the whole body due to their high affinity for, notably, triglycerides and cholesterol esters and their lower affinity for phospholipids and cholesterol (Lan ting et al., 1998a). In the liver PCBs interfere with enzyme systems im portant for detoxification (Matthews and Anderson, 1975). 6.7. The most recent PCB accident Despite the various international and national regulations and legislation, releases, whether accidental or deliberate, are inevitable. In Belgium, in January 1999, chickens were found to have been fed a m ixture o f anim al fats m ixed with 8-50 litres o f PCBs and furans. Although the source of PCBs has not been proven it was strongly suspected that they derived from the illegal disposal o f old transform ers. Levels in chickens and eggs were high, varying for 2,3,4,7,8 PCDF from 1 299 picogram s (one million-millionth o f a gram ) per gram in chicken fat to 1 103 p g /g in egg yolk. Total dioxin (toxic equivalent level or TEQ ) was respectively 958 and 685 p g /g fat (Hens, 1999). According to WHO guidelines the tolerable daily intake or TDI for people is 1-4 picogram s dioxin per kg bodyweight per day (WHO, 1999). The discovery was only m ade because the levels were so high as to cause chick edem a (swelling o f tissue due to increased fluid content). It is likely that other food contamination incidents have occurred at lower levels before an d will continue to occur in future. This was underlined in 2000, a year later, when in another part o f Belgium more anim al food was found to be contam inated with PCBs an d dioxins. In 1999 Belgian toxicologists published an analysis o f the incident and suggested that it was very unlikely that the isolated episode o f contamination in Belgium would have caused health effects in the general population. Without providing data on the background levels o f PCBs in the Belgian population, they speculated that a two- to three-fold increase would be com parable with levels in the 1980s of those regularly eating contaminated seafood (B ern ard et al., 1999). This fails to take into account Dutch studies showing the effects o f background levels o f PCBs and dioxins on unborn babies. Since the PCB levels in the Belgian population are com parable to those o f the Dutch, their statem ent is som ew hat surprising and appears reminiscent o f Sanford Brown o f the Halowax Corporation in 1937 when endeavouring `not to alarm the workforce'. One can legitimately wonderju st how much we have learned since then. 6.8. Conclusion By the 1930s there was already evidence, som e at a low level o f proof, that PCBs could poison people. This inform ation was largely retained within the industry, and does not appear to have been widely circulated am ongst policy-makers or other stakeholders. The application o f the precautionary principle at that time would have prevented the toxic legacy that now exists. Thirty years later, by the end o f the 1960s, there was a high level o f proof, mainly due to the Yusho accident, that in certain circumstances PCBs, or their breakdown products, such as dibenzofurans, could cause serious harm to hum an health. The findings o f S 0ren Je n se n also offered a high degree o f p roof that PCBs did bioaccum ulate and were presen t in the Baltic fo od chain. A lower, but still substantial level o f p roo f o f the adverse reproductive effects on m arine m am m als was also available from the Baltic seal studies. H ad precautionary action at a level o f proof less than `beyond reasonable dou bt' been acceptable to, an d applied by, policy-makers o f that era, their action would still have resulted in a m ore m anageable, less costly problem than we are faced with today. Many years of use o f PCBs would have been avoided. The 1960s evidence, with attendant worries o f future liabilities, is likely to have been a factor that influenced M onsanto to reform ulate som e o f its PCB products in 1971. Unfortunately this action was based on interpretation o f the then current, but incomplete, scientific knowledge of how PCBs could cause harm. During the 1970s evidence increased both o f the environmental transport o f PCBs to the rem otest parts o f the world and o f the potential o f PCBs to cause harm. PCBs were becom ing widespread in the environment even though generally not being intentionally spread. It took until the early 1970s for any governm ent to act and even then these actions only affected new `open ' uses of PCBs. By the late 1970s some governments had accepted that there was a greater risk of harm and enacted legislation to stop new `closed' uses o f PCBs. At this 72 Late lessons from early warnings: the precautionary principle 1896-2000 point there was no action by any governm ent to address the problem o f existing uses or the cleaning up o f contam inated sites. It is probable that the technical difficulties and costs o f such actions were the reasons behind this h alf m easure. By the late 1970s, a few countries had called for all production to end. By this time alternatives to PCBs in closed uses were available. In the 1980s a deeper understanding developed as to how PCBs caused harm. This did m uch to increase the level o f certainty because som e discrepancies between studies could now be explained as due to different PCB congeners. During this time the first evidence that PCBs could affect the unborn child was published. The North Sea states agreed that there was a danger and in 1987 adopted the political aspiration to reduce inputs to the marine environment by `the order o f' 50 %, by 1995. Also in 1987 the OECD accepted that the level o f concern had increased and current legislation had not been effective. At this point all OECD countries agreed to end all new uses o f PCBs by 1989. This time the risk posed by existing uses o f PCBs was fully acknowledged and the introduction of `controls' was recom m ended, as well as the removal o f PCB equipm ent in certain circumstances. Nevertheless, production of PCBs still continued in several countries. would have allowed an earlier resolution. However, it could also be argued that the call for more science can also be used as a reason to delayjustifiable action. The resistance o f PCBs to degradation means they are expensive to destroy. As a result o f the many delays, a large percentage o f historic PCB production has escaped beyond our control into the environment. Wide dispersal, and bioaccumulative properties, m eans that in many cases PCBs are in places where their recovery and destruction is not possible. Also, in many countries where they are still in use, the facilities for safe destruction do not exist. In 1999, during the negotiations for a global convention to phase out all POPs, attention was focused on PCBs still in use, particularly those posing a risk due to accident or leakage. While this focus on prevention o f further releases is essential, it is also necessary to pay greater attention to our response to the high levels o f PCBs that are already in the environment. Behavioural problems and respiratory diseases affecting children, two o f today's m ost im portant problem s in paediatrics, could be due, in substantial part, to intoxication with PCBs. T here is an urgen t n eed to find ways o f reducing current body burdens of these chemicals in people. O f course, no equivalent action is possible for other species. Although nine states within the North Sea catchm ent agreed in 1990 to phase out the use o f all PCBs, it took until the mid-1990s, and the UNEP W ashington Declaration, for a significant global response to emerge, acknowledging the problem and agreeing the need to act on PCBs in use. Even today som e o f the science is still being debated. M uch o f this is to do with the different effects o f commercial PCB mixes and those found in the environment, especially those that have been bioaccumulated. It could be argued that further scientific study at an earlier time PCBs also highlight issues such as whoju dges what risks are acceptable, and whether all stakeholders are fairly represen ted in this debate. Both PCBs and dioxins are fetotoxic. Is it acceptable to tolerate risks for involuntary exposure to unborn babies, when this may affect their future capabilities and those of their offspring? At alm ost every stage governm ent action was taken only when there was a high level o f scientific proof. The non-application o f the precautionary principle has left us with a legacy, the total effects and costs o f which can only be guessed at. PCBs and the precautionary principle 73 PCBs: early warnings and actions 1899 1929 1936 1937 1966 1968 1970s 1972 1976 1979 1980s 1990s 1996 1999 Chloracne identified in w orkers in chlorinated organic industry Mass production of PCBs for commercial use begins More workers affected by chloracne and liver damage Chloracne and liver dam age observed in exp erim en ts with rats. Results did not gain attention from policy-makers but both labour regulators and manufacturers were made aware of the concerns surrounding PCBs Jen sen discovers unknown m olecules in sea eagles in Sw eden -- only in 1969 w as he able to demonstrate that they were PCBs Poisoning of 1 800 people who had ingested PCB-contam inated rice oil in Jap an gives rise to a new Jap anese w o rd : Yusho -- rice oil disease, and to the first well-publicised warning that PC B s are harmful to humans High levels of PC B s found in infertile seals of three different species Sweden bans 'open' uses of PCBs T o xic Substances Control A c t (United States) -- PC B s to be used only in a 'totally enclosed m anner' 2 000 people again poisoned, in Taiw an, by polluted rice oil. Follow-up research showed that 25 % of children born of poisoned mothers died before the age of four years Evidence of PCB contam ination of breast milk PC B s associated with IQ and brain effects in children exposed in utero to m others' PCBcontaminated diets. Fetotoxicity represents a new paradigm for toxicology EU directive to eliminate PCBs, with phase-out by 2010 C hicken food contam inated with PC B s is found in Belgium 6.9. References AMAP, 1997. Arctic pollution issues: A state of the Arctic environment report, Arctic M onitoring and Assessment Programme, Oslo, at http/ / :w w w .am ap.no/am ap.htm (polychlorinated biphenyls, lead and cadmium) in the areas of Prague (Czech Republic) and Katowice (Poland). C om parison with referen ce values in the N etherlan ds', Central European Journal of Public Health Vol. 2, pp. 73-76. AMAP, 1998. `Persistent organic p ollutan ts' in de M arch et al. (eds), AMAP assessment report: Arctic pollution issues, AMAP Secretariat, Oslo. Bouwm an, C. A., 1994. Modulation of vitamin K dependent blood coagulation by chlorinated biphenyls and dioxins in rats, Thesis (ISBN : 90 393-0581-1), University o f Utrecht, Utrecht. ASCH, 1997. A position paper of the American Council on Science and Health: Public health concerns about environmental polychlorinated biphenyls (PCBs) (update o f the 1991 report by L. T. Flynn an d C. F. K leim an), A cadem ic Press, at http://w w w .acsh.org/publications/ reports/pcupdate2.htm l Aulerich, R. J . et al., 1986. `A ssessm ent o f primary versus secondary toxicity o f Aroclor 1254 in m in k', Arch. Environ. Contam. Toxicol. Vol. 15, pp. 393-399. Bernard, A., H erm ans, C., Broeckaert, F., De Poorter, G., De Cock, A. an d H ouins, G., 1999. `Food contamination by PCBs and dioxin s', Nature Vol. 401, pp. 231-232. Bernes, C., 2001. Will time heal every wound, Swedish Environmental Protection Agency B oersm a, E. R. et al., 1994. `C ord blood levels o f potentially neurotoxic pollutants Bowman, R. E. an d H eironim us, M. P., 1981. `Hypoactivity in adolescent monkeys perinatally exposed to PCBs and hyperactive as ju v en iles', Neurobehav. Toxicol. Teratol. Vol. 3, pp. 15-18. CBS, 1980. PCBs in Nederland, Centraal Bureau voor de Statistiek, Staatsuitgeverij, `s-Gravenhage. DoE, 1987. Summary of the ministerial declaration from the International Conference of the Protection of the North Sea, D epartm ent o f the Environment, London. DoE, 1990. Summary of the ministerial declaration from the International Conference of the Protection of the North Sea, D epartm ent o f the Environment, London. Dessens, A., Cohen-Kettenis, P., M ellenbergh, G., van de Poll, N., K oppe, J . an d Boer, K., 1998. `Prenatal exposure to anticonvulsant Table 6.1. 74 Late lessons from early warnings: the precautionary principle 1896-2000 drugs an d spatial ability in ad u lth o o d ', Acta Neurobiol. Exp. Vol. 58, pp. 221-225. Drinker, C. K. et al., 1937. `T he problem o f possible systemic effects from certain chlorinated hydrocarbons',Journal of Industrial Hygiene and Toxicology Vol. 19 (Septem ber), pp. 283-311. Francis, E., 1998. Conspiracy of silence -- how three corporate giants covered their toxic trail, at http: / / www.planetwaves.net/silence2.html (includes responses from the companies involved at http://www.planetwaves.net/ respon se.h tm l) H uism an, M., 1996. Effects of early infant nutrition and perinatal exposure to PCBs and dioxins on neurological development: A study of breast-fed and formula-fed infants, D issertation (ISBN 90-3670688-2), University o f Groningen, Groningen. H uism an, M., Koopm an-Esseboom , C., Lanting, C. I., van der Paauw, C. G., Tuinstra, L. G. M. Th., Fidler, V., W eisglas-Kuperus, N., Sauer, P. J . J., Boersm a, E. R. an d Touwen, B. C. L., 1995. `N eurological condition in 18month-old children perinatally exposed to polychlorinated biphenyls an d dioxin s', Early Hum. Dev. Vol. 43, pp. 165-176. Fuerst, P., Fuerst, C. an d Wilmers, K., 1992. Bericht ber die Untersuchung von Frauenmilch a u f Polychlorierte Dibenzodioxine, Dibenzofurane, Biphenyle sowie organochlorpestizide 1984-1991, Chem isches Lan desu ntersuchu n gsam t NRW, Mnster. Greve, P. A. an d W egman, R. C. C., 1983. `PCB residues in anim al fats, hum an tissues, duplicate 24-hours diets, eel and sedim ents', Proceedings ofPCB seminar, Scheveningen, The Hague, Netherlands, 2 8 -3 0 September 1983, pp. 54-65. Guo, L. Y, 1999. `H um an health effects from PCBs and dioxin-like chemicals in the rice-oil poisonings as com pared with other exposure ep isod es', Organohalogen Compounds Vol. 42, p p .241-243. H arada, M., 1976. `Intra-uterine poisoning: Clinical and epidem iological studies and significance o f the p rob lem ', Bull. Inst. Const. Med. Kumamoto University Vol. 25 (Su ppl.), pp. 1-69. Jacobson, J. L. and Jacobson, S., 1996. `Intellectual im pairm ent in children exposed to polychlorinated biphenyls in utero', N. Engl. J. Med. Vol. 335, N o 11, pp. 783-789. Jaco b so n , J . L., Jaco b so n , S. W. an d Humphrey, H. E., 1990. `Effects o f exposure to PCBs and related com pounds on growth and activity in ch ildren ', Neurotoxicol. Teratol. Vol. 12, pp. 319-326. Jen sen , S., Johnels, A. G., Olsson, M. and Otterlind, G., 1969. `DD T and PCB in m arine anim als from Swedish w aters', Nature Vol. 224, p. 247. Kim burgh, R. D. et al., 1987. `H um an health effects of polychlorinated biphenyls (PCBs) and polybrom inated biphenyls (PBBs)', Ann. Rev. Pharmacol. Toxicol. Vol. 27, pp. 87-111. Koopm an-Esseboom , C., 1995. Effects of perinatal exposure to PCBs and dioxins on early human development, Dissertation, University o f Rotterdam, Rotterdam. Hens, L., 1999. `Dioxines en PCBs in Belgische eieren en kippen ', Milieu 1999/4, p p .220-225. H olden, A. V., 1970. `M onitoring organochlorine contamination of the marine environment by the analysis o f residues in seals' in Ruivo, M. (ed.), Marine pollution and sea life, pp. 266-272, Fishing News Books Ltd, England. Hornshaw, T. C. et al., 1983. `Feedin g G reat Lakes fish to mink: Effects on mink and accum ulation and elim ination o f PCBs by m in k', J. Toxicol. Environ. Health Vol. 11, pp. 933-946. Koopm an-Esseboom , C., Morse, D. C., W eisglas-Kuperus, N., Lutkeschipholt, I. J., van der Paauw, C. G. and Tuinstra, L. G ., 1994. `Effects of dioxins and polychlorinated biphenyls on thyroid horm one status of pregn an t women an d their infants', Pediatr. Research Vol. 36, No 4, pp. 468-473. Koppe, J. G., Pluim, H. J. and Olie, K., 1989. `Breastmilk, PCBs, dioxins and vitamin K deficiency',J . Royal Soc. ofMed. Vol. 82, pp. 416-20. Lanting, C. I., 1999. Effects ofperinatal PCB and dioxin exposure and early feeding on child development, D issertation (ISBN 90-3671002 2), University o f Groningen, Groningen. PCBs and the precautionary principle 75 Lanting, C. I., H uism an, M., M uskiet F. A. J., van der Paauw, C. G., Essed, C. E. and Boersm a, E. R, 1998a. `Polychlorinated biphenyls in adipose tissue, liver and brain from nine stillborns of varying gestational a g e s', Pediatr. Res. Vol. 44, pp. 1-4. Lanting, C. I., Patandin, S., Fidler, V., W eisglas-Kuperus, N., Sauer, P. J . J., Boersm a, E. R. and Touwen, B. C. L., 1998b. `N eurological condition in 42-month-old children in relation to pre- and postnatal exposure to polychlorinated biphenyls and dio xin s', Early Hum. Dev. Vol. 50, pp. 283 292. M asuda, Y, 1994. `The Yusho rice oil poison in g in cid en t' in Schecter, A. (ed.) Dioxins and health, pp. 633-659, Plenum Press, New York and London. Matthews, H. B. an d A nderson, M. W., 1975. `Effect o f chlorination on the distribution and excretion of polychlorinated biphenyls', DrugMetab. Dispos. Vol. 3, No 5, pp. 371-380. OECD, 1973. Decision-recommendation of the Council on Protection of the Environment by control ofpolychlorinated biphenyls, C (73) 1 (Final). OECD, 1987. Decision-recommendation of the Council concerningfurther measures for the protection of the environment by control of polychlorinated biphenyls, C (87) 2 (Final). Oliver, B. G. and Niimi, A. J., 1988. `Trophodynamic analysis o f polychlorinated biphenyl congeners and other chlorinated hydrocarbons in the Lake Ontario ecosystem ', Environ. Sci. Technol. Vol. 22, pp. 388-397. Patandin, S., 1999. Effects of environmental exposure to polychlorinated biphenyls and dioxins on growth and development in young children, PhD thesis (ISBN 90-9012306-7), Erasmus University, Rotterdam. Pluim, H. J., Koppe, J. G., Olie, K., van der Slikke, J . W., Kok, J . H., Vulsma, T., van Tijn, D. and de Vijlder, 1992. `Effects o f dioxins on thyroid function in newborn babies', Lancet Vol. 339, 23 May, p. 1303. Schwartz, T. R. et al., 1987. `Are polychlorinated biphenyl residues adequately described by Aroclor mixture equivalents? Isomer-specific principal com ponent analysis o f such residues in fish and turtles', Environ. Sci. Technol. Vol. 21, pp. 72-76. Seegal, R. F. and Schantz, S. L., 1994. `Neurochemical and behavioural sequelae of exposure to dioxins an d P C B s' in Schecter, A. ( ed.), Dioxins and health, Plenum Press, New York and London. Swedish Environmental Protection Agency, 1998. Persistent organic pollutants: A Swedish view of an international problem, text by Claes Bernes, M onitor 16. Theelen, R. M. C. an d Lie, A. K. D., 1997. Dioxins: Chemical analysis, exposure and risk assessment, Thesis (ISBN 90-393-2012-8), University o f Utrecht, Utrecht. United States Court o f Appeal, 1980. Environmental Defense Fund, Inc., petitioner, v. Environmental Protection Agency, respondent, Ad Hoc Committee on Liquid Dielectrics of the Electronic Industries Association et al., Joy M anufacturing Company, Edison Electric Institute et al., and Aluminum Company ofAmerica, Intervenors, N os 79-1580, 79-1811 an d 79 1816, United States Court o f Appeals, District o f Columbia Circuit, argued 6 Ju n e 1980, decided 30 October 1980, at h ttp :// w w w .m an hattan .edu/w cb/schools/ ENGINEERING/envl/w m atysti/2/files/ edfcase.htm W HO, 1999. Dioxins and their effects on human health, Fact Sheet No. 225, World H ealth Organization, Geneva, at h ttp :// www.who.int/inf-fs/en/fact225.htm l 76 Late lessons from early warnings: the precautionary principle 1896-2000 Figure 7.1. Source: EEA 7. Halocarbons, the ozone layer and the precautionary principle Joe Farman 7.1. Overview In the second half of the last century some 23 million metric tonnes of chlorofluorocarbons (C FC s), about 11 m illion tonnes o f methyl chloroform , 2.5 m illion tonnes o f carbon tetrachloride and 4 million tonnes of a hydrofluorocarbon (HCFC-22) were released into the atm osphere. All o f these h alocarbons have atm ospheric lifetim es (4) long enough for them to be transported to the stratosphere. The am ount o f chlorine in the stratosphere today is between six and seven times what it was in 1950. T here is an ozone hole over Antarctica from Septem ber to D ecem ber each year, less systematic but nevertheless large ozone losses in the Arctic, and m oderate ozone losses in the middle latitudes of both hem ispheres. The evidence which allows us to conclude, beyond reasonable doubt, that halocarbons are responsible for the dam age to the ozone layer, is set out in a series o f five reports prepared by the World M eteorological O rganization (WMO) for the Parties to the United Nations Vienna Convention and M ontreal Protocol (WMO, 1985, 1989, 1991, 1994 and 1999). As a result o f this dam age, there will be serious im pacts including an increase in skin cancer (see Figure 7.1.). This graph illustrates the approxim ate tim e lags betw een C F C production, the resulting depletion of the stratosp heric ozone layer and subsequent extra penetration of UV radiation and the im pact this will eventually have on increasing the background rate o f skin cancer, given th e 3 0-40 y e a r ave ra g e late n t p erio d fo r such ca n c e rs. R e ality is far m ore c o m p le x than th is sch e m atic illustration. For exam ple, there are other ozone-depleting chem icals (H C FC s, H FC s and methyl brom ide); the ozone hole varies with latitude, tim e of the year and m eteorological conditions; the increased UV radiation varies betw een different w avelengths and with latitude and cloud cover; and the skin cancer excess com es on top of a rising background rate of skin cancer, with differential effects on the d ifferent types of skin cancer, such as m alignant m elanom a and non-m alignant skin cancers. Human b eh avio u r is also a d ete rm in in g e ffe ct in skin cancer. H ealth e ffe cts also in clu d e cataracts and im m une re sp o n se su p p ressio n . How ever, the figure illustrates the main relationships and tim e lags betw een C F C production and skin cancer, and the 'success' in sto p p in g C F C p ro d u ctio n and avertin g m uch m ore skin can ce r from o zon e d ep le tio n than w h a t is n ow e x p e c te d . (Slap er, e t al., 1996). (4) Lifetim es are given as e-folding tim es. The rate of removal of a constituent is proportional to its concentration, so that, in the absence of em issions, the residual concentration after N lifetim es is e-N tim es the initial concentration. Halocarbons, the ozone layer and the precautionary principle 77 The end o f what m ight be called the CFC episode o f industrial history is alm ost in sight. The consumption (consum ption = production + imports -- exports) and most production o f CFCs, halons and methyl chloroform have already ceased in developed countries. Exceptions include feedstock use, a few so-called essential uses (5) an d an allowance to supply the needs o f developing countries whose own phase-out schedule began with a freeze in 1999 and culminates in cessation o f supply in 2010. The reductions in releases are the basis on which the M ontreal Protocol (signed in Septem ber 1987 and coming into force on 1January 1989) and its later am endm ents and adjustments have been hailed as a major success. However the cessation o f releases is not the end o f the story for the atm osphere and for the ozone layer. Methyl chloroform has the shortest atmospheric lifetime, about 5 years, of the m ajor ozone depleting substances (O D S), and only about 2 % o f the current atm ospheric concentration will rem ain after four lifetimes, or 20 years. CFC-12, on the other hand, has a lifetime o f about 100 years and the concentration in the atm osphere in 2100 will be at least 37 % o f what it is today. The last statem ent is qualified because even in developed countries where production has ceased releases continue from the bank of past production held in equipm ent and plastic foam s; this was estim ated to be 791 kilotonnes in 1995. The production peaked in 1988 and ceased in developed countries at the end o f 1993. Whilst there is som e production in China, India and Korea, due to cease in 2002, the current releases come predom inantly from the bank in developed countries, estimated at 70 kilotonnes in 1995 and bein g run down at abo u t 4 % per year. This halon has an atm ospheric lifetime of about 65 years, and its atm ospheric concentration will continue to increase until about 2020, unless the protocol is am ended to require the destruction o f banked halons. Historical and projected total chlorine and bromine loadings of the troposphere Figure 7.2. Source: DETR, 1999 Year Reactive brom ine is scaled in pptv on the right, and in ppbv of eq uivalent chlorine on the left. The com bined loading is expressed as eq uivalent chlorine. The line at 2.5 ppbv of equivalent chlorine indicates the loading in the late 1970s. (5) Th ese include m edical use of C F C s in m etered dose inhalers (now largely replaced by H FC s or PFC s perfluorocarbons) and the use of halons in fire protection system s in m ilitary equipm ent. 78 Late lessons from early warnings: the precautionary principle 1896-2000 As a guide to past and projected dam age to the ozone layer the sim plest indicator is the combined loading of chlorine and bromine in the troposphere, expressed as equivalent chlorine. M olecule for m olecule, brom ine is som e 58 times more effective than chlorine in destroying ozone: equivalent chlorine = chlorine + 58 x brom ine. Estimates o f the historical and projected loadings are shown in Figure 7.2. O pinion is divided over ju st how far back in time dam age to the ozone layer can be traced, but it has become conventional to regard a level o f 2.5 ppbv (parts per billion by volum e) o f equivalent chlorine, first reached in the late 1970s, as significant. U nder the latest amendm ents (1 9 9 7 ) to the protocol a return to this level is expected between 2050 and 2060. For com parison it can be noted that under the original protocol, the London amendments o f 1990 and the Copenhagen am endm ents of 1992, there was no guarantee that the equivalent chlorine loading would ever return to 2.5 ppbv. The adjustm ents in Vienna in 1995 did finally achieve a projected return, and the am endm ents in M ontreal in 1997 brought the return a few years nearer. The failure to act m ore decisively in the early negotiations can be seen perhaps as the Parties to the protocol setting consensus before effectiveness; in retrospect it seems that the negotiators consistently took the view that it would be easier to get agreem ent at the n ext m eeting ( 6). Should the CFC episode have been avoided? Could it have been avoided? Three sketches o f industrial and environmental history may help to put these questions in perspective. 7.2. Early history The first episode is set at the end o f the 19th century. The chlorinated derivatives o f the simpler hydrocarbons had been studied, finding incidentally many com pounds for which there are no significant sources outside the laboratory. Most ODS are com pounds o f that sort. Carbon tetrachloride was being produced industrially; it was used as a solvent, for dry cleaning and in fire extinguishers. Methyl chloroform was known but, surprisingly in view o f developm ents after 1950, foun d no application. A Belgian chemist, Swarts, had laid the foundation of the study of fluorinated hydrocarbons, m aking CFC-11 and CFC-12 by the catalysed reaction of hydrofluoric acid with carbon tetrachloride. At the turn o f the century ozone was an im portant industrial chemical. In an episode, now largely ignored in textbooks, it was widely used in applications ranging from the chemical laboratory to m unicipal services. It was the m ost powerful oxidant known, left no objectionable residue, an d was easy, if somewhat expensive, to make. It was used to sterilise public water supplies and to purify the air in the Central London underground railway system. As a bleaching agent and a deodorant o f oils, waxes and fats, it was widely used in the m anufacture o f linen, cotton, paint, varnish, linoleum and perfum es. It was used as a preservative for food, and refrigerated warehouses and ships were often equipped with portable ozonisers. Little was known about ozone as a natural constituent o f the atm osphere. M easurements at Rouen in 1857 had shown that the m ixing ratio near the ground was abo u t 10-8 (10 ppbv) (H ouzeau, 1857), and quite dense networks of stations undertaking such measurements had been set up, most notably in Belgium and France. It had been inferred from the cut-off at short wavelengths o f the solar spectrum that m uch larger mixing ratios must occur in the upper atm osphere. At the turn o f the century there was a comfortable feeling am ong m eteorologists that the thermal structure of the atm osphere was sim ple: tem perature decreased with heigh t at a rate that was well understood, and the decrease should continue to the limit o f the atm osphere. This complacency was shaken in 1901 when Teisserenc de Bort (see, for exam ple, Goody, 1954) reported that the decrease stopped, often quite abruptly, at a height o f about 11 km, an d that from there to abo u t 14 km, the greatest height attained by his instruments, the tem perature rem ained alm ost constant. At first he called his discovery the isotherm al layer, but later suggested the words troposphere, to denote the well-mixed lower region, and stratosphere for the region above, in (6) Paragraph 9(c) of A rticle 2 of the M ontreal Protocol states 'In taking such decisions, the Parties shall make every effort to reach agreem ent by consensus. If all efforts at consensus have been exhausted and no agreem ent reached, such decisions shall, as a last resort, be adopted by a two-thirds majority vote of the Parties present and voting representing at least 50% of the total consumption of the controlled substances of the Parties.' No issue has yet been put to a vo te . Paragraph 11 of the sam e article states 'N otw ithstanding the provisions contained in this A rticle, Parties may take more stringent m easures than those required by this A rticle .' It has been invoked only once, by the European Union in March 1991, suggesting that, on the w hole, the negotiators have been quite content to proceed at a pace compatible with consensus. Halocarbons, the ozone layer and the precautionary principle 79 recognition o f its high static stability. Routine soundings o f the stratosphere had to await cheap portable radio, lightweight instruments and improved balloons. A start was m ade in 1933, but the successive effects o f economic recession and world war m ade progress slow until the 1950s. 7.3. The 1930s -- the CFC industry is born The second sketch is for the decade centred on 1930. By then chlorine had largely replaced ozone as an industrial bleaching and sterilising agent. It was cheap, as large amounts were produced as a by-product of the electrolysis o f rock salt, a process introduced to satisfy the growing dem and for caustic soda for preparing sheet metal for painting. A Norwegian, Eric Rotheim, applied for a patent for the aerosol container (spray can) principle in 1926. (Th ere is no record o f use on a large scale until the Secon d World War, when insecticide sprays were needed in the Pacific sector. Mass production started in the U nited States in 1947 and in Germany in 1953.) In 1929 Thom as Midgely, at General Motors in the United States, conducted a survey of chemicals suitable for use as refrigerants, and recom m ended CFC-12 and CFC-11 as efficient, non-toxic and non-flammable. Du Pont started industrial production o f CFC-12 in 1930 and o f CFC-11 in 1934. By this time, the ozone layer had been discovered and, considering the resources available, quite comprehensively investigated. The annual variation in ozone am ount had been m easured at stations ranging from the Arctic Circle to New Zealand, largely through the efforts of Dobson and his collaborators at Oxford, following up the pioneer m easurem ent o f Fabry and Buisson in France. The broad features of the vertical distribution o f ozone had been established, by ground-based m easurem ents o f light scattered from the zenith sky at twilight (Gotz, M eetham and D obson), and by the Regeners at Stuttgart, who sent a small solar spectroscope up on a balloon to 31 km, at which point it had passed through about 70 % o f the ozone layer. T hese two very different techniques gave results that agreed remarkably well. To com plete this sketch it should be m entioned that in 1930 Chapman published the first theoretical treatm ent of the formation of atm ospheric ozone (for more on all this see, for exam ple, Goody, 1954). Within its limitations, equilibrium photochemistry in a static atm osphere, it gave a fair account o f the upper part o f the ozone layer (above 30 km ), but failed completely to explain the observed variations o f ozone with latitude an d season. M ore than 50 years were to pass before realistic air motions could be incorporated into models o f the ozone layer, an d even today there remains a fascinating interplay between uncertainty in the chemistry and uncertainty in the dynamics o f the atm osphere. 7.4. The 1970s -- the seeds of doubt Concern about the effect o f hum an activities on the ozone layer becam e an international issue in 1970. Initially fears were expressed about the emissions o f nitrogen oxides, carbon m onoxide and water from the projected commercial fleets of supersonic stratospheric transport (SST) aircraft. These were heeded -- the U nited States abandoned developm ent o f SSTs, and the com bined total of Anglo-French Concordes and Russian Tupolev-144s never exceeded 40 aircraft. Attention then shifted to the CFC industry, which had expanded enormously throughout the 1960s. A survey, using a recently developed device capable o f m easuring small traces of CFCs, had shown that these gases, released largely in the northern hemisphere, h ad spread worldwide (Lovelock et al., 1973). Moreover their concentrations were such as to indicate that processes removing CFCs from the troposphere m ust be extremely slow -- virtually all o f the am ounts released were still in the atm osphere. In 1974 sem inal papers by American scientists pointed out that CFCs were so stable that they would eventually reach the stratosphere, that chlorine would be released there by photolysis, and that an ozone-destroying chain reaction would ensue (Molina and Rowland, 1974; C icerone et al., 1974). A vigorous debate followed, which was to continue for over a decade. The first significant move was the agreem ent in 1977 for a research-oriented `world plan o f action on the ozone layer', and the establishment by the United Nations Environment Programme (UNEP) of a Coordinating Committee on the Ozone Layer, to oversee the plan. It is noteworthy that representatives of non-governmental organisations (NGOs) were given the right to attend m eetings o f the committee. In the United States an effective public campaign, stimulated by the earlier debate on the im pact o f supersonic aircraft, resulted in 80 Late lessons from early warnings: the precautionary principle 1896-2000 regulations in 1977 prohibiting the use of CFCs as aerosol propellants. Canada, Norway and Sweden took similar action. European environmental groups were at the time concentrating on the issue o f acid rain, and there was little public pressure on governments over CFCs. However in 1980 European Council Decision 8 0 /3 7 2 was passed, aim ing to freeze the production capacity for CFC-11 and CFC-12, and to reduce their use in aerosols by at least 30 % from 1976 levels by the end of 1981. The effect o f these m easures on releases is not easily seen. R eleases h ad fallen sharply a few years before the m easures came into force, largely as a result o f a downturn in world trade following a crisis in the oil market. The reduction in use in the aerosol sector was quickly offset by increased use in the foam blowing sector. T he E u ro p ean freeze was little m ore than a token gesture, since most plan t was running well below its full capacity. For all that, these m easures appear to be the first, and last, unequivocal application o f the precautionary principle in the ozone story. It was the only action taken before the evidence against CFCs had become compelling. The 1977 US Clean Air Act explicitly states that `no conclusive proof... but a reasonable expectation' o f harm ful effects is sufficient to justify action (U S EPA, 1987). This was in stark contrast to the `wait an d se e ' attitude ad op ted by industry. Du Pont, the w orld's first and largest producer of CFCs, had set this out in a full page in the New York Times (30 Ju n e 1975). It declared that: `Should reputable evidence show that some fluorocarbons cause a health hazard through depletion o f the ozone layer, we are prepared to stop production o f the offending com pounds.' It was to deny the existence of reputable evidence until 1986. Nevertheless, it should be acknowledged that the industry, through the Chemical M anufacturers Association (an international funding agency), throughout this time gave substantial financial support to institutes and individuals engaged in research into the ozone problem. In the early 1980s, the problem o f the ozone layer seem ed to have settled down. Com putational m odels predicted only small long-term reductions o f ozone, an d this was in reason able accord with observations, which showed no significant trend. Negotiations for an international convention on the ozone layer, started by UN EP in 1981, p roceeded slowly. While the U n ited States favoured a range o f controls on the use of CFCs in various applications, Europe argued for a cap on production capacity. The differences could not be bridged, and the Vienna Convention for the Protection o f the Ozone Layer, agreed in M arch 1985, contained only pledges to cooperate in research and m onitoring, to exchange information on CFC production, and to pass control protocols only if and when justified. By this time the United States was not prepared to act unilaterally, claiming that as a result o f its earlier action, its production of CFC-11 and CFC-12 had fallen from 46 % o f the world total in 1974 to 28 % in 1985. T here was som e consolation for those advocating controls; the Parties to the Convention gave UNEP im m ediate authority to convene working group negotiations for a protocol, to be signed if possible in 1987. 7.5. The Montreal Protocol and the ozone hole T he M ontreal Protocol was, in deed, signed on 16 Septem ber 1987, after a hectic last week o f controversy. Richard Benedick, the ch ief U S negotiator, later claim ed that this too was an application o f the precautionary principle. A sim pler interpretation would be that the working group had been overtaken by events. Ozone depletion over Antarctica h ad been reported in Nature (Farm an et al., 1985) in May 1985, m uch m ore severe than any prediction, and confirm ed by NASA in O ctober 1985. In reporting the NASA results, the Washington Post gave the world the expressive term `ozone hole'. Du Pont, rem inded o f its statem ent in 1975, wrote to its CFC customers in Septem ber 1986, declaring that it now accepted the need for som e controls (Cagin and Dray, 1993, p. 308). Also in 1986, the US National Ozone Expedition (NOZE) to M cM urdo Station in Antarctica had produced m uch evidence to su p p ort the view that the depletion was chemically driven. In Septem ber 1987 attention was again focused on Antarctica, with press releases expected from NOZE II, and from the US Airborne Antarctic Ozone Experim ent (AAOE) whose planes had flown from Punta Arenas in Chile into and under the ozone hole. The timing o f the announcem ent o f the protocol makes sense only as a pre-emptive move astutely designed to preserve som e credibility for the negotiators, and to give industry time for orderly reorganisation. Halocarbons, the ozone layer and the precautionary principle 81 Success, compromise, m uddle, failure -- all these terms have been used to describe the 1987 protocol. T here can be no d ou bt that it was a psychological breakthrough. However the terms were tem pered by what was thought practicable, the ultimate objective was not clearly defined, and releases increased rapidly in the period before the next full m eeting o f the Parties. The protocol was ratified and came into force on 1January 1989, in line with the timetable. The review procedure was set in m otion at once. By then a consensus on the main scientific issues had been reached, NGOs had fought vigorous cam paigns for public awareness, and industry was responding m uch faster to the problem than had initially seem ed possible. The London amendm ents in 1990, although substantially stronger than the original protocol, nevertheless were disappointing by com parison with statem ents m ade by m ost of the Parties in preparatory meetings. However, in addition to the changes in the controls, two im portan t decisions were taken. It was agreed that there should be full reviews every two years instead o f every four, an d it was agreed (in a new Article 10) to establish a Multilateral Fund for the Implementation of the M ontreal Protocol (MFMP) to help developing countries reduce their dependence on ODS. The main concern in these negotiations was to replace CFCs quickly with new chemicals, hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs) being the options preferred by industry (7). Som e 75 % o f the global production o f CFCs was in the hands o f 13 groups o f com panies, who were quite content to close down old CFC plant if the protocol would allow reasonable time for the industry to profit from investment in HCFC and HFC production. The negotiators readily accepted this; these transitional substances were m ade subject to guidelines rather than controls and their future was left open-ended, as consensus could not be reach ed on a phase-out date (8). In my view, this approach was deeply flawed. Technical surveys had already shown that large quantities of CFCs and halons were released unnecessarily by poor working practices; the quantities of replacem ents needed were much less than current consumption. More emphasis should have been placed on prudent long-term goals, with active encouragem ent o f the development of halocarbon-free and energy-efficient technologies, to protect the ozone layer, to slow down the forcing o f climate change and to reduce the cost o f raising living standards in developing countries. The MFMP has been spent largely on the replacem ent o f CFCs by HCFCs, and m ore m oney is now needed to phase-out HCFCs. It would surely have been better to stimulate m ore radical changes in technology from the outset. The m ost recent full m eeting o f the Parties was in Beijing, China, in early Decem ber 1999, when 129 governments were represented. Two noteworthy events had taken place during the preparations for the meeting. In March 1999 the MFMP approved USD 150 million to fund the complete closure over the next 10 years o f CFC production in China, now the w orld's largest producer and consumer o f CFCs and halons. A week before the m eeting the MFMP approved USD 82 million to lead to the complete phase-out o f CFC production in India, the w orld's secon d largest prod u cer o f CFCs. At the m eeting the Parties agreed to raise an additional USD 440 million for the MFMP; the fund has disbursed m ore than USD 1 billion since 1991. New controls were adopted for HCFCs, including a ban on trade in these chemicals with countries that have not yet ratified the 1992 am endm ent (for phase-out of H CFCs). A complete phase-out by 2002 o f the production o f the recently marketed chemical, brom ochlorom ethane, was agreed. A Beijing D eclaration was adopted, calling for continued efforts to halt illegal trade in ODS. There was one brief glimpse o f the precautionary principle: the Scientific and Econom ic Assessment Panels were asked to propose m eans to prevent the development and marketing of new ozonedepleting chem icals (9). Im portan t issues that were not resolved include the possible destruction of halon banks, an earlier phase out for methyl brom ide and reducing the current scale of CFC and halon consumption in Russia and other countries with econom ies in transition. (7) H C FC s are ozone depleting substances, but much less harmful than C F C s. H FC s have zero ozone depletion potential, but they are powerful greenhouse gases. (8) Th ere w ere sharp differences of opinion on the date by which phase-out could be achieved. Targets ranged from 2010 to 2040. (9) For more information on the Beijing meeting visit the w eb site: http://w w w .iisd.ca/ozone/m op11/ 82 Late lessons from early warnings: the precautionary principle 1896-2000 7.6. Late lessons The report in 1985 of rapid and severe seasonal destruction of ozone over Antarctica was an outcom e o f systematic long-term measurements, begun solely for scientific exploration. It took everyone by surprise, including the authors. The accepted view was that the effects o f CFCs should be seen first at high altitudes (30-50 km) in the tropics, and that any changes in the lower stratosphere in polar regions w ould be very slow. As it h appen ed , two other groups were close to challenging the accepted view -- Jap an ese scientists had presented anom alous ozone profiles from Antarctica, but without identifying a systematic trend, and NASA scientists were re-exam ining very low ozone am ounts flagged as `suspect' by their software in initial retrievals o f satellite data. This was before the Internet, and the absence of effective collaboration, whilst regrettable, was not uncom m on in institutions with stretched resources. It should also be noted that the ozone community was ju st com pleting the first o f the WMO reports -- an onerous task for authors and reviewers. The funding o f long-term m onitoring remains a serious problem . It is simply not practicable to provide open-ended funding at the level needed to cope with uncovering a m ajor environmental issue. Recession and war restricted the growth of the CFC industry. From 1930 to 1948 the cumulative releases o f CFC-12 and CFC-11 were only 25 and 5 kilotonnes respectively. T he n ext two decades brough t spectacular growth -- in 1970 the annual releases were 300 and 207 kilotonnes. People had been p ersu ad ed to buy CFCs an d throw them away (aerosols). Industry's dream o f `w onder chemicals' was apparently coming true. This was the time when it m ight reasonably have been asked whether such developm ent was sustainable. There can be little doubt that a conventional risk assessment, in say 1965, would have concluded that there were no known grounds for concern. It would have noted that CFCs were safe to handle, being chemically very inert, non-flammable, and having very low levels o f toxicity. They were good heat insulators, and som e were excellent solvents, m ixing readily with a wide range of other organic substances. The assessment m ight have pointed out that it was not known what happens to CFCs when they are released to the atm osphere, but would no doubt have added that they had been released for m ore than 30 years with no apparent harm being done. There would probably have been a long list o f recom m endations for further research. In retrospect, it is easy to see that these should have included investigating the ultraviolet and infrared spectra of CFCs, measuring their atm ospheric concentrations, and identifying their decom position products. As it happens, there were hints o f the risk to the ozone layer available (Weigert, 1907; Norrish and Neville, 1934 (10)) if it h ad been recognised that atomic chlorine and fluorine m ight be released from CFCs in the upper atm osphere. But they would surely have been dismissed on the grounds that there would be at least 10 000 times m ore ozone than CFCs at the relevant altitudes. It is ju st possible that such an assessm ent m ight have accelerated the research needed to make out a plausible case against CFCs. But, as history shows, a plausible case brought only limited action. Serious negotiations did not begin until severe depletion had occurred, and strong evidence found to link it with CFCs. The protracted, and frequently amended, schedule for phase out o f ODS suggests that, even at this stage, precautionary m anagem ent o f the environment was not upperm ost in the m inds o f policy-makers. The recent decision (2001) o f the US adm inistration not to ratify the Kyoto Protocol on Clim ate Change will have disappointed many who were hoping that the experience gained with the M ontreal Protocol would facilitate getting agreements for further protection of the global environment. One o f the lessons should have been that neither governments nor multinational companies have a mandate for global experiments, even when such experiments consist solely of `business as usual'. The CFC issue provides a stark warning. Practices that appear to be reasonable when introduced (in this case when there were huge gaps in the understanding o f atmospheric processes) may later (as understanding improves) be seen to be leading to a m ajor global problem that can neither be avoided, nor rapidly (10) These two papers report laboratory experim ents on the decomposition of ozone photosensitised by chlorine. T h e y are cited in a system atic review of stratospheric chem istry, dated 1977, but there is nothing to suggest that they had any influence on w orkers in the 1960s and early 1970s. Halocarbons, the ozone layer and the precautionary principle 83 alleviated. There is a deep-seated paradox here. Short-term safety appears to dem and that synthetic chemicals in everyday use should be non-reactive. It has taken a long while for it to be realised that this m eans that they will be extrem ely persistent. It should not be assum ed that environmental science has reached the stage where all hazards can be foreseen. All too often technology outstrips the science needed to assess the risks involved. If there is to be a solution, policy-makers must learn to recognise, m uch m ore rapidly than they have in the past, when ignorance has been replaced by understanding, however rudimentary. Halocarbons: early warnings and actions 1907 1934 1973 1974 1977 1977 1980 1985 1985 1987 1990s 1997 1999 Laboratory experiments by Weigert on the decomposition of ozone photosensitised by chlorine D itto by Norrish and Neville Global survey of C F C s by Lovelock e t al. showing their distribution in the atmosphere worldwide Molina and Rowland publish their theoretical arguments that C F C s would be destroying the ozone layer United States bans C F C s in aerosols based on 'reasonable expectation' of damage, followed by Canada, Norway and Sweden. Research-oriented 'world plan of action on the ozone layer' agreed, overseen by UNEP European decision restricting use of C F C s in aerosols, but rising use in refrigerators, etc. marginalises this restriction UNEP Vienna Convention for the protection of the ozone layer agrees research, monitoring, information exchange and restrictions if and when justified Farman, Gardiner and Shanklin publish results showing hole in ozone layer over Antartica Montreal Protocol on protection of the ozone layer is signed, with phasing out of ozone depleting substances for both developed and developing countries within different timescales Increasing finance to developing countries to help them reduce their dependence on ozone depleting substances Amendments to the Montreal Protocol in order to restore levels of chlorine by 2050-60 Beijing Declaration calling for efforts to stop illegal trade in ozone depleting substances 7.7. References Cagin, S. and Dray, P., 1993. Between earth and sky, Pantheon Books, New York. Cicerone, R. J., Stolarski, R. S. and Walters, S., 1974. Science Vol. 185, p. 1165. DETR, 1999. `Stratospheric ozone 1999', DETR Reference No 99EP0458, United K ingdom Stratospheric O zone Review Group, D epartm ent o f the Environment, Transport and the Regions. Farm an, J . C., Gardiner, B. G. and Shanklin, J . D., 1985. Nature Vol. 315, p. 207. H ouzeau, A., 1857. Comptes Rendus Acad. Sci. Paris Vol. 45, p. 873. Lovelock, J. E., M aggs, R. J. and Wade, R. J., 1973. Nature Vol. 241, p. 194. M olina, M. J . and Rowland, F. S., 1974. Nature Vol. 249, p. 810. Norrish, R. G. W. an d Neville, G. H. J., 1934. J. Chem. Soc., p. 1684. U S EPA, 1987. Protection o f stratospheric ozone, Fed. Register 52, p. 47491. Weigert, F., 1907. Ann. Physik. Vol. 24, p. 243. Goody, R. M., 1954. The physics of the stratosphere, C am bridge University Press, C am bridge. WMO, 1985, 1989, 1991, 1994 and 1999. Global Ozone Research and Monitoring Project, Reports Nos 16, 20, 25, 37 and 44, World M eteorological Organization, Geneva. Table 7.1. 84 Late lessons from early warnings: the precautionary principle 1896-2000 8. The DES story: long-term consequences of prenatal exposure Dolores Ibarreta and Shanna H. Swan 8.1. Introduction In 1970, Herbst and colleagues reported an unprecedented finding; they had diagnosed a rare vaginal cancer (vaginal clear-cell adenocarcinom a) in seven young women, a cancer that had never before been seen in this age group in this hospital (H erbst and Scully, 1970). T he following year, these authors published the startling finding that seven of the eight cases (one m ore had been identified), but none o f 32 matched controls, had been prenatally exposed to the synthetic oestrogen diethylstilboestrol (DES) (Herbst et al., 1971). Seven m onths after this publication, the US Food and Drug Adm inistration (FDA) withdrew approval for use o f DES by pregnant women, for whom this potent synthetic oestrogen had been prescribed since 1947 in the mistaken, but widespread, belief that it prevented spontaneous abortion (m iscarriage). In fact, this apparently innocent treatm ent proved to be a time bom b for the infants exposed during the first third o f pregnancy. The discovery that DES was a transplacental carcinogen (and was shown after 1971 to also be a teratogen -- causing birth defects) occurred only 10 years after the awful discovery (James, 1965) that the use of thalidomide by women in early pregnancy caused severe limb reductions. These tragic lessons forced scientists to abandon their commonly held view o f the foetal environment as a safe place, protected by the placental `b arrie r', an d replace it with an understanding o f the extrem e vulnerability o f the developing foetus, for whom any maternal exposure that occurred during critical periods had the potential to alter development, even long after birth. 8.2. Optimistic beginnings Oestrogens are steroid hormones, made primarily in the female ovaries and the male testes, that act by binding to oestrogen receptors. O estrogens, together with the other so-called sex horm ones (progestins and androgens), are required for the regulation o f reproduction and the developm ent of secondary characteristics in both males and females. In 1938, Charles Dodds and colleagues form ulated DES, the first orally active synthetic oestrogen (D odds et al., 1938). This synthetic (non-steroidal) oestrogen has been estim ated to be five times as potent as oestradiol, the m ost potent naturally occurring oestrogen in mammals (N oller and Fish, 1974). D ES is inexpensive and simple to synthesise and the developing pharm aceutical industry quickly began worldwide production; DES was m arketed under more than 200 brand names. Like other pharm aceuticals produced at that time, DES underwent very limited toxicological investigations. It rapidly becam e popular in a wide variety o f treatments, including those for m enopausal symptoms and prostate cancer. Other therapeutic uses were suppression of lactation, post-coital contraception (morning-after pill) and post m enopause syndrome (Noller and Fish, 1974). It was later used as a growth prom oter in chicken, sheep and cattle (Aschbacher, 1976). A m ong the multiple uses o f DES was that for the prevention o f spontaneous abortion (m iscarriage). It was believed in the 1940s that spontaneous abortion was the result o f a decrease in oestrogen levels, now recognised as a consequence, not a cause (Smith, 1948). As early as 1941, Karnaky and others began experim enting with the use o f very high doses (100 m g administered intramuscularly into the cervix daily) in women `threatening to abort', and deem ed the drug effective for this purpose, stating that there were no adverse consequences to mother or foetus (Karnaky, 1942). The use o f DES for prevention o f m iscarriage was further prom oted by the work o f Olive and George Smith, who conducted multiple (uncontrolled) trials o f DES for use in pregnancy throughout the 1940s and published extensively regarding its effectiveness (Sm ith, 1948; Smith et al., 1946). Paradoxically, Charles Dodds, the scientist who first synthesised DES, a discovery for The DES story: long-term consequences of prenatal exposure 85 which he was later knighted, speculated years later about the minimal testing o f new drugs at that time. Reviewing the history of stilboestrol (another comm on term for DES), suggested by D odds h im self (D odds et al., 1938), he noted that no long-term toxicity tests o f this synthetic oestrogen had been conducted. He stated: `I suppose we have to be very thankful that it (DES) did prove to be such a non-toxic substance.' (Dodds, 1965) Unfortunately, when Dodds wrote this the long-term consequences of DES had not been discovered. 8.3. Tragic consequences Reports that DES increased cancer incidence in laboratory animals appeared as early as 1938 (Lacassagne, 1938; Geschickter, 1939; Shimkin and Grady, 1941; Greene and Brewer, 1941). Since those early reports, extensive research has dem onstrated an increased incidence o f cancer o f the mammary glands, cervix and vagina in multiple rodent species in response to DES. However, until 1971 there was little evidence suggesting that maternal use of DES could result in cancer and reproductive abnorm alities in humans, 20 years after exposure. The 1971 case-control study o f H erbst and colleagues identifying prenatal exposure to DES as the likely cause o f a rare and often fatal vaginal cancer in young women was startling and unprecedented. This casecontrol study was followed within a few months by a second, corroborative, casecontrol study (Greenw ald et al., 1971). The relationship between DES use by pregnant women and the occurrence o f vaginal clear cell adenocarcinom a is clearly apparent in Figure 8.1. As can be seen, there is a remarkable correspondence between sale of DES (represented here by m arket sales figures produced by a large and representative US manufacturer during litigation) and diagnosis o f this cancer 20 years later (M elnick et al., 1987). Market sales of 25 mg DES vs. cases of clear-cell cancer by year of diagnosis Figure 8.1. Source: D. Ibarreta and S. H. Swan 86 Late lessons from early warnings: the precautionary principle 1896-2000 This shocking finding finally raised a red flag concerning the dangers o f DES use by pregnant women, as well as providing convincing evidence o f the potential hazards that exposure o f the pregnant woman to pharmaceuticals (and other chemicals) pose to the foetus. As a result o f this study, the FDA released a bulletin in Novem ber 1971 stating that DES use was contraindicated (must not be used) during pregnancy (US D epartm ent o f Health, E. and Welfare, 1971). Tragically, the use o f DES by pregnant women continued even after 1971 outside the U nited States and in som e countries was reported many years later (Direcks and t'H oen, 1986). The num ber o f women who took DES is known imprecisely, but the National Institutes of Health has estimated that it was used by 4.8 m illion pregnant women in the United States alone. Total exposure (to mothers, sons and daughters) worldwide has been estim ated to be 10 million (see Box 8.1.). Later studies, including follow-up o f a large cohort o f DES-exposed women (DESAD cohort, for DES-adenosis) showed that these `DES daughters' exhibited a wide range of reproductive tract abnormalities. These included epithelial changes (such as vaginal adenosis) as well as structural changes such as cervical stenosis (narrowing of the cervical opening) and uterine malformations. Unlike vaginal clear-cell adenocarcinom a, which was extremely rare am ong DES-exposed women, these abnormalities were common, particularly in women whose m others had been prescribed DES in early pregnancy. These changes, at least some o f which were found to be prevalent even before birth (Joh n son et al., 1979), resulted in serious reproductive consequences for those affected. The risk o f all adverse pregnancy outcomes, including ectopic pregnancy (embryo attachm ent outside the uterus), spontaneous abortion and premature delivery, were significantly increased (Swan, 1992). While studies on `DES sons' have been m uch m ore limited, several have demonstrated increased rates of genital abnormalities in these men (Goldberg and Falcone, 1999). 8.4. DES ineffective for prevention of miscarriage After the first claims o f the effectiveness of DES for the prevention o f miscarriage, several studies were carried out to assess its efficacy, with m ixed results. As these studies becam e m ore rigorous, the support for the use o f D ES declined. Finally, in the early 1950s, two random ised, placebo-controlled trials were conducted. The larger of these trials, conducted by Dieckm ann et al. (1953), reported no statistically significant differences in adverse pregnancy outcomes when DES was com pared to a placebo. The authors concluded (as did the authors o f the second, smaller trial) that DES was ineffective in the prevention o f m iscarriage and complications in late pregnancy. In spite of this, the drug continued to be prescribed even to women without previous pregnancy problem s. An advertisement for one o f the DES congeners (DesPlex) read, `R ecom m ended for routine prophylaxis in all pregnancies' (see Figure 8.2.). A reanalysis o f D ieckm an n 's data was published in 1978, after the long-term carcinogenic and teratogenic effects of DES were known (Brackbill and Berendes, 1978). T he authors reanalysed D ieckm an n 's published data and found that DES actually increased the risk o f the endpoints it was sold to prevent. The m ethods used in the reanalysis were n ot new, an d h ad been available in 1953. As the authors noted, had the data been properly analysed in 1953, nearly 20 years o f unnecessary exposure to DES might have been avoided. The fact that this drug was prescribed for two decades after its lack o f efficacy was clearly dem onstrated illustrates a massive failure o f the system. In fact, it was not the lack o f efficacy that triggered the end o f DES marketing for use in pregnancy, but a fortuitous accident. If the seven cases originally detected by H erbst and colleagues had been diagnosed in several different medical centres, rather than at Massachusetts General Hospital (where DES use had been high because the Smiths had conducted their early experim ents on DES there), the dangers o f DES m ight well have gone unrecognised. The cancer that DES caused in young women (vaginal clear-cell aden ocarcin om a) is extrem ely rare. It is estimated to have occurred in less than one in every 1 000 exposed daughters (M elnick et al., 1987). In fact, the registry established by H erbst in 1972 to actively register all cases o f adenocarcinom a o f the genital tract in young women has identified fewer than 800 cases worldwide. Thus, this cancer and its link to DES could well have gone undetected. H ad this occurred, it is unlikely that DESassociated genital tract changes, which can only be identified by a physician conducting The DES story: long-term consequences of prenatal exposure 87 a special DES examination, and the multiple reproductive consequences of DES exposure, would ever have been identified. 8.5. Assessing the extent of the damage and close to 1 000 unexposed controls for long-term follow-up. This DESAD has proved invaluable for identifying the range of adverse consequences of DES exposure. Unfortunately, no comparable cohort of `DES sons' has ever been established. The young age o f the women affected m ade the need for early detection o f vaginal clear cell adenocarcinom a critical. It was o f crucial importance to identify the exposed population so that screening and early detection could help prevent fatal consequences. To assemble data on this rare cancer, H erbst and co-workers established the Registry o f Clear-Cell A denocarcinom a of the Genital Tract in Young Females in 1972 (H erbst et al., 1972). B ecause it becam e evident, from several studies in which DES cohorts had been screened, that DES also had caused extensive non-cancerous DES changes o f the fem ale genital tract, in 1976 the National Cancer Institute established a cohort o f over 3 000 DES-exposed women The exact num ber o f individuals exposed to D ES in utero is unknown, bu t estim ates range from 2-10 million (see Box 8.1.). A study funded by the European Commission and published in 1991 (Direcks et al., 1991) attem pted to estimate the prevalence o f DESaffected women in Europe. Eighteen countries were surveyed by m eans of questionnaires. This study found that use was greatest in Great Britain, France and the Netherlands, followed by Belgium, Ireland, Portugal and Spain. The survey concluded that DES had been prescribed during pregnancy, to a greater or lesser extent, in all countries surveyed except Sweden and Hungary. Box 8.1. Use of DES to prevent miscarriage in Europe and the United States Country Period prescribed Approximate number of pregnancies Belgium Czechoslovakia France 1950-65 1958-76 1950-77 63 000 200 000 60 000-240 000 Source: Direcks e t al., 1991, unless otherwise indicated Pons e t al., 1988 Direcks e t al., 1991 G erm any Ireland Italy N eth erla n d s Norway Portugal Spain United Kingdom United States -1977 1950-76 - 1960 (?) 1947-75 1948-72 1960-70 1953-77 (1983?) 1940-71 (1973?) 1943-71 200 000 189 000-378 000 25 000 7 000-8 000 2-6 million W ingfield, 1992 H anselaar e t al., 1991 Palmlund et al., 1993 Garcia-Alonso e t al., 1988 Kinlen et al., 1974 Goldberg and Falcone, 1999 The story o f D ES is far from over. Many o f the youngest w om en exp osed in utero are n ot yet 30. The known consequences o f its use will continue to appear as this cohort ages. In addition, the cohort m ust be followed to identify unknown consequences, including a possibly increased cancer risk with age. The use of experimental animal models may help address som e o f these concerns about the future since the m ouse appears to be a good m odel in which to study the effects o f developmental exposure to DES (McLachlan, 1993). Such studies on ageing mice may be useful in predicting DES-effects as the exposed hum an population ages, as will continued follow-up o f exposed cohorts. W hat will h appen to subsequen t generations o f the D ES-exposed wom en? T here is considerable interest in m onitoring the grandchildren of DES-treated women. Studies in mice have shown an increased 88 Late lessons from early warnings: the precautionary principle 1896-2000 susceptibility to tum our form ation in the third gen eration (New bold et al., 1998), suggesting the DES grandchildren may also be at increased cancer risk (Miller, 1999). This is a question that will take m any years to resolve. 8.6. Lessons from the DES story It is worthwhile asking how this tragic event might have been prevented and what lessons can be learn ed from it. First, the D ES story demonstrates that long-term and hidden effects o f horm onal exposure (as was learned from the thalidomide story and other chemical exposures) are possible, and that such consequences may be devastating. Therefore, extreme caution should be taken before exposing pregnant women to substances that may alter the endocrine system, especially during foetal development. The DES episode also dem onstrated that the absence o f visible and immediate teratogenic effects cannot be taken as p roo f o f the absence o f reproductive toxicity. DES represents not only the first transplacental hum an carcinogen, but also the clearest example o f hum an endocrine disruption. Although the doses o f DES adm inistered to pregnant women were far higher than usual environmental levels o f synthetic chemicals, DES remains a clear warning o f the consequences o f perturbing the endocrine system through synthetic chemicals. What precautionary action might have prevented the use o f DES by pregnant women? First, no long-term toxicity tests were ever carried out. DES is a synthetic oestrogen and the carcinogenicity potential of oestrogens was already known even during the initial experim ental use of DES (Cook and Dodds, 1933). Adverse effects of DES on animals had been published (Shimkin and Grady, 1941; Gardner, 1959; Dunn and Green, 1963), but they were largely disregarded. It was also known at the time that chemicals could cross the placental `barrier'. There were early indications of adverse effects o f exposure to oestrogens during developm ent (G reene et al., 1939) and Karnaky in his early experim entation with D ES n oted darkened areolae (the area around the nipple) and linea alba (the line running from the pubis to the navel) in exposed newborns, indicating the oestrogenic action of DES on the female offspring (Karnaky, 1945). This evidence rem ained largely ignored, although some m odest attempts to pinpoint the risks can be found in the literature. For exam ple, in 1948, Bernard Laplan, a physician who believed in the efficacy o f DES, expressed reservations about possible long-term effects: `The possibility o f som e latent effects on the reproductive and endocrine system o f the infant resulting from the high dosage oestrogen therapy m ust be borne in mind as indicated by the work o f Greene, Burrill, and Ivy in experim ental anim als. Furtherm ore, we would avoid this therapy in any patient with a family history o f m alignant d isease.' (Laplan, 1948) Second, it m ight have been recognised in prem arket testing that DES was never effective for the prevention o f miscarriage. The m ethods used in the clinical trials conducted in 1951-52 that dem onstrated this had been available before 1947 when DES was put on the m arket for use in pregnancy without adequate p ro o f o f efficacy. A properly conducted and analysed clinical trial m ight have avoided the entire episode. The fact that DES was not patented, and very inexpensive to produce, m ade it extremely profitable. This, plus the absence o f obvious acute toxic effects, undoubtedly contributed to the rapidity with which its use spread worldwide. The widespread use o f D ES was undoubtedly furthered by the faith in the advances o f science and the belief in man m ade solutions to n atu re's problem s that were prevalent during the post-Second World War era. Undoubtedly those prescribing DES believed it was safe and effective, and both `m odern and scientific'. In the Netherlands, for exam ple, the use o f DES was particularly prom inent, aided in part by its endorsem ent by the Q u een 's gynaecologist (Braham s, 1988). Consequently, there was strong pressure on physicians to prescribe DES. In 1952, Robinson and Shettles found DES ineffective when com pared to untreated cases and therefore discouraged its use. They discussed in their article the `peer pressure' that doctors were subjected to: `The synthetic oestrogen, diethylstilbestrol, has more recently becom e a popular form o f therapy in threatened abortion. The public has been so frequently told o f the virtue o f this drug through articles appearing in layjourn als that it now requires a courageous physician to refuse this m edication. The mass of pharm aceutical literature, extolling the wonders o f this drug, has also rendered m ost practitioners am en able to his patien t's dem ands. This situation, together with the understandable desire to do som ething The DES story: long-term consequences of prenatal exposure 89 positive toward rescuing a teetering pregnancy, has resulted in the widespread use o f DES in threatened abortion.' (Robinson and Shettles, 1952) Advertisement for DES Figure 8.2. Source: Des Plex, anonymous The role o f another im portant stakeholder in the wide use o f DES, the pharm aceutical industry, should be exam ined. Pharmaceutical retailers and advertising prom oted the effectiveness and safety o f DES to doctors and consumers. The pharm aceutical industry ignored the lack o f efficacy o f their product and failed to assess its adverse health effects. In fact, some manufacturers prom oted it as a panacea for use in all pregnancies (see Figure 8.2.). The eagerness o f the pharm aceutical com panies to sell this profitable product was com pounded by the failure o f medical and regulatory agencies to react rapidly to the em erging evidence. While the FDA announced that DES was contraindicated for use in pregnancy only m onths after H erb st's publication in 1971 (US D epartm ent of Health, Education and Welfare, 1971), in Europe the delay to withdrawal was as long as 12 years. It was not until 1974 that the Dutch Ministry o f H ealth advised that DES should not be given to pregnant women (Palmlund et al., 1993); in France, it was n ot until 1977 that national authorities included the contraindication for pregnancy in the labelling o f the drug (Epelboin and Bulwa, 1993). In some countries, such as Spain, there are reports of use as late as 1983 (Direcks et al., 1991). In addition, underdeveloped countries remained an open m arket for DES for many years. As recently as 1985, DES products were used in maternity care in Brazil, Costa Rica, Kenya, Mexico, Peru, Rwanda and Zaire. In the early 1990s, it was still prescribed during pregnancy in Mexico, U ganda and Poland (Palmlund, 1996). DES continued to be used for a variety o f therapeutic uses for years after the evidence o f its adverse effects had been widely publicised. This included use as a post-coital contraceptive. Since DES is less than 100 % effective in terminating pregnancy, the use o f a chemical potentially carcinogenic to the foetus was not permissible. In 1973, the FDA discouraged the use o f DES as a contraceptive, restricting it to emergency situations such as rape or incest (Mills, 1974). The use o f DES as a growth prom oter (see chapter on H orm ones as growth prom oters) began in the 1950s and ended after a heated battle in 1979. Gradually use for other therapeutic purposes decreased and in the spring o f 1997 Eli Lilly, the last, and predom inant, m anufacturer o f DES in the United States ceased production (Pat Cody, pers comm.). There has been significant progress in most key areas o f drug assessm ent since 1970, in part spurred by the consequences of DES and thalidomide. In fact, the entire field o f teratology was established in response to these episodes. Toxicity testing is m uch m ore comprehensive and thorough and sound evidence o f efficacy is now required before a drug is m arketed. Regulatory authorities are also m uch m ore alert to reporting o f adverse drug reactions and m ore inclined to take action than they were in the 1960s and 1970s. Nonetheless several areas of concern remain: drugs that were m arketed prior to current m ore stringent regulations may not have been adequately evaluated with respect to efficacy or toxicity; monitoring and surveillance in patients taking drugs is neither consistent nor comprehensive. Moreover, it is difficult to demonstrate an association between a drug and a health outcom e that is delayed a n d / or prevalent; formalised risk/benefit analyses o f new drugs (as well as other chemicals) are not widely utilised; drug surveillance (and evaluation of hum an health risks from other chemicals) should not be limited to detection of gross 90 Late lessons from early warnings: the precautionary principle 1896-2000 Table 8.1. malformations and cancers; such limited surveillance would likely have m issed most of the adverse effects of DES. It is clear, in hindsight, that the concerns expressed by scientists regarding the m arketing o f a drug with known carcinogenic potential should have been heeded. Failure to do so reflects the prevalent attitudes towards health risks, and an absence of precautionary thinking. The lack o f a timely response by m anufacturers and regulators after DES was dem onstrated to be a transplacental hum an carcinogen in 1971 can n ot be excused so easily. Even after the tragic consequences o f DES use in pregnancy were confirmed, economic interests predom in ated, resulting in a `wait an d se e ' approach. Those considering the wisdom of precautionary action today would do well to consider the history o f DES. This drug was a known animal carcinogen, a suspect human carcinogen and a drug that had been shown to produce observable changes in the offspring o f women exposed in pregnancy. Even if marketing in 1947 could have been justified on the basis of insufficient evidence o f harm, after DES was proven to be ineffective for use in pregnancy in 1953 a review o f its risks and benefits should have resulted in immediate contraindication of this use. At that point there was no longer anyjustification for subjecting consum ers to any carcinogenic risk in the absence o f any benefit. H ad DES been withdrawn for use in pregnancy at that time, the unnecessary and tragic exposure of millions of mothers, sons and daughters could have been avoided. DES: early warnings and actions 1938 1938 1939 1942 1942 1947 1948 1953 1970 April 1971 N ovem ber 1971 1972 1978 1985 DES synthesised First report of increased cancer incidence in anim als after D ES adm inistration First report of DES administered to patients Approval of DES by the Am erican Council of Pharmacy and Chemistry First report of DES used for prevention of abortion US Food and Drug Administration (FDA) approves DES for the treatm ent of threatened or habitual abortion Use of D ES increases following publication of large-scale study in the US First large placebo-controlled random ised trial shows D ES ineffective in the prevention of m iscarriag e Published report of seven cases of vaginal clear-cell adenocarcinom a in young wom en Prenatal D ES exposure is linked to vaginal clear-cell adenocarcinom a FD A withdraws approval of DES for use by pregnant women R eg istry of C lear-Cell Adenocarcinom a of the G enital Tract in Young Fem ales is established Reanalysis of 1953 Dieckmann data shows that DES actually increased the risk of miscarriage and other adverse pregnancy outcom es Last reported use of DES by pregnant women world-wide 8.7. References Aschbacher, P. W., 1976. `Diethylstilbestrol m etabolism in food-producing animals , J. Toxicol. Environ. Health Suppl. Vol. 1, pp. 45 59. Bishop, P. M. F., Boycott, M. an d Zuckerm an, S. 1939. `The estrogenic properties of stilbestrol (diethylstilbestrol)', Lancet Vol. 2, p p .5-11. Brackbill, Y and Berendes, H. W., 1978. `Dangers o f diethylstilbestrol: review o f a 1953 p a p e r', Letter to the Editor, Lancet Vol. 2, No 8088, p. 520. Braham s, D., 1988. `Latent drug injury: failure of Dutch diethylstilbestrol action', Lancet Vol. 2, N o 8616, p. 916. Cook, J . W. an d D odds, E. C., 1933. `Sex horm ones and cancer-producing co m po u n ds', Nature Vol. 131, pp. 205-206. Dieckm ann, W. J., Davis, M. E., Rynkiewics, L. M. and Pottinger, R. E., 1953. `Does the administration of diethylstilbestrol during pregnancy have therapeutic value?', Am. J. Obstet. Gynecol. Vol. 66, N o 5, pp. 1062-1081. Direcks, A., Figueroa, S., Mintzes, B. and Banta, D., 1991. R eport from DES Action the Netherlands for the European Commission The DES story: long-term consequences of prenatal exposure 91 Program m e Europe Against Cancer, p. 69, DES Action the Netherlands, Utrecht. Direcks, A. and t'H oen, E., 1986. `DES: The crim e con tin ues', in M cDonnell, K., (ed.), Adverse effects of drugs: women and the pharmaceutical industry, pp. 41-49, International Organization of Consumers Unions, regional office for Asia and the Pacific, Penang. D odds, E. C., G oldberg, L., Lawson, W. and Robinson, R., 1938. `Estrogenic activity o f certain synthetic co m p o u n d s', Nature Vol. 141, pp. 247-248. D odds, C., 1965. `Stilboestrol an d after', Sci. Basis Med. Annu. Rev. 1965, pp. 1-16. D unn, T. B. an d Green, A. W., 1963. `Cysts o f the epididymis, cancer o f the cervix, granular cell myoblastoma, and other lesions after estrogen injection in new born m ice ', J. Nat. Cancer Inst. Vol. 31, No 2, pp. 425-455. Epelboin, S. and Bulwa, S., 1993. `Drug surveillance and reproduction: the bad exam ple o f distilbene', Contracept. Fertil. Sex. Vol. 21, No 9, pp. 658-672. Garcia-Alonso, F., Rodriguez-Pinilla, E. and Martinez-Frias, M. L., 1988. `The use o f diethylstilbestrol in S p ain ', Medicina Clinica Vol. 91, No 11, p. 436. Gardner, W. U ., 1959. `C arcinom a o f the uterine cervix and upper vagina: induction under experim ental conditions in m ice', Ann. NY Acad. Sci. Vol. 75, pp. 543-564. Geschickter, C. F., 1939. `M am m ary carcinom a in the rat with m etastasis induced by estrogen ', Science Vol. 89, pp. 35-36. G oldberg, J . M. an d Falcone, T., 1999. `Effect o f diethylstilbestrol on reproductive fu n c tio n , Fertil. Steril. Vol. 72, No 1, pp. 1-7. Greenwald, P., Barlow, J . J., N asca, P. C. an d Burnett, W. S., 1971. `Vaginal cancer after m aternal treatm ent with synthetic estrogens , V. Engl. J . Meid. Vol. 285, N o 7, pp. 390-392. Hanselaar, A. G., Van Leusen, N. D., De Wilde, P. C. an d Vooijs, G. P., 1991. `Clear cell adenocarcinom a of the vagina and cervix. A report o f the Central Netherlands Registry with em phasis on early detection and progn osis', Cancer Vol. 67, N o 7, pp. 1971 1978. H erbst, A. L. an d Scully, R. E., 1970. `A den ocarcin om a o f the vagina in adolescence. A report of 7 cases including 6 clear-cell carcinomas (so-called m eso n ep h ro m as)', Cancer Vol. 25, No 4, pp. 745-757. Herbst, A. L., Ulfelder, H. and Poskanzer, D. C., 1971. `A den ocarcin om a o f the vagina: Association of maternal stilbestrol therapy with tum or appearan ce in young w om en , N. Engl. J. Med. Vol. 284, No 15, pp. 878-881. Herbst, A. L., Kurm an, R. J., Scully, R. E. and Poskanzer, D. C., 1972. `Clear-cell adenocarcinom a o f the genital tract in young fem ales: Registry re p o rt', N. Engl. J. Med. Vol. 287, No 25, pp. 1259-1264. Jam es, WH, 1965. `Teratogenetic properties o f th alidom ide', Br. Med. J . 1965 O ct 30; 5469:1064. Jo h n so n , L. D., Driscoll, S. G., H ertig, A. T., Cole, P. T. an d N ickerson, R. J., 1979. `Vaginal adenosis in stillborns and neonates exposed to diethylstilbestrol and steroidal estrogens and progestin s', Obstet. Gynecol. Surv. Vol. 34, No 11, pp. 845-846. Karnaky, K. J., 1942. `T he use o f stilbestrol for the treatm ent of threatened and habitual abortion and prem ature labor: A preliminary re po rt', South M J. Vol. 35, p. 838. Greene, B. R., Burrill, M. W. an d Ivy, A. C., 1939. `Experim ental intersexuality: the paradoxical effects of estrogen on the sexual developm ent o f the fem ale rat ', The Anatomical Record Vol. 74, N o 4, pp. 429-438. Greene, R. R. and Brewer, J . I., 1941. `R elation o f sex horm ones to tum ors o f the fem ale reproductive system ', Amer. J. Roentgen. Rad. Ther. Vol. 45, N o 3, pp. 426 445. Karnaky, K. J., 1945. `Prolonged adm inistration o f diethylstilbestrol' J. Clin. EndocrinolJu ly 1945 5, pp. 279-284. Kinlen, L. J., Badaracco, M. A., Moffett, J. and Vessey, M. P., 1974. `A survey o f the use o f oestrogens during pregnancy in the United Kingdom and of the genito-urinary cancer mortality and incidence rates in young people in En glan d an d W ales', J. Obstet. Gynaecol. Br. Commonw. Vol. 81, No 11, pp. 849-855. 92 Late lessons from early warnings: the precautionary principle 1896-2000 Lacassagne, A., 1938. `The appearance o f mammary adenocarcinom as in male mice treated by a synthetic estrogenic substance', Weekly Reports of the Society of Biology and of its Affiliates, Vol. 12, pp. 641-643. Laplan, B., 1948. `Diethylstilbestrol in the treatment of idiopathic repeated abortion', NY StateJ Med. Vol. 48, pp. 2612-14. M cLachlan, J . A., 1993. `Functional toxicology: a new approach to detect biologically active xenobiotics', Environ. Health Perspect. Vol. 101, N o 5, pp. 386-387. Pat Cody, DES Action, personal com m u nication . Pons, J. C., Goujard, J., Derbanne, C. and Tournaire, M., 1988. `O utcom e o f pregnancy in patients exposed in utero to diethylstilbestrol. Survey by the National College of French Gynecologists and O bstetricians', J. Gynecol. Obstet. Biol. Reprod., Paris, Vol. 17, N o 3, pp. 307-316. Robinson, D. and Shettles, L., 1952. `The use of diethylstilbestrol in threatened abortion', Am. J. Obstet. Gynecol. Vol. 63, pp. 1330-1333. M elnick, S., Cole, P., A nderson, D. and Herbst, A., 1987. `Rates and risks o f diethylstilbestrol-related clear-cell adenocarcinom a o f the vagina and cervix: An u p d a te ', N. Engl. J. Med. Vol. 316, No 9, pp. 514-516. Miller, M., 1999. `D ES research heats up again after breast cancer finding',J. Nat. Cancer Inst. Vol. 91, No 16, pp. 1361-1363. Mills, D. H., 1974. `Prenatal diethylstilbestrol an d vaginal cancer in offsprin g', JAMA Vol. 229, No 4, pp. 471-472. Newbold, R. R., H anson, R. B., Jefferso n , W. N., Bullock, B. C., H asem an, J. and M cLachlan, J . A., 1998. `Increased tumors but uncom prom ised fertility in the fem ale descendants o f mice exposed developmentally to diethylstilbestrol', Carcinogenesis Vol. 19, No 9, pp. 1655-1663. Noller, K. L. and Fish, C. R., 1974. `Diethylstilbestrol usage: Its interesting past, im portant present and questionable future', Medical Clinics of North America Vol. 58, No 4, p p .793-810. Shimkin, M. B. and Grady, H. G., 1941. `Toxic and carcinogenic effects o f stilbestrol in strain C3H m ale m ice',J. Nat. Cancer Inst. Vol. 4, p. 55. Smith, A. E., 1942. `Council o f pharm acy and chemistry: Diethylstilbestrol', JAMA Vol. 119, N o 8, pp. 632-635. Smith, O. W., 1948. `Diethylstilbestrol in the prevention and treatment of complications in pregn an cy', Am. J. Obstet. Gynecol. Vol. 56, p p .821-834. Smith, O.W., Smith, G.v.S. an d Hurtwitz, D. 1946. `Increased excretion o f pregnandiol in pregnancy from diethylstilbestrol with especial reference to the prevention o f late pregnancy acciden ts', Am. J . Obstet. Gynecol. Vol. 51, p. 411-415. Swan, S. H., 1992. `Pregnancy outcom es in D ES daugh ters', in Giusti, R. (ed.), Report of the N IH Workshop on Long-term Effects of Exposure to Diethylstilbestrol (DES), pp. 42-49, US Departm ent of Health and Human Services, Public Health Service, National Institutes o f Health, Washington, DC. Palm lund, I., 1996. `Exposure to a xenoestrogen before birth: the diethylstilbestrol exp erien ce', J. Psychosom. Obstet. Gynaecol. Vol. 17, No 2, pp. 71-84. US Departm ent o f Health, Education and Welfare, 1971. Food and Drug Administration, Fed. Register (21537 21538). Palmlund, I., Apfel, R., Buitendijk, S., Cabau, A. and Forsberg, J . G., 1993. `Effects o f diethylstilbestrol (DES) m edication during pregnancy: Report from a symposium at the 10th international congress o f ISPO G ',J. Psychosom. Obstet. Gynaecol. Vol. 14, N o 1, pp. 71-89. W ingfield, M., Donnelly, V. S., Kelehan, P., Stronge, J . M., Murphy, J . an d Boylan, P., 1992. `D ES clinic -- the first six m onths', Ir Med. J Vol. 85, N o 2, pp. 56-58. Antimicrobials as growth promoters: resistance to common sense 93 9. Antimicrobials as growth promoters: resistance to common sense Lars-Erik Edqvist and Knud Borge Pedersen 9.1. Introduction An antibiotic is defined as a substance produced by m icroorganism s that inhibits or kills other m icroorganism s. Synthetic antimicrobial substances are referred to as chemotherapeuticals. The word `antim icrobial' (as a noun) is often used to encom pass any substance of natural, semi synthetic or synthetic origin that kills or inhibits the growth o f a microorganism. The antim icrobial era began in 1910 with the introduction o f Salvarsan into the therapy of syphilis and expanded in the 1930s when the first chemotherapeuticals (sulphonam ides) m ade their way into clinical use in hum an m edicine. The antibiotic, penicillin, was discovered by Alexander Flem ing (Fleming, 1929) but was not introduced for therapy until 1941. Shortly after its introduction for use in humans, penicillin was used in animals for the treatm ent of various bacterial diseases. Antimicrobials are probably the single m ost im portant discovery in the history o f medicine. They were considered miracle drugs. Over the years they have saved millions o f lives by killing bacteria that cause some o f the worst infectious diseases in man and animals. 9.1.1. Antimicrobials for growth prom otion The growth prom oting properties of antimicrobial agents in farm animals were discovered in the late 1940s. Trials where ferm entation waste from tetracycline production was fed to chickens as a source o f vitam in B 12revealed that the chickens fed the waste grew m ore rapidly than did the controls. It was soon found that this effect was not due to the vitamin content o f the feed but to residual tetracycline (Stokstad and Jukes, 1949). The practice o f feeding sub-therapeutic doses o f antimicrobials over long periods o f time was readily adopted and soon becam e an integrated part o f the production systems developed in industrialised animal husbandry. Apart from increased growth rate an d /o r increased feed conversion, exam ples o f other observed effects o f antim icrobials at low doses are improved egg production in laying hens, increased litter size in sows and increased m ilk yield in dairy cows. Antimicrobial growth prom oters are freely accessible and sold over the counter, while antimicrobial drugs for therapy in m ost cases are available on prescription only. T he use o f antimicrobial growth prom oters over four decades has undoubtedly contributed to the development of current animal production systems. Use o f antim icrobial growth prom oters apparently provides some protection against certain diseases prom oted by intensification and allows, for exam ple, unphysiological early weaning and high stocking rates, raising questions on animal ethics and anim al welfare. 9.1.2. Development o f resistance Most antim icrobials are produced by m icroorganism s and existed long before they were used as medicinal drugs. Microorganisms produce antimicrobial substances to kill other m icroorganism s in order to create a foundation for their own survival and propagation. The bacteria exposed to antimicrobials develop strategies for survival; developm ent o f resistance is one such strategy. This follows the classical con cept o f survival o f the fittest an d it is thus not surprising that antimicrobial resistance has probably existed for as long as have bacteria. Concern about penicillin resistance m ore or less accom panied its introduction as a medical drug in the early 1940s. The discoverer of penicillin Alexander Fleming, in a new spaper interview by the New York Times (1945), w arned that m isuse o f penicillin could have the result that `m icrobes are educated to resist penicillin'. Increasing resistance was noted in several im portant bacteria known to cause infectious disease. By the early 1950s the problem of antim icrobial resistance was well acknowledged in the medical, veterinary and pharmaceutical press. 94 Late lessons from early warnings: the precautionary principle 1896-2000 Originally it was thought that acquired resistance in a bacterium only occurred through mutation in existing genes, which would m ean that the resistance trait would be confined to the m utant clone and spread o f resistance confined to that clone (vertical transmission). In the 1960s it was shown that resistance, in addition to mutation, could also be developed through the uptake of existing genes. In this case, the resistance trait through mobile genetic elements can also spread to other bacterial clones, to other bacterial species and even to other genera (horizontal transmission) (for a review see, for exam ple, Amabile-Cuevas and Chicurel, 1992). 9.2. The first early warning 9.2.1. The Swann Committee In the mid-1960s growing concern over foodborne infections with multi-drug-resistant salm onella in the United Kingdom led the government to establish an independent advisory committee in 1968. The task o f the committee, chaired by Professor Michael Swann, was to exam ine the issue of transferable antim icrobial resistance and the consequences for human and animal health arising from the use o f antimicrobials for growth prom otion and in veterinary medicine. The Swann Committee (Swann, 1969) ju d g e d the data available `a sufficiently sound basis for action' and the principal recom m endations on antimicrobial growth prom oters were that: `perm ission to supply and use an antibiotic without prescription for adding to anim al feed should be restricted to the antibiotics which: are o f econom ic value in livestock production under United Kingdom farming conditions; have little or no application as therapeutic agents in m an or animals; will n ot im pair the efficacy o f a prescribed therapeutic antibiotic or antibiotics through the development o f resistant strains of organism s.' Recom m endations on specific drugs were also included, for exam ple: `tylosin should not be available without prescription for use as a `feed' antibiotic.' Another recom m endation was the establishment of a single perm anent committee which `should have overall responsibility for the whole field o f use of antibiotics and related substances whether in man, animals, food preservation, or for other purposes'. 9.3. Subsequent action or inaction 9.3.1. Im plem entation and dilution o f the Swann Report The recom m endations in the Swann Report were based on less than full scientific certainty and created m uch debate and cries for more research, and faced strong opposition from the pharm aceutical industry and farm ing community in the United Kingdom. But m ost o f the main recom m endations were adopted in the United Kingdom and later on in the European Union (EU). However, subsequent governments in the United Kingdom gradually diluted the recom m endations of the Swann Report. The recom m endation to establish an overarching, strong and perm anent committee responsible for the whole field of antim icrobial usage was not fully implemented. For example, no epidem iological studies to m onitor antimicrobial resistance development were set up. Against the Swann recom m endations, the EU accepted the m acrolides tylosin and spiramycin as growth prom oters in 1975. This has probably been one of the m ajor reasons for the widespread m acrolide resistance in, for example, enterococci and campylobacter from pigs. Use o f the antim icrobial growth prom oter avoparcin was extended to other species, such as adult cattle, against the Swann recom m endations and its use increased from the mid-1970s at about the sam e time that its m edical equivalent, vancomycin, started to come into hospital use. One of the scientific arguments put forward to support this use o f antimicrobials for growth prom otion was that the low dose presents a special case in selecting for resistance. For exam ple, Walton (1988) stated: `In practical terms the use o f a sub lethal or a sub-inhibitory antibiotic concentration is therefore unable to select resistant strains from a bacterial population, an d in this respect the Swann R ep o rt's conclusion and recom m endations were in error.' However, the recent bans on avoparcin, virginiamycin and tylosin followed the publication of studies demonstrating that Antimicrobials as growth promoters: resistance to common sense 95 this view (W alton's am on gst others) was wrong. Worldwide, there are great differences in the regulatory control o f antimicrobials for therapy, prophylaxis and growth prom oting purposes. In som e countries, such as the U nited States, low doses o f tetracycline and penicillin are still used as feed additives for prophylaxis and growth prom otion without veterinary prescription, while therapeutic antimicrobials are often prescription-only m edicin es. treatment, prevention or cure o f diseases, m eaning that their use for growth prom otion should not be allowed. The basis cited for this am endm ent was the risk o f increased resistance, especially the risk of cross resistance to other substances and the risk of increased susceptibility o f animals to salmonella and other enteric pathogens. The governm ent also stressed that `there is uncertainty on the long-term effects o f the continuous use of feed containing chem otherapeutics' (Government Bill 1984/85). 9.3.2. The Swedish ban Similar to the situation in other countries, som e Swedish scientists viewed the practice of routine addition o f antimicrobials to anim al feeds with scepticism. Following the recom m endations o f the Swann Committee in the U nited Kingdom , a broader debate was initiated, which eventually led to a reassessm ent o f the use o f antimicrobials as feed additives (LBS, 1977). The Feedingstuffs Act was passed by parliam ent in November 1985 and came into force in January 1986. Since then antimicrobials, whether in feed or adm inistered otherwise, have only been allowed for therapy and on veterinary prescription, and as a result the total consum ption o f antimicrobials was greatly reduced from around 50 tonnes in 1985 to around 20 tonnes in 1996 (SOU, 1997). A working group o f the Board o f Agriculture concluded, am ong other things, that: `the use o f antibiotic feed additives entails a risk o f increased resistance in bacteria but as the substances in use are mainly active against gram-positive bacteria from which resistance is not transferred, the im pact o f such developm ent is n egligib le.' O n the other hand, a negative attitude to all kinds of additives am ong consum ers was noted by the group. The benefits, in terms o f increased production and prevention o f certain diseases, were also acknowledged (LBS, 1977). Legislative changes, especially in the requirem ents for approval, were proposed in order to m itigate possible risks. At the same time, farm ers were growing increasingly sceptical towards feed antim icrobials. They were concerned that the continued use of antimicrobials might harm consumer confidence. The Federation of Swedish Farmers (LRF) m ade a policy statement, declaring that Swedish agriculture aim ed towards a m ore restricted and controlled use o f antibiotics. In a letter to the Ministry o f Agriculture in 1984, the LRF requested a ban on the use of antibacterials as feed additives. In response to the above, the Ministry of Agriculture drafted a new Feedingstuffs Act (Government Bill 1984/85). Among other things, the draft proposed that the use of antim icrobials in feed should be restricted to During the accession negotiations with the EU, Sweden was granted a temporary derogation from European legislation concerning the use o f antibiotics as growth prom oting feed additives. In support o f the Swedish view the Ministry o f Agriculture appointed a com m ission to collect and review scientific data on antibiotic growth promoters. In 1997 the commission presented its report and am ong other points noted that `antimicrobial feed additives can at levels perm itted in feeding stuffs, be used for treatment or prevention of animal disease, which is in violation o f Council Directive 7 0 /5 2 4 /E E C ' (SO U, 1997). According to this directive, as am ended in 96/5 5 /E E C , Article 3a, authorisation of an additive shall be given only if at the level permitted, treatment or prevention of animal disease being excluded. The commission concluded that `the risk for increased resistance associated with the gen eral use o f antibiotic growth prom oters is far from negligible and the potential consequences are serious for both animal and human health' (SOU, 1997). T he m agnitude o f the risk is difficult to fully establish because of the complexity o f the problem and the lack o f pertinent data. The report outlined the research required to be undertaken in a 17-step causal chain, assessed the minimum time required to undertake the research to be 5-10 years, and also that the 96 Late lessons from early warnings: the precautionary principle 1896-2000 research had to be undertaken for each resistance gene and for each antimicrobial substance, with subsequent updates o f the risk assessment. The report went on to ask who would bear the costs of waiting to do further research, or o f taking action then to restrict antim icrobials -- the risk-maker or the risktaker? T he com m ission finally urged: `As the risks involved are o f uncertain m agnitude, the decisions on risk m anagem ent are particularly difficult. The risk can obviously not be excluded with certainty, nor can it be determ ined as acceptable. Scientists may declare that the inform ation is inadequate for decision making, but for the policym akers, failure to take action is n ot a neutral position but represents a positive decision to do nothing. In a climate of uncertainty it is preferable to show caution .' (SOU, 1997) 9.3.3. The ban o f avoparcin In March 1995 , when the first information on the occurrence of avoparcin- and vancomycin-resistant enterococci in pigs and poultry h ad becom e available (Klare et al., 1995; Bates et al., 1994; A arestrup, 1995), the D anish farm ers' organisations agreed with the feeding industry that there would be a voluntary cessation in the use o f avoparcin in anim al feed to reduce the spread of antim icrobial resistance. This was followed by a governm ental ban im plem ented on 20 May 1995 and reported to the European Commission as requested by the safeguard clause o f Council Directive 70/524/E E C . According to this a M ember State can, as a result of new information, temporarily suspend the use o f an approved feed additive if its use constitutes a danger to anim al or hum an health. The scientific background for the Danish ban was the demonstration: o f cross-resistance between avoparcin and vancomycin; that the resistance is transferable; that the use o f avoparcin as a growth prom oter selects for vancomycin-resistant enterococci and that vancomycin-resistant enterococci can be transferred to humans via the food chain (DVL, 1995). In Norway the use o f avoparcin was suspended in Ju n e 1995, and in Germany the governm ent issued a ban on avoparcin in January 1996. In May 1996 the EU Scientific Committee on Animal Nutrition (SCAN) concluded that further evidence was required to establish a risk to hum an health, anim al health or the environment caused by avoparcin. However, the committee accepted that serious questions concerning the safety of avoparcin had been raised and stated that the feedadditive use of avoparcin should be reconsidered at once if it were shown that transfer o f resistance was possible from anim al to hum an. The European Com mission, however, proposed that in the climate o f uncertainty and to avoid taking any risk, a tem porary ban should be placed on the use o f avoparcin as a feed additive in all EU M em ber States. This was agreed by a qualified majority vote o f the Standing Committee on Feedingstuffs in Decem ber and the ban came into force on 1 April 1997. 9.3.4. Danish ban o f virginiamycin On 16 January 1998, the Danish governm ent banned all use o f virginiamycin as a growth prom oter in Denmark, due to a risk o f selection of streptogramin-resistant enterococci in pigs and poultry (A arestrup et al., 1998). T he step was taken to protect hum an health and to preserve the lifespan o f Synercid, which was then undergoing hospital trials but which has now been licensed for the treatm ent o f certain multi drug-resistant infections in humans. 9.3.5. EU bans four antimicrobial growth prom oters On 14 Decem ber 1998, the agriculture ministers o f the EU M em ber States voted in favour of a proposal to ban the use o f four antimicrobial growth prom oters from July 1999: virginiamycin, bacitracin zinc, tylosin phosphate and spiramycin. The ban was subm itted by the European Com m ission as `a precautionary m easure to minimise the risk o f developm ent o f resistant bacteria and to preserve the efficacy o f certain antibiotics used in hum an m edicine'. The pharmaceutical industry protested against the decision and called for further scientific facts about the risks involved in the use of antimicrobial growth prom oters. Afterwards, the decision was challenged before the European Court ofJustice by the m anufacturer o f virginiamycin, who called for an annulm ent o f the entire decision. Final ruling in the case is not expected before the end of the year 2001. Antimicrobials as growth promoters: resistance to common sense 97 9.3.6. Avilamycin The latest exam ple of an antimicrobial growth prom oter from the EU list of approved products showing cross-resistance with a potential hum an drug (everninomycin) is avilamycin (Aarestrup, 1998). However, the m anufacturers of everninomycin have recently withdrawn it from hospital trial worldwide. 9.3.7. Scientific reports and recom m en dation s In 1997 the World H ealth Organization (WHO) organised a scientific m eeting which addressed the m edical im pact o f the use of antimicrobials in food animals. The meeting concluded that `the m agnitude o f the medical and public health impact of antimicrobial use in food animal production is not known' (WHO, 1997). The recom m endations from the m eeting stated: `Increased concerns regarding risks to public health resulting from the use o f antimicrobial growth prom oters indicate that it is essential to have a systematic ap proach towards replacing growth prom oting antim icrobials with safer non-antim icrobial alternatives.' The Invitational EU Conference on The Microbial Threat held in Copenhagen, Denmark, in 1998, reached a similar conclusion and stated: `Most o f those at the conference considered the use of antim icrobials for growth prom otion was not ju stified and that it was essential to have a systematic approach towards replacing growth prom otin g antim icrobials with safer non-antimicrobial alternatives including better farming practice.' (The Copenhagen Recommendations, 1998) The conference was organised by the chief medical officers o f the EU M ember States and included both medical and veterinary authorities and researchers, representatives o f farm ers' organisations, the pharmaceutical industry and the anim al feed industry. Because o f concern over the implications for hum an and animal health o f the rapidly increasing rate o f development of antimicrobial resistance, the European Commission (DGXXIV) asked the Scientific Steering Committee (SSC) to evaluate the current position regarding the prevalence and development o f antimicrobial resistance, and to exam ine its im plications for hum an and anim al health, particularly with regard to the development and m anagem ent of infections. In 1999 the SSC concluded in its report that `action needs to be taken promptly to reduce the overall use of antim icrobials in a balan ced way in all areas: human medicine, veterinary medicine, animal production and plant protection' (SSC, 1999). In relation to antibiotics for animal growth the SSC recom m ends that `the use o f agents from classes which are or may be used in human or veterinary medicine should be phased out as soon as possible and ultimately abolished'. In the newly issued `G lobal Principles for the Containment o f Antimicrobial Resistance due to Antimicrobial Use in Animals Intended for Food', WHO recom m ends that the use o f antimicrobial growth prom oters that belong to classes o f antim icrobial agents used (or subm itted for approval) in hum ans should be terminated or rapidly phased out in the absence o f risk-based evaluations (WHO, 2000). 9.4. Advantages and disadvantages of the use of growth promoters Over the entire period during which antimicrobial agents have been used as growth prom oters there has been great emphasis on the advantages o f these feed additives as a m eans to improve farm anim al production and productivity. Much less attention has been directed towards possible side-effects, and the num ber of independent scientific studies is relatively sm all com pared to studies and congress contributions supported by the pharm aceutical industry. Not until recently, when the occurrence of vancomycin-resistant enterococci in animals was found to be associated with the use o f avoparcin as an additive to animal feed and the possible consequences for hum an health were m ade clear, did the num ber of in d epen d en t research studies dealing with these aspects increase significantly. In the last few years substantial scientific evidence has shown that the use of antim icrobial growth prom oters in food anim als contributes to the problem s o f antim icrobial resistance in humans. This has m ost convincingly been shown for vancomycin-resistant enterococci. Although the widespread use of antimicrobials in human medicine undoubtedly is o f m ore im portance for the emerging antimicrobial resistance problems 98 Late lessons from early warnings: the precautionary principle 1896-2000 in hum ans, this cannotjustify ignorance of potential hum an health risks related to the use of antimicrobials in food animals. The continuous use o f antim icrobials in feed is one of the m ajor sources of overuse and misuse of antimicrobials in animal farming. This ongoing debate has m ade it clear that the usage of antimicrobial agents as feed additives is a com plex issue, with im plications not only for hum an and anim al health but also for anim al welfare, fo od safety, environmental aspects and for the developm ent o f production systems, feeding practices and m anagem ent. Som e o f the positive and negative effects o f antimicrobial growth prom oters are listed in Table 1. field and the possible consequences for widespread antimicrobial resistance. In other words science that embraces complexities, uncertainties an d unknowns with m ore humility and less hubris is needed. Furtherm ore, scientific committees which are responsible for the evaluation of confidential inform ation from industry, should consist of a broad panel o f independent and up-to-date experts, with relevant expertise and experience from all o f the disciplines that are im plicated in a broad assessm ent o f the risks, benefits and technological options involved. In this case, expertise from hum an m edicine would have been particularly valuable for every risk assessment. 9.5. Conclusions and lessons for the future The original early warning in the Swann Report on the risk o f antimicrobial resistance spreading amongst animals, and from animals to hum ans, was based on a low level of scientific proof, but on a competent m icrobiological assessment that foresaw possible adverse consequences o f the continuous use of therapeutic antimicrobial agents in anim al feed. The recom m endations were clearly precautionary, although the word `precau tion ' was n ot actually used in the Swann Report. Subsequent scientific research, developed primarily during the 1990s, has shown that transferable resistance is not restricted to certain bacteria (gram-negative), but is much m ore widespread within the m icrobiological universe, and that genes can easily move not only between closely related bacterial species, but also even between genera. These results confirm that the Swann R eport was both accurate in its evaluation o f data at the time and far-sighted in its assessm ent o f future trends. T he ju stification for the later dilution o f its conclusions and com prom ises on its recom m endations was based mainly on narrow considerations o f what was precisely known rather than on taking account o f what was not known, o f the ignorance within the The experience from this case suggests that assessm ents o f risk n eed to be m uch wider, taking into account both positive and negative impacts, the long-term m icrobiological and ecological effects on hum an and anim al health and the environment, and alternative options, such as better animal husbandry (see Table 9.1.). As the risks involved are o f uncertain m agnitude, the decisions on risk m anagem ent are particularly difficult. The risk can obviously n ot be excluded with certainty, nor can it be determ ined as acceptable. In a clim ate o f uncertainty it is preferable to show caution. In this situation decision-making needs to involve precaution, particularly when it is unacceptable, inhum an and unethical to wait for ultimate proof, when hum an fatalities could be involved. A nother clear lesson from this case study is that stakeholders other than regulatory bodies, such as the farm ers and their organisations, can take voluntary steps in advance o f legislation to stop the use o f products which cause concern and loss of confidence am ongst consum ers. In this case they, as well as the Swann Com m ittee, have been vindicated by history. Com m on sense and far-sighted use of good scientific evidence which can predict serious impacts should not be ignored whilst waiting for ultimate proof. Antimicrobials as growth promoters: resistance to common sense 99 Positive and negative effects of antimicrobial growth promoters on some broad issues of animal production Table 9.1. Broad issue Anim al health Human health Anim al welfare Enviro nm ent Animal husbandry Production system Animal feed Positive effect Certain diseases -- primarily enteric -- may be controlled to some extent None Alleviates and dampens signs of disease Better utilisation of animal feed ; less manure Increased production and improved productivity Lo w er labour dem and due to possibility of more intense production methods None Negative effect Lim its treatm ent possibilities due to developm ent of antimicrobial resistance; masks sub-clinical disease and infection; limits incentives for hygienic improvements Transfer of resistance to humans with increased societal costs for health care; shortens econom ic life of medical antim icrobials; occupational hazards through exposure to aerosol and dust contaminated with antimicrobials Cam ouflages stress associated with sub-clinical disease; allows higher stocking rates Increases the environmental pool of antibiotic resistance genes; antibiotic residues Stim ulates increased intensification of animal production Hampers the developm ent of animal-friendly production systems Cam ouflages bad feed quality; hampers im provem ents in feed form ulation and developm ent of alternatives Source: L-E. Edqvist and K. B. Pedersen 1945 1950s 1960s 1969 1970s 1975 1977 1984 1985 1997 1997 1998 1999 1999 2000 Antimicrobials: early warnings and actions A lexand er Flem m ing warns against misuse of penicillin as 'm icrobes are educated to resist' Antibiotic resistance widely recognised -- vertical transmission. Horizontal transmission recognised Swann C om m ittee recom m ends severe restrictions on antim icrobials in animal feed M ost Swann recom m endations initially im plem ented in the United Kingdom and EU Swann recom m endations are relaxed : tolysin and spiram ycin still perm itted as growth prom oters as human equivalents; vancom ycin com es into use Swedish Agriculture Board considers potential risk of antibiotic resistance, but concludes it is negligible Swedish farm ers ask for governm ent ban on antim icrobials in animal feed because of health and consumer concerns Swedish ban on grounds of antibiotic resistance in anim als and 'uncertain' long-term effects Swedish report concludes that risk of antibiotic resistance in humans is 'far from negligible' W H O scientific meeting concludes that it is 'essential to replace growth prom oting antim icrobials' EU bans four antim cirobials in animal feed as 'precautionary' m easure. EU Scientific Steering C om m ittee recom m ends phase-out of antim icrobials that may be used in human/animal therapy Pharmaceutical industry opposes EU bans and takes EU to the European Court; judgem ent expected end 2001 W H O recom m ends ban on antim icrobials as growth prom oters if used in human therapy and in absence of risk-based evaluation. 9.6. References A arestrup, F. M., 1995. `O ccurrence o f glycopeptide resistance am on g Enterococcus faecium isolates from conventional and ecological poultry farm s', Microbiology Drug Research Vol. 1, pp. 255-257. A arestrup, F. M., 1998. `A ssociation between decreased susceptibility to a new antibiotic for treatment of hum an diseases, everninomycin (SCH 27899), and resistance to an antibiotic used for growth prom otion in anim als, avilamycin', Microbiology Drug Research Vol. 4, pp. 137-141. A arestrup, F. M., Bager, F., Je n se n , N. E., M adsen, M. and Wegener, H. C., 1998. `Surveillance o f antim icrobial resistance in bacteria isolated from food animals to growth Table 9.2. 100 Late lessons from early warnings: the precautionary principle 1896-2000 prom oters and related therapeutic agents in D en m ark', APMIS Vol. 160, pp. 606-622. Amabile-Cuevas, C. F. an d Chicurel, M. E., 1992. `Bacterial plasm ids an d gene flu x ', Cell Vol. 70, pp. 189-199. Bates, J., Jo rd en s, J . Z. an d Griffiths, D. T., 1994. `Farm animals as a putative reservoir for vancomycin-resistant enterococcal infection in m an ', Journal ofAntimicrobial Chemotherapy Vol. 34, pp. 507-516. DVL, 1995. `The effect o f avoparcin used as feed additive on the occurrence of vancomycin resistant Enterococcus faecium in pig and poultry production', Report from the Danish Veterinary Laboratory. New York Times, 1945. Ju n e 26, p. 21. SO U , 1997. `A ntim icrobial feed additives', Swedish Government Official Reports, p. 132. SSC, 1999. `O pinion o f the Scientific Steering Committee on antimicrobial resistance', European Commission, Directorate-General XXIV. Stokstad, E. L. R. an d Ju kes, T. H., 1949. `Further observations on the anim al protein factor', Proceedings of the Society ofBiological and Experimental Medicine Vol. 73, pp. 523-528. Swann, M. M., 1969. Report, Joint Committee on the use ofAntibiotics in Animal Husbandry and Veterinary Medicine, HM SO, L on don . Fleming, A., 1929. `On the antibacterial action o f cultures o f penicillin, with special reference to their use in isolation o f B. influenzae', British Journal ofExperimental Pathology Vol. 10, pp. 226-236. Government Bill, 1984/85. Regeringens proposition 1984/85:149 om lag om foder, Sweden. Klare, I., Heier, H., Claus, H., Reissbrodt and Witte, W., 1995. ` VanA-mediated high-level glycopeptide resistance in Enterococcusfaecium from anim al husbandry', FEMS Microbiology Letters Vol. 125, pp. 165-172. LBS, 1977. `Utredning avseende utnyttjandet av specialam nen sasom antibiotika/ kemoterapeutika, vilka i tillvaxtbeframjande syfte inblandas i fo d e r', Lantbruksstyrelsen, Jonkoping. The Copenhagen Recom mendations, 1998. Report from the Invitational EU Conference on the Microbial Threat, Copenhagen, September. Walton, J . R., 1988. `Antibiotic resistance: An overview', Veterinary Record Vol. 122, pp. 249 251. WHO, 1997. `The m edical im pact o f the use of antimicrobials in food anim als', Report of a WHO meeting, Berlin. WHO, 2000. `Global principles for the containment of antimicrobial resistance due to antimicrobial use in animals intended for food', at http://www.who.int/emcdocum ents/zoonoses/ w h ocdscsraph 20004c.h tm l. Sulphur dioxide: from protection of human lungs to remote lake restoration 101 10. Sulphur dioxide: from protection of human lungs to remote lake restoration Arne Semb In Decem ber 1952 a dense smog -- a mixture of fog and coal smoke -- descended on L on d on . R ecords from L o n d o n 's hospitals showed that m ore than 2 000 people died from the exposure to air pollutants during one week, and there was strong pressure that this should never happen again. Yet the 1952 London sm og incident was not unique (Brim blecom be, 1987; Ashby, 1981). Worse sm og incidents had occurred previously in London, and the air quality there, as well as in other cities in the U nited K ingdom was actually improving, not deteriorating. What was new was the political and social set-up after the war, which m ean t that the public was no longer willing to accept the situation. The incident was also well docum ented, both with respect to health effects and air quality. The m easurem ents m ade by the scientists at St Bartholom ew 's H ospital showed that the concentrations of smoke particles and sulphur dioxide reached several milligrams per cubic metre. A parliamentary commission prepared a special report nam ed after its Chairm an, Sir H ugh Beaver. The remedies that the Beaver Report could propose were m odest. London was entirely dependent on coal for heating and energy purposes, and coal contains sulphur which can only partly be removed prior to use. So, the subsequent Clean Air Act instead considered the smoke com ponent o f the air pollution, allowing the local councils to establish `smoke-free areas', in which the infam ous coal fireplaces were to be replaced with m ore efficient com bustion units, electricity and `smokeless fuel'. Emission sources other than domestic fireplaces were to be `as far as possible' smokeless. A revision o f the Alkali Act in 1958 provided the Alkali Inspectorate with m ore authority in relation to industrial pollution sources, but did not specify emission limits or standards. `Best practicable m eans' was the governing principle. In practice this m eant that the main m ethod o f reducing sulphur dioxide (SO 2) concentrations at grou n d levels was the use o f tall chimneys, in proportion to the emitted amounts. These were obvious steps in the right direction, bu t slow, an d a second sm og incident that occurred in 1962 with about 800 additional deaths further dem onstrated the need for action. Oil from the Middle East becam e available in large quantities after 1950, and was to becom e the main source o f energy in western E urope during the following two decades (Mylona, 1996). As refining was relatively primitive, a substantial fraction o f the production was heavy or residual fuel oil, with a sulphur content o f 2.5-3 %. Gas oils with less than 1 % sulphur were also m arketed as fuel oil for use in small boilers and in the domestic market. Heavy fuel oil was used in the new electrical power generating plants in many countries in Europe, particularly in countries with no indigenous coal deposits, but also in the United Kingdom . It was also used in boilers for central heating in large building complexes, hospitals and other service buildings, and in industry. However, growing environmental awareness in the 1960s led to concern in cities across Europe about air quality, an d restrictions with respect to the use of different qualities of fuel oil and their sulphur contents began to appear. In fact, som e cities had experienced black smoke and SO 2 concentration levels com parable to those in London, but caused by the use of sulphur-containing fuel oil with im proper combustion. The Organisation for Economic Co operation and Development (OECD) had set up an Environmental Section in Europe in the late 1960s and after 1971 representatives o f the m em ber countries would m eet in an Air M anagem ent Sector Group to discuss com m on problem s and experience. In turn, this group took up the problem s o f urban air pollution, acid rain and photochemical oxidant formation, and m ade useful contributions to their solution. The question o f air quality standards, or guidelines, was particularly useful in connection with the planning o f new industries and licensing o f combustion sources. There was a good deal o f reluctance 102 Late lessons from early warnings: the precautionary principle 1896-2000 to accept fixed limits, which were mainly introduced to Europe under the influence o f the exam ple o f the U S Clean Air Act (United States o f America, 1970). However, Professor Lawther o f St Bartholom ew 's had already determ ined (WHO, 1972), on the basis o f the deaths during the London smog, that health damage occurred whenever concentrations o f SO 2 and black smoke, respectively, rose above 500 and 250 m icrogram s per cubic m etre (Mylona, 1996). Since concentrations o f pollutants are extremely dependent on weather conditions, air quality standards have to be fixed at considerably lower average levels, in order to safeguard the public against harmful exposures. Countries have had different attitudes to the establishment o f air quality standards, and particularly to their enforcem ent in relation to emission controls. Nevertheless, by collecting the available docum entation about health effects, the World H ealth Organization (WHO) has since 1979 recom m ended air quality guidelines that set a standard for the pollution abatem ent work in Europe (WHO, 1979 and 1987). 10.1. Dead fish, dying forests Increasing energy consumption in Europe during the 1960s resulted in a large increase in em issions o f SO 2. M uch o f the increased energy consum ption was in the form of electricity, produced in large therm oelectric power plants, and the SO 2 was released from chimneys m ore than 100 m etres tall. As a consequence o f this, as well as from the increased chimney heights in industry and in cities, surface air quality was gradually improved in spite o f the increased emissions. Optimistic representatives o f the electricity generating industry were confident that the emissions could be diluted and dispersed to levels that were not harmful. This was not so. While British and other European countries were striving to improve the urban air quality, energy consum ption and total emissions o f sulphur dioxide in Europe increased dramatically, as shown in figure 10.1. In Sweden the soil scientist Hans Egner had initiated measurements of air and precipitation chemistry across Europe (E gn er et al., 1955), supported by the m eteorologist C. G. Rossby. Their primary interest was the supply o f plant nutrients by precipitation, and the link with meteorology. However, in 1968 one o f their colleagues exam ined the results and concluded that the precipitation was becom ing m ore acidic as a result o f increasing SO2 emissions, and coupled this to observations o f acidification o f Swedish rivers. Svante O den chose to present his results with sweeping statements o f their implications, not only in a bulky report but also in a newspaper article (Oden, 1968 and 1967). Sweden chose also to give this as a case study at the 1972 UN environm ent conference in Stockholm , with evidence o f the dispersion, deposition and effects of SO 2and `acid rain' (Sweden, 1971). As intended, the creative presentation o f this information caught the interest o f the international community. Was it really possible that releases o f SO2 in England could change the water quality in Scandinavia, causing fish deaths and reducing forest production? Clearly, m ore docum entation was needed. Sulphur dioxide: from protection of human lungs to remote lake restoration 103 Annual em issions of sulphur dioxide in Europe 1 8 8 0-19 90 Figure 10.1. Source: Arne Semb (based on Mylona, 1996) Plans for a m ore comprehensive study o f the transport of air pollutants across national boundaries in Europe had been discussed in the OECD Air M anagem ent Sector Group since 1969. In 1972 a cooperative study was launched, with the participation o f 11 m em ber countries. O f the European OECD countries only Spain, Portugal and Italy decided not to participate, because they felt that the issue was not relevant to their situations. By arrangem ent within the Nordic countries, the Norwegian Institute for Air Research (NILU) was given the task of coordinating this long-range transport o f air pollutants program m e. The report was completed in 1977, establishing exports and im ports within the countries o f Europe and determining the relationship between emissions and deposition of sulphur com pounds (OECD, 1977). It was now quite clear that acidification in Scandinavia could be quantitatively related to em issions in several European countries, and that these countries were responsible for the damage which occurred from acid rain. A very ambitious Norwegian research program m e had been started in 1972 on acid precipitation and its effects on forests and fish, and by 1976 an interim report was prepared which showed widespread acidification o f rivers and lakes, and strong documentation of the declining stocks of both trout and salmon in such areas (Braekke, 1976). The report was not released, however, as the project was presented to an international audience in Ju n e 1976 (Knabe, 1976; Ambio, 1976). The suspected effects on forests were, however, much m ore difficult to demonstrate. The SNSF project was finalized in 1980 (Overrein et al., 1980; D rabl0s, 1980). Although strong scientific evidence was now available, resistance to change was strong. An editorial in the UK-based scientific jou rn al Nature (Nature, 1977) claim ed that acid rain was a million-dollar problem with a billion dollar solution, referring to the value o f the fish and the costs o f installing and operating flue gas cleaning on electric power plants. The reply from Erik Lykke (Lykke, 1977) at the Norwegian Ministry o f the Environm ent stated that this was obvious, but that the costbenefit analysis was incomplete. Emitter countries such as the United Kingdom should also consider the dam ages caused by releases o f SO 2 in their own countries, to health, materials, agriculture and to natural ecosystems. If all these dam ages were considered, the economic benefits of 104 Late lessons from early warnings: the precautionary principle 1896-2000 reducing the emissions would also be obvious. It was n ot a case o f E u ro pe vs. Scandinavia, it was Europe versus itself. A cost-benefit study carried out, again by the OECD (OECD, 1981), confirm ed that the value o f the lost fish populations was small in econom ic terms. The perceived costs in terms of timber production were also not particularly significant, but the damage caused to buildings and construction m aterials was shown to entail com parable costs to that o f reducing emissions by flue gas cleaning and fuel desulphurisation. In the event, the oil com panies would also benefit from fuel desulphurisation, because o f the h igher profits from light an d lowsulphur distillation products, com pared to heavy fuel oil. The oil com panies were well aware o f this, and had already started restructuring the refinery industry to produce less heavy fuel oil, and m ore light and m edium distillates, removing as m uch as 70 % o f the crude oil's sulphur content in the process. This move was also a result o f the high oil prices following the 1973 oil crisis, which m ade western European countries aware o f their reliance on the oil supply for electricity production. This resulted in the conversion o f oil-fired power plants to coal and building of new coal-burning power plants, while France switched to the use o f nuclear energy for electricity production. The situation in central and eastern Europe was different. In the Germ an Democratic Republic (GDR), Czechoslovakia and Poland, energy and industrial production had been increased mainly by the use o f large am ounts o f low quality coals with high sulphur contents, as the old, high-quality coal seams were depleted. The results were dramatic as the SO2 emissions in these countries increased by a factor o f 10 from 1960 to 1985, when these three countries alone emitted more SO 2 than all other countries in northwest Europe together. To combat acid rain and SO2 emissions without the participation o f the Eastern bloc countries was clearly not possible. East-West political relationships were tense during the Cold War period, and the discussions in the Conference o f Security and C ooperation in Europe in Helsinki in 1975 were difficult. In fact, the two factions could reach agreem en t on one issue only: cooperation on the issue o f airborne pollutants. Countries from both sides agreed to establish cooperation under the Econom ic Commission for Europe (E C E ), in Geneva, with participation from the World M eteorological O rganization and the United Nations Environment Programme. A cooperative programm e for monitoring and evaluation o f long-range transmission o f airborne pollutants in Europe (EM EP), was initiated in 1976, and was followed by the Convention on Longrange Transboundary Air Pollution (CLRTAP), in 1979. This convention expressed the signatories' intentions to investigate the problem , to exchange inform ation about emissions as well as effects, and to reduce, as m uch as possible, emissions causing harmful effects. The convention also set up a num ber o f task forces and cooperative program m es in addition to EMEP, which were to prepare assessments information for future protocol negotiations. By this time, forest decline had becom e a hot issue in Europe. Following releases by Germ an forest scientists at the University o f G ottingen (U lrich et al., 1980), a series o f articles in the Germ an popular jo u rn al Der Spiegel exposed the issue o f `n euartiger W aldschaden', or new forest damage. It appeared that forest dam age was widespread across Germany, but mainly along the border with the GD R an d Czechoslovakia. It was also m ost pronounced in freely exposed trees, in hilly areas and on ridges. The symptoms were quite general, and m ight be interpreted as general stress, although needle yellowing im plied m agnesium deficiency at som e sites. Different schools o f thought offered at least three different explanations, but none satisfied a stringent cause-effect relationship. Both soil acidification and direct action o f air pollutants were held to be responsible, continuing the division o f opinion dating back to when concerns first em erged in Germany in the 19th century. Meanwhile, reports of even m ore serious forest diebacks came from Bohem ia and Poland. In the border areas between the GDR and Czechoslovakia, several square kilometres of dead spruce forest were present in the areas now identified as those with the highest concentrations and deposition of airborne sulphur com poun ds an d acidity (M oldan et al, 1992). By this stage, public opinion and the press had been alerted. Acid rain an d fish-kills in rem ote lakes and rivers were identified in North Am erica as well as in Europe. When forests and terrestrial ecosystems were so Sulphur dioxide: from protection of human lungs to remote lake restoration 105 apparently endangered, the situation seem ed obvious to the public, even if the scientists did not agree on the mechanisms. In March 1983 the official Council of Environmental Advisers in Germany issued a report to the Germ an Bundestag relating to forest damage (Council o f Environmental Advisers, 1983). In their opinion, flue gas desulphurisation should already have been installed at coal-fired power plants where emissions exceeded 400 milligrams per cubic m etre o f SO 2. B ut the council was n ot convinced by the scientific explanation o f the new forest damage, and cautioned that SO 2 emission reductions m ight not improve this particular situation. Action justified on this basis alone m ightjeopardise the credibility of the environmental policy if subsequently found to be incorrect. In its own words: `As lon g as the preventive m easures aim at reducing the em issions o f the pollutants in question within the scope o f a long-term strategy an d in appropriate relation, there is hardly any danger of their being classified as unreasonable. This is also valid if the immission situation in clean air areas can only be im proved slowly as a result an d if positive influences on the vegetation can initially only be surmised. A balanced reduction of the emissions means that one m ust proceed from the objective criteria which speak for greater or lesser hazard potential o f the individual pollutants, and that one m ust reduce the total spectrum proportional to the apprehension, not in an unsystematic m anner or only under the influence of certain popular fears.' Since m uch o f the dam aged forest was located near the borders with G erm any's neighbouring countries, international cooperation in research and policy developm ent was also necessary. 10.2. The 1985 CLRTAP Protocol and beyond The 1985 CLRTAP Protocol, committing countries to 30 % reduction in sulphur em issions, was carried through by the socalled 30 % club, consisting o f the Nordic countries, the Netherlands, Switzerland and Austria, and jo in ed by the Federal Republic o f Germany. Two countries did not sign the protocol, Poland and the United Kingdom. A somewhat superfluous research programme carried out by the UK Royal Society and the research academ ies in Norway and Sweden concluded that acid rain did indeed kill fish (Mason, 1992), and the Chairm an o f the Central Electricity Generating Board in the U nited Kingdom was im pelled to explain to the UK Prime Minister that the accusations m ade by the Norwegian government regarding the long-range im pact o f the U nited K in gdom 's coal-burning power stations were scientifically well founded. Several severe winters with prolonged anticyclonic conditions and easterly flow over northern Europe provided strong reminders o f the deterioration o f the environm ent in the Eastern bloc countries during the 1980s. Over large areas of Germany, including western Germany, levels o f SO 2 and black smoke were observed that far exceeded the air quality guidelines recom m ended by WHO (Bruckm ann et al., 1986). This experience again dem onstrated that tall stacks did not necessarily solve the problem o f air pollution with the in creased em issions o f SO 2 that had occurred. H ealth warnings were issued to the inhabitants o f West Berlin, which were gleefully com m ented on by the official news services in the GDR. But the source was East Germany, together with the adjoining parts o f Poland an d Czechoslovakia, in the socalled `black triangle'. The collapse o f the com m unist regim es which followed was in m any ways related to the uneconom ic development o f energy-consuming heavy industries in these countries, and to the collapse o f the Soviet empire which had paid for the products with expensive fuel oil and petroleum products. N egotiations for a follow-up protocol on sulphur emissions took advantage o f this new situation (ECE, 1994). In addition, Germany and some other countries had started on an ambitious program m e of emission reductions, installing flue gas desulphurisation and denitrification in both new and old coal-burning electric power plants. In addition to desulphurised oil products, natural gas from Russia and from the North Sea was available to replace coal an d sulphur-containing fuel oil. By now, also, the scientific knowledge accum ulated in the various bodies and task forces under the CLRTAP could be used to ensure a costefficient use o f the emission reductions in relation to environmental damage. O ne o f the m ain reasons for the wish to reduce the emissions continued to be forest dam age, which had been dem onstrated in many countries through series o f national 106 Late lessons from early warnings: the precautionary principle 1896-2000 inventories o f tree vitality, involving assessments o f crown thinning and leaf and needle yellowing. However, there was still no consensus with respect to the cause, nor was there any indication o f quantitative causeeffect relationships. The problem, therefore, had to be attacked from another angle, more precautionary than the prevailing approach, which came to be known as the `critical load ' con cept (N ilsson et al., 1988). By definition, the critical load is the highest deposition o f a pollutant that does not change the ecosystem in an unacceptable way. For soil acidification this im plies a deposition o f sulphur (and nitrogen) which is balanced in relation to the release o f base cations (such as calcium) from the weathering o f soil minerals. In the case o f lakes and rivers the acid input m ust be less than the leaching o f base cations in the watersheds. The critical loads may be derived from soil surveys, or surveys o f lake chemistry, and have been extensively m apped all over E u rope (H ettelingh et al., 1991), providing a rational basis for the negotiations of the 1994 sulphur protocol. Because the sulphur deposition exceedances were mainly located in northern Europe, and because o f the state o f technical developments, emission reductions were largely called for in this area. It also proved to be too difficult to satisfy the m ost stringent requirem ents with respect to critical loads: only a 60 % gap closure was found to be achievable for the grids where the relative exceedances o f the critical loads were the highest. These international negotiations also have im plications for urban air quality. It has long been realised that reducing air pollutant concentrations in the cities will have ad d ed benefits for hum an health as well as in terms o f reducing material damage. New and stringent limit values for SO2 and other air pollutants are now included in European Union legislation, taking into account both the benefits o f improved air quality and the desirability o f reducing emissions (European Com m ission, 1999). It is interesting, however, to see that the short-term limit for protection o f hum an health, which is set at 350 m icrogram s per cubic m etre for one hour, is still not very m uch lower than the values proposed on the basis of the experience from the L o n d o n sm og. Lim its are also set for 24hour exposure, at 125 m icrogram s per cubic metre. It would seem that progress has been slow, as this lim it was exceeded for 34 % o f E u ro p e 's popu lation in 1990 (Stanner et al., 1995). Dam age due to air pollutants has been substantial to materials and historical monuments, particularly to masonry and iron constructions. As concentrations rise above 10-20 m icrogram s per cubic m etre the dam age to building materials and historical monuments significantly exceeds the natural degradation occurring in clean environments (C oote et al., 1991). D isfigurem ent o f medieval ornaments and other building details in carved sandstone, sculptures and other works o f art, and stained-glass windows, are am ong the costs not easily quantified. However, a num ber o f retrospective studies have shown that the costs o f reducing the emissions have m ore than been recovered in terms of generally reduced maintenance costs for buildings and steel constructions in the urban environment. Sulphur dioxide emissions over northern Europe have now been reduced by more than 50 % since the m aximum in 1980. Through the use of clean fuels, particularly natural gas, concentration levels in the urban environments have been even more reduced. In the United Kingdom, average urban concentration levels have been reduced by m ore than 70 %. A new and m ore stringent protocol to the CLRTAP has been negotiated (ECE, 1994), which will lead to further S O 2 em ission reductions. T he m ain em phasis is shifting towards interactions between pollutants such as nitrogen oxides, am m onia and volatile organic hydrocarbons, the combined acidifying effects of sulphur and nitrogen deposition, the form ation of photochem ical oxidants, and the eutrophication effect o f transboundary transport of nitrogen compounds. T he acid lakes an d stream s show a slow, but steady, recovery (Skjelkvale et al., 1998). Even forest vitality is improving, according to the latest surveys (ECE and EC, 1997), although the evidence is not unam biguous. It seem s that the trout and salm on populations, however insignificant in economic terms, have played an im portant educational part in m aking Europe understand the full implications o f the emissions o f air pollutants. 10.3. Late lessons So what role has precaution played in this story? Generally action has only been taken on the basis of p roof beyond reasonable doubt, and protagonists have felt compelled to justify action on such a basis. Only when the issue was taken to the international level Sulphur dioxide: from protection of human lungs to remote lake restoration 107 could significant change occur. It was relatively easy for the Scandinavian countries to argue for action, as the initial perception was that they were bearing m ost o f the costs, but few o f the benefits, o f fossil fuel com bustion in other countries. The United Kingdom regarded itself as bearing high costs but gaining little benefit from action to reduce long-distance transportation of sulphur emissions; arguably it was only when it was appreciated that acid rain had costs closer to hom e that the political climate changed. In the case o f eastern Europe, the countries could not afford the short-term costs involved, no matter what the merits o f action. Throughout the early history o f SO 2 pollution and acid rain, the precautionary approach was altogether absent, mainly because decision-makers had limited understanding o f the problem s. The tall stack approach was a clear exam ple o f the reluctance to accept effects beyond the immediately obvious. All this was to change perceptibly when the issue came to be the subject o f an ECE convention, which was also seen as an im portant instrum ent in the im provem ent of political relations between eastern and western European countries in the 1980s. Another m ajor elem ent was the political pressure o f the G reen Party in Germany. It is noteworthy that the Germ an Council of Environmental Advisers was concerned that statements about forest damage, couched unjustifiably as a scientific fact, established in terms o f proof beyond doubt, could underm ine the wider basis for action if these should subsequently be demonstrated as unfounded. If there was general acceptance o f the need for action based on a lower standard of proof -- a precautionary approach -- such potential problem s could be side-stepped. In reality, the coun cil's argum ent was for the record. Politically, the decision had already been taken. At the international level, the machinery o f the CLRTAP had to find a rational approach to negotiations about the limitation and reduction o f SO 2 emissions. In principle, the task was to find an approach which m inim ised the adverse effects, while distributing the costs o f emission reductions evenly am ong the different countries. The many task forces and committees within the convention provided substantial scientific advice and knowledge for this task, and served also to guarantee that the signatories' national interests were taken into account. W hat the task force on effects cam e up with was the critical load concept. This can be regarded as related to the precautionary approach because it aims at a situation where no dam age or adverse effects are expected to occur. This has the effect o f shifting the burden of proof, from demonstrating and calculating in economic terms the damage which could occur in the future from decreased growth o f forests, or loss o f species, to the relatively sim ple task o f determining which level o f deposition can be accepted without perceptible chemical changes in soil composition. With respect to water acidification, the critical loads calculated also represent som e simplifications, which are accepted only because o f reasonably wide safety m argins regarding the survival of sensitive species. Thus, while not exactly representing an application of the precautionary principle, the application o f critical loads represents a rational way o f dealing with uncertainties regarding the environmental effects o f acid rain. 108 Late lessons from early warnings: the precautionary principle 1896-2000 Table 10.1. Sulphur dioxide: early warnings and actions 1952 1952 1962 1968 Late 1960s 1972 1972 1972 1977 1979 1979 1980s 1985 1988 1994 A dense smog kills more than 2 000 people In London The Beaver Report prepared by a parliam entary commission proposes m odest remedies Second smog incident kills 800 people in London A cidification of precipitation and rivers in Sw eden is linked to sulphur dioxide em issions in other countries An O E C D A ir M anagem ent Sector G roup is set up, contributing to solving air problem s Further evidence of acidification of Swedish lakes presented to the UN environm ent conference, Stockholm O E C D acid rain study is launched O E C D program m e on long-range transport of air pollutants is launched O E C D report is published, determ ining the relationship betw een em issions and depositions of sulphur compounds Convention on Long-range Transboundary A ir Pollution (C LR TA P ) is agreed W H O recomm ends air quality guidelines, setting a standard for the pollution abatem ent work in Europe Forest death from 'acid rain' in G erm any, Poland, Czechoslovakia and North A m erica C LR T A P Protocol agrees 30 % reduction in sulphur em issions The EU directive on large com bustion plants is published, and am ended in 1988. Th ese were very efficiently used by Germ any against the British opposition to emission controls Second sulphur protocol, based on the critical load concept that had already been m apped all over Europe 10.4. References Ambio, 1976. Vol. 5, pp. 200-252. Ashby, E. an d A nderson, M., 1981. The politics of clean air, O xford University Press. Lon don . Braekke, F. (e d .), 1976. Impact of acid precipitation on forest and freshwater ecosystems in Norway, N orges landbruksvitenskapelige forskningsrad, Oslo-As. Brim blecom be, P., 1987. The big smoke: A history of air pollution in London since medieval times, R outledge, Lon don . dam age and air pollution), Kohlhammer, Mainz. D rabl0s, D. an d Tollan, A., (ed s.), 1980. `Ecological im pact of acid precipitation. Proceedings of an international conference', Sandefjord, Norway, M arch 11-14, 1980. `Editorial', 1977. Nature Vol. 268, p. 89. ECE, 1994. Protocol to the 1979 Convention on Long-range Transboundary Air Pollution on further reduction of sulphur emissions, United Nations, Geneva. Details may be found at www.ece.org. Bruckm ann, P., Borchert, H., Klske, S., L acom be, R., Lenschow, P., Mller, J., and Vitze, W., 1986. Die Smog-Periode imJa n u a r 1985. Synoptische Darstellung der Luftbelastung in der Bundesrepublik Deutschland. Bericht des Lnderausschusses f r Immissionsschutz, Ministerium fr Umwelt, Dsselsdorf. Coote, A. T., Yates, T. J . S., Chakrabarti, S., Biland, D. J., R iland, J . P., an d Butlin, R. N., 1991. UN/ECE international cooperative programme on materials, including historic and cultural monuments. Evaluation of decay to stone tablets, Building R esearch Centre, Garston, Watford, UK. Council of Environmental Advisers (Rat von Sachverstndigen fr Umweltfragen), 1983. Waldschden und Luftverunreinigungen (Forest EC E an d EC, 1997. Forest Condition in Europe, Executive Report, Federal R esearch Centre for Forestry and Forest Products, Hamburg, Germany. European Commission, 1999. Council Directive 1999/30/E C o f 22 April 1999 relating to limit values for sulphur dioxide, nitrogen dioxide and oxides o f nitrogen, particulate m atter an d lead in am bien t air, Off. J. Eur. Communities L163, 29.6.1999, pp. 0041-0060. Egner, H., Brodin, G., and Johansson, O., 1955. `Sampling technique and chemical examination o f air and precipitation', Kungliga Lantbrukshogskolan Ann. Vol. 22, pp. 369-410. Sulphur dioxide: from protection of human lungs to remote lake restoration 109 H ettelingh, J.-P., Downing, R. J., an d de Smet, P. A. M., 1991. M apping critical loadsfor Europe, National Institute o f Public Health and Environmental Protection, Bilthoven, Netherlands. Knabe, W., 1976. Internationale Konferenz ber die Wirkung saurer Niederschlge in Telemark, Norw egen, 14 bis 19 Ju n i 1976. Lykke, E., 1977. `Europe versus itself', Letter to the Editor, Nature Vol. 269, p. 372. M ason, B. J., 1992. Acid Rain. Its Causes and its Effects on Inland Waters, C larendon Press, Oxford. M oldan, B. and Schnoor, J., 1992. `Czechoslovakia. Exam ining a critically ill environm ent', Environmental Science and Technology, Vol. 26, pp. 14-21. Mylona, S., 1996. `Sulphur dioxide emissions in Europe 1880-1991 and their effect on sulphur concentrations and depositions', Tellus Vol. 48B, pp. 662-689. Nilsson, J . an d Grennfelt, P., 1988. Critical loads for suplhur and nitrogen -- reportfrom a workshop held at Skokloster, Sweden, 19-24 March 1988, Nordic Council o f Ministers, Copenhagen. O den, S., 1967. Dagens Nyheter, 24 O ctober 1967, Stockholm. O den, S., 1968. `N ederbrdens och Luftens Frsurning, dess Orsaker, Frlopp och Verkan i Olika M iljer' (The acidification of precipitation an d air, causes and effects in different environm ents), Statens Naturvetenskapliga Forskningsrad, Ekologikomiteen, Bulletin No. 1. OECD, 1968. Methods for measuring air pollution, O rganisation for Econ om ic Co operation and Development, Paris. OECD, 1977. The OECD programme on long range transport of air pollutants. Measurements and findings, Organisation for Econom ic Co operation and Development, Paris. OECD, 1981. The costs and benefits of sulphur oxide control, O rganisation for Econom ic Co operation and Development, Paris. Overrein, L. N., Seip, H. M., and Tollan, A., 1980. `Acid precipitation -- effects on forest and fish', final report of the SNSF project 1972-1980, Oslo-As. Stanners, D. and B ordeau, P., 1995. E u ro p e 's Environment. The Dobris Assessment. European Environm ent Agency, Copenhagen. Skjelkvale, B. L., Wright, R. F., and Henriksen, A., 1998. Norwegian lakes show widespread recovery from acidification; results from national surveys o f lakewater chemistry 1986-1997, Hydrology and Earth System Sciences, Vol. 2, pp. 555-562. Sweden, 1971. `Air pollution across national boundaries. The im pact on the environment o f sulfur in air an d precipitation ', Sw eden's case study for the UN conference on the hum an environment, Royal Ministry for Foreign Affairs, Royal Ministry of Agriculture, Stockholm. Ulrich, B., Mayer, R., an d Khanna, P. K., 1980. Chemical changes due to acid precipitation in a loess derived soil in Central E urope, Soil Science, Vol. 130, pp. 193-199. United States o f America, 1970. `Clean Air Act'. WHO, 1972. A ir quality criteria and guides for urban air pollutants, R eport o f a WHO expert committee, Technical report series No 506, World Health Organization, Geneva. WHO, 1979. Sulphur oxides and suspended particulate matter, Environm ental health criteria No 8, World H ealth O rganization, Geneva. WHO, 1987. A ir quality guidelines for Europe, WHO regional publications, European series No 23, World H ealth Organization, Regional Office for Europe, Copenhagen. 110 Late lessons from early warnings: the precautionary principle 1896-2000 11. MTBE in petrol as a substitute for lead Martin Krayer von Krauss and Poul Harremoes 11.1. Introduction A core notion o f sustainability is that o f futurity, whereby developm ent in the presen t should not com prom ise the ability o f future generations to m eet their needs. Although it may not be possible for us to predict exactly what the needs o f the future will be, we can respond by developing procedures for making robust and flexible decisions. T he goal o f this case study is to exam ine how foresight, em bodied in the precautionary principle, can lead to robust and flexible decisions that m eet the needs of the present, but that can be adapted to m eet the changing needs of the future. Recently there has been much debate over the use o f methyl tert-butyl ether (M TBE) in petrol. Concerns in particular over the potential for contamination of groundwater supplies have prom pted regulative authorities in western nations to reassess the risks associated with the use o f MTBE. After presenting the background, this case study will exam ine whether, at the time o f M T B E 's introduction, it w ould have been possible for the petrochem ical industry and regulatory agen cies to foresee that M T B E 's physical, chemical and m icrobiological properties m ight eventually be deem ed to be undesirable characteristics. T he analysis will then be pursued to illustrate how foresight, in the form o f som e essential precautionary questions, can be applied to arrive at robust and flexible decisions. The editing o f this contribution was finished on 10 May 2001. Undoubtedly, the debate on M TBE will continue after that date an d som e aspects o f the paper may be quickly outdated, but it is h op ed that the p ap er will be a valuable contribution to the debate on the precautionary principle. 11.2. Lead in petrol The toxicity o f lead has been known to man since ancient times. The Rom ans were aware o f the fact that the lead they used to make their utensils was poisoning people. The field of occupational health formally established the toxicity o f lead in the 1800s. Already in the early 1920s, at the very beginning o f the autom obile industry, the wisdom o f using lead in petrol was questioned. The debate was reignited in the 1960s. There developed a widespread consensus that a long-term trend of rising airborne lead concentrations was undesirable and it was eventually decided by m ost western nations to phase out lead in petrol. The gradual phase-out o f lead in petrol began in the late 1970s and by the mid-1990s, m ost western countries had phased out lead in petrol. As a substitute to lead, the petrochem icals industry chose MTBE. 11.3. The MTBE case The choice of MTBE as an engine anti knocking agent to replace lead was m ade on the basis o f a variety o f advantageous properties: it is a low-cost and easily produced chemical with favourable transfer and blending characteristics; it can be produced at the refinery; it blends easily without separation in petrol; and the blend can be transferred through existing pipelines (Squillace et al., 1996). Com m ercial production o f MTBE started in Europe in 1973 and in the United States in 1979 (DeWitt & Com pany Inc., 2000). Italy introduced MTBE into petrol in the late 1970s, followed by other European countries, mainly in the mid-1980s. Total worldwide annual production was about 21.4 million tonnes in 1999. Roughly 3.3 m illion tonnes of M TBE were produced in the European U nion (EU) in 1999. Som e 2.3 million tonnes were used domestically, 1.1 m illion tonnes were exported and 0.2 million tonnes were im ported (DeWitt & Com pany Inc., 2000). In 1995 M TBE was the third m ost produced organic chemical in the United States, at approxim ately 8.0 m illion tonnes p e r year (Joh n son et al., 2000). M TBE is therefore a very high production-volume ch em ical.' In the United States, the use o f MTBE greatly increased following the 1990 am endm ents to the Clean Air Act requiring petrol `oxygenates', such as M TBE or ethanol, to be MTBE in petrol as a substitute for lead 111 added to petrol to reduce atmospheric concentrations of carbon m onoxide (CO) or ozone. In January 1995 a special blend of petrol (dubbed `reform ulated gasoline', RFG) containing m ore M TBE was introduced for use in m etropolitan areas in the United States that had severe ozone pollution problems. The present average MTBE content in RFG in the U nited States is aroun d 11 % (N STC, 1997; EN D S Env. Daily, 2000). T here is also a large volum e o f conventional petrol sold in the United States, with a 2-3 % M TBE content. In the EU, the application o f M TBE was regulated by a 1985 directive (8 5 /5 3 6 /E E C ) on additives to petrol, by which no M ember State can prevent the use o f organic oxygenates at concentrations lower than regu lated by the directives. For M TBE this is a m axim um o f 15 % by volum e. In 1999 in the EU , the average was 2.1 %, calculated from the total consum ption figures on motor petrol and MTBE consumption. Actual concentrations o f M TBE in petrol are highly variable, from 0 to 15 %, d ep en d in g on petrol grade, oil company and country (Finnish EPA, 2001). In the following sections we will attem pt to summarise the relevant background inform ation. As the aim is to exam ine the potential use o f the precautionary principle, particular attention will be focused on areas o f uncertainty within current scientific knowledge. although the use of RFG results in significant reductions in automotive pollutant emissions relative to conventional petrol, studies have shown that CO, NOx and volatile organic com pound (VOC) emissions are not significantly affected by including MTBE in petrol (NRC, 1996). In a more recent report investigators concluded that the use of oxygenates in RFG had a negligible impact on am bient ozone concentrations (NRC, 1999). It has also been reported that the benefits of MTBE are decreasing, because these mostly apply to older vehicles, which are progressively being replaced by vehicles with m ore effective em ission controls (Keller et al., 1998b). 11.5. The impacts of MTBE 11.5.1. Persistency in groundwater Concerns regarding the use o f MTBE began to arise following studies indicating that MTBE moves at about the same speed as groundwater, making it a very m obile chem ical (Barker et al., 1990). These concerns becam e widespread following a report published by the U S Geological Survey in 1996. O f VOCs analysed in samples o f shallow am bient groundwater collected from eight urban areas during 1993-94, M TBE was the second m ost frequently detected chem ical (Squillace et al., 1996). Its solubility is in the order o f 50 000 m g /l (milligrams per litre). A level o f 10 % M TBE in petrol gives an equilibrium concentration in water o f the order o f 5 000 m g/l. 11.4. The benefits of MTBE The original purpose o f adding M TBE to petrol was to prevent `knocking'. In addition M TBE or other oxygenates allow more com plete combustion, which in turn leads to a reduction of CO emissions and a reduced emission of precursors for ozone formation (K oshland et al., 1998). M TBE also achieves some elevation in the octane rating that alternative com ponents o f fuel such as benzene (a class 1 carcinogen) and other aromatics provide. The role o f MTBE in reducing air emissions has however been questioned. In this respect, it is im portan t to distinguish between the general effects o f using RFG and those associated specifically with the presen ce o f MTBE in RFG. During the 1990s the United States saw a significant decrease in CO , lowlevel ozone an d nitrogen oxide (N O x) pollution in m ajor cities using RFG. But, Persistency can be discussed in terms of physical, chemical or microbiological persistency. Uncertainties concerning the persistency o f MTBE in soil and groundwater still rem ain. M TBE is rapidly ph otod egrad ed in the atm osphere. The available data in the sub-surface is limited, there being few welldocum ented reports concerning attenuation o f M TBE in the field. Two studies have shown that biodegradation could not be detected in the groundwater under field conditions (B orden et al., 1997; Schirm er and Barker, 1998). However, in subsequent laboratory studies, a very slow but m easurable biodegradation was observed (Landm eyer et al., 1998). Each o f these three investigations occurred under aerobic conditions. The results o f field investigations have shown that MTBE can be degraded anaerobically (Sufflita and M orm ile, 1993; M orm ile et al., 1994; Yeh and Novak, 1994; H u rt et al., 1999). But significant rates o f MTBE degradation, either biologically (Eweis et al., 1998a an d b) 112 Late lessons from early warnings: the precautionary principle 1896-2000 or chemically have only been observed under laboratory conditions. Thus, because o f its high solubility, mobility and persistency, M TBE poses a potential risk to groundwater. 11.5.2. Aesthetics The aesthetic im pact of MTBE contam ination o f groundw ater is considerable. The potent terpene-like odour in water is detectable by people at very low concentrations. The odour threshold was determ ined to be 180 g / l (m icrogram s per litre) in D enm ark (Larsen, 1997), but as low as 5 g / l in California (CAL-EPA, 1999). Based on current monitoring data, MTBE has been shown in som e cases to be present in drinking water at concentrations exceeding taste and odour thresholds. Relatively small am ounts o f M TBE may thus render large reserves o f groundwater useless. When the taste and odour threshold in water is exceeded, contam inated drinking water is normally not used, requiring the utilisation o f alternative supplies. When a large and im portan t groundw ater reservoir is contaminated the consequences can be remarkable, both in terms o f cost and disruption. Leaking underground storage tanks and spillage from overfilling the tanks are the main cause o f groundwater contamination (Finnish EPA, 2001). T he severity o f the consequences may vary greatly between countries depending on, for exam ple, the level o f groundwater utilisation for drinking water and the condition o f petrol station underground storage tanks. The num ber of pollution cases varies significantly from country to country in the EU. The EU risk reduction report (Finnish EPA, 2001) states that `M TBE concentrations are currently not routinely m on itored in groundwater. T here is only limited data on MTBE concentration trends'. In a recent report to the European Commission (Arthur D. Little Limited, 2001) it is stated that `little public inform ation is available across M ember States regarding monitoring of groundwater contamination by M TBE'. The report refers to unpublished data as a significant source of information, com bined with data available from six M ember States (Denmark, Finland, France, Germany, Sweden and the United K ingdom ). It concludes that `none o f the findings indicated widespread or serious groundwater contamination by M TBE on the same scale as the USA'. However, according to the EU risk assessm ent study, the `docum ented cases provide sufficientjustification for concern that MTBE poses a risk for the aesthetic quality o f drinking water from groundwater supplies'. The EU risk reduction report states: `It isju stified to conclude that M TBE is causing a risk for the aesthetic quality o f drinking water.' (Finnish EPA, 2001) 11.5.3. Cancer The classification of the carcinogenicity of M TBE is the focus o f a continuing controversy (see Box 11.1.). The recent EU risk assessm ent concluded that M TBE was a borderline case for classification as a carcinogen (Finnish EPA, 2001). Nonetheless, the European Chemicals Bureau has decided that MTBE should not be classified as a carcinogen (Dixson-Decleve, 2001). For the purpose o f this case study, it is noteworthy that the issue of carcinogenicity was not addressed until long after the marketing o f MTBE as a high productionvolume chemical: no comprehensive pre marketing carcinogenicity testing was carried out. Even today, after in-depth investigations, uncertainties about the carcinogenicity of MTBE remain. The ongoing debate in the United States and the EU over the classification o f MTBE illustrates the difficulty o f dealing with uncertainty and borderline results. 11.5.4. Asthma A sim ilar debate is taking place regardin g whether the use o f M TBE in petrol has contributed to an increase in urban asthma. In an editorial publish ed in the Archives of Environmental Health in February 2000 (Joseph, 2000), concerns were raised over epidemiological studies indicating an increase in the occurrences o f asthm a and other conditions of respiratory an d /o r inflammatory natures in the years following the introduction o f M TBE to the cities o f Philadelphia (Joseph, 1997) and New York (Crain et al., 1994; Leigh ton et al., 1999). The past five years have seen reports o f alarmingly high rates (above 20 %) o f asthma am ong children in several large cities on the east coast o f the United States where M TBE was in use (M angione et al., 1997; M cBride, 1996; Hathaway, 1999; Leigh ton et al., 1999). It has been suggested that the increase in asthm a is due to an unknown exhaust product of M TBE combustion entering the am bient air (Joseph, 1999; L eik au f et al., 1995). The University o f California, Davis report states that it is plausible that combustion products of MTBE could exacerbate or even MTBE in petrol as a substitute for lead 113 cause asthma, but that there have been no studies to date designed to address this issue. They conclude that `there is little evidence at present either to implicate or exonerate M TBE as a cause or exacerbating factor in asth m a' (Keller et al., 1998b). 11.5.5. Other im pacts There are a few investigations indicating that MTBE may function as an endocrine disrupter (Williams et al., 2000; M oser et al., 1998; Day et al., 1998). However, the mechanism s by which M TBE may disrupt endocrine systems remain unclear and further research is required. In this respect, the EU risk assessment concluded: `Because the data is not considered to be sufficient, no N O A EL (n o observed adverse effect level) is assigned.' (Finnish EPA, 2001) Box 11.1. Carcinogenicity -- complex science, borderline results and conflicting interests Comprehensive investigations of carcinogenicity were initiated in the 1990s. Evidence from three animal bioassay studies demonstrated that chronic exposure to M TBE -- either by inhalation or orally -- causes cancers in animals. M TBE inhalation exposure caused an increased incidence of kidney and testicular tumours in male rats (Bird e t al., 1997; Chun e t al., 1992), and liver tumours were identified in mice (Burleigh-Flayer e t al., 1992). Oral administration of M TBE produced a statistically significant increased incidence of lymphomas and leukaemias in female rats and testicular tumours in male rats (Belpoggi e t al., 1995 and 1998). Based on these results, it would seem that M TBE induces benign and malignant tumours at multiple sites, in multiple species, and via multiple routes of exposure. However, the results have been criticised on the grounds that the 'conduct and reporting in the carcinogenicity studies is inadequate, interpretation of the obtained results is difficult and a relation of the observed increases in tumour incidences to M TBEadministration is questionable. M oreover, rodentspecific mechanisms of action have been established for several of the tumours induced by M TBE, thus making the results inapplicable to man.' (Dekant, 2000) According to the International Agency for Research on Cancer (IARC) of the W orld Health Organization (W HO), no analytical epidem iological studies have addressed a possible association of M TBE with human cancer. IARC concluded that because there is 'inadequate evidence in humans' for carcinogenicity, and only 'limited evidence' for experimental animals, 'M TB E is not classifiable as to its carcinogenicity to humans' (IARC, 1999). The US National Toxicology Program's (NTP's) Board of Scientific Counselors voted 6 to 5 not to list M TBE as being 'reasonably anticipated to be a human carcinogen' (BSC, 1998). The International Programme on Chem ical Safety (IPCS), a joint programme of the International Labour Organisation, the United Nations Environment Programme and W HO , echoed the IARC's conclusion, recommending that in order to provide quantitative guidance on relevant limits of exposure and to estim ate risk, the acquisition of additional data in several areas remains necessary (IPCS, 1999). O ther institutions and individual scientists have decided that the evidence provided by the animal bioassay studies is sufficient to conclude that M TBE is likely to cause cancer in humans and should therefore be labelled a probable human carcinogen (Mehlman, 2000). A review of the existing studies by the University of California (Keller e t al., 1998b), indicated that M TBE is a potential human carcinogen, although further studies would be needed to make a conclusive determ ination. The W hite House National Science and Technology Council concluded that 'there is sufficient evidence that M TBE is an animal carcinogen' and that 'the w eight of evidence supports M TBE as having carcinogenic hazard potential for humans' (NSTC, 1997). The US Environmental Protection Agency (EPA) concluded that 'the weight of evidence indicates that M TBE is an animal carcinogen, and M TBE poses a carcinogenic potential to humans' (US EPA , 1997). The most recent in-depth review of the available carcinogenicity data was performed by the Finnish E P A in the context of the EU Risk Assessm ent on M TBE (Finnish EP A , 2001). In the draft risk assessm ent report, the conclusion on carcinogenicity states: 'There are indications of carcinogenicity in two species. However the treatm ent relation of the occurred tumours is equivocal in some studies (mouse adenoma) and the relevance of the mode of action is questionable in others (Leydig cell). M oreover, the tumours appear mostly at very high and system ically toxic doses, and M TBE is not genotoxic In vitro or In vivo. On the other hand, the human relevance of the testicular interstitial adenom as observed in rats on two separate rat strains cannot be neglected. In addition, some uncertainty remains as to the significance of the lymphatic tumours found, especially in the light of the limitations of the study and somewhat inadequate study quality reporting. The rapporteur considers M TBE as a borderline case between non-classification and Carcinogenicity Category 3.' (Finnish EPA, 2001) The classification and labelling of M TBE for hazard was discussed by the W orking Group on the Classification of Dangerous Substances at the European Chemicals Bureau in N ovem ber 2000. A proposal to classify M TBE as a carcinogen in Category 3, with an R40 risk phrase was rejected (R40 states that a substance poses 'possible risk of irreversible effects') (Dixson-Decleve, 2001). The probable basis for this decision is that it is generally accepted that there is a threshold concentration below which no effects can be observed for carcinogenic substances that are not mutagenic. For M TBE, several In vitro m utagenicity assays with m etabolic activation are positive, while In vivo studies are negative. It is believed that the positive responses occur when cells In vitro are exposed extracellularly to the M TBE metabolite formaldehyde, and that form aldehyde will not exert this effect when it is generated intracellularly, in the intact organism . For this reason, M TBE is regarded as non-mutagenic (Finnish EPA, 2001). Thus the level of concern for M TBE is not as high as it would be for a mutagenic carcinogen w here, in theory, effects can be expected at all doses, right down to infinitely small concentrations. For non-mutagenic carcinogens, the effect is often connected to some other damage to the target organ, for instance the kidney or liver. The high regeneration and growth of cells that then takes place in connection with tissue regeneration in the damaged organ can be accompanied by an increased tumour production. For the primary dam age on the target organ, there will be a threshold value. This threshold value is the no observed adverse effect level (N O AEL). By remaining below the threshold value, damage to the organs is prevented, as is the formation of tumours (stergaard, 2001). Thus, the risk of carcinogenicity posed by such substances is considered m anageable. 114 Late lessons from early warnings: the precautionary principle 1896-2000 11.5.6. M TBE in drinking water In late 1995, the authorities in the city o f Santa M onica in California detected MTBE in a well that supplied the city with drinking water. ByJu n e the following year the problem h ad escalated an d the city authorities were forced to shut down som e o f the potable water supply groundwater wells. A review o f known and suspected petroleum spill sites identified about 10 potential sources o f contam ination within 1 kilometre o f the well field (Joh n son et al., 2000). As a result o f the M TBE contam ination the city lost 71 % o f its local water supply. About h alf its total water consumption must now be bought from outside sources at a cost o f USD 3.5 m illion a year (Rodriguez, 1997). Some MTBE plumes have originated from very small spills. For example, a spill resulting from a single autom obile accident in Standish, ME, led to M TBE transport through m ore than 0.7 kilometres o f fractured rock and the contamination of m ore than 20 domestic wells (Joh n son et al., 2000; H unter et al., 1999). By now there are many docum ented cases o f MTBE pollution o f groundwater in the United States, the num ber is increasing and where pollution occurs the consequences can be serious. Assuming treatm ent by granular activated carbon, the cost of purifying MTBE contaminated drinking water sources could be enormous, due to the extremely low concentration limits for MTBE in drinking water (in the order of 5 to 30 g / l ) . The cost o f dealing with underground storage tank and pipeline leaks and spills could be o f the order o f tens to hundreds o f m illions o f dollars per year in California alone (Keller et al., 1998a). This however could change in light o f yet unpublished results from work indicating that MTBE could potentially be biologically degraded in ordinary waterworks using sand filtration (Arvin, 2001). 11.6. Responses In Denmark, a report on M TBE was published in 1998 ( M i j og Energiministeriet, 1998). The introduction acknowledges: `In 1990 The Danish EPA received information from Professor Erik Arvin, Institute for Environm ental Science and Engineering, Technical University o f Denmark, that there m ight be a problem in relation to groundwater, because o f the mobility and apparent persistency in groundwater.' The Danish EPA did not consider that to be a problem at the time, because leakage from tanks was considered a m inor problem , and `because problem s with gasoline com ponents were rarely found in groundwater at the tim e'. In 1998, the Danish EPA introduced a tentative limit for drinking water at 30 g /l. In California, the first US state to suffer the consequences o f MTBE, the California Legislature created a task force within the University o f California. The task force evaluated the available evidence, produced an assessment of health and environmental impacts and m ade several recom m endations to the Governor and the Legislature, including the phase-out o f M TBE (Keller et al., 1998b). As a result, a G overnor's Executive Order in March 1999 recom m ended developing `a timetable by Ju ly 1, 1999 for rem oval o f M TBE from petrol at the earliest possible date, but no later than Decem ber 31, 2002' (Davis, 1999). The University o f California study also concluded that: `A lesson to be learn ed from the M TBE story is that addition o f any chemical com pound to the environment in quantities that constitute a significant fraction o f the total content o f gasoline may have unexpected environmental consequences. Therefore, we recom m end a full environmental assessment of any alternative to M TBE.' The intention here is to emphasise that any alternative to MTBE should u n dergo a full assessm ent prior to its introduction. This m easure is in particular due to the `potential adverse health effects associated with incom plete com bustion products of ethanol'. The report also observed that: `The current structure o f state agencies which focus on specific m edia (land, air, water), leads to fragm en ted and incomplete environmental impact assessm en ts.' A `Blue Ribbon Panel' was established by the U S EPA in 1998 `in response to the growing concerns from State and local officials and the public' (Blue Ribbon Panel, 1999). The report o f Septem ber 1999 concluded inter alia: `The Panel agreed broadly that, in order to minimise current and future threats to drinking water, the use o f M TBE should be reduced substantially. Several m em bers believed that the use of MTBE should be phased out completely.' One lesson, the pan el concluded, is that: `In order to prevent future such incidents... EPA should conduct a full, m ulti-m edia assessm ent (o f effects on air, soil, and water) o f any m ajor new additive to gasoline prior to introduction.' In March 2000, the US EPA announced that steps MTBE in petrol as a substitute for lead 115 would be taken to `significantly reduce or elim inate' MTBE as a petrol additive, as a `possible carcinogen' and in order `to protect America's drinking water supplies' (ENDS Env. Daily, 2000). 11.7. Present trends T he future in the U nited States is likely to see a shift towards other oxygenates such as ethanol, as well as equivalent non-oxygenated petrols, which m eet the strict emission requirem ents (Keller et al., 1998b; CEC, 1998). In Europe, the comparatively low use of MTBE has m eant that concerns and possible problem s associated with M TBE have developed later than in the United States (M orgenroth and Arvin, 1999). The EU regulations introduced on fuel quality covering both petrol and diesel in 1998 (9 8 /7 0 /E E C ) call for com pounds such as benzene (class 1 carcinogen) and overall arom atics to be reduced in steps by the years 2000 and 2005. This could result in an increase in the use o f M TBE in Europe. In August 2000 the Danish EPA placed MTBE on its list o f undesirable substances and announced that it planned to introduce economic instruments to discourage the use of MTBE. The recent EU risk assessment of M TBE concluded that because o f the risk that M TBE poses to groundwater reservoirs, risk reduction m easures are justified (Finnish EPA, 2001). A report to the E u ropean Commission on risk reduction for underground storage tanks was released on 19 A pril 2001 (Arthur D. Little Lim ited, 2001) as a supplem ent to the risk reduction report written in connection with the risk assessm ent (Finnish EPA, 2001). The reports stress the difference between the US and EU situations. T he anticipation is that `M TBE octane requirem ents will settle out in the 1-4 percent volume range, depending on the available octane, an d will still be well below the 10-15 percent currently used in reform ulated gasoline in the USA'. Many EU countries have recently adopted regulatory measures against leakage from underground storage tanks to ensure that the risk of groundwater contam ination `rem ains low in the future'. The report points at the risks posed by existing underground storage tanks not complying with the new regulations in the interim period until full enforcement. The regulations are anticipated to have taken full effect by 2005. The report em phasises the need for `strong enforcem ent o f the underground storage tank specifications' and suggest `penalties to prove an effective deterrent'. Accordingly, it seem s that the position in the EU tends to be that oxygenates such as MTBE should have their place to improve combustion and help reduce emissions of CO and organic compounds. The increased risk associated with leaking underground storage tanks is seen as a technical problem that can be m anaged by risk reduction. However, studies in the United States have shown that even sophisticated, double-walled underground storage tanks with detectors may leak undetected due to im proper installation (Couch and Young, 1998). This explains the push for `strong enforcem ent' and follow-up m onitoring o f groundwater in the report to the European Commission (Arthur D. Little Limited, 2001). 11.8. Discussion in relation to the precautionary principle O ur analysis will first take a historical perspective in order to highlight the questions that could have been asked at the time o f the introduction o f M TBE in the early 1980s. The objective here is to illustrate the potential o f applying foresight as part o f the precautionary principle. As early as 1954, studies were published indicating the very low biodegradability o f the ether family (Mills and Stack, 1954). By 1960, the resistance o f ethers to biological degradation had become textbook knowledge (Sawyer, 1960). O n this basis we consider it reasonable to believe that at the time o f the introduction o f MTBE, competent chemists and microbiologists could have anticipated that MTBE might be persistent in groundwater. Docum entation for this argum ent has however not been found. It appears that the first warnings against persistency o f M TBE in groundwater, based on the first experim ental facts (although only at the laboratory scale), were publish ed in 1990 (Barker et al., 1990; Je n se n and Arvin, 1990). However, the regulatory agencies did not react to these first warnings. It was not until the full-scale results em erged in 1996 that the regulatory system was alerted to the seriousness o f the problem . That delay is docum ented in the United States and Denmark. On the other hand, it is reasonable to believe that there were no indications o f any obvious harm ful effects associated with M TBE at the 116 Late lessons from early warnings: the precautionary principle 1896-2000 time it was chosen to replace lead in petrol. Accordingly, the fundam ental issue is not whether a persistent and harmful chemical poses an unacceptable risk. The fundam ental issue is whether or not the prevailing knowledge was adequate at the time to foresee that persistency in itself might eventually be considered problematic. 11.8.1. The historical perspective W idespread concern over the persistency o f chemicals first arose following the publication o f the lan dm ark book Silent spring (Carson, 1962). In her book, Carson described the disastrous effects o f large-scale applications o f pesticides on insects, birds and ecosystems. At the time, the debate was closely connected to what was known as the `dirty dozen' o f pesticides, now referred to as persistent organic pollutants, or POPs. In connection to this debate, research on biodegradability and mobility, as well as other characteristics of pesticides, has been ongoing since the early 1960s. A similar debate em erged in Europe in the early 1960s, following incidences o f foam ing in rivers caused by the use o f non biodegradable detergents. The soap and detergent industry becam e actively involved in efforts to develop criteria and m ethodology for testing the persistency/ degradability o f detergents (ASDA, 1965; Swisher, 1987) as did the O rganisation for Econom ic Co-operation and Development (OECD, 1970). It would therefore appear that industry and regulators would have had reason to suspect that persistency could entail complications when presen t in com bination with detrim ental properties. T he experien ce with pesticides and detergents provided scientific evidence that in com bination with certain other chemical properties, persistency could cause lasting undesirable effects, such as persistent toxicity or foam ing rivers. Based on this, we may further ask if at the time o f the introduction o f MTBE there were any indications that society was formulating and im plem enting responses to address the issue o f chem ical persistency com bined with other negative characteristics. In 1978/79, an OECD working group on test methods for degradation and accumulation was form ed to pioneer the developm ent of the specific detergent methods into more general m ethods for m easuring persistency (OECD, 1981). At these workshops, representatives from governments, industry and academ ia discussed the technical details o f the already accepted concept of persistency. As well as the OECD, a num ber of governmental, industrial and non governm ental organisations were active in developing standard test m ethods for biodegradability during the 1970s and the early 1980s (U S EPA, 1979; EC ETO C , 1982; ISO, 1984). Persistency appeared early in the development of EU environmental legislation as an undesirable property of a chemical (European Commission, 1967 and 1984). According to the sixth am endm ent (European Commission, 1979) to the EC classification and labelling directive for new substances (European Commission, 1967) adopted in Septem ber 1979, all new substances intended to be m anufactured or m arketed in the EU m ust be notified to the com petent authority in one o f the M em ber States. This pre-marketing notification must be accom panied by a set o f data on the identity, use an d properties o f the substance. T he size o f the data set is d epen d en t on the foreseen production-volume of the substance involved. T ogether with toxicity, bioaccumulative potential and suspected carcinogenicity, persistency could trigger a classification as `particularly dangerous to the environm en t' . However, M TBE was not legally bound to these `new substance' regulations, which only applied to chemical substances that were not m arketed within the EU in the 10 years prior to 18 Septem ber 1981. Because M TBE had been m arketed prior to this date, it was classified as an `existing substance' in the EU. Regulations for the evaluation and control o f existing substances were adopted in 1993 (European Commission, 1993). The Finnish Environm ental Institute is currently conducting the risk assessment of MTBE required under this regulation. Draft versions o f the risk assessment, already referred to, becam e available by the end o f 2000. The final version is expected during 2001. The large am ount o f effort dedicated to developing test m ethods for persistency indicates that at the time society did in fact deem it im portant to identify persistency in a chemical. The subsequent incorporation into law o f persistency as bein g an undesirable characteristic o f a chemical further indicates that society was indeed prepared to MTBE in petrol as a substitute for lead 117 im plem ent responses to address the impact of chem ical persistency in com bination with other negative characteristics. At the time it could have been expected that the projected high production-volume of M TBE would have given rise to concern. These concerns could have prom pted further investigations, although there was no requirem ent at the time. Such investigations would have revealed the persistency of MTBE to decision-makers. H ad industry and regulators taken account of the persistency of M TBE, it is reason able to assum e that they could have considered the following question: does MTBE exhibit any other characteristic which, when combined with persistency, requires a response by society due to the risks involved? As has been m entioned previously, the taste and odour problem s associated with MTBE have been known for a long time. These were, however, overlooked during the screening o f MTBE. It seem s that the possibility that M TBE could pose a threat to groundwater reservoirs was never considered. Furtherm ore, the carcinogenicity of MTBE did not start to be investigated in detail until the 1990s and the classification o f M TBE with regard to carcinogenicity rem ained uncertain long after the introduction o f the chemical as a high production-volume chemical. There is no doubt that the introduction of a high production chemical displaying a combination o f persistency, taste and odour characteristics, potential carcinogenicity and other adverse effects could have sparked concern in the early 1980s. Studies should therefore have been done at that time. The choice o f MTBE as a long-term replacem ent to lead in petrol was not grounded in precaution, as the possible threats associated with the choice were not adequately characterised. By applying foresight, it could have been possible for the petrochem icals industry and regulative authorities to identify that the persistency of M TBE could eventually becom e problem atic. 11.8.2. The modern-day perspective O f all the negative features of MTBE, persistency is unique in that it implies irreversibility. A persistent chemical remains presen t in the environm ent lon g after its use has been discontinued, thus dramatically increasing the risk associated with a false diagnosis of known detrimental effects, and with other detrim ental effects o f which science is still unaware. Indeed, the release of persistent chemicals into the environment begs the following questions: should a response be im plem ented on the sole basis of persistency, and should persistency motivate systematic, comprehensive and thorough investigations o f perceivable negative effects? These questions are still pending today, but the high production volume and the high solubility o f M TBE should definitely give rise to concern. The list o f known negative features o f chemicals, other than persistency, is still expanding. O nly now, for exam ple, are we beginning to investigate the m echanism s that induce endocrine disruption. Should new information becom e available associating any new negative effect to a persistent chemical, the financial and technological resources required for rem ediation are likely to be out o f reach once this chemical is spread throughout the environment. In this case the interpretation o f the precautionary principle is that the risk of being wrong and the potential consequences should always be considered. In the case o f persistent chemicals, that risk is significantly greater than with other chemicals due to irreversibility. Lack o f full scientific understanding o f cause-effect relationships cannot be used as an excuse to postpone action. Moreover, scientific uncertainty and ignorance warrant precautionary measures. 11.8.3. Alternatives The precautionary approach invokes the principle of substitution, seeking out safer alternatives to potentially harm ful activities. There is no doubt that M TBE is a reasonably better option than lead. However, because we have reason to believe that there could be irreversible detrim ental effects associated with the use o f M TBE, we also have to investigate substitution as an alternative to risk reduction. There are alternatives to MTBE and it is likely that m ore could be discovered if sufficient research were conducted in this direction. Thus, the im portant follow-up question in the implementation o f the precautionary approach is: what is the status o f the alternatives to M TBE? The m ost recent accoun t is presen ted in the E U report on risk reduction drafted in connection with the risk assessm ent (Finnish EPA, 2001) 118 Late lessons from early warnings: the precautionary principle 1896-2000 The most commonly considered alternatives to MTBE include other oxygenates, improved engine technology and altered petrol com position (M orgenroth and Arvin, 1999; CEC, 1998). Other oxygen-containing compounds include ethanol, ETBE (ethyl tert-butyl ether), TAME (tert-amyl methyl ether), and DIPE (di-isopropyl ether). However, other oxygenates could also have adverse impacts on the environment. For example, the area o f agricultural land required to grow the plant matter for the production of biological ethanol would be significant and the ecological im pact of the fertilisers and pesticides required to produce this crop need to be considered. Moreover, ethanol production requires considerable amounts of energy and this option may not lower greenhouse gas emissions significantly when the entire p roduction cycle is considered. Other ethers could have health effects similar to those o f M TBE which, as we have noted, are not fully understood. A nother alternative is the introduction o f vehicles with im proved engine technology, such as gasoline direct injection (GDI). At part load, GDI does not require as high an octane rating as conventional engines. At full load however the differences are not as im portant. With the reduction o f the arom atics in petrol being called for in Europe, the addition o f oxygenates would rem ain necessary even with im proved engine technology. Another widely considered alternative could be to improve the composition o f the petrol. An increase in the octane rating o f the petrol could be achieved by changing the structure o f the `branch alkanes' in petrol. This option may however be opposed by the petrochem icals industry, as it would imply process changes at the refinery, entailing m ajor investment costs. The use o f M TBE as a petrol additive to reduce emissions o f pollutants from motor vehicles is indicative o f a deeper issue of which MTBE is only the symptom. As countries like India and China pursue the process o f industrialisation, it is clear that future generations cannot afford to operate a m ass transit system resulting in the emission o f VOCs, CO, carbon dioxide, NOx and other pollutants. Although the above m entioned alternatives could improve the current situation, it is n ot likely that the benefits they provide will be significant enough to m eet the needs o f future generations. Thus, the precautionary principle would now beckon us to invest seriously in the developm ent of alternatives entirely different from the ones considered above. In this respect, som e European governments have been advocating a response directed at the very driving forces that lead to environmental degradation. In essence, they are questioning the basic need for a mass transit system. For exam ple, the position of the current governm ent in D enm ark is that with advances in com m unication technologies, personal vehicles are increasingly becom in g a `w ant' versus a `n eed '. They have therefore enacted various economic policies that aim to discourage citizens from driving personal vehicles. A nother poin t o f view is that society will always have a n eed for a m eans o f rapid transportation, but that a radical shift in engine technology can eliminate the pressures exerted on the environment by personal transport. One example of such a technological intervention is the hydrogen fuel cell. The emission by-products o f this system are limited to oxygen and water. Pilot projects using hydrogen-based fuel cell buses have been set up, for example, in Chicago and Vancouver. A feature o f all these options is that none o f them has been investigated to the point where they could be im plem ented on a wide scale in the short term. As we are currently realising the problem s associated with the use o f M TBE, foresight would beckon us to actively research and develop alternatives to this option. 11.8.4. Cost-benefit analysis A cost-benefit analysis should ideally incorporate all costs and all benefits -- or a reasonable approxim ation thereof. In the face o f ignorance, a cost-benefit analysis is of only limited value to decision-makers and cannot, o f itself, form the sole basis for a policy decision. Nevertheless, a systematic and comprehensive assessm ent o f all the options can serve as an im portant tool for their relative assessment, even under situations of partial ignorance. A cost-benefit analysis on the use o f M TBE in California was conducted in 1998 (Keller et al., 1998a). T he analysis exam ined the hum an health benefits derived from MTBE in petrol as a substitute for lead 119 controlling air pollution, and then analysed the costs associated with the use o f M TBE across the following categories: hum an health costs due to an increase in cancer cases caused by air pollution from M TBE and its com bustion by-products (aldehydes); costs o f averting hum an health damages through water treatm ent or by using alternative water supplies; direct costs paid by the consum ers in the form o f increased petrol prices as well as decreased petrol efficiency due to the oxygenates; m onitoring costs; recreational costs; ecosystem damage. The analysis com pared the costs and benefits o f using MTBE, o f using ethanol, and o f the use o f non-oxygenated petrol with either toluene or iso-octane. Conventional petrol sold prior to the introduction o f RFG was used as a baseline comparison. The air quality benefits associated with the three options were essentially the sam e, as all formulations achieve approxim ately the same reduction in CO and ozone precursors emissions. The air quality benefits o f the three options were estimated to be relatively minor, between USD 14 million and USD 78 m illion p er year. T he water treatm ent costs associated with M TBE m ade it the most expensive option, when all costs were considered. Based on an estimate o f the num ber o f groundwater supplies and surface water reservoirs currently contam inated with M TBE, the estim ated aggregate cost o f water treatment in California was foun d to be between USD 340 m illion an d USD 1 480 m illion p er year. These costs were based on the prem ise that contam inated water would be treated to a concentration below 5 g /l using granular activated carbon. As described earlier, there may be cheaper methods. For nonoxygenated petrol and ethanol blends, the differential cost o f rem ediation a n d /o r water treatm ent relative to conventional petrol would be small. Other im portant costs associated with the use o f M TBE included the direct price paid by consum ers for oxygenated petrol (USD 435 million to USD 1 055 million per year), and the potential loss o f recreational boating activities in order to prevent this source o f M TBE contamination o f surface water reservoirs (USD 160 million to USD 200 million per year). The report concluded that non-oxygenated petrol achieves air quality objectives at the least cost, followed by ethanol-based petrol fo rm u lation s. Given the ignorance surrounding our understanding o f the health effects o f MTBE, a cost-benefit analysis will always be provisional. It is possible that many potential costs still rem ain unknown to us. For example, if new information became available establishing a clear link between the use o f MTBE and an increase in asthma occurrences, the health costs associated with the option of MTBE would increase significantly. On the other hand, if new information showed that MTBE could be degraded in m unicipal waterworks, the costs associated with M TBE could decrease significantly. But on the basis o f what is currently known with reason able certainty, water treatm ent costs make MTBE a very expensive option. Furtherm ore, if we assum e a worst-case scenario for the health effects o f using MTBE, the cost o f correction o f damage would increase even further, far outweighing the cost o f prevention. It is widely agreed that in the pursuit o f the precautionary principle, the cost o f any precautionary m easures taken should be proportionate to the benefits that can be achieved by them. The potentially very high costs associated with M TBE would therefore justify a proportionately costly response. It rem ains unclear who should be held responsible for the costs incurred as a result o f the use o f MTBE. Our analysis suggests that both industry and society (represented by regulators) were short-sighted in their decision to replace lead in petrol with M TBE. Because o f this, it is difficult to single out who will bear the cost for the response. Who is the polluter, who should pay? Until now, the m ajority o f the costs associated with the use o f MTBE have been borne by society in the form of risk reduction, water treatm ent or alternative water supply costs, costs to industry carried over to consum ers and human health costs. The consequences o f any decision relating to our mass transit system will be far reaching. Because o f this, foresight would now beckon society to select those alternatives to M TBE that are based on best environmental practices, even if these alternatives are m ore costly in the short term. The alternatives selected must be the best for 120 Late lessons from early warnings: the precautionary principle 1896-2000 the needs o f the present day and for the needs o f the future. 11.9. Conclusion By applying foresight at the time of the introduction o f MTBE, it could have been possible to predict that the persistency of MTBE would eventually be deem ed problem atic by society. The persistency, mobility and high production volume of M TBE should have prom pted systematic, comprehensive investigations into perceivable adverse characteristics o f the substance. This could have identified taste and odour problems and suspicions of carcinogenicity o f MTBE at an earlier stage, and could have acquitted the chemical or confirmed suspicions. The risk to groundwater due to persistency and the strong odour and taste o f M TBE was ignored for a long time but has now been recognised. T he classification o f M TBE with regards to its carcinogenicity rem ains uncertain, long after its introduction as a high production-volume chemical. A state of near ignorance characterises our knowledge of the possible endocrine disrupting and asthm a inducing effects of MTBE. Additional research is still recom m ended in these areas. Because o f its persistency in groundwater, the use o f MTBE poses an ongoing, everlasting risk o f having irreversible adverse effects. These adverse effects could come in the form o f known effects such as taste and odour in contam inated drinking water, as well as in the form o f presently unknown adverse effects. T he key question o f principle is: does persistency alone, without indication of other adverse effects, give reason to apply the precautionary principle? It seems reasonable to conclude that systematic, comprehensive and thorough investigations into all known possible adverse effects should be exercised before any release o f large volumes of a persistent chemical into the environment. These investigations should be reopened and pursued as new categories of adverse effects are discovered. It also seems reasonable to answer that, because persistent chemicals expose us and future generations to a variety o f possible negative effects o f which we are still ignorant, alternatives to the use o f persistent chemicals should be sought out whenever possible. There are now a considerable num ber o f examples of persistent chemicals (including chlorofluorocarbons, polychlorinated biphenyls and tributyltin) that caused unwelcom e `surprises' with serious consequences. Foresight beckons us to investigate alternative decisions thoroughly and to reconsider these alternatives when we realise the limitations o f the current decisions. Best environmental practices, like risk reduction, should be pursued. Due to the possibility o f risk reduction being insufficient, research and development into alternatives should be actively encouraged in order to render them feasible as soon as possible. Such thorough investigations ensure that decisions are robust and flexible so that they may be adapted to unforeseen circumstances in the future. Thus, foresight, em bodied in the precautionary principle, is the guarantor o f futurity within the concept o f sustainability. MTBE in petrol as a substitute for lead 121 MTBE: early warnings and actions 1954 1960 1990 1990 1990 1995 1996 1997 1998 1999 2000 2000 2000 2000 2001 2001 First scientific paper dem onstrates low biodegradability of the ether fam ily in w ater Inform ation about taste and odour in w ater and low biod egradability available in textb ook First indications of the potential for groundwater pollution with M TBE at laboratory scale Significant increase in use of M TBE in the United States due to am endm ents to the C lean A ir A ct C om prehensive investigations of carcinogenicity initiated in the 1990s M TBE detected in w ells that supplied drinking w ater in Santa M onica, California. Th ese had to be closed and the city lost 71 % of its local w ater supply Concerns become widespread following a US Geological Survey report Field studies dem onstrate that M TBE is highly soluble, mobile and persistent and therefore a potential risk to groundwater A Danish E P A report acknow ledges that the governm ent w as informed in 1990 that M TBE might give rise to groundw ater problem s and presents an action plan for M TBE rem ediation and risk reduction California recomm ends removal of M TBE from petrol as soon as possible, but not later than end 2002 Indications that M TBE m ight be linked to asthma in som e US cities Indications that M TBE m ight be an endocrine disrupter US EP A announces that steps will be taken to significantly reduce or eliminate M TBE as a petrol additive Danish E P A places M TBE on its list of undesirable substances EU reports on analysis of risks and risk reduction related to M TBE. European Chem icals Bureau decides that M TBE should not be classified as a carcinogen The debate goes on 11.10. References Arthur D. Little Lim ited, 2001. `M TBE and the requirements for underground storage tank construction and operation in M ember States, Report to the European Commission, March 2001', Ref. E N V .D .1/E T U /2000/ 0089R. Arvin, E., 2001. Pers. comm., D epartm ent of Environment and Resources, Technical University o f Denmark. ASDA (American Soap and Detergent Association, Biodegradation Subcommittee), 1965. `Procedure for determ ination o f biodegradability o f A BS an d L A S',J . Am. Oil Chemists Soc. Vol. 42, pp. 986-993. Barker, J. F., H ubbard, E. an d Lem on, L. A., 1990. `The influence o f m ethanol and MTBE on the fate o f persistence o f m onoarom atic hydrocarbons in groundw ater', Proceedings of Petroleum Hydrocarbons and Organic Chemicals in Groundwater Prevention, Detection, and Restoration, 31 October-2 November 1990, pp. 113-127, American Petroleum Institute and Association of Groundwater Scientists and Engineers, Houston, Texas. Belpoggi, F., Soffritti, M. and M altoni, C., 1995. `Methyl-tertiary butyl ether (M TBE) -- a gasoline additive -- causes testicular and lym phohaematopoietic cancers in rats Toxicol. Ind. Health Vol. 11, pp. 119-49. Belpoggi, F., Soffritti, M. and M altoni C., 1998. `Pathological characterization o f testicular tumours and lymphomasleukemias, and of their precursors observed in Sprague-Dawley rats exposed to methyltertiary butyl ether (M T B E )', Europ.J. Oncol. Vol. 3, pp. 201-206. Bird, M. G., Burleigh-Flayer, H. D., Chun, J. S., Kneiss, J. J. and Andrews, L. S., 1997. `Oncogenicity studies of inhaled methyl tertiary butyl ether (M TBE) in CD-1 m ice and F344 rats'.J. Appl. Tox. Vol. 17, S45-S55. BSC, 1998. Report on carcinogenesis, B o ard o f Scientific Counsellors, National Toxicology Program. Borden, R. C., Daniel, R. A., LeBrun, L. E. and Davis, C. W., 1997. `Intrinsic biodegradation of MTBE and BTEX in a gasoline-contam inated aq u ife r', Water Resources Research Vol. 33, N o 5, pp. 1105-- 1116. Blue R ibbon Panel, 1999. Executive summary and recommendations, Blue R ibbon Panel on Oxygenates in G asoline, U S EPA. Table 11.1. 122 Late lessons from early warnings: the precautionary principle 1896-2000 Burleigh-Flayer, H. D., Chun, J. S. and Kintigh, W. J., 1992. `Methyl tertiary butyl ether: Vapour inhalation oncogenicity study in CD-1 m ice', Bushy Run R esearch Center, Export, PA, BRRC report 91N0013A, O ctober 15, Union Carbide Chemicals and Plastics Co., Inc., submitted to the U S EPA under TSCA Section 4 Testing Consent Order 40 CFR 799.5000 with cover letter dated 29 October 1992, EPA/OPTR#42098. CAL-EPA, 1999. Public health goal for methyl tertiary butyl ether (MTBE) in drinking water, Pesticide and Environmental Toxicology Section, Office o f Environmental Health Hazard Assessment, California Environmental Protection Agency, Sacram ento, CA. Carson, R., 1962. Silent spring, H oughton Mifflin, Boston, MA. CEC, 1998. `Evaluating the cost and supply of alternatives to M TBE in C alifornia's reformulated gasoline', Draft report, February 1998, California Energy C om m ission. Chun, J. S., Burleigh-Flayer, H. D. and Kintigh, W. J., 1992. `Methyl tertiary butyl ether: Vapor inhalation oncogenicity study in Fischer 344 rats', Bushy Run Research Center, Export, PA, BRRC report 91N 0013B, 13 November, U n ion Carbide Chem icals and Plastics Co., Inc., submitted to the U S EPA under TSCA Section 4 Testing Consent O rder 40 CFR799.5000 with cover letter dated November 1992, EPA/OPTS#42098. Couch, A. an d Young, T., 1998. `Failure rate o f u n d ergroun d storage tanks', in Health and environmental assessment ofM TBE Vol. 3, UC Toxics Research and Teaching Program, University o f California, Davis. Crain, E. F., Weiss, K. B., Biijur, P. E., 1994. `An estim ate o f the prevalence o f asthm a an d wheezing am ong inner-city children', Pediatrics Vol. 94, pp. 356-362. Davis, Gray, 1999. G overnor o f the State o f California, Executive O rder D-5-99, 25 M arch 1999, at http://www.calepa.ca.gov/ program s/M TBE/eotask.htm Day, K. J., De Peyster, A., Allgaier, B. S., Luong, A. and M acGregor, J . A., 1998. `Methyl t-butyl ether (M T BE): Effects on the male rat reproductive endocrine axis', Abstract #861, Society of Toxicology: Abstracts of the 37th annual meetingVol. 42, p. 174. Dekant, W., 2000. `T he m echanistic toxicology o f MTBE does not support a classification as Category 3 carcinogen', E xpert review, Dixson-Decleve, S., 2000. Pers. comm. DeWitt & Company Inc., 2000. MTBE and oxygenates 2000 annual, H ouston, Texas. Dixson-Decleve, S., 2001. Pers. comm., Environmental Consultant, Brussels office, R ef to Meeting of the Commission Working Group on the Classification and Labelling of Dangerous Substances, European Chemicals Bureau, Ispra, 1-17 November 2000, EC B I/ 7 6 /0 0 08. ECETOC, 1982. `Biodegradation m ethodology with em phasis on readybiodegradibility test', Report o f sub-group, European Chemical Industry Ecology and Toxicology Centre, Brussels. EN D S Environm ent Daily, 2000. `EU backs MTBE despite US prohibition', Wednesday 22 March, at http://w w w .ends.co.uk/ envdaily. EN D S Environm ent Daily, 2001, `EU `should take action' on MTBE risks', Wednesday 24 January, at http://w w w .ends.co.uk/envdaily. European Commission, 1967. Council Directive 6 7 /5 4 8 /E E C o f 16 A ugust 1967 on the approxim ation o f the laws, regulations and administrative provisions relating to the classification, packaging and labelling of dan gerou s substances, O ff.J . Eur. Communities L196. European Commission, 1979. Council Directive 7 9 /8 3 1 /E E C o f 18 Septem ber 1979 am ending for the sixth time Directive 6 7 / 5 4 8 /E E C on the approxim ation o f the laws, regulations and administrative provisions relating to the classification, packaging and labelling o f dan gerous substances, O ff.J . Eur. Communities L259. European Commission, 1984. Directive 79 831, A nnex V, Part C: M ethods for the determination o f ecotoxicity, Test m ethods C 3,4,5,6,7,8,9, O ff.J . Eur. Communities L251. European Commission, 1993. Commission Directive 9 3 /6 7 /E E C o f 20 July 1993, laying down the principles for the assessment of MTBE in petrol as a substitute for lead 123 risks to m an and the environment o f substances notified in accordan ce with Council Directive 6 7 /5 4 8 /E E C , O ff.J . Eur. Communities L277. M TBE and gasoline com pounds in w ater', in Arendt, F., Hinsenveld, M. an d Van den Brink, W. J . (eds), Contaminated soil, pp. 445 448, Kluwer, Dordrecht. Eweis, J . B., Ergas, S. J., Chang, D. P. and Schroeder, E. D., 1998a. Bioremediation principles, McGraw-Hill. Eweis, J. B., W atanabe, N., Schroeder, E. D., Chang, D. P. Y an d Scow., K. M., 1998b. `MTBE biodegradation in the presence o f other gasoline com pounds', National G round Water Association Conference on M TBE and Perchlorate, Anaheim , CA, 3-4 June. Finnish EPA, 2001. `Risk assessm ent: Tertbutyl methyl ether', CAS-No1634-04-4, EINECS-No 216-653-1, draft o fJanuary 2001; `Risk reduction strategy: Tert-butyl methyl eth er', CAS-No 1634-04-4, EINECS-No 216 653-1, draft o f 5 M arch 2001. Hathaway, W., 1999. `Asthma: An epidem ic am on g H artfo rd 's ch ildren ', in Hartford Courant. H artford, CT. Hunter, B. et al., 1999. `Im pact o f sm all gasoline spills on groundw ater', presented at the Maine Water Conference, Augusta, ME, 15 April 1999, M aine D epartm en t o f Environmental Protection, Augusta, ME. Hurt, K. L., Wilson, J . T., Beck, F. P. an d Cho, J . S., 1999. `A naerobic biodegradation o f M TBE in a contam inated aquifer', In-Situ and On-Site Biorem ediation Conference, San Diego, CA, 19-22 April 1999, U S EPA, Ada, OK. IARC, 1999. International Agency for R esearch on Cancer. `M T B E ', Chem. Abstr. Serv. Reg. N o 1634-04-4(73), p. 139. IPCS, 1999. `Environm ental H ealth Criteria 206: Methyl tertiary-butyl ether', International Programme on Chemical Safety, at h ttp://w w w .w h o.in t/pcs/docs/ ehc_206. ISO , 1984. `Water quality: Evaluation in an aqueous m edium o f the `ultim ate' aerobic biodegradability o f organic compounds: M ethod by analysis o f dissolved organic carbon (D O C )', International Organization for Standardization. Jen sen , M. H. and Arvin, E., 1990. `Solubility and degradability o f the gasoline additive Jo h n so n , Pankow, 2000. Environ. Sci. and Tech. 1 May (News section). Jo se p h , P. M., 1997. `C hanges in disease rates in Philidelphia following the introduction o f oxygenated gasoline', Annual meeting paper 97-TA34.02, Ju n e 10 1997, pp. 1-15, Air and Waste M anagem ent Association, Pittsburg, PA. Joseph , P M., 1999. `New hypotheses for MTBE combustion products', Annual m eeting p ap er 99-885, 22 Ju n e 1999, pp. 1 10, Air and Waste M anagem ent Association, Pittsburg, PA. Jo se p h , P. M., 2000. `Is urban asthm a caused by methyl tertiary butyl ether (M TBE)?', Archives ofEnvironmental Health Vol. 55, N o 1, pp. 69-70. Keller, A. A., Fernandez, L. F., Hitz, S., Kun, H., Peterson, A., Smith, B. an d Yoshioka, M., 1998a. `An integral cost-benefit analysis o f gasoline formulations m eeting California Phase II Reform ulated Gasoline Requirem ents', Report to the Governor of California, UC Toxic Substances Research and Training Program. Keller, A. A., Froines, J., Koshland, C., Reuter, J., Suffet, I. an d Last, J., 1998b. `H ealth and environmental assessment of MTBE, summary and recom m endations', R eport to the Governor o f California, U C Toxic Substances Research and Training Program. Koshland, C. P., Sawyer, R. F., Lucas, D. and Franklin, P., 1998. `Evaluation o f autom otive M TBE com bustion byproducts', in Health and environmental assessment ofM TBE Vol. 2, U C Toxic Substances Research and Training Program, University o f California, Davis. Landm eyer, J. E., Chapelle, F. H., Bardley, P. M., Pankow, J . F., Church, C. D. an d Tratnyek, P. G., 1998. `Fate o f M TBE relative to benzene in a gasoline-contaminated aquifer (1 9 9 3 -9 8 )', Ground Water Monitoring Remediation (Fall), pp. 93-102. Larsen, P. B., 1997. Evaluation of health hazards by exposure to M TBE and estimation of limit values in ambient air, soil and drinking water, 124 Late lessons from early warnings: the precautionary principle 1896-2000 Institute o f Toxicology of the National Food Agency of Denmark. Leighton, J., Matte, T., Findley, S., 1999. `Asthm a prevalence am on g school children in a B ron x com m unity ', A sthm a Conference, February 9 1999, p. 26, Centers for Disease Control and Prevention, Atlanta, GA. endocrine alterations in mice are not m ediated through the estrogen receptor', Toxicol. Sci. Vol. 41, pp. 77-87. NRC (National Research Council), 1996. Toxicological and performance aspects of oxygenated motor vehiclefuels, pp 75-115, National Academy Press, Washington, DC. Leikauf, G. D., Kline, S., Albert, R. E., Baxter, C. S., Bernstein, D. I. an d Buncher, C. R., 1995. `Evaluation of a possible association of urban air toxics and asthm a', Environmental Health Perspectives Vol. 103, N o 6, pp. 253-271. M angione, S., Papastam elos, C., Elia, J, 1997. `Asthm a prevalence an d absenteeism am on g innercity school children: A survey o f two Philadelphia m iddle schools ', Am. J. Respir. Crit. Care. Med. Vol. 155. M cBride, A. D., 1996. Outcome of asthma survey of kindergertners entering public school system in 1996. Stamford Health Department, Stam ford, CT. M ehlman, M A., 2000. `M isclassification o f carcinogenic methyl tertiary butyl ether (M TBE) by the National Toxicology Program Board: Sm okescreen in, science out?', Archives ofEnvironmental Health Vol. 55, No 1, pp. 73-74. NRC (National Research Council), 1999. Ozoneforming potential of reformulated gasoline, pp 1-212, National Academy Press, Washington, DC. N STC, 1997. Interagency Assessment of Oxygenated Fuels, Office o f Science and Technology Policy, N ational Science and Technology Council, Washington, DC. OECD, 1970. Pollution by detergents: Determination of the biodegradability of anionic synthetic surface active agents, O rganisation for Econom ic Co-operation and Development, Paris. OECD, 1981. OECD chemicals testing programme: Expert group degradation/ accumulation, fin a l report, O rganisation for Econom ic Co-operation and Development, Paris; Umwelt Bundesamt, Bundesrepublik Deutschland, Berlin; Government o fJapan, Tokyo (joint publication). Milj0 og Energiministeriet, 1998. M ilj0styrelsen: H andlingsplan for M TBE, Ju n i (in Danish). Mills, E. J . an d Stack., V. T., 1954. `Proc. 8th Ind. Waste Conf., Purdue Univ., Ext. Ser. 83:492', in Bon d, R. G. an d Straub, C. P., 1974, Handbook of environmental control Volume IV: Wastewater treatment and disposal, CRC Press, Ohio. 0stergaard, G., 2001. Pers. comm. Rodriguez, R., 1997. `M TBE in groundwater an d the im pact on the city o f Santa M onica drinking water supply', in Technical Papers of the 13thAnnual Environmental Management and Technology Conference West, 4 -6 November 1997. Sawyer, C. N., 1960. Chemistry for sanitary engineers, Mcgraw Hill. M orgenroth, E. and Arvin, E., 1999. `The European perspective to MTBE as an oxygenate in fuels', International Congress on Ecosystem Health -- M anaging for Ecosystem Health, Sacram ento, CA, 15-20 August 1999. M orm ile, M. R., Liu, S. and Suflita, J . M., 1994. `A naerobic biodegradation o f gasoline oxygenates: Extrapolation o f inform ation to m ultiple sites and redox conditions', Environ. Sci. Technol. Vol. 28, pp. 1727-1732. Schirmer, M. an d Barker, J . F., 1998. `A study o f long-term M TBE attenuation in the B orden aquifer, O ntario, C an ad a', Ground Water Monitoring Remediation (sprin g), pp. 113-122. Squillace, J . P., Zogorski, J . S., Wilber, W. G. and Price, C. V., 1996. `Prelim inary assessment of the occurrence and possible sources o f M TBE in groundwater in the U nited States, 1993-1994', Env. Sci. Tech. Vol. 30, p p .1721-1730. Moser, G. J., Wolf, D. C., Sar, M., Gaido, K. W., Jan szen , D. an d Goldsworthy, T. L., 1998. `Methyl tertiary butyl ether induced Suflita, J . M. and M ormile, M. R., 1993. `A naerobic biodegradation o f known and potential gasoline oxygenates in the MTBE in petrol as a substitute for lead 125 terrestrial su b su rface', Environ. Sci. Technol. Vol. 27, pp. 967-978. Swisher, R. D., 1987. Surfactant biodegradation, M arcel Dekker Inc, New York. U S EPA, 1979. `Investigations o f Biodegradability and Toxicity o f Organic C om pounds', EPA-600/2-79-163, US Environmental Protection Agency, Cincinnati. U S EPA, 1997. `Drinking water advisory: Consum er acceptability advice and health effects analysis on methyl tertiary butyl ether (M TBE) ', EPA822-F-97-009;ODW 4304, Health and Ecological Criteria Division, Office of Science and Technology, Office of Water, U S Environm ental Protection Agency, Washington, DC. Williams, T. M., Cattley, R. C. an d B orghoff, S. J., 2000. `A lterations in endocrine responses in male Sprague-Dawley rats following oral administration o f methyl tert butyl ether', Toxicol. Sci. Vol. 54, pp. 168-176. Yeh, C. K. an d Novak, J . T., 1994. `A naerobic biodegradation o f gasoline oxygenates in soils', WaterEnvironment Research Vol. 66, N o 5, pp. 744-752. 126 Late lessons from early warnings: the precautionary principle 1896-2000 12. The precautionary principle and early warnings of chemical contamination of the Great Lakes Michael Gilbertson 12.1. The first significant early warnings The growth o f chemical m anufacturing and use o f organochlorine com pounds in the 20th century has resulted in global contam ination with a wide variety o f toxic and persistent organochlorine residues. Probably the most significant `early warning' was the publication o f Silent spring by Rachel Carson (1962). This book com piled the existing evidence o f previous early warnings o f the effects o f organochlorine pesticides on fish and wildlife and warned particularly of the threat posed by these chemicals in relation to cancer in humans. One area o f North America where the scientific and policy implications of organochlorine contamination have been intensively studied, particularly since the publication o f Silent spring, is the G reat Lakes basin. The United States and Canada share a boundary in four o f the five Great Lakes. Early warnings o f the contam ination o f this large ecosystem came from a variety of sources, not only as chemical analytical observations o f the presence o f residues of these organochlorine compounds, but also as observations o f effects on populations o f wild organisms, and particularly predatory birds. The first analytical results of the presence of organochlorine com pounds in organisms in the Great Lakes were published by Dr Joseph Hickey and co-workers and concerned the bioaccumulation of DDT (dichlorodiphenyl trichloroethane) and metabolites, and dieldrin in a Lake M ichigan food chain (Hickey et al., 1966). A dditional organochlorine pesticides including lindane, heptachlor, aldrin and endrin were identified in sam ples o f Lake Erie water (Pfister et al., 1969). After the discovery o f polychlorinated biphenyls (PCBs) in a white-tailed sea eagle in Sweden (Jensen, 1966), analytical methods were developed to detect the presence o f PCBs (Reynolds, 1969) in Great Lakes samples. Subsequent application of these m ethods led to the finding of hexachlorobenzene in samples of eggs of comm on terns from Hamilton Harbour, Ontario (Gilbertson and Reynolds, 1972). The flame retardant and pesticide m irex was subsequently identified in fish sam pled from Lake Ontario (Kaiser, 1974). The first reports of the effects of organochlorine compounds on Great Lakes populations of birds were those of Dr Joseph H ickey's students and co-workers (Keith, 1966; Ludwig and Tomoff, 1966) who docum ented the effects o f DDT and metabolites, and dieldrin on reproduction and mortality in Lake M ichigan herring gulls. Previous observations o f the decline in the population and reproductive failure of Florida bald eagles had been published (Broley, 1952; Broley, 1958) but research on Great Lakes bald eagles did not start until the mid-1960s, by which time m ost o f the popu lation had been destroyed (Spru n t et al., 1973). The first observation o f changes in eggshell quality was reported by naturalists visiting Pigeon Island in Lake Ontario in 1963 (Edw ards et al., 1963) when two softshelled eggs were found in the nest of a pair o f double-crested cormorants. The first published observation o f a deform ed chick of a Great Lakes fish-eating bird was o f a juvenile bald eagle (Grier, 1968). Systematic studies o f the incidence o f deform ities in the chicks o f various species o f colonial fish eating bird were undertaken in the early 1970s (G ilbertson et al., 1976). In the mid1960s, there was an outbreak o f adult and kit mortality in ranch m ink that fed on fish from the Great Lakes that had serious effects on this econom ic activity (H artsough, 1965). Concerns about the possible effects of organochlorine compounds on human health were first addressed in 1974 in a cohort of Great Lakes fishermen (Humphrey, 1983). Consum ption o f contam inated fish resulted in elevated levels o f PCBs in humans but was not associated with any recognised acute effects in fish eaters. These concerns were heightened in the spring o f 1978, when Lois Gibbs, a resident living in a housing developm ent built next to the Love Canal, N iagara Falls, New York started to investigate The precautionary principle and early warnings of chemical contamination of the Great Lakes 127 the incidence of diseases in her community and the possible relationship to the 20 000 tonnes of toxic wastes that had been disposed o f in the canal by the H ooker Chemical Com pany during the previous 20 years. These inquiries set off a series of psycho social dynamics within families, am ong workers, and between communities and institutions such as the local university, hospitals, churches and the medical profession (Levine, 1982). Out of desperation, Lois Gibbs organised her neighbours into the Love Canal Homeowners Association and struggled for m ore than two years for relocation. O pposin g the g ro u p 's efforts, though, were the chemical manufacturer, Occidental Petroleum which had bought the H ooker Chemical Company, and local, state and federal government officials who insisted that the leaking toxic chemicals, including dioxin, the m ost toxic chemical known, were not the cause o f high rates of birth defects, miscarriages, cancers and other health problem s. Finally, in O ctober 1980, President Jim m y Carter delivered an Emergency Declaration which moved 900 families from this hazardous area and signified the victory o f this grassroots movement. In 1981, Lois Gibbs responded to the growing public perception o f the pervasive nature o f the contamination and created the Center for Health, Environment and Justice (formerly Citizens Clearinghouse for Hazardous Wastes), an organisation that has assisted m ore than 8 000 grassroots groups with organisational, technical and general information nationwide. reference group (Jacobson and Jacobson, 1996). The strongest effect was on mem ory and attention, and the m ost highly exposed children were three times m ore likely to have low average IQ scores and twice as likely to be at least two years behind in reading co m p reh en sion . 12.2. Date and nature of subsequent action or inaction The information about the presence and effects o f organochlorine chemicals in the Great Lakes in the 1960s had been foreshadowed not only by the publication of Silent spring in 1962, bu t also by surveys in the United Kingdom o f the risks to workers applying pesticides (1951), residues in food (1953) and risks to wildlife (1955) (cited in Cook, 1964). In 1961 the authorities decided on a voluntary ban on the use o f aldrin, dieldrin and heptachlor for seed dressing for spring-sown cereals. Further restrictions were decided in 1964 based on the observations of high mortality o f seed-eating birds and the widespread decline of the peregrine falcon (Ratcliffe, 1972). These moves were opposed by the Shell Chemical Com pany which stated (Robinson, 1967): `the correlations between the time o f usage o f aldrin/dieldrin and the declines in populations are difficult to assess as there is insufficient quantitative data which can be used to establish precise relationships for any o f the raptor species. It is impossible from these surveys to establish, in a rigorous manner, the nature of the relationship between the usage o f aldrin/dieldrin and the breeding status of the peregrine falcon.' In 1980, based on the concerns raised in the preliminary studies o f elevated levels o f PCBs in fish eaters as well as the concerns at the Love Canal, a cohort o f infants was established to investigate the developmental effects o f maternal consumption o f Lake M ichigan fish contam inated with persistent toxic substances. At birth, infants exposed to the highest levels o f PCBs weighed less and h ad sm aller h ead circum ferences (Fein et al., 1984) and exhibited one or more behavioural deficits (Jacobson et al., 1984). Subsequent assessment and testing indicated that the growth retardation was irreversible, and that there were effects on short-term m em ory at seven m onths and at four years, an d effects on attention (Jacobson et al., 1990; Jacobson and Jacobson, 1993). Testing at 11 years o f age indicated that the children m ost highly exp osed in utero had IQ scores that were m ore than six points below the This scepticism was m irrored on the other side o f the Atlantic. Linda Lear (1997), in her biography o f Rachel Carson, has detailed the intense controversy surrounding the use o f organochlorine pesticides in the 1950s and 1960s and the partisan review o f the US National Academy of Sciences National Research Council in its highly controversial report entitled `Pest control and wildlife relationships'. The serialised publication of parts o f Silent spring in the New Yorker in 1962 resulted in a letter from general counsel of the Velsicol Chem ical Com pany to H oughton Mifflin, publisher o f Silent spring, threatening legal action unless the last instalment in the New Yorkerwas cancelled (Lear, 1997). After publication o f Silent spring, the N ational Agricultural Chemicals Association (NACA) produced and distributed a critical information booklet entitled `Fact and fancy' and sent it with letters to editors o f m agazines 128 Late lessons from early warnings: the precautionary principle 1896-2000 and newspapers indicating that future advertising revenues might be affected if Silent springwere to receive favourable reviews (Lear, 1997). In November 1963 there was a massive fish kill in the lower Mississippi River that was eventually traced to the organochlorine pesticide endrin and its chemical m anufacturing by Velsicol Chemical Company at Memphis. The political response was to stimulate the production o f a draft Clean Water Bill and a precautionary ruling by Stewart U dall as Secretary o f the D epartm ent o f the Interior prohibiting the use o f pesticides where there was a reasonable doubt about safety (Lear, 1997). But even four years later, the Shell Chemical Company (1967) published a report that stated: `The implication o f endrin in the 1963 M ississippi River fish kill has not been verified by recent studies'. The authorities in the United States were reluctant to take action and the US Departm ent o f Agriculture responsible for registration o f pesticides tended to align itself with the pesticide m anufacturers and the farmers. A suit brought by Victor Yannacone in 1966 on beh alf o f the Brookhaven Town Natural Resources Committee to stop the use o f DDT by the Suffolk County (Long Island, New York) M osquito Control Commission, led to the form ation o f the Environmental D efense Fund in 1967. This group o f lawyers, scientists and economists eventually brought a suit against the US Departm ent of Agriculture, first in Wisconsin and subsequently in Washington, DC, to cancel the registration of DDT. Responsibility for pesticide registration was transferred to the newly form ed U S Environm ental Protection Agency, and the Environmental Defense Fund suit eventually led to judicial administrative decisions to suspend the registrations o f DDT, in 1972, and dieldrin, in 1973. In Canada, Professor Donald Chant (1969), on beh alf o f the newly form ed Pollution Probe at the University o f Toronto, successfully petitioned the Minister o f Health and Welfare and the Minister o f Agriculture to ban DDT and related pesticides. These national decisions had an immediate effect on the concentrations of organochlorine pesticides in the Great Lakes environment that was reflected in gradual im provem ents in the status and reproduction o f bald eagles (Grier, 1982). Other national and international decisions were being m ade that affected the concentrations o f organochlorine com pounds in the Great Lakes. The 1972 Federal Water Pollution Control Act laid the groundwork for regulating pollutant discharges into U S waters. In Canada, the 1970 revisions to the Fisheries Act were being used to im plem ent federal controls on water pollution. In 1970, Monsanto, as the sole m anufacturer of PCBs in North America, announced that it was restricting the sale of PCBs for open uses, such as adhesives, sealants, chlorinated rubber, special paints and fire-resistant hydraulic fluids (Monsanto, 1970). At the regional level, the governments o f the United States and of Canada had referred the matter o f the pollution o f the lower Great Lakes to the International Jo in t Commission (IJC). While the primary concern in the 1960s was eutrophication o f the Great Lakes, particularly Lake Erie, the report o f the IJC (1969) drew attention to the presence of organochlorine com pounds in fish and wildlife. The report led to the negotiation o f the 1972 Great Lakes Water Quality Agreem ent signed by President Nixon and Prime Minister Trudeau. In the 1970s, new information about the presence and effects of persistent toxic substances in the Great Lakes was affecting the political and regulatory processes in both countries. For exam ple, the finding o f m irex in Lake Ontario fish (Kaiser, 1974) led to a New York State prohibition o f the possession o f fish caught in Lake Ontario. This precautionary approach was, however, m et with such flagrant violation o f the regulation that it was rescinded after only a few months. Similarly, when the G reat Lakes Water Quality A greem ent was renegotiated in 1978, it included a precautionary policy which stated that `the discharge o f any or all persistent toxic substances be virtually elim inated'. However the politics o f the 1980s were very different, in both countries, from those o f the 1970s when the Great Lakes Water Quality Agreem ent was renegotiated. The statements about the effects o f persistent toxic substances on Great Lakes wildlife an d hum ans were m et with scepticism by governm ent officials and demands for proof o f a causal relationship before `m assive' appropriations and expenditures o f public or private funds on rem edial works. The precautionary principle and early warnings of chemical contamination of the Great Lakes 129 12.3. Consequences of institutional responses The past 30 years have been a period in which these institutional responses have led to significant improvements in water quality in the Great Lakes and particularly in relation to decreases in the sources, loadings and concentrations of persistent toxic substances. Early in the process, indicator organisms, such as lake trout and the eggs o f herring gulls (Pekarik and Weseloh, 1998) were selected and analysed to follow the trends in concentrations. The evidence from these analyses dem onstrate that the concentrations decreased markedly between abo u t 1975 an d the early 1980s (Stow et al., 1999). The trends in concentrations have generally followed a first order decline, and while technically the decreases continue at the sam e logarithmic rates, in practice the curves have become non-zero asymptotes at concentrations that are still o f toxicological significance. For example, the present concentrations of PCBs in water are abo u t two orders o f m agnitude higher than the established water quality criterion for the protection o f hum an health based on a cancer risk assessment. The changes in the politics in both countries in the 1980s do not seem to have led to the level o f control o f releases o f organochlorine com pounds or to the clean-up o f non-point sources, such as contam inated landfill sites, sediments and atm ospheric emissions, that are required to protect hum an health. The question remains whether present levels are affecting human development and whether the effects are serious enough to warrant im plem enting the stringent requirem ents of the Great Lakes Water Quality Agreement, particularly concerning restoration of water quality. In 1990, the late Dr H elen Daly set up a team to replicate the Lake Michigan epidem iological studies by establishing a cohort centred on Oswego, New York and comprised o f infants whose mothers had eaten Lake Ontario fish. While the researchers did not find the same relationship between m aternal fish consumption prior to pregnancy and effects on body weight and head circum ference, the same behavioural effects were observed (Lonky et al., 1996). In addition, based on rat studies (Daly, 1993), it was shown that the infants o f the m others with the highest fish consum ption could not adapt to frustrating events. Subsequent chemical analytical determinations have shown the specific relationship between these behavioural anom alies and prenatal exposures to the highly chlorinated biphenyls (Stewart et al., 2000). To address the pervasive scepticism o f the 1980s and 1990s, a small group o f scientists started applying new m ethods (Fox, 1991) for integrating evidence to infer causal relationships between the observed injury to wildlife and hum an health and exposures to persistent toxic substances. This process has resulted in a series o f case studies linking the effects in a variety o f organisms, including hum ans, with specific chemical causes (reviewed in Gilbertson, 1996. There has, however, been a variety o f objections to this approach. For exam ple, O 'Brien (1994) has objected that the approach is not precautionary. The precautionary principle is, by definition, applicable to circum stances in which there is a high degree o f uncertainty. The application o f this a posteriori process is designed to reduce uncertainty using all the existing evidence to infer a causal relationship, thereby precluding the special need for applying the precautionary principle. Similarly, the application o f post-normal science (Funtowicz and Ravetz, 1993) to the im plem entation o f the Great Lakes Water Quality Agreement, based on the prem ises of system complexity and uncertainty, has recently resulted in a singular legitimacy for multi-causal ecological statements (Shear, 1996; H artig et al., 1998; D onahue, 1999). This supposed complexity and uncertainty has not been inconvenient to those interests reluctant to im plem ent the costly rem edial policies contained in the Great Lakes Water Quality Agreement. In contrast, the causal statements o f the toxicologists are based on the integration o f a diversity o f evidence drawn from traditional science based on sim ple linear systems. These yield, first, a high degree o f certainty, and thus dispense with the need to apply the precautionary principle. Second, the causal relationships could be used as a reliable and scientifically defensible basis for remedial action to restore water quality. This is particularly valuable in situations, such as the Great Lakes, where water quality has been chronically im paired by releases o f persistent toxic substances and where organisms, including humans, have already been injured over a long period of time. Rem edial actions, based on these 130 Late lessons from early warnings: the precautionary principle 1896-2000 statements, to address the extensively contam inated sediments and to secure the leaking chemical landfill sites, could be precautions against the production of another generation of infants prenatally exposed to chemicals that profoundly alter structural and functional development (C olborn an d Clem ent, 1992; C olborn et al., 1998). In this sense, there is a powerful cognitive dissonance between the scientific information produced during the past 20 years on the injury to health (Joh n son et al., 1998 and 1999) and the response of governments to im plem enting the policies contained in the Great Lakes Water Quality Agreement. 12.4. Costs and benefits What have been the costs and benefits of these actions and their distribution across time? There have been significant improvements in the past 20 years in the development and application o f methods for estimating costs and benefits in relation to chemical pollution (see, for example, Swanson and Vighi, 1998). A few cost and benefit analyses have been undertaken o f the effects o f persistent toxic substances in the Great Lakes and of their removal from discharges, contam inated landfill sites and sediments. For example, Burtraw and Krupnick (1999) have described m ethods to m easure health benefits in monetary and non-monetary terms in the context of reductions in pollutants as part of a program m e to improve Great Lakes water quality. In the developm ent o f the Great Lakes Water Quality Guidance (U S EPA, 1993) under the Clean Water Act, there was an analysis o f the costs and benefits. Incremental costs were calculated for additional construction of treatm ent facilities and process changes including pollution prevention and waste minimisation programmes; for additional monitoring programmes; and for pre treatm ent program m es. These were evaluated for 316 m ajor municipal dischargers, 272 m ajor industrial dischargers and 3 207 minor dischargers using four different scenarios. Non-point sources were not covered. The total annual costs of compliance in the United States were estimated in 1992 to be between USD 80 and USD 200 million (ca. EUR 80-200 million) depending on the scenario. T he D etroit Wastewater Treatm ent Plant is one of the largest point sources to the Great Lakes and discharges m ore than 100 kilogram s o f PCBs per year. Since 1971, nearly USD 1 billion has been spent in upgrading the sewer infrastructure and USD 120 million was recently spent in improving the operation through construction of a pum ping station. There are plans for additional improvements costing USD 1 billion, including an increase in the primary treatm ent capacity. W hether these costs will result in improvements in the PCB discharge from the plan t is uncertain. There are also a few studies o f the costs and benefits of rem edial actions for chemical landfill sites. One study was undertaken (Sudar and Muir, 1989) concerning the leaching of organochlorine pollutants from four o f the largest dum p sites located next to the N iagara River, upstream from Lake Ontario. The following three options were examined: no action; containment; and rem oval an d therm al destruction, with consideration of short and long time horizons, and from the viewpoint of who pays an d who benefits, with different discount rates. The cheapest option is containment, through capping the waste site and collecting and treating the contaminated groundwater, involving costs o f USD 100 m illion with a 10year horizon, and nearly USD 300 million with a 35-year horizon. This option essentially leaves the problem to another generation. The m ost expensive option is also containm ent with leaks that, over a 100-year horizon, w ould cost m ore than U SD 19 billion. The removal and thermal destruction option for the four sites would entail capital expenditures of about USD 50 million and annual operating costs o f abo u t USD 75 m illion for 15 years. T he no action option, over 35 years and 100 years, respectively, would cost society m ore than USD 1 billion and 16 billion. These sam e costs to society would be involved if industry spent about USD 300 million over 35 years or USD 3 billion over 100 years and was unsuccessful in stopping the plum es from reaching the face o f the Niagara Gorge. These estimates were for four dum p sites, but there are several hundred around the Great Lakes basin that are leaking persistent toxic substances. The Great Lakes Science Advisory Board o f the International Joint Commission recently toured nine o f the hazardous waste sites in the Niagara Falls area. Extensive engineering work has been successfully undertaken to The precautionary principle and early warnings of chemical contamination of the Great Lakes 131 intercept and treat contaminated groundwater leachate from the hazardous waste sites and prevent movem ent to the N iagara River and to Lake Ontario. However, these are costly systems to construct and maintain. For example, at the Occidental Chem ical C o rp oration 's Hyde Park hazardous waste site between 1998 and 2000, som e 100 million gallons of water (ca. 375 million litres) were intercepted and treated, and between 1989 and 2000, nearly 300 000 gallons of non-aqueous phase liquids (ca. 1.135 m illion litres) were collected and destroyed. The annual operation and m aintenance costs at this site alone are USD 2 million. Many o f the rem edial costs borne by the industries are unavailable, but available rem edial costs to industry and governments, incurred to date for the New York landfill sites, are at least USD 370 million. Rem ediation costs at New York hazardous waste sites are expected to total over USD 630 million (US EPA and NYSDEC, 2000). T here is a n eed for estim ates o f the benefits that will accrue from rem edial actions at these and at other sites around the Great Lakes, or public and political support for these costly schem es is likely to wane. The costs o f rem ediation of contaminated sedim ent in the U S part o f the Great Lakes have been detailed (U S EPA, 1998) with about USD 580 million spent on 38 sediment remediation projects since 1985. However this only represents a small fraction o f work necessary. Costs o f rem edial work for the O utboard Marine Corporation site at W aukegan, in Illinois, were USD 21 m illion to remove 136 000 kilograms of PCBs from OMC property. A further 900 kilograms of PCBs is in the navigational channel an d will be removed starting in 2002, at an estimated cost o f USD 12-14 m illion which will include construction o f a confined disposal facility. Similarly, estimates for treatm ent costs for rem edial work on the sediments o f Ham ilton H arbour range from CAD 60 million to 1 billion. There seem to be few benefits analyses to supplem ent these cost estimates. 12.5. Conclusions and lessons for the future The Great Lakes have provided a valuable if unwitting laboratory for studying the effects o f organochlorine com pounds, not only on the health o f wildlife and humans, but also on the political responses to pollution o f large ecosystems with persistent toxic substances. The record indicates the extraordinary lengths o f time between the introduction o f a new technology, the detection of an effect, the demonstration of a causal relationship, and the appropriate and sufficient response from the authorities (Lawless, 1977). It is only after m ore than half a century o f exposures to organochlorine com pounds from the Great Lakes that scientists are beginning to com prehend the scale of dam age that occurred to hum an health and to wildlife. Over this long period o f time, scientists have been faced with a dilem m a o f whether, on the one hand, to investigate injury to populations o f fish, wildlife and humans suspected to have been caused by persistent toxic substances, or to make a special plea to stop the sources and exposures even though they originally had little m ore than a suspicion. Ironically, even though a causal relationship has now been proven between the injury to health and exposures to persistent toxic substances, Great Lakes scientists have found it difficult to comm unicate their scientific evidence within the social, econom ic and political contexts in which their work is undertaken to effect remedial actions. These various case studies demonstrate that the length o f time between the introduction o f a new technology, produ ct or undertaking, the discovery o f its deleterious effects, and the regulatory, judicial or administrative action to reduce exposures, is seldom less than 25 years. The existing evidence concerning the persistence of the organochlorine com pounds released into the Great Lakes during the past 60 years indicates that it will probably be several m ore decades before the necessary rem edial actions will have reduced concentrations sufficiently to protect hum an reproduction and development from chemically in duced injury, particularly from consumption o f contaminated Great Lakes fish. The scientific aspects are characterised by a high degree o f certainty. In essence, we are engaged in the trans-generational transmission not only of the legacy of contamination and the associated dilemmas, but also o f chemically induced injury to the structural and functional development of exposed infants. But for the health, fisheries and environmental researchers and administrators there is a com plex array of social, economic and political dilemmas. For example, should health administrators draw attention to the contamination and recommend that children and women of 132 Late lessons from early warnings: the precautionary principle 1896-2000 Table 12.1. child-bearing age should not eat Great Lakes fish, or would this advice jeopardise commercial, sport and tribal fisheries interests? Similarly, should health and environmental researchers report on the injury to fisheries and wildlife populations, and on the injury to hum an health from the continuing exposures to elevated concentrations o f persistent toxic substances? Or would the requisite response to this information by the environmental administrators represent an apparently impossible financial burden for the U S and Canadian taxpayers? Great Lakes: early warnings and actions 1962 1963 1966 1969 1972 1974 1978 1978 1980 1984 1996 2000 2000 S ile n t sp rin g by Rachel Carson Is the significant early warning of the effects of organochlorlne pesticides on fish and wildlife as well as the threat of cancer in humans First observation of changes in eggshell quality on Pigeon Island in Lake O ntario H ickey e t al. publish the first analytical results of the presence of organochlorine com pounds in organism s in the G re at Lakes D D T and related pesticides are banned in Canada D D T is banned in the United States (dieldrin is banned in 1973) and gradual im provem ents in w ildlife begin in the G reat Lakes Concerns of possible effects of organochlorine com pounds on human health A ssociation of incidence of diseases (high rates of birth d efects, m iscarriages, cancers, etc.) in Love Canal, Niagara Falls with the disposal of toxic w astes (including dioxin) denied by Hooker Chemical Company Renegotiation of the Great Lakes W ater Q uality Agreem ent, including a precautionary policy, but it is not properly im plem ented President's Em ergency Declaration moves 900 fam ilies from the hazardous Love Canal area Studies show that, at birth, infants exposed to high levels of PC Bs (maternal consumption of Lake M ichigan contam inated fish) w eighed less and had sm aller heads Studies of Lake O ntario affected children published that observed sam e behavioural effects as in affected children from Lake Michigan The specific relationship between behavioural anomalies and prenatal exposures to the highly chlorinated biphenyls is determ ined Reluctance to undertake costly rem edial actions even after causal relationship is proven 12.6. References Broley, C. L., 1958. `Plight o f the Am erican bald eagle', Audubon Magazine Vol. 60, pp. 162-171. Broley, M. J., 1952. Eagle man, Pellegrini and Cudahy, New York. Burtraw, D. and Krupnick, A., 1999. Measuring the value of health improvementsfrom Great Lakes cleanup, Discussion p ap er 99-34, Resources for the Future, Washington, DC. Carson, R. 1962. Silent spring, H oughton Mifflin, Boston, MA. Chant, D. W., 1969. A petition to the Ministers o f H ealth and Welfare and of Agriculture, 25 Septem ber 1969. Colborn, T. an d Clem ent, C. (eds), 1992. Chemically induced alterations in sexual and functional development: The wildlife/human connection, Advances in M odern Environm ental Toxicology Vol. 21, Princeton Scientific Publishing, Princeton, New Jersey, 403 pages. Colborn, T., vom Saal, F. an d Short, P., 1998. Environmental endocrine-disrupting chemicals: neural, endocrine, and behavioral effects, Princeton Scientific, New Jersey, 418 pages. Cook, J . W., 1964. Review of the persistent organo-chlorine pesticides, H M SO , Lon don . Daly, H. B., 1993. `Laboratory rat experim ents show consum ption o f Lake Ontario salmon causes behavioral changes: support for wildlife and hum an research results', Journal of Great Lakes Research Vol. 19, pp. 784-788. Donahue, M. J., 1999. `The case for good governm ent: Why a comprehensive review of the G reat Lakes Water Quality A greem en t is n e e d e d ', Toledo Journal of Great Lakes Law, Science and Policy Vol. 2, pp. 1-11. Edwards, M. H., Quilliam, H. R. and Webb, D., 1963. `Uninhabited island in Lake The precautionary principle and early warnings of chemical contamination of the Great Lakes 133 O n tario', Audubon Field Notes Vol. 17, pp. 510 511. Fein, G. G., Jaco b so n , J . L., Jaco b so n , S. W., Schwartz, P. M. an d Dowler, J . K., 1984. `Prenatal exposure to polychlorinated biphenyls: Effects on birth size and gestational a g e ',Journal ofPediatrics Vol. 105, p p .315-320. Fox, G. A., 1991. `Practical causal inference for ecoepidem iologists',Journal of Toxicology and Environmental Health Vol. 33, pp. 359 373. Funtowicz, S. and Ravetz, J., 1993. `Science for the post-norm al a g e ', Futures Vol. 25, pp. 739-755. Gilbertson, M., 1996. `Great Lakes forensic toxicology and the implications for research an d regulatory p ro g ram s', Toxicology and Industrial Health Vol. 12, pp. 563-571. Gilbertson, M., M orris, R. D. an d Hunter, R. A., 1976. `A bnorm al chicks an d PCB residue levels in eggs o f colonial birds on the lower G reat Lakes (1 9 7 1 -1 9 7 3 )', The Auk Vol. 93, pp. 434-442. Gilbertson, M. an d Reynolds, L. M., 1972. `Hexachlorobenzene (HCB) in the eggs of common terns in Hamilton Harbour, O n tario', Bulletin ofEnvironmental Contamination and Toxicology Vol. 7, pp. 371 373. Grier, J . W., 1968. `Im m ature bald eagle with an abn orm al b e ak ', Bird Banding Vol. 39, pp. 58-59. Grier, J . W., 1982. `Ban o f D D T an d subsequent recovery o f reproduction in bald eagles', Science Vol. 218, pp. 1232-1235. Humphrey, H. E. B., 1983. `Population studies o f PCBs in M ichigan residents' in D'Itri, F. M. an d Kam rin, M. A. (eds), PCBs: Human and environmental hazards, Butterworth, Boston, MA, pp. 299-310. International Join t Commission. 1969. Pollution o f Lake Erie, Lake Ontario and the International Section o f the St. Lawrence River. (IJC: Ottawa and W ashington). 150 pp. Jaco b so n , J . L. an d Jaco b so n , S. W., 1993. `A four-year followup study o f children born to consumers of Lake Michigan fish', Journal of Great Lakes Research Vol. 19, pp. 776-783. Jaco b so n , J . L. an d Jaco b so n , S. W., 1996. `Intellectual im pairm ent in children exposed to polychlorinated biphenyls in u tero ', New England Journal ofMedicine Vol. 335, pp. 783 789. Jaco b so n , J . L., Jaco b so n , S. W., Fein, G. G., Schwartz, P. M. an d Dowler, J . K., 1984. `Prenatal exposure to an environmental toxin: A test of the multiple effects m odel', Developmental Psychology Vol. 20, pp. 523-532. Jaco b so n , J . L., Jaco b so n , S. W. and Humphrey, H. E. B., 1990. `Effects o f exposure to PCB and related com pounds on growth and activity in children', Neurotoxicology and Teratology Vol. 12, pp. 319 326. Jensen, S., 1966. `R eport o f a new chemical h azard', New Scientist Vol. 15, Decem ber, p. 612. Jo h n so n , B. L., Hicks, H. E. an d DeRosa, C. T., 1999. `Key environm ental h um an health issues in the G reat Lakes and St. Lawrence River basin s', Environmental Research Section A Vol. 80, pp. S2-12. H artig, J . H., Zarull, M. A. an d Law, N. L., 1998. `An ecosystem ap p roach to G reat Lakes m anagem ent: Practical steps',Journal ofGreat Lakes Research Vol. 24, pp. 739-750. H artsough, G. R., 1965. `Great Lakes fish now suspect as m ink fo o d ', American Fur Breeder Vol. 38, pp. 27-35. Hickey, J . J., Keith, J . A. an d Coon, F. B., 1966. `An exploration o f pesticides in a Lake M ichigan ecosystem. Pesticides in the environment and their effects on wildlife', Journal ofApplied Ecology Vol. 3 (Su p p l.), pp. 141-154. Johnson, B. L., Hicks, H. E., Jon es, D. E., Cibulas, W. an d DeRosa, C. T., 1998. `Public health implications o f persistent toxic substances in the Great Lakes and St Law rence B asin s', Journal of Great Lakes Research Vol. 24, pp. 698-722. Kaiser, K. L. E., 1974. `Mirex: an unrecognized contaminant o f fishes from Lake O n tario', Science Vol. 185, pp. 523-525. Keith, J.A. 1966. Reproduction in a population of herring gulls (Larus argentatus) contam inated by DDT.Jou rn al of Applied Ecology. 3(Suppl):57-70. 134 Late lessons from early warnings: the precautionary principle 1896-2000 Lawless, E. W., 1977. Technology and social shock, R utgers University Press, New Brunswick, NJ, 616 pages. Lear, L., 1997. Rachel Carson: Witnessfor nature, H enry Holt, New York, NY Levine, A. G., 1982. Love Canal,: Science, politics and people, L exin gton Books, Toronto, 263 pages. Lonky, E., R eih m an ,J., Darvill, T., Mather, Sr., J . and Daly, H., 1996. `N eonatal behavioral assessment scale perform ance in humans influenced by m aternal consum ption of environmentally contaminated Lake Ontario fish ',Journal of Great Lakes Research Vol. 22, pp. 198-212. Ludwig, J.P. and C.S. Tomoff. 1966. Reproductive success and insecticide residues in Lake M ichigan herring gulls. Jack-Pine Warbler. 44:77-84. M onsanto, 1970. `Monsanto cites actions taken on environm ental issue', 16 July, St Louis, 4 pages. O'Brien, M. H., 1994. `Com m ent on preface to the special section on cause-effect linkages. Causality: The missing link between science an d policy', Journal o f Great Lakes Research Vol. 20, pp. 590-592. Pekarik, C. and W eseloh, D. V., 1998. `Organochlorine contaminants in herring gull eggs from the Great Lakes 1974-1995: Change point regression analysis and short term regression', Environmental Monitoring and Assessment Vol. 53, pp. 77-115. Pfister, R. M., D ugan, P. R. an d Frea, J . I., 1969. `M icroparticulates: Isolation from water and identification of associated chlorinated p esticides', Science Vol. 166, pp. 878-879. Ratcliffe, D. A., 1972. `The peregrine population o f Great Britain in 1971', Bird Study Vol. 19, pp. 117-156. the science and technology o f residual insecticides in fo od production with special reference to aldrin and dieldrin, Shell Chemical Company, 26 October 1967. Shear, H., 1996. `The developm ent and use o f indicators to assess the state o f ecosystem health in the G reat L ak e s', Ecosystem Health Vol. 2, pp. 241-258. Shell Chem ical Company, 1967. `Aldrin, dieldrin, endrin: A status report', Agricultural Chemicals Division. Sprunt, A. I. V., R obertson, Jr., W. B., Postupalsky, S., Hensel, R. J., Knoder, C. E. an d Ligas, F. J., 1973. `Com parative productivity o f six bald eagle populations', Transactions of the North American Wildlife and Natural Resources Conference Vol. 38, p. 96. Stewart, P., Reihm an, J., Lonky, E., Darvill, T. and Pagano, J., 2000. `Prenatal PCB exposure and neonatal behavioral assessment scale (N BAS) p erfo rm an ce', Neurotoxicology and Teratology Vol. 22, pp. 21-29. Stow, C. A., Jack so n , L. J . an d Carpenter, S. R., 1999. `A m ixed-order m odel to assess contam inant declines', Environmental Monitoring and Assessment Vol. 55, pp. 435 444. Sudar, A. an d Muir, T., 1989. `Costs and consequences o f uncontrolled toxic waste sites along the N iagara River', Water Pollution Research Journal of Canada Vol. 2, pp. 279-297. Swanson, T. and Vighi, M., 1998. Regulating chemical accumulation in the environment. Cam bridge University Press, 278 pages. U S EPA, 1993. Water quality guidance for the Great Lakes system and correction: Proposed rules, Fed. Register 58, pp. 20802-21047. U S EPA, 1998. Realizing remediation: A summary of contaminated sediment remediation activities in the Great Lakes Basin, G reat Lakes National Program Office. Reynolds, L. M., 1969. `Polychlorobiphenyls (PCBs) and their interference with pesticide residue analysis', Bulletin ofEnvironmental Contamination and Toxicology Vol. 4, pp. 128 143. Robinson, J., 1967. `Residues of aldrin and dieldrin in wildlife in Britain', Symposium on U S EPA an d NYSDEC, 2000. Reduction of toxics loadings to the Niagara Riverfrom hazardous waste sites in the United States, U nited States Environmental Protection Agency and New York State Departm ent of Environmental Conservation. Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex 135 13. Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex David Santillo, Paul Johnston and William J. Langston 13.1. Introduction There is little doubt that, without som e form o f control, the accumulation o f marine fouling communities on vessels and m an m ade structures at sea increases drag and, in the case o f vessels, fuel consum ption, with substantial consequences in terms of econom ics and emissions. There is no doubt, also, that tributyltin (TBT) com pounds (and the less widely used triphenyltins) are extremely effective and relatively economical as antifouling biocides, contributing to the rapid take-up o f organotin-based paints by the shipping industry and small boat owners in the 1970s. These two argum ents have formed the basis o f the defence o f TBT antifouling formulations since the first undesirable consequences o f their use becam e apparent in the late 1970s, and indeed are still cited today by proponents of their continued suitability (for example, see Evans, 2000; Abel, 2000). What is missing from such evaluations, o f course, is p rop er consideration not only o f the quantifiable financial losses suffered by the aquaculture industry and im posed upon harbour authorities as a result o f the widespread application o f TBT paints, but also o f the broader environmental `costs' which led initially to restrictions on use and which now underlie the decision for a global phase-out. T he T B T story presents, in som e ways, a rath er unusual case study. Firstly, the widespread use o f TBT paints is a relatively recent development, with its origins in the late 1960s. Secondly, initial concerns about adverse effects were raised during the period in which its use becam e m ost popular, concerns that led rapidly to som e national and regional restrictions. Thirdly, the adverse effect m ost comm only associated with TBT, that o f im posex in marine gastropod m olluscs resulting from interference with steroid horm one metabolism, represented a highly sensitive, chemical-specific phenom enon. This factor contributed greatly to the early acceptance o f a direct causal relationship and, in turn, o f the need for controls. The paint m anufacturers, in concert with other interested parties, m ounted challenges to restrictions im posed in the 1980s, but the strength o f the evidence for severe effects on biota, including regional extermination o f some species, could hardly be denied. The prohibition of the use o f TBT paints on vessels under 25 m etres in length, effective in France from 1982, in the United Kingdom from 1987 and m ore widely from the end of the 1980s and early 1990s, did m uch to improve the situation within marinas and sheltered harbours where use on leisure craft had predom inated. Some regional recovery of affected mollusc populations has since been recorded. Through the late 1980s and 1990s, however, a picture o f m ore w idespread TBT contamination and population-level effects em erged, coincident with improved m onitoring and understanding o f the properties and environmental distribution of organotins. Vos et al. (2000) estim ate that im posex has now been docum ented in the wild for as many as 150 species o f m arine prosobranch snails worldwide. Evidence relating the prevalence o f im posex to density o f shipping traffic, along with poor recovery of affected populations in some areas and widespread accum ulation o f butyltin residues in m arine mammals, led to renewed calls for prohibitions to be extended to use on all vessels, irrespective o f size. As we stand on the brink ofju st such a prohibition, developed under the auspices o f the International M aritime O rgan ization 's M arine Environmental Protection Committee (M EPC), what can we learn from the process o f its evolution? 13.2. The emergence of the TBT problem Organotin com pounds were first developed as moth-proofing agents in the 1920s, and only later used m ore widely as bactericides an d fungicides (M oore et al., 1991). Dibutyltin and tributyltin com pounds have been produced since the late 1940s (Laughlin and Linden, 1985), although use o f TBT in marine antifouling paints dates only from the 1960s (Balls, 1987; ten HallersTjabbes, 1997) and then initially as a booster biocide in copper-based formulations. As a result o f its superior effectiveness over 136 Late lessons from early warnings: the precautionary principle 1896-2000 co p p er (Wade et al., 1988a), the use o f TBT paints by private and comm ercial users accelerated greatly in the 1970s, during which these formulations captured a m ajor proportion o f the antifouling m arket (Evans, 2000). `Free association' paints, which released the biocide rapidly at first but dem anded frequent reapplication, were gradually replaced by `self-polishing copolym er' formulations which ensured a more constant release of biocide and reduced repainting frequency. At the sam e time as its explosive increase in use, the first observations were m ade o f TBT effects on non-target organisms. While toxicity to fouling organisms was intentional, its propensity for wider impacts on the m arine environment had been grossly underestimated. An early focus on acute effects, especially mortality (Laughlin and Linden, 1987), failed to identify sub-lethal consequences o f prolonged exposure in som e taxa. For example, im posex (the developm ent o f m ale sexual structures in females) can be initiated in some gastropod m olluscs by TBT in the low n g /l (parts per trillion) range (Bryan et al., 1986; Alzieu, 1998), concentrations also known to cause shell deform ity an d larval m ortality (Alzieu et al., 1986). The phenom enon o f im posex was first described by Smith (1971), from studies of the A m erican mud-snail (Nassarius obsoletus) in the vicinity o f harbours on the U S east coast. Around the same time, Blaber (1970) recorded the appearance o f a penis in female dogwhelks (Nucella lapillus) in Plymouth Sound, United Kingdom, at much greater prevalence close to the harbour than further away. D espite the severity o f the phenom enon, however, the causal agent rem ained unknown. Only as analytical capabilities improved in the late 1970s and early 1980s was the connection with shipping m ade and the extent o f dam age already done recognised. Two regional case histories were instrum ental in identifying low-dose effects o f TBT and initiating development o f the first regional controls; the collapse o f the shellfish industry in Arcachon Bay (Atlantic coast o f France) and the reporting o f widespread im posex in dogwhelks from southern U K coastal waters. 13.3. Arcachon Bay Until the mid-1970s, Arcachon Bay had been an im portan t area for oyster ( Crassostrea gigas) culture, with production o f 10 000 15 000 tonnes per year (Evans, 2000) covering substantial areas of the tidal mud flats. The bay was also popular with leisure craft, with vessel num bers increasing from 7 500 in the mid-1970s to 15 000 at the start o f the 1980s. Estimated inputs o f TBT to the bay peak ed at aroun d 8 kg p er day (Ruiz et al., 1996). Im posex was first observed in the bay in 1970, affecting the predatory gastro pod Ocenebra erinacea (oyster drill), leadin g rapidly to its near extirpation from the bay (Gibbs, 1993). TBT was identified as the responsible agent only in the early 1980s. H ad the adverse effects been limited to the loss o f this species, considered a pest within the shellfish industry for its dam age to oyster stocks, little if any action may have followed. However, this early warning was quickly followed by failure o f the oyster stocks themselves. Despite a norm al spawning event in sum m er 1976, few o f the larvae survived. Larval settlem ent largely failed through the late 1970s and into the 1980s, resulting in massive financial losses by the shellfish industry. By 1981, annual oyster production h ad fallen to only 3 000 tonnes (Ruiz et al., 1996). In addition to reproductive failure, adult oysters were rendered unsaleable by shell deform ation leading, in severe cases, to `ball-shaped' specim ens (Alzieu et al., 1989). Such observations pre-dated analytical techniques sensitive enough to describe in detail environmental distributions of TBT. Alzieu et al. (1986) provided the first reliable survey o f organotins in the waters o f A rcachon Bay, while sedim ent data were not available until the 1990s (Sarradin et al., 1994). Nevertheless, the severity o f impacts on the ecology o f A rcachon Bay, m anifest in heavy financial losses, was sufficient to stim ulate relatively swift action by the French government. Acting on the best information available linking the oyster collapse to the presence o f TBT, France was the first country to introduce legislation prohibiting the application o f TBT paints to small (< 25 m etre) vessels, in 1982 (Michel and Averty, 1999a). T hese controls undoubtedly markedly reduced TBT inputs from marinas throughout France. In the case o f Arcachon, im plem entation was probably aided by the Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex 137 local provenance o f many boat-owners and their interest in preserving a local industry. Their effectiveness in addressing more widespread TBT contamination remains questionable, however (M ichel an d Averty, 1999a). 13.4. UK harbours and coastal waters Following B lab er's (1970) observations in Plymouth Sound, other research began to identify im posex as a m ore widespread problem . Frequency o f im posex in N. lapillus was greatest close to ports and harbours, although evidence was also growing for the spread of the phenom enon through coastal waters o f the southern United Kingdom (Bryan et al., 1986 and 1987). In addition, Waldock and Thain (1983) linked the failure o f attempts to introduce Pacific oyster culture into the United Kingdom in the early 1980s to TBT exposure rather than to fine sedim ent particles as previously thought. Reproductive failures and shell deformities in the United Kingdom showed many parallels with those record ed in A rcachon Bay. Based on em ergent concerns, the UK government introduced controls on the sale o f TBT paints for use on small vessels in 1985. These included controls on retail sales, guidelines for handling paints and the setting o f an environmental quality target (EQT) of 20 n g /l TBT (8 n g /l as tin), a level considered at the time to be sufficiently protective o f m arine biota (W aldock et al., 1987). TBT concentrations in many coastal waters consistently exceeded the EQT. Moreover, as understanding developed through the 1980s, it became clear that concentrations well below the EQ T could cause severe effects. Sterilisation of female dogwhelks occurs at T BT concentrations as low as 3-5 n g /l TBT (G ibbs et al., 1988), with alm ost all fem ales affected once concentrations reach 10 n g /l and more fundamental changes at higher concentrations (including testes development and suppression o f egg production). During the 1980s, butyltin concentrations exhibited strong seasonal trends relating to boating activity and m aintenance. Sum m er time concentrations in m arinas regularly exceeded 100 n g /l and occasionally 1 000 n g /l (W aldock et al., 1987). L an gston et al. (1987) reported similar concentrations in m arina waters within Poole H arbour (southern United Kingdom ). These marinas regularly served 5 000 leisure craft, engaged in hull m aintenance and berthing. Even beyond the marinas, TBT concentrations above 100 n g /l were sometimes encountered. The overlap between field concentrations and those known to have severe developmental effects highlights the extent of the problem faced, and illustrates the inadequacy of the EQ T as a protective in stru m en t. Evidence continued to accumulate during the 1980s. Bryan et al.'s (1986) study o f dogwhelks in southwest England was particularly influential, concluding: `that the incidence o f `im p o sex '... is w idespread, that all populations are affected to som e degree... Populations close to centres of boating and shipping activity show the highest degrees o f im posex'. The study also noted that the rapid increase in im posex in Plymouth Sound strongly coincided with contem porary increases in TBT applications, and revealed a strong correlation with tissue concentrations o f organotin residues. The significance o f this study was profiled in Septem ber the same year in the jo u rn a l Marine Pollution Bulletin ( `T B T linked to dogwhelk decline', 1986), in many ways m arking wider acceptance o f the central role o f TBT in observed declines. The geographical extent of the TBT problem in Europe had been recognised by the late 1980s (Bailey and Davies, 1988a and 1989; Gibbs et al., 1991). Balls (1987) reported that use o f TBT paints on salmon-farming cages had resulted in contamination o f a Scottish sea loch, and noted the generally higher prevalence o f im posex within sea lochs used for aquaculture. Cleary and Stebbing (1987) highlighted the additional concern of accumulation o f organotins in the lipid-rich sea-surface microlayer, a threat considered particularly relevant to buoyant fish larvae (because o f resulting exposure during sensitive developmental periods) and intertidal organisms (exposed to the microlayer during tidal ebb and flow) (Cleary et al., 1993). C oncerns over persistence and impacts in sediments were also em erging at this time (Lan gston et al., 1987; L an gston and Burt, 1991). Faced with the accum ulation o f research, the UK government considered options for further retail restrictions and, in January 138 Late lessons from early warnings: the precautionary principle 1896-2000 1987, reduced the maximum permissible content o f organotins in applied paint, helped by technological developments (Side, 1987) . These limits were rapidly superseded, however, by the introduction in May 1987 o f a total ban on the retail sale of TBT paint (W aldock et al., 1987 an d 1988) for use on vessels under 25 metres and on fish cages (earlier voluntary m easures having failed adequately to address this latter application). At this time, work by the U K Water Research Centre was under way to recom m end environmental quality standards (EQS) for organotin com poun ds in water (Zabel et al., 1988) , involving consideration o f toxicity and environmental behaviour of a range of organotins. It became apparent that for TBT, in particular, no-effect levels could not be achieved in the environment without use restrictions. Two years later, and in recognition o f the low-dose effects, the 1985 EQ T was replaced by an EQ S o f only 2 n g /l (Cleary, 1991). In Ju n e 1987, the Paris Commission (PARCOM), responsible for administering the Paris Convention (1978), recognised that the use o f T B T paints was causing `serious pollution in the inshore areas o f the Convention waters (NE A tlantic)' (PARCOM, 1987). PARCOM Recommendation 87/1 called for a harm onised ban on retail sales for application to pleasure boats and fish cages and, furthermore, for consideration of restrictions for seagoing vessels and underwater structures. Though strong in intent, it quickly becam e apparent that PARCOM could not achieve restrictions within the com m ercial shipping sector. As this issue had already been referred to the International Maritime Organization (IM O ), PARCOM instead turned its attention to inputs from docking activities, especially hull maintenance operations, in PARCOM Recom m endation 88/1 (PARCOM, 1988). The effectiveness o f such measures remains difficult to evaluate, though it is indisputable that shipyards and docks remain important point sources of organotins. 13.5. A global pollutant Concerns over the impacts of TBT soon extended worldwide, leading to a series of national and regional m easures towards the end o f the 1980s. The annual O ceans symposia, organised by the U S Marine Technology Society, did m uch to disseminate the growing body o f research (Wade et al., 1988b; Krone et al., 1989). The NOAA (U S National Oceanic and Atmospheric Administration) Mussel Watch program m e highlighted the ubiquitous distribution of butyltins in bivalves from U S coasts (Wade et al., 1988a an d b; U h ler et al., 1989), while work in New Zealand confirm ed accum ulation in sedim ents (King et al., 1989). Prohibitions addressing small vessels were adopted in the U nited States in 1988, followed in 1989 by similar m easures in Canada, New Zealand and Australia, and by European legislation in 1991 harm onising controls throughout the EU (Evans, 2000). 13.6. Effectiveness of controls on small vessels Retail restrictions undoubtedly resulted in a m ajor shift away from the use o f T B T paint on leisure craft (although the extent of illegal application remains unknown) and substantial reductions in inputs. In some regions at least, this was m anifest in partial recovery o f mollusc populations. In A rcachon Bay, m onitoring o f the effectiveness o f regulations in reducing water colum n concentrations o f butyltins was limited by the high detection limits o f early analytical m ethods. While Alzieu et al. (1986) were able to demonstrate that concentrations fell from 900 n g /l tin in 1983 to below 100 n g /l by 1985, it took until the late 1980s to confirm that concentrations had fallen below 10 n g /l in m ost parts o f the bay. Hydrodynamic disturbance has largely prevented reconstruction of sediment concentration trends (Ruiz et al., 1996). Recovery o f the oyster beds, and resum ption of commercial operations, in the mid-1980s provided the first indirect evidence o f the effectiveness o f the ban. Indeed, the recovery o f the bay is frequently cited as an illustration o f how the limited controls of the 1980s `solved' the T B T problem (N icholson and Evans, 1997; Evans, 2000), a view which has m et substantial disagreement. Evidence for partial recovery o f severely affected mollusc populations certainly exists from several parts of Europe. For example, Minchin (1995) reported recovery of the flame shell (Lima hians) and associated flora in Mulroy Bay (Ireland), previously decimated through use of TBT on salmon cages between 1981 and 1985. In the sheltered inlet o f Sullom Voe, Shetland Islands, and adjacent Yell Sound, im posex in dogwhelks was significantly less prevalent in Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex 139 1995 than in 1991 (H ardin g et al., 1997) although overall prevalence rem ained high. Evans et al. (2000a) suggest that there has been a more general recovery in UK dogwhelk populations over the last decade. O ther studies provide less cause for optim ism . M inchin et al. (1997) stressed that recovery of dogwhelk populations in Ballybegs (Ireland) had been slower than expected, and recoveries in Sullom Voe and Yell S oun d occurred against a continuing high background of imposex, and continued absence o f dogwhelks from worst affected areas. In C anada, St-Jean et al. (1999) noted that butyltin concentrations in blue mussels (Mytilus edulis) an d sedim ents from the southern G ulf of St Lawrence remained remarkably high eight years after retail restrictions. Despite initial declines in French coastal waters, dissolved T B T concentrations have since stabilised and rem ain at levels often well in excess o f those known to cause adverse effects in some species (Michel and Averty, 1999a). In som e areas o f A rcachon Bay, concentrations rem ained high enough to cause im posex in sensitive species 10 years after the 1982 regulations (Ruiz et al., 1996). Similar concerns have been voiced for UK estuarine and coastal locations (Cleary, 1991; L an gston et al., 1994). The underlying causes of continued contam ination and failure o f ecosystems fully to recover undoubtedly differ from location to location. In som e cases (e.g. H u et et al., 1996), isolated but significant illegal use of T B T paints on small vessels has been im plicated, while other researchers have stressed the importance o f release from historically contaminated sediments (W aldock et al., 1990; L an gston et al., 1994). H igh larval sensitivities (Gibbs, 1993) and lon g life histories (Rees et al., 1999) undoubtedly also contribute to the slow rate o f recovery of some species. Potentially the m ost significant single contributor to continued widespread presence o f TBT in the marine environment, however, and currently the m ost fiercely debated, is use on seagoin g vessels. 13.7. The significance of seagoing vessels At the time o f the first national and regional prohibitions, the use o f TBT paints on comm ercial and military ships was viewed as a lesser concern because, it was argued, those vessels spent m ost o f their operational lives on the open seas. Nevertheless, these ships use inshore waters and port facilities on a regular basis. The scale of TBT inputs from hull m aintenance of large vessels has been recognised for som e time. W aldock et al. (1988) recorded TBT in wash water from a naval frigate at approximately 1 million times the lowest biologically effective concentrations, and estimated total input from cleaning a single vessel at 100 g of freely available TBT. Including T B T bound to paint-chips, which m ight act as a long-term reservoir o f butyltins to the environment, led to estimates o f alm ost 1 kg TBT per vessel per cleaning operation. Guidelines to control inputs from such operations have been in use in many countries for som e time, although their effectiveness is difficult to evaluate. Studies surrounding the Sullom Voe oil terminal (Bailey and Davies, 1988b) provide evidence for significant inputs o f organotins from norm al transport operations of TBTcoated ships. Although uses on small service vessels and m arker buoys contributed to early inputs, more recent inputs arise predom inantly from movements o f oil tankers themselves. Significant contributions from heavy commercial traffic have also been highlighted in the G ulf o f St Lawrence (StJe a n et al., 1999). While ongoing inputs from comm ercial shipping to coastal waters are now generally accepted, the significance o f such sources to remote coastal and offshore areas remains the subject o f intense debate (ten HallersTjabbes, 1997; Evans, 2000). This is despite the increasing body o f evidence supporting a link between density of shipping traffic and occurrence of biological effects (predom inantly im posex) in offshore waters. Am ong the first to highlight im posex in the com m on whelk (Buccinum undatum) from the open North Sea and describe a positive correlation with shipping intensity were ten H allers-Tjabbes et al. (1994). Sim ilar correlations have since been reported in the Strait o f Malacca, connecting the Bay of B en gal with the South China Sea (Sw ennen et al., 1997; H ashim oto et al., 1998), and in rem ote parts o f the G alician coast (Ruiz et al., 1998). Cadee et al. (1995) speculated that organotins might be responsible for local destruction o f whelks in the Dutch Wadden Sea. M ore recently, Davies et al. (1998) estim ated total annual inputs o f 68 tonnes T B T to the N orth Sea from shipping, with 140 Late lessons from early warnings: the precautionary principle 1896-2000 continuous leaching from pain ted hulls when various infectious diseases (K annan et al., in port or under way the m ajor contributor. 1998). Nicholson and Evans (1997) provided further evidence supporting the pioneering work o f ten H allers-Tjabbes et al. (1994) but questioned the significance o f what they termed `m ild' im posex against a background o f overexploitation by shell fisheries. T h ere is little doubt, however, that detection o f population-level effects in offshore regions was an im portant factor underlying calls for the extension o f restrictions in the mid1990s. D uring the sam e period, the truly global nature o f the T BT problem was acknowledged (Ellis and Pattisina, 1990; K annan et al., 1995a, b an d c). T hough data remain limited for areas outside Europe and North America, recent studies in Jap an and the Philippines (H arin o et al., 1998a, b, c and 1999; Prudente et al., 1999) have gon e som e way towards redressing the balance, confirming (as suspected) the occurrence of similar relationships. Recent research has also highlighted the w idespread accum ulation o f organotins in organism s higher in the food chain, including cetaceans. Iwata et al. (1995) were am ong the first to determine concentrations in marine mammals, suggesting that the high levels recorded (up to 10 parts per m illion (ppm ) in porpoise liver) reflected a low potential for metabolism of these com poun ds. Tanabe et al. (1998) later extended the data set, including species from North Pacific and Asian coastal waters. K annan et al. (1996) described butyltin accumulation in dolphin, tuna and sharks from the Mediterranean, noting markedly different ratios of T BT breakdown products within the different taxa. First reports of butyltin residues in European cetaceans and seals, including in pelagic cetaceans that feed in rem ote offshore waters, were published by A riese et al. (1998) and Law et al. (1998 and 1999). Although uncertainty remains, evidence suggests that accum ulation o f butyltins in top predators m ight adversely effect the immune system. K annan et al. (1997) reported high levels o f TB T and its breakdown products (primarily dibutyltin) in bottlenose dolphins associated with mortality events on the US Atlantic and Gulf coasts. Similar correlations have been reported for Californian southern sea otters (Enhydra lutris nereis) exhibiting H um an intake o f organotins from seafood, especially from fish farm ed in TBT-treated cages, has long been recognised (Davies and McKie, 1987; Ishizaka et al., 1989). Only m ore recently, however, has such intake been evaluated. Cardwell et al. (1999) reported that residues were widely detectable in U S seafood but that estimated intakes were significantly below levels judged to be o f concern for human health. In contrast, B elfroid et al. (2000) concluded that average intakes o f TBT from seafood could lead to exceedance o f tolerable daily intakes based on m ore sensitive im m unotoxicological end points for some products retailed in North America, Europe and Asia, while stressing that, for the majority o f countries, data were simply unavailable. 13.8. Progress towards a global phase-out Despite uncertainties regarding `far-field' effects, contamination o f the marine environm ent with organotins is clearly a persistent and pervasive problem . In 1995, and in spite o f a decision by MEPC in 1994 that no further controls were necessary, ministers at the fourth Conference on the Protection o f the North Sea in Esbjerg agreed: `to undertake concerted action within IMO aim ing at a worldwide phase-out o f the use o f TBT on all ships' (MINDEC, 1995). Their preference for global action within IMO, in com m on with the position taken by the OSPAR (Convention for the Protection o f the M arine Environm ent o f the Northeast Atlantic) Commission, recognised the limits o f regional m easures in addressing a transboundary problem . On the basis of em erging research, M EPC agreed in 1996 to look again at the need for a global TBT ban (ten Hallers-Tjabbes, 1997). The possibility o f such a ban was explicitly recognised at M EPC 's 40th session (M EPC, 1997), m arking a sea change in the thinking o f that committee. Draft mandatory regulations were duly adopted the following year (MEPC, 1998). The deadlines now incorporated in the draft International Convention on the Control of Harm ful Anti-fouling Systems (2003 for phase-out o f application o f organotin paints, 2008 for their presence on ship hulls) were Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex 141 adopted under IMO Assembly Resolution A.895(21) in November 1999, and formal adoption o f the convention is expected in 2001. 13.9. The question of alternatives With the advent o f such a progressive m easure, attention is focusing increasingly on the availability o f effective and economically viable alternatives to TBT (Evans et al., 2000b). It is clear that som e o f the preparations in use prior to the widespread introduction o f TBT (including those containing mercury or arsenic com pounds) are not acceptable alternatives. Copper-based systems generally require booster biocides in order to be effective, biocides which themselves may present additional problem s. Those containing the triazine herbicide Irgarol 1051, used extensively in som e areas as a replacem ent for TBT on leisure craft, are commonly cited exam ples. W idespread occurrence o f this h erbicide in certain estuaries (Scarlett et al., 1997) , coupled with direct effects on plant growth in the field (Dahl and Blanck, 1996; Scarlett et al., 1999), in som e ways m irror early findings with respect to TBT. Some commentators use such examples as justification against substitution o f TBT in comm ercial shipping applications (Abel, 2000; A bbott et al., 2000). This is a som ewhat negative argument. The m ore appropriate question is, can fouling control be achieved without recourse to such highly toxic, persistent and bioaccumulative substances? Research into use of natural antifouling chemicals is developing rapidly (Clare, 1998) , though commercial application may rem ain som e time away. Perhaps the m ost prom ising contem porary solution is biocidefree, `non-stick' coatings which simply present a physical barrier to settlement. Such coatings have been commercially available for many years, and retain a significant m arket share for leisure craft. Their viability and perform ance on larger vessels remain under evaluation, although applicability for fast-moving craft has been demonstrated. Although significant technical issues remain to be resolved, advances towards antifouling mechanism s which do not release hazardous substances to the sea would seem desirable. 13.10. Late lessons from the TBT story The above discussions indicate the degree to which the hazards o f TBT were initially underestim ated. It was initially assum ed, for exam ple, based on acute toxicity tests that concentrations in the u g /l (micrograms per litre) range were necessary to initiate biological effects and that, as a consequence, an EQ T of 20 n g /l would be sufficiently protective. It quickly becam e apparent that, even if the EQ T could be m et (and it was greatly exceeded in som e areas), severe biological effects could be expected. H ad the endocrine-mediated mechanisms underlying im posex (Matthiessen and Gibbs, 1998) been identified earlier, the potential for low-dose effects would have been clear. O f course, much o f our knowledge o f organotin toxicity has been obtained through hindsight an d is likely to continue to develop (Langston, 1996; Bouch ard et al., 1999; M orcillo and Porte, 2000). In the meantime, it may be h o p ed that the lessons gain ed will allow earlier identification and, wherever possible, avoidance o f future problem s in advance. The persistence of organotins, permitting their accumulation and wider dissemination, was also underestim ated. Early predictions that TBT would degrade rapidly in surface waters (see reviews by Sim m onds, 1986 and L ee et al., 1989) failed to account for the high lipophilicity and sediment-binding properties o f organotins in seawater. Half-life estimates in the order o f days for eutrophic surface waters m ust be contrasted with up to several years for residues in nutrient-poor waters (M ichel and Averty, 1999b) an d m arine sedim ents (de M ora et al., 1989), especially anaerobic sediments. The persistence of TBT, and toxicity to sedim ent communities, raises the prospect of delayed recovery of dam aged ecosystems (Lan gston et al., 1994; D ah llof et al., 1999) and represents a substantial legacy to be borne by authorities responsible for dredging operations. This problem is only now receiving form al recognition within international conventions (London Convention 1972, OSPAR Convention 1992) regulating the dum ping of dredge materials at sea. Organotin bioaccum ulation was also underestimated, and the specific role of biofilms in enhancing bioaccumulation and toxicity has only recently been recognised (Labare et al., 1997). A ccum ulation in top predators was simply not envisaged. 142 Late lessons from early warnings: the precautionary principle 1896-2000 H indsight is a w onderful thing, o f course. Nevertheless, increasing awareness of pathways and interactions within the marine environm ent, coupled with an appreciation o f their complexity and indeterminate nature (Santillo et al., 1998), should equip us to reduce or even avoid the potential for future TBT-style scenarios involving other persistent organic pollutants. Two key factors contributed to early ignorance o f the geographical extent o f the TBT problem : low sensitivity o f analytical m ethods and absence, in m ost regions, o f adequate baseline data on the distribution and ecology of non-commercial species. The first o f these was subject to an unavoidable process o f m ethods development, partly overcome by the high specificity o f im posex as a biomarker for TBT exposure. The second limitation illustrates a more general concern which could have been avoided through greater focus on the collection of `baseline' data. Where such data were available (e.g. southwest England), they were invaluable in the early detection o f effects on non-commercial species. Though frequently underrated, baseline studies play a vital role in the early detection o f adverse trends and may consequently serve the application o f precaution. Equally im portant is the need for continuing long-term monitoring of TBTaffected areas, as this will provide us with unique insights into recovery of pollutantdam aged environments. The continuation o f the TB T problem in coastal waters o fJa p a n (Iwata et al., 1995; H arin o et al., 1998a an d b ), despite a national ban on all applications o f TBT for marine antifouling, clearly illustrates the transboundary nature o f the problem . It would seem that universal, global restrictions are the only way to ad d ress the totality o f the TBT problem and to achieve environmental quality standards in all coastal systems. Finally, though the selection o f alternatives to h azardous chem icals is never a sim ple task, this does notjustify inaction. Several o f the existing TBT `alternatives' carry their own problem s, though thankfully, on a global scale, none has yet proved as dam aging as TBT. In selecting appropriate alternatives, those which do not rely on the release o f hazardous substances to the marine environment should, perhaps, be viewed m ost favourably. Broader consideration of problem s may give rise to m ore beneficial solutions than simple `chemical for chemical' substitution. 13.11. Conclusions: precaution or retrospective action? The use of organotin antifoulants has and continues to result in widespread and sometimes severe environmental effects. Unusually in the field o f ecotoxicology, the evidence linking cause and effect (with respect to TBT-induced im posex) is irrefutable. Vos et al. (2000) refer to the phenom enon as `the best example of endocrine disruption in invertebrates that is causally linked to an environmental pollutant'. It would be difficult to argue, therefore, that any o f the actions to address T BT to date have been precautionary, resulting as they have from extensive docum entation of ecological impacts. Actions have undoubtedly contributed towards rem ediating the m ost severe problem s, but this is not precaution. The 1987 UK restrictions, for example, stemmed from a realisation that environmental concentrations were already well above those known to cause chronic effects and had already resulted in widespread declines in gastropod populations. It could be argued, similarly, that agreem ent o f the IMO convention which will bring about elimination of organotin antifoulants has come only after the consequences of continued use have been well docum ented. Although the significance o f this global, legally binding treaty should not be underestim ated, it too is fundamentally retrospective in action. Complete phase-out of organotin paints from the global shippin g fleet by 2008 will m ark the closure o f an im portant chapter in the T B T story. Persistence in sedim ents an d longlived biota will rem ain to be addressed, but widespread inputs from shipping should, at least, be a thing o f the past. Organotin inputs will continue, o f course, through their use as additives in a wide range o f consum er products. Whether efforts to address these em ergin g challenges will draw on the lessons o f the past rem ains to be seen. Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex 143 TBT: early warnings and actions Early 1970s 1976-81 1982 1985 1986 January 1987 May 1987 June 1987 1988 1989 1991 1994 1995 1997 1998 November 1999 2001 Rapid increase in the use of T B T antifouling paints on vessels of all sizes and first reports of im posex in marine snails (Blaber, 1970; Sm ith, 1971) Repeated failure of larval settlem ent leads to near collapse of oyster fishery, Arcachon Bay, France France introduces legislation prohibiting the use of T B T paints on small vessels First controls introduced in United Kingdom limiting concentrations of T B T in paints Bryan e t al. (1986) report w idespread im posex in dogwhelks on southern coast of United Kingdom , linked to T B T United Kingdom announces further restrictions on T B T content of applied antifouling paint United Kingdom introduces ban on retail sale of T B T paint for use on vessels < 25 m and on fish cages PARCO M Recom m endation 87/1 calls for sim ilar ban over entire convention area (Northeast Atlantic) United States introduces restrictions. W aldock e t al. (1988) highlight significance of inputs from shipyards Restrictions introduced in C anad a, A ustralia and N ew Zealand Harmonised ban on retail sale of T B T paint introduced at European Union level Early reports of im posex in w helks from offshore areas of North Sea linked to shipping activity Ministerial declaration of fourth North Sea conference (Esbjerg) commits to working for global phase-out of T B T paint within IMO Concept of global phase out of organotin containing paints agreed at M EPC 's 40th session Draft mandatory regulations aimed at such a phase-out adopted. O SPA R (Convention for the Protection of the Marine Environm ent of the Northeast Atlantic) prioritises organotins for action to cease all releases. Cessation of all releases of organotins to marine environm ent, under O S P A R 's hazardous substances strateg y in 2020 Deadlines for phase-out adopted under IMO Assem bly Resolution A.895(21) Te xt of International Convention on the Control of Harmful Anti-fouling System s to be finalised. In 2003 w orldw ide prohibition on new application of organotin antifoulants to all vessels and in 2008 the existing organotin antifouling coatings will be replaced on all vessels worldwide 13.12. References Abbott, A., Abel, P. D., A rnold, D. W. an d M ilne, A., 2000. `Cost-benefit analysis o f the use of TBT: The case for a treatment a p p ro a ch ', The Science of the Total Environment Vol. 258, pp. 5-19. Abel, P. D., 2000. `T B T -- towards a better way to regulate pollu tan ts', The Science of the Total Environment Vol. 258, pp. 1-4. Alzieu, C., 1998. `Tributyltin: case study o f a chronic contaminant in the coastal environm ent', Ocean and Coastal Management Vol. 40, pp. 23-26. Alzieu, C., Sanjuan, J., Deltriel, J.P. and Borel, M., 1986. `Tin contam ination in A rcachon bay: Effects on oyster shell anom alies, Marine Pollution Bulletin Vol. 17, pp. 494-498. Alzieu, C., Sanjuan, J., Michel, P., Borel, M. and D reno, J . P., 1989. `M onitoring and assessment o f butyltins in Atlantic coastal w aters', Marine Pollution Bulletin Vol. 20, pp. 22-26. Ariese, F., van H attum , B., H opm an, G., Boon, J. and ten Hallers-Tjabbes, C., 1998. Butyltin and phenyltin compounds in liver and blubber samples of sperm whales (Physeter m acroceph alus) stranded in theNetherlands and Denmark, R eport W 98-04, M arch 1998, Institute for Environm ental Studies, Vrije Universiteit, Amsterdam. Bailey, S. K. an d Davies, I. M., 1988a. `Tributyltin contam ination in the Firth o f Forth (1 9 7 5 -1 9 8 7 )', The Science of the Total Environment Vol. 76, pp. 185-192. Bailey, S. K. an d Davies, I. M., 1988b. `Tributyltin contam ination around an oil terminal in Sullom Voe (Shetland) ', Environmental Pollution Vol. 55, pp. 161-172. Bailey, S. K. an d Davies, I. M., 1989. `The effects o f tributyltin on dogwhelks (Nucella lapillus) from Scottish coastal waters', Journal of the Marine Biological Association, UK Vol. 69, p p .335-354. Balls, P. W., 1987. `Tributyltin (T B T ) in the waters o f a Scottish sea loch arising from the Table 13.1. 144 Late lessons from early warnings: the precautionary principle 1896-2000 use o f antifoulant treated netting by salm on farm s', Aquaculture Vol. 65, pp. 227-237. Belfroid, A. C., Purperhart, M. an d Ariese, F., 2000. `O rganotin levels in seafood', Marine Pollution Bulletin Vol. 40, N o 3, pp. 226-232. Blaber, S. J. M., 1970. `The occurrence o f penis-like outgrowth behind the right tentacle in spen t fem ales o f Nucella lapillus ( L .) ', Proceedings of the Malacological Society of London Vol. 39, pp. 231-233. Bouchard, N., Pelletier, E. and Fournier, M., 1999. `Effects o f butyltin com pounds on phagocytic activity o f hemocytes from three m arine bivalves', Environmental Toxicology and Chemistry Vol. 18, N o 3, pp. 519-522. Bryan, G. W., Gibbs, P. E., H um m erstone, L. G. and Burt, G. R., 1986. `The decline o f the gastro p od Nucella lapillus aroun d south-west England: Evidence for the effect of tributyltin from antifouling p ain ts',Journal of the Marine Biological Association, UK Vol. 66, pp. 611-640. Bryan, G. W., Gibbs, P. E., H um m erstone, L. G. and Burt, G. R., 1987. `Copper, zinc and organotin as long-term factors governing the distribution o f organisms in the Fal estuary in southwest E n g lan d ', Estuaries Vol. 10, N o 3, p p .208-219. Cadee, G. C., Boon, J . P., Fischer, C. V., M ensink, B. P. an d ten H allers-Tjabbes, C. C., 1995. `Why the whelk (Buccinum undatum) has becom e extinct in the Dutch Wadden S e a ', NetherlandsJournal ofSea Research Vol. 34, No 4, pp. 337-339. Cardwell, R. D., Kiethly, J . C. and Sim monds, J., 1999. `Tributyltin in U.S. market-bought seafood and assessment of human health risks', Human and Ecological Risk Assessment Vol. 5, N o 2, pp. 317-335. near-shore waters', CM 1993/E:28, International Council for the Exploration of the Sea, Marine Environm ent Quality Com mittee, 14 pages. Cleary, J. J. and Stebbing, A. R. D., 1987. `Organotin in the surface microlayer and subsurface waters of southwest England', Marine Pollution Bulletin Vol. 18, No 5, pp. 238-246. Dahl, B. and Blanck, H., 1996. `Toxic effects o f the antifouling agent Irgarol 1051 on periphyton communities in coastal water m icrocosm s', Marine Pollution Bulletin Vol. 32, No 4, pp. 342-350. Dahllof, I., Blanck, H., Hall, P. O. J . and M olander, S., 1999. `Long-term effects o f trin-butyl-tin on the function o f a m arine sedim ent system ', Marine Ecology Progress Series Vol. 188, pp. 1-11. Davies, I. M., Bailey, S. K. an d H arding, M. J. C., 1998. `Tributyltin inputs to the North Sea from shipping activities, and potential risk of biological effects', ICESJournal ofMarine Science Vol. 55, pp. 34-43. Davies, I. M. an d McKie, J . C., 1987. `A ccum ulation o f total tin and tributyltin in the muscle tissue of farm ed Atlantic salm on', Marine Pollution Bulletin Vol. 18, No 7, pp. 405-407. de Mora, S. J., King, N. G. and Miller, M. C., 1989. `Tributyltin and total tin in m arine sediments: profiles and the apparent rate of T B T d egrad atio n ', Environmental Technology Letters Vol. 10, pp. 901-908. Ellis, D. V. an d Pattisina, L. A., 1990. `W idespread n eo gastro pod im posex: A biological indicator of global contamination', Marine Pollution Bulletin Vol. 21, pp. 248-253. Clare, A. S., 1998. `Towards nontoxic an tifoulin g', Journal ofMarine Biotechnology Vol. 6, pp. 3-6. Cleary, J. J., 1991. `O rganotin in the m arine surface microlayer and sub-surface waters of south-west England: Relation to toxicity thresholds and the U K environmental quality stan d ard ', Marine Environmental Research Vol. 32, pp. 213-222. Cleary, J. J., McFadzen, I. R. B. and Peters, L. D., 1993. `Surface microlayer contam ination and toxicity in the North Sea and Plymouth Evans, S. M., 2000. `M arine antifoulants' in Sheppard, C. (ed.), Seas at the millennium: An environmental evaluation, Vol. III: Global issues and processes, pp. 247-256, Elsevier Science Ltd, Oxford. Evans, S. M., Birchenough, A. C. and Fletcher, H., 2000a. `The value and validity o f community-based research: TBT contam ination o f the North S ea', Marine Pollution Bulletin Vol. 40, No 3, pp. 220-225. Evans, S. M., Birchenough, A. C. and Brancato, M. S., 2000b. `The T B T ban -- out Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex 145 o f the frying pan into the fire?', Marine Pollution Bulletin Vol. 40, N o 3, pp. 204-211. Marine Environmental Research Vol. 45, No 2, pp. 169-177. Gibbs, P. E., 1993. `A m ale genital defect in the dog-whelk, Nucella lapillus (N eogastropoda), favouring survival in a TBT-polluted a re a ', Journal of the Marine Biological Association, UK Vol. 73, pp. 667-678. Huet, M., Paulet, Y M. an d Glem arec, M., 1996. `Tributyltin (TB T) pollution in the coastal waters as indicated by im posex in Nucella lapillus', Marine Environmental Research Vol. 41, pp. 157-167. Gibbs, P. E., Pascoe, P. L. an d Burt, G. R., 1988. `Sex change in the fem ale dog-whelk, Nucella lapillus, in duced by tributyltin from antifouling p ain ts', Journal of the Marine Biological Association, UK Vol. 68, pp. 715-731. Gibbs, P. E., Bryan, G. W. an d Pascoe, P. L., 1991. `TBT-induced im posex in the dogwhelk, Nucella lapillus. G eographical uniformity o f the response and effects', Marine Environmental Research Vol. 32, pp. 79 87. Ishizaka, T., N em oto, S., Sasaki, K., Suzuki, T. and Saito, Y, 1989. `Simultaneous determ ination o f tri-n-butyltin, di-n-butyltin and triphenyltin com pounds in marine p rod u cts', Journal ofAgriculture and Food Chemistry Vol. 37, pp. 1523-1527. Iwata, H., Tanabe, S., M izuno, T. and Tatsukawa, R., 1995. `High accum ulation o f toxic butyltins in marine m am m als from Ja p a n e se coastal w aters', Environmental Science and Technology Vol. 29, pp. 2959-2962. H arding, M. J . C., Rodger, G. K., Davies, I. M. and M oore, J . J., 1997. `Partial recovery o f the dogwhelk (Nucella lapillus) in Sullom Voe, Shetland from tributyltin contam ination', Marine Environmental Research Vol. 44, N o 3, pp. 285-304. H arino, H., Fukushim a, M., Yamamoto, Y, Kawai, S. and Miyazaki, N., 1998a. `O rganotin com pounds in water, sedim ent and biological sam ples from the Port o f Osaka, Ja p a n ', Archives ofEnvironmental Contamination and Toxicology Vol. 35, pp. 558-564. H arino, H., Fukushim a, M., Yam am oto, Y., Kawai, S. an d Miyazaki, N., 1998b. `Contam ination o f butyltin and phenyltin com pounds in the marine environment of O tsuchi Bay, Ja p a n ', Environmental Pollution Vol. 101, pp. 209-214. H arino, H., Fukushim a, M., Kawai, S. and M egum i, K., 1998c. `M easurem ent o f butyltin contamination o f water and sedim ent in O saka Bay, J a p a n ', Applied Organometallic Chemistry Vol. 12, pp. 819-825. H arino, H., Fukushim a, M. and Kawai, S., 1999. `Temporal trends o f organotin com pounds in the aquatic environment o f the Port o f Osaka, Ja p a n ', Environmental Pollution Vol. 105, pp. 1-7. H ashim oto, S., W atanabe, M., N oda, Y., Hayashi, T., Kurita, Y , Takasu, Y an d Otsuki, A., 1998. `Concentration and distribution o f butyltin com pounds in a heavy tanker route in the Strait o f M alacca and in Tokyo Bay', Kannan, K., Tanabe, S., Iwata, H. and Tatsukawa, R., 1995a. `Butyltins in muscle and liver o f fish collected from certain Asian and O ceanian countries', Environmental Pollution Vol. 90, No 3, pp. 279-290. Kannan, K., Tanabe, S., Tatsukawa, R. and Williams, R. J., 1995b. `Butyltin residues in fish from Australia, Papua New Guinea and the Solom on Islands', InternationalJournal of Environmental Analytical Chemistry Vol. 61, pp. 263-273. Kannan, K., Yasunga, Y, Iwata, H., Ichhashi, H., Tanabe, S. an d Tatsukawa, R., 1995c. `Concentrations o f heavy metals, organochlorines and organotins in h orsesh oe crab, Tachpleus tridentatus, from Ja p a n e se coastal w aters', Archives of Environmental Contamination and Toxicology Vol. 28, pp. 40-47. Kannan, K., Corsolini, S., Focardi, S., Tanabe, S. and Tatsukawa, R., 1996. `A ccum ulation pattern o f butyltin com pounds in dolphin, tuna and shark collected from Italian coastal w aters', Archives ofEnvironmental Contamination and Toxicology Vol. 31, pp. 19 23. Kannan, K., Senthilkumar, K., Loganathan, B. G., Takahashi, S., Odell, D. K. and Tanabe, S., 1997. `Elevated accum ulation o f tributyltin and its breakdown products in bottlenose dolphins ( Tursiops truncatus) found stranded along the U.S. Atlantic and G u lf coasts', Environmental Science and Technology Vol. 31, pp. 296-301. 146 Late lessons from early warnings: the precautionary principle 1896-2000 Kannan, K., Guruge, K. S., Thom as, N. J., Tanabe, S. an d Giesy, J . P., 1998. `Butyltin residues in southern sea otters (Enhydra lutris nereis) foun d dead along California coastal w aters', Environmental Science and Technology Vol. 32, pp. 1169-1175. King, N., Miller, M. and de M ora, S., 1989. `Tributyltin levels for sea water, sedim ent and selected marine species in coastal Northland an d Auckland, New Z ealan d', New Zealand Journ al ofMarine and Freshwater Research Vol. 23, pp. 287-294. K rone, C. A., Burrows, D. G., Brown, D. W., Chan, S.-L. an d Varanasi, U., 1989. `Tributyltin contamination o f sedim ent and English sole from Puget S o u n d ', Proceedings of Oceans '89, Vol. 2: Ocean Pollution, IEEE Publication No 89CH2780-5, pp. 545-549. Labare, M. L., Coon, S. L., Mathias, C. and Weiner, R. M., 1997. `M agnification o f tributyltin toxicity to oyster larvae by bioconcentration in biofilm s o f Shewanella colwelliana', Applied and Environmental Microbiology Vol. 63, N o 10, pp. 4107-4110. phocoena) and grey seals (Hlichoerus gryfrns) from the coastal waters of England and W ales', Marine Pollution Bulletin Vol. 36, N o 3, p p .241-247. Law, R. J., Blake, S. J . an d Spurrier, C. J . H., 1999. `Butyltin com pounds in liver tissues of pelagic cetaceans stranded on the coasts of En glan d an d W ales', Marine Pollution Bulletin Vol. 38, No 12, pp. 1258-1261. Lee, R. F., Valkirs, A. O. an d Seligm an, P. F., 1989. `Im portance o f m icroalgae in the biodegradation o f tributyltin in estuarine w aters', Environmental Science and Technology Vol. 23, No 12, pp. 1515-1518. M atthiessen P. an d Gibbs, P. E., 1998. `Critical appraisal o f the evidence for tributyltinm ediated endocrine disruption in m ollusks', Environmental Toxicology and Chemistry Vol. 17, N o 1, pp. 37-43. MEPC, 1997. `R eport o f the 40th session o f the Marine Environmental Protection Committee o f the International Maritime Organization, 18-25 Septem ber 1997.' Langston, W. J., 1996. `R ecent developm ents in T B T ecotoxicology', Toxicology and Ecotoxicology News Vol. 3, N o 6, pp. 179-187. Langston, W. J . an d Burt, G. R., 1991. `Bioavailability and effects of sedimentbound TBT in deposit-feeding clams, Scrobicularia p lan a', M arine Environmental Research Vol. 32, pp. 61-77. MEPC, 1998. `R eport o f the 41st session o f the Marine Environmental Protection Committee o f the International Maritime Organization, 30 M arch-3 April 1998.' Michel, P. an d Averty, B., 1999a. `Contam ination o f French coastal waters by organotin com pounds: 1997 update', Marine Pollution Bulletin Vol. 38, No 4, pp. 268-275. Langston, W. J., Burt, G. R. an d M ingjiang, Z., 1987. `Tin an d organotin in water, sediments and benthic organisms of Poole H arb o u r', Marine Pollution Bulletin Vol. 18, No 12, p p . 634-639. Michel, P. an d Averty, B., 1999b. `Distribution and fate o f tributyltin in surface and deep waters o f the northwestern M editerranean', Environmental Science and Technology Vol. 33, pp. 2524-2528. Langston, W. J., Bryan, G. W., Burt, G. R. and Pope, N. D., 1994. Effects of sediment metals on estuarine benthic organisms, RandD N ote 203, N ational Rivers Authority, 141 pages. Laughlin, R. B. and Linden, O., 1985. `Fate an d effects o f organotin co m p o u n d s', Ambio Vol. 14, N o 2, pp. 88-94. Laughlin, R. B. and Linden, O., 1987. `Tributyltin -- contem porary environmental issues', Ambio Vol. 16, N o 5, pp. 252-256. Minchin, D., 1995. `Recovery o f a population o f the flam e shell, Lim a hians, in an Irish Bay previously contam inated with T B T ', Environmental Pollution Vol. 90, No 2, pp. 259 262. Minchin, D., Bauer, B., Oehlm ann, J., Schulte-Oehlmann, U. and Duggan, C. B., 1997. `Biological indicators used to m ap organotin contamination from a fishing port, Killybegs, Irelan d ', Marine Pollution Bulletin Vol. 34, No 4, pp. 235-243. Law, R. J., Blake, S. J., Jo n e s, B. R. an d R ogan, E., 1998. `Organotin com pounds in liver tissue o f h arbour porpoises (Phocoena MINDEC, 1995. Ministerial Declaration of the Fourth International Conference on the Tributyltin (TBT) antifoulants: a tale of ships, snails and imposex 147 Protection o f the North Sea, 8-9 Ju n e 1995, Esbjerg, Denmark. Environmental Pollution Vol. 93, No 2, pp. 195 203. M oore, D. W., Dillon, T. M. an d Suedel, B. C., 1991. `Chronic toxicity o f tributyltin to the m arine polychaete worm, Neanthes arenaceodentata', Aquatic Toxicology Vol. 21, pp. 181-198. Morcillo, Y and Porte, C., 2000. `Evidence o f endocrine disruption in clams -- Ruditapes decussata -- transplanted to a tributyltinpolluted environm ent', Environmental Pollution Vol. 107, pp. 47-52. Nicholson, G. J. and Evans, S. M., 1997. ` Anthropogenic impacts on the stocks o f the com m on whelk Buccinum undatum (L .)', Marine Environmental Research Vol. 44, N o 3, p p .305-314. OSPAR, 1998. OSPAR strategy with regard to hazardous substances, OSPAR 9 8 /1 4 /1 Annex 34, OSPAR Convention for the Protection o f the Marine Environm ent o f the Northeast Atlantic. PARCOM, 1987. PARCOM Recommendation 87/1 on the use o f tributyl-tin com pounds, 3 Ju n e 1987, Paris Convention for the Prevention o f Marine Pollution from LandBased Sources. PARCOM, 1988. PARCOM Recommendation 88/1 on m easures to reduce organotin com pounds reaching the aquatic environm ent through docking activities, 17 Ju n e 1988, Paris Convention for the Prevention o f Marine Pollution from LandBased Sources. Prudente, M., Ichihashi, H., Kan-atireklap, S., W atanabe, I. an d Tanabe, S., 1999. `Butyltins, organochlorines and m etal levels in green mussel, Perna viridis L. from the coastal waters o f the Ph ilippin es', Fisheries Science Vol. 65, No 3, pp. 441-447. Ruiz, J. M., Quintela, M. an d Barreiro, R., 1998. `U biquitous im posex and organotin bioaccum ulation in gastropods Nucella lapillus from Galicia (NW Sp ain ): A possible effect o f nearshore sh ippin g', Marine Ecology Progress Series Vol. 164, pp. 237-244. Santillo, D., Stringer, R., Jo h n sto n , P. and Tickner, J., 1998. `The precautionary principle: Protecting against failures of scientific m ethod and risk assessm ent', Marine Pollution Bulletin Vol. 36, No 12, pp. 939-950. Sarradin, P. M., Astruc, A., Sabrier, R. and Astruc, M., 1994. `Survey o f butyltin com pounds in Arcachon Bay sedim ents', Marine Pollution Bulletin Vol. 28, pp. 621-628. Scarlett, A., Donkin, P., Filem an, T. W. an d M. E. Donkin (1997. `Occurrence of the marine antifouling agent Irgarol 1051 within the Plymouth Sound locality: Implications for the green m acroalga Enteromorpha intestinalis', Marine Pollution Bulletin Vol. 34, N o 8, pp. 645-651. Scarlett, A., Donkin, P., Filem an, T. W., Evans, S. V. and Donkin, M. E., 1999. `Risk p osed by the antifouling agent Irgarol 1051 to the seagrass, Zostera marina', Aquatic Toxicology Vol. 45, pp. 159-170. Side, J., 1987. `Organotins -- not so good relation s', Marine Pollution Bulletin Vol. 18, No 5, pp. 205-206. Sim m onds, M., 1986. `The case against tributyltin', Oryx Vol. 20, No 4, pp. 217-220. Smith, B. S., 1971. `Sexuality o f the Am erican m ud snail, Nassarius obsoletus (Say )', Proceedings of the Malacological Society ofLondon Vol. 39, pp. 377-378. Rees, H. L., W aldock, R., M atthiessen, P. and Pendle, M. A., 1999. `Surveys o f the epibenthos o f the Crouch Estuary (UK) in relation to T B T con tam in ation ',Journal of the Marine Biological Association, UK Vol. 79, pp. 209-223. St-Jean, S. D., Courtenay, S. C., Pelletier, E. and St-Louis, R., 1999. `Butyltin concentrations in sediments and blue m ussels (Mytlus edulis) o f the southern G u lf o f St Lawrence, C an ada', Environmental Technology Vol. 20, pp. 181-189. Ruiz, J . M., Bachelet, G., Caum ette, P. and D onard, O. F. X ., 1996. `T hree decades o f tributyltin in the coastal environm ent with em phasis on A rcachon Bay, F ran ce', Swennen, C., Ruttanadakul, N., A rdseungnern, S., Singh, H. R., Mensick, B. P. and ten H allers-Tjabbes, C. C., 1997. `Im posex in sublittoral and littoral gastropods from the gulf o f Thailand and 148 Late lessons from early warnings: the precautionary principle 1896-2000 Strait o f M alacca in relation to shipping', Environmental Technology Vol. 18, pp. 1245 1254. Tanabe, S., Prudente, M., M izuno, T., Hasegawa, J., Iwata, H. and Miyazaki, N., 1998. Butyltin contamination in marine mammals from north Pacific and Asian coastal w aters', Environmental Science and Technology Vol. 32, No 2, pp. 193-198. `TBT linked to dogwhelk decline', 1986. Marine Pollution Bulletin Vol. 17, No 9, p. 390. ten Hallers-Tjabbes, C. C., 1997. `Tributyltin and policies for antifouling', Environmental Technology Vol. 18, pp. 1265-1268. Wade, T. L., Garcia-Rom ero, B. an d Brooks, J. M., 1988a. `Tributyltin contam ination in bivalves from United States Coastal E stuaries', Environmental Science and Technology Vol. 22, pp. 1488-1493. Wade, T. L., Garcia-Rom ero, B. an d Brooks, J. M., 1988b. `Tributyltin analyses in association with NOAA's N ational Status an d Trends M ussel Watch P rogram ', Proceedings of the Oceans '88 Conference, Baltimore, Maryland, October 31-November 2 1988, pp. 1198-1201. Waldock, M. J. and Thain, J. E., 1983. `Shell thickening in Crassostrea gigas: O rganotin antifouling or sedim ent induced?', Marine Pollution Bulletin Vol. 14, pp. 411-415. ten Hallers-Tjabbes, C. C., Kem p, J . F. and Boon , J . P., 1994. `Im posex in whelks (Buccinum undatum) from the open North Sea: Relation to shipping traffic intensities', Marine Pollution Bulletin Vol. 28, No 5, pp. 311-313. Uhler, A. D., C oogan, T. H., Davis, K. S., Durell, G. S., Steinhauer, W. G., Freitas, S. Y and Boehm , P. D., 1989. `Findings o f tributyltin, dibutyltin and monobutyltin in bivalves from selected U S coastal waters', Environmental Toxicology and Chemistry Vol. 8, p p .971-979. Vos, J . G., Dybing, E., Greim, H. A., L ad efoged , O., Lam bre, C., Tarazona, J . V., Brandt, I. an d Vethaak, A. D., 2000. `H ealth effects of endocrine-disrupting chemicals on wildlife, with special reference to the E u ro p ean situation', Critical Reviews in Toxicology Vol. 30, N o 1, pp. 71-133. Waldock, M. J., Thain, J. E. and Waite, M. E., 1987. `The distribution and potential toxic effects of TBT in UK estuaries during 1986', Applied Organometallic Chemistry Vol. 1, pp. 287-301. Waldock, M. J., Waite, M. E. and Thain, J . E., 1988. `Inputs o f T B T to the m arine environm ent from shipping activity in the U K ', Environmental Technology Letters Vol. 9, p p .999-1010. Waldock, M. J., Thain, J. E., Smith, D. and Milton, S., 1990. `The degradation o f TBT in estuarine sedim en ts', Proceedings of the third International Organotin Symposium, Monaco, 17-20 April 1990, International Atomic Energy Authority, pp. 46-48. Zabel, T. F., Seager, J . an d Oakley, S. D., 1988. Proposed environmental quality standards for List II substances in water: Organotins, Water Research Centre technical report TR255, Marlow, UK. Hormones as growth promoters: the precautionary principle or a political risk assessment? 149 14. Hormones as growth promoters: the precautionary principle or a political risk assessment? Jim W. Bridges and Olga Bridges 14.1. Introduction Oestrogenic steroid horm ones (`oestrogens') have a crucial role in cellular regulation in all vertebrate species. The levels required to bring abo u t such changes are very low, around 0.1-1 p g /m l (picogram s per millilitre) o f serum. It has been known for over five decades that oestrogens affect the development of the male reproductive system o f mammals (Zuckerman, 1940). Nonetheless, oestrogenic steroid hormones are often called `fem ale horm ones'. The fem ale reproductive system initially develops independently of the horm one regulatory system. This m eans that by default an animal is fem ale if horm one stimulation is absent (Wilson and Lasnitzki, 1971). Nonetheless, in both sexes oestrogens are n eed ed for fertility. At a level above the physiological one, natural oestrogens produced by mammals, such as oestradiol-17p, have a lasting effect on males. Experim ental studies have shown, for exam ple, that the administration o f oestradiol-17p to mice, rats, guinea pigs and rabbits, during both foetal and perinatal life, can result in significant defects in the pituitary-hypothalamic function in males. This, in turn, may disrupt testicular function during adulthood (Takasugi, 1979; OrgebinCrist et al., 1983; Davies an d Danzo, 1981; Brown-Grant et al., 1975). For maximum growth a combination o f both oestrogens and androgens (male hormones) is required. The growth prom oting effects are attributed principally to the ability of combinations o f oestrogens and androgens to increase the retention o f dietary nitrogen through protein synthesis through several mechanisms (European Commission, 1996). After the Secon d World War, the recognition o f the growth prom oting properties o f oestrogens, either alone or in combination with an drogen s, led to their introduction as a tool to increase m eat production. Diethylstilboestrol (DES), as a cheap, better absorbed analogue o f the natural horm one oestradiol-17p, becam e the favoured growth prom oter for cattle, sheep and poultry in many countries (Schmidely, 1993). As with many steroid growth prom oters, DES was adm inistered as an im plant under the skin o f young animals or as a feed additive. In the early 1970s concerns abo u t its safety were raised when DES was confirm ed to be a human carcinogen. In the scientific community, however, consensus was reached that the health risk was insignificant. The DES residues in m eat were very low (below the limit o f analytical detection) com pared to those which individuals were exposed to when DES was used as a drug. Use o f DES as a growth prom oter continued in some M ember States of the European Union (EU) longer than in the U nited States. It was finally banned for this use throughout the EU in 1987 because of uncertainty as to whether there was a definable `no effect' dose for its potential tum our inducing effects in humans (European Commission, 1996), although some Member States had introduced an earlier ban. In the U nited States a totally different pattern of events took place. DES was banned initially as a growth prom oter in 1972 on the grounds that it was a carcinogen, violating the so-called 1958 Delaney Clause. This clause prohibits food for human consumption that contains carcinogenic substances, but was very difficult to apply in practice because the majority of foods contain trace amounts o f carcinogenic substances. However, public opinion ensured that the clause rem ained on the statute books. The regulators had been forced to refer to `minim um detectable levels' in justifying their inaction regarding the Delaney Clause in the case o f DES. The US Food and Drug Administration (FDA) estimated that the economic burden to the consumers from a DES ban would be approximately USD 500 million per annum. The calculation o f this estimate involved a num ber of questionable assumptions. The estim ated health risk was 1 cancer in 133 years (Jakes, 1976). In 1974 the use o f DES as a growth prom oter was tem porarily reinstated because o f 150 Late lessons from early warnings: the precautionary principle 1896-2000 procedural deficiencies in the original bill banning DES in the United States. The farming lobby m ade strong claims regarding the serious economic consequences of a further ban. These claims were m ade despite the fact that there were alternative growth prom oters already on the U S market (see Table 14.1.). The `breathing sp ace' enabled the pharm aceutical industry to develop additional horm onal growth prom oters. At the sam e time a scientific debate was taking place on what residue level o f DES could be regarded as without significant risk to hum an health. In 1976 the FDA set the minim um detectable level o f DES (the regulatory level) at 2 ppb (parts per billion). The FDA estim ated that levels o f DES in m eat were o f the order o f 0.5 ppb (McMartin, 1978), but was n ot able to verify this estimate by measurements. The safety of oestrogens in the oral contraceptive pill and the high levels o f natural oestrogens in pregnant women were cited as crucial evidence o f low residues o f DES in food being without risk to the consumer. This argum ent did not take into accoun t that young children with low natural levels o f oestrogens were the likely `at risk' group (McMartin, 1978). The FDA also omitted the fact that DES has many structural differences from both oestradiol and the oestrogenic com ponents o f the oral contraceptive pill. In 1979 D ES was finally banned because there were no toxicological grounds for identifying a residue level below which a carcinogenic effect would not occur (Jakes, 1976). The concerns o f the high cost to the consum er from a DES ban were probably groundless. When the ban was eventually im plem ented there was little evidence o f a sustained increase in the costs of m eat production. In the case o f the U nited States the lack o f increase in costs could be explained by the availability o f alternative growth prom oters and, in part, by the wrong assumptions m ade in the preliminary calculation o f costs. It is worth noting that the FDA has continued to support the use of other oestrogenic com pounds as growth prom oters in cattle, (including oestradiol, trenbelone and zeranol) because o f the perceived im portance for the economic efficiency of m eat production. In 1982, an EU expert working group (the Lamming Committee) comprising members o f the Scientific Committee for Food and the Scientific Committee for Animal Nutrition (the lead committee for issues concerning growth prom oters), reached an interim conclusion that oestradiol and several other growth prom oters with a horm on al action were safe as a growth prom oting agent in cattle. This decision was clearly unpopular with the EU officials. Further work was carried out by the Lam m ing Committee over the n ext few years, bu t this did n ot change its opinion. The com m ittee was disbanded in 1987 and the EU did not publish its interim conclusions. However, m em bers o f the committee published their opinions independently of the EU in the scientific literature (Lam m in g et al., 1987). In 1988, a risk assessment by the Jo in t Expert Committee on Food (JECFA) o f the World Health Organization (WHO) and the Food and Agriculture Organization reached a similar conclusion to that o f the Lam m ing Committee. JECFA used the standard approach for risk assessments still employed by scientific advisory comm ittees today. JECFA only considered the following (JECFA/W HO, 1988): risk when the growth prom oters are used according to authorised use (though at the time and since there have been indications of significant accidental or deliberate m isuse, which m ight be expected to lead to higher meat residues) ; individual growth prom oters (rather than their combinations); data provided by the manufacturers. Shortly after the publication o fJE C FA 's conclusions the EU issued a ban not only on the use of oestradiol but also on the use of other natural and synthetic steroid horm ones as growth prom oters. This ban was first adopted in 1985 but was disputed before the European Court by the United Kingdom. It was annulled because o f procedural deficiencies. The ban was finally agreed in 1988. It encom passed a ban on the use of oestradiol 17-0, testosterone, progesterone, zeranol, trenbelone acetate and melengestrol acetate within M em ber States. In 1989 it was extended to imports from third world countries except where such growth prom oters were already banned or countries were operating hormone-free cattle export programmes. This action could be regarded as an application of the precautionary principle, although the principle was not form alised at the time. Hormones as growth promoters: the precautionary principle or a political risk assessment? 151 It is im portant to analyse the reasons why the European Commission decided to overrule the views o f both com m ittees. T hree factors appear to have had a particular influence on the C om m ission 's decision regard in g D E S 's use as a growth promoter: firstly, the scientific evidence that DES, which had been extensively used as a growth prom oter, caused vaginal clear-cell adenom a in young women (H erbst and Bern, 1988); secondly, increasing public concern about the health risks from horm ones generally. The first generation oral contraceptive pill was blam ed for increased incidence of breast cancer and thrombosis; thirdly, several epidem iological studies published at the time claimed that oestrogenic environmental contamination could result in anom alies in growth, sexual development and puberty. In Puerto Rico over 10 000 cases o f anom alous sexual development, including premature development of breasts and body hair and pseudo precocious puberty, were reported (Perez-Comas, 1988). These changes were associated by the authors with high serum total oestrogens. However, the source o f the oestrogen contamination was not clearly identified. Sim ilar adverse effects were observed in Italy, believed to result from accidental contamination of food by D ES (Fara et al., 1979). In addition, in 1980, analysis o f Italian baby food m ade with hom ogenised veal showed the presence of DES at significant levels. This was claimed to result from implants that were not removed after slaughtering o f animals. It is evident that the hum an risk associated with the use o f oestradiol and oestradiolrelated com poun ds is governed by many interrelated factors: the nature o f the growth prom oter(s) used, the site and dose adm inistered to cattle and the time period which has elapsed between its adm inistration and the slaughter o f the an im al; the am ount of m eat and m eat products derived from slaughtered animals treated with the growth prom oters which is consum ed by an individual over an extended period o f time; indirect contact with substances with oestrogenic properties through environmental contamination or other forms of exposure; the susceptibility o f the individual con sum er. In addition, there exists a potential for accidental or deliberate misuse o f oestradiol and other hormonally related substances in the cattle industry. This misuse may take several forms: use of a dose higher than considered acceptable; a complex mixture of oestrogenic steroids; an inappropriate injection site; failure to remove an injection site or an im plant (likely to have m uch higher horm one levels than elsewhere) from a slaughtered animal; a shortened withdrawal period; use of illegal substances. Illegal use o f growth prom oting substances in M em ber States was the subject o f the 1989 Pimenta Report. The report found no evidence in the use o f oestradiol-l7p, but nonetheless endorsed the ban because it facilitated controls and consumer confidence in meat. There have been claims that the EU ban on steroid hormones has led to illegal use, not only o f the `safer' steroids but also the m ore toxic ones, such as D ES (Loizzo et al., 1984). In other words, the ban may have led to an increased risk to the consumer, rather than a reduction in risk. In the absence of a substantial regular monitoring program m e, the extent o f such misuse and the con sequ en t increased risk, if any, to the consum er is hard to gauge. It is noted, however, that D ES was detected last year in U S m eat im ported into Switzerland. In most reported cases of accidental contamination of food by oestrogenic substances, the anom alous effects have been considered to be transient and reversible. However, the long-term effects o f exposure of prepubescent children to oestrogenic substances are as yet unknown. The absence of a demonstrable threshold concentration, below which there is no effect, ad d s to this uncertainty (European Commission, 1996). 14.2. Impacts of oestrogenic compounds on wildlife This issue was not considered by either the Lam m ing Com mittee or JECFA. Two sources o f information have em erged on possible impacts o f oestrogenic com pounds on the environment, namely: 152 Late lessons from early warnings: the precautionary principle 1896-2000 effects o f natural and synthetic oestrogens themselves on endocrine function in wildlife; effects o f non-steroid chemicals which im pact wildlife through endocrine disruption. No comprehensive studies were conducted until the late 1980s. However, there was sufficient information generated in the 1970s to raise concern about the environmental im pacts o f com pounds with oestrogenic properties. For exam ple, it was reported in 1970 that dichlorodiphenyl trichloroethane (DDT) decreased oestradiol blood levels and deposition of m edullar bone (Scientific Committee on Veterinary M easures Relating to Public Health, 1999). It was suggested that the oestrogenic property o f DDT was responsible for the observed effect. Growth reduction was observed in catfish exposed to DES as long ago as 1972, indicating that D ES was likely to affect a wide range o f species in the environment (Peakall, 1970).Jo h n sto n e et al. (1978) subsequently described a significant suppression o f both length and weight o f rainbow trout following dietary administration of oestradiol-170 (Bulkey, 1972). A num ber o f chemicals widely present in the environment (DDT, polychlorinated biphenyls (PCBs) and alkylphenols) disrupt oestrogen receptor function in wildlife (Johnstone et al., 1978; M ueller an d Kim, 1978; Reijnders, 1986; Bergm an and Olsson, 1985; A ulerich et al., 1985). Fem inisation o f male birds (having both ovarian and testicular tissue) was induced by ethinyloestradiol-17a (D elong et al., 1973). The studies cited above were ignored until the late 1980s, when concerns regarding the possible im pact on the environment o f veterinary drugs and growth prom oters were raised. This was discussed in the 1989 Collins Report, but no in-depth evaluation resulted. This may be ascribed to the lack o f interest of drug agencies in the environment and to the widely held assum ption that any excreted drugs or growth prom oters would be in a very diluted form and quickly degrade in the environment. The extent o f the environmental im pact o f the use of oestrogenic growth prom oters remains to be established. The recent international workshop on horm ones and endocrine disrupters in food and water (in Copenhagen in 2000) confirm ed concerns about this issue. 14.3. What were the uncertainties regarding the use of oestrogenic growth promoters for human health? First, oestrogenic growth prom oters were introduced to improve efficiency in cattle production. Hum an health issues, possible environmental impacts and farm animal welfare were not given significant attention. Histological changes in the prostate and bortholinic gland have been detected in cattle following oestradiol-17p administration. However, the physiological significance o f these changes remains unclear. Second, residues o f the com pounds in m eat were subsequently m easured and shown to be low (in the case o f oestradiol-17p within the `physiological range'), although what constitutes the physiological range continues to be controversial. The other limitation in these studies was the suitability o f the analytical m easurem ents. O ne aspect o f this is the failure to assay m etabolites with poten tial oestrogenic activity, for exam ple oestradiol esters. The EU has continued to encourage scientific evaluation o f the safety o f these growth prom oters. In 1995 in Brussels the European Commission organised an international conference on growth prom oters and m eat production, involving the m ajor stakeholder groups. No definitive conclusions were drawn. The expert opinions o f the Scientific Committee on Veterinary M easures Relating to Public Health have been published since. The conclusion of both opinions is that further attention needs to be given to the exposure o f sensitive populations to these substances because of the possible effects on the im m une system, the endocrine system an d cancer. T he EU is also sponsoring a number o f ongoing research projects in this area. JECFA continues to hold to its original view that each o f the growth horm ones with steroid like action is safe for the consumer. Prepubertal boys are identified as being at risk. The reliability o f m easurem ents o f their endogenous levels and production o f oestrogens is very questionable since these levels were close to the limits o f detection. It means that any additional exogenous oestrogen represents a relatively high percentage o f the total body oestrogen. This is particularly relevant in the context o f the Hormones as growth promoters: the precautionary principle or a political risk assessment? 153 safety criterion applied by the FDA, that intake of any horm one from food should constitute less than 1 % o f the individual's daily endogenous production. Neither possible hum an exposure via release o f oestrogenically active excreta from cattle into the environment, nor possible effects on wildlife, were properly taken into account. Public concern in Europe was initiated by the association o f D ES with vaginal clear-cell adenom a and increasing worries about the carcinogenic effects o f som e horm ones. It is pertinent in this regard that recent research has demonstrated that oestradiol-l7p (although a natural horm one), is a genotoxic carcinogen. This finding has further fuelled the debate on what is a safe level. 14.4. Has the approach adopted by the European Commission proved to be sound? There is a growing consensus that cancer of the breast in women and prostate cancer in men may be prom oted by high oestrogen exposure. North Am erica has one o f the highest breast cancer rates in the world, while Asia and Africa have m uch lower rates. The pattern is sim ilar for prostate cancer. It has been noted that women with high levels of oestrogen, particularly free oestrogen, have a higher risk o f breast cancer developm ent (Adkins, 1975). Studies on m igrant populations indicate this risk is principally due to environmental rather than genetic factors. T h ere is, so far, no g o o d evidence that the EU ban on the use o f growth prom oters has protected the health o f the public. However, it can also be argu ed that in those countries where these growth prom oters have continued to be used, reliable evidence has not been accum ulated on their safety to the consumer. Although potential risks to the environment have been identified, no conclusive evidence that growth prom oters constitute a significant environmental risk has been produced. The case for DES use as a growth prom oter was a purely econom ic one. There was no benefit either for the cattle (or other species to which DES was adm inistered) or for the environment. Agricultural specialists continue to debate the extent o f the benefits of anabolic steroids as growth promoters. The econom ic consequences o f this parting o f the ways between the U nited States and the EU with regard to anabolic steroids as growth prom oters are very considerable and have yet to be fully evaluated. The divergence of opinions has led to a trade dispute between the EU and the United States (H en derson et al., 1988). From 1989, the United States took unilaterally retaliatory m easures in the form o f a 100 % ad valorem duty on a variety o f EU exports at a value of USD 93 million (ca. EUR 93 million) per annum. In 1996 these m easures were withdrawn as n ot com patible with World Trade O rganization (W TO) law. In 1997 the legal basis for the EU ban was challenged by the United States and Canada in front of a WTO panel. The panel found largely against the EU. On appeal, several of the grounds were reversed. The only aspect where the p a n e l's view was u ph eld was that the EU had in its ban not focused specifically enough on residue in meat o f each growth hormone. The dispute raises the issue of where the benefit o f doubt should lie. An issue o f concern is that attem pts to resolve the problem are being m ade in the absence of any formal mechanism s for trading risks and benefits for the public. Potential environmental impacts or anim al welfare issues have not been considered in this dispute. The EU is currently experiencing sanctions against its exports o f the order o f GBP 100 million (ca. EUR 160 million) per annum. The success o f the United States in the WTO hearings on anabolic steroid growth prom oters has encouraged further actions on other products where the EU has adopted a precautionary approach on health grounds. The question is whether, if the EU applies the precautionary principle on health grounds to any p ro d u ct(s), it w ould prove acceptable to the WTO. Som e issues have, however, been clarified by the WTO appeal, namely a sanitary protection m easure (including a ban) can be perm itted if supported by a risk assessm ent, even if this: is not necessarily quantitative in nature; takes into account real world issues such as difficulties in control measures; is based on `qualified and respected sources'(W T/DS, 1997), even if these are in the minority. 154 Late lessons from early warnings: the precautionary principle 1896-2000 Table 14.1. Source: Henderson et al., 1988 US FDA approval dates for anabolic agents Trade name Synovex-S Synovex-H Ralgo MGA Anabolic agent 200 mg progesterone/ 20 mg oestradiol benzoate 200 mg testosterone propionate/ 20 mg oestradiol benzoate 36 mg zeranol O r 12 mg zeranol 0.25-0.50 mg/day M GA Approved date 20/2/1956 16/7/1958 5/11/1969 3/6/1977 14.5. Overall conclusions The EU took action in 1985 and again in 1988 to ban growth prom oters with a steroid like action, principally in response to public concern about involuntary and unnatural horm one exposure. This action was not at the time sup p orted by the E U 's own scientific committee (the Lam m ing Committee) nor that o f W HO (JECFA). It is appropriate to apply the precautionary principle in those situations where the science base is inadequate to confirm the safety o f a product. Neither comm ittee was required to properly characterise the degree of uncertainty in their assessment. As this question was never asked, the original EU ban was, in reality, a political risk assessm ent. M ore recent scientific research, however, probably justifies the application by the EU o f the precautionary principle to continue the ban. The handling of the issue of steroid horm ones as growth prom oters has wider implications in determining acceptability of the use o f chemical substances, namely that: it is very im portant, in arriving at their conclusions, that scientific committees are requested to identify the uncertainties in their assessments; rigorous and transparent mechanisms must be developed for evaluating risks against benefits. Table 14.2. Hormones as growth promoters: early warnings and actions 1970s 1972 1972 1974 1976 1979 1982 1985 1987 1988 1988 1989 1989 1989-96 1995 1999 2000 2001 Concerns about growth promoters' safety, as DES confirmed a human carcinogen Peakal publishes that D ES likely to affect a w ide range of sp ecies in the environm ent (wildlife) but this w as ignored until the late 1980s D ES banned as a hormone growth prom oter in the United States Use of D ES reinstated in the United States US Food and Drug Adm inistration (FDA) sets the minimum detectable level of DES DES banned again on the grounds of the impossibility of identifying levels below which it would not be carcinogenic EU expert working group (Lamming Com m ittee) concludes that some growth prom oters are safe First EU ban is ad opted, ignoring results from the Lamming C om m ittee because the scope of their assessm ents had not been broad enough Lamming Com m ittee disbanded by EU and their results were not published Ban of several growth prom oters throughout the EU based on uncertainty of their effects on humans W H O /FA O Joint Expert Com m ittee on Food, using standard risk assessm ents, reaches same conclusions as Lamming Comm ittee EU ban extended to other growth prom oters and to imports from third world countries Pim enta R eport finds illegal use of growth prom oters in som e M em ber States USA takes unilateral retaliatory measures on EC exports European Com m ission organises an international conference on growth prom oters and meat production where uncertainties remain regarding effects on the immune system , endocrine system and cancer The EU Scientific Com m ittee on Veterinary M easures Relating to Public Health publishes a report concluding that no threshold levels can be defined for six growth prom oters International workshop on horm ones and endocrine disrupters in food and w ater confirm s im pacts on the environm ent (wildlife) of veterinary drugs EU still suffers from sanction to its exports of around EUR 160 million per year Hormones as growth promoters: the precautionary principle or a political risk assessment? 155 14.6. References Adkins, E. K., 1975. `H orm on al basis o f sexual differentiation in the Japanese quail', Journal of Comparative and Physiological Psychology Vol. 89, pp. 61-71. Aulerich, R. J., Bursian, S. J., Breslin, W. J., Olson, B. A. and Ringer, R. K., 1985. `Toxicological manifestations o f 2,4,5,2,4,5 2,3,6,2,3,6- and 3,4,5,4,5,-hexachlorobiphenyl and Aroclor 1254 in m ink',Journal of Toxicology and Environmental Health Vol. 15, pp. 63-79. Bergm an, A. and Olsson, M., 1985. `Pathology of Baltic ringed seal and grey seal fem ales with special reference to adrenocortical hyperplasia: Is environmental pollution the cause o f a widely distributed disease syndrom e?', Finn. Game Research Vol. 44, pp. 47-62. Brown-Grant, K., Fink, G., G reig F. and Murray, M. A., 1975. `A ltered sexual development in male rats after oestrogen administration during the neonatal period', Journal ofReproduction and Fertility Vol. 44, pp. 25-42. Bulkey, R. V., 1972. `Diethylstilbestrol in catfish fe e d ', Trans American Fisheries Society Vol. 101, pp. 537-599. Davies, J. and Danzo, B. J., 1981. `Horm onally responsive areas o f the reproductive system o f the male guinea pig II: Effects of estrogen s', Biological Reproduction Vol. 25, pp. 1149-1158. D elong, R. L., Gilm artin, W. G. an d Sim pson, J. G., 1973. `Prem ature births in California sea lions: Association with high organochlorine pollutant residue levels', Science Vol. 181, pp. 1168-1170. the Richard and H inda Rosenthal Foundation Award lecture', Cancer Research Vol. 48, pp. 246-253. Herbst, A. L. and Bern, H. A., 1988. Developmental effects of diethylstilbestrol (DES) in pregnancy, Thiem e-Stratton, New York. Jak es, T. H., 1976. `Diethylstilboestrol in b e e f production: W hat is the risk to the con sum er?', Preventive Medicine Vol. 5, pp. 438-453. JE C FA /W H O , 1988. Technical Report Series N 7 / 63, Geneva. Jo h n sto n e, R., Sim pson, T. H. an d Youngson, A. F., 1978. `Sex reversal in salm onid cultures', Aquaculture Vol. 13, pp. 115-134. Lam m ing, G. E., Ballarini, G., Bauliieu, E. E., Brooks, P., Elias, P. S., Ferrando, R., Galli, C. L., Heitzman, R. J., Hoffm an, B., Karg, H., Meyer, H. H. D., Michel, G., Poulsen, E., Rico, A., van Leuw en F. X . R. an d White, D. S., 1987. `Scientific report on anabolic agents in anim al p rod u ction ', Veterinary Records Vol. 121, p p .389-392. Loizzo, A., Gatti, G. L., Macri, A., Moretti, G., Ortolani, E. and Palazzesi, S., 1984. `Italian baby food containing diethylstilboestrol three years later', Lancet Vol. 5, pp. 1014 1015. McMartin, K. E., 1978. `Diethylstilboestrol: A review o f its toxicity and use o f a growth prom otant in food producing anim als', Journal ofEnvironmental Pathological Toxicology Vol. 1, pp. 279-313. Mueller, G. G. and Kim, U. H., 1978. `D isplacem ent o f estradiol from oestrogen receptors by simple alkylphenols', Endocrinology Vol. 102, pp. 1429-1435. E u ropean Com m ission, 1996. Scientific Conference on Growth Promoters and Meat Production Proceedings, O ffice for O fficial Publications o f the EU Communities, Luxem bourg. Fara, G. M., Del Corvo, G., Bernuzzi, S., Bigatello, A., Di Pietro, C., Scaglioni, S. and Chium ello, G., 1979. `Epidem ic o f breast enlargem en t in an Italian sch ool', Lancet Vol. 11, pp. 295-297. Orgebin-Crist, M. C., Eller, B. C. and Danzo, B. J., 1983. `The effects of estradiol, tamoxifen and testosterone on the weights and histology o f the epididymis and accessory sex organs of sexually immature rabbits', Endocrinology Vol. 113, pp. 1703-1715. Peakall, D. B., 1970. `p ,p '-DDT: Effect on calcium m etabolism and concentration of estradiol in b lo o d ', Science Vol. 168, pp. 592 594. H enderson, B. E., Ross, R. and Bernstein, L., Perez-Comas A., 1988. `Prem ature sexual 1988. `Estrogens as a cause o f h um an cancer: developm ent in Puerto R ico', Boletin 156 Late lessons from early warnings: the precautionary principle 1896-2000 Asociacion Medica de Puerto Rico Vol. 80, pp. 85-90. R eijnders, P. J . H., 1986. `Reproductive failure in comm on seals feeding on fish from polluted coastal w aters', Nature Vol. 324, pp. 456-457. Schmidely, Ph., 1993. `Revue bibliographique quantitative sur l 'utilisation des horm ones anabolisantes action steroidienne chez les ruminnants en production de viande. Perform an ces zootech n iques', Annales de Zootechnie Vol. 42, pp. 333-360. Takasugi, N., 1979. `D evelopm ent o f permanently proliferated and cornified vaginal epithelium in mice treated neonatally with steroid horm ones and the im plication in tum origen esis', National Cancer Institute Research Monographs Vol. 51, pp. 57-66. Wilson, M. E. and Lasnitzki, I., 1971. `Dihydrotestosterone found in foetal tissues o f the rabbit an d ra t', Endocrinology Vol. 89, p p .659-668. World Trade Organization, 1997. W T /D S2 6 / R/U SA and W T/DS48/R/CAN. Scientific Committee on Veterinary M easures Relating to Public Health, 1999. Assessment o f potential risks to hum an health from horm one residues in bovine m eat and m eat products'. Zuckerman, S., 1940. `The histogenesis o f tissues sensitive to estrogen s', Biological Review Vol. 15, pp. 231-271. 'Mad cow disease' 1980s-2000: how reassurances undermined precaution 157 15. 'Mad cow disease' 1980s-2000: how reassurances undermined precaution Patrick van Zwanenberg and Erik Millstone 15.1. Introduction Many o f the U K policy-makers who were directly responsible for taking policy decisions on bovine spongiform encephalopathy (BSE) prior to March 1996 claim that, at the time, their approach exem plified the application of an ultra precautionary approach and of rigorous science-based policy-making. ( 11) We argue that these claims are not convincing because government policies were not genuinely precautionary and did not properly take into account the implications o f the available scientific evidence. The BSE saga is enorm ously com plex and this accoun t is necessarily selective. It is, however, essential to appreciate that U K public policy-making was handicapped by a fundamental tension. The departm ent responsible for dealing with BSE has been the Ministry o f Agriculture, Fisheries and Food (MAFF), and it was expected simultaneously to prom ote the economic interests o f farm ers and the food industry whilst also protecting public health from food-borne hazards. The evidence cited here suggests that because MAFF was expected sim ultaneously to m eet two contradictory objectives it failed to m eet either. 15.2. A new cattle disease The first cases o f BSE were officially acknowledged in November 1986. The pathological characteristics o f the new cattle disease closely resembled scrapie, a transmissible spongiform encephalopathy (TSE) that is endem ic in the U K sheep population. TSEs are a group of very poorly understood, untreatable and invariably fatal brain diseases that afflict both animals and hum ans. Creutzfeldt-Jakob disease (CJD ) is the best known hum an TSE. MAFF scientists suspected that BSE had been caught from sheep infected with scrapie and was being transmitted through contam inated feed. The rendered remains o f sheep, cattle and other animals were routinely incorporated into animal feedstuffs. Contam inated feed was quickly confirm ed as the principal vector o f the disease but whether BSE had in fact derived from scrapie, or from a spontaneous TSE in cattle, or from another source, remains unclear. There was no evidence that eating sheep m eat from scrapie-infected animals could cause CJD, but unfortunately policy-makers could not be sure that the agent which caused BSE had in fact derived from scrapie. Moreover, even if the scrapie pathogen had ju m ped species into cattle, policy-makers could not be sure that BSE would subsequently have the sam e transmission characteristics as scrapie. Experimental evidence indicated that it was then not possible to predict what the host range o f a given strain o f scrapie would be once it had ju m p e d to anoth er species (Kim berlin et al., 1987). Even if policy-makers assum ed that BSE was pathogenic to hum ans they could n ot quantify the risk. No one knew, for exam ple, which cattle tissues, if any, w ould be free o f the infectious agent, or what the levels o f infectivity in the various tissues would be, or how this could vary over the period of incubation, and no one knew if there might be a threshold of hum an exposure below which the risk would be negligible. In the late 1980s there was no test that could reliably detect the path ogen in live anim als before clinical symptoms appeared. Asymptomatic cattle could not be identified nor differentiated from cattle which were uninfected. As soon as the first cases o f BSE had been diagnosed senior officials realised that BSE posed a possible risk to hum an health (BSE Inquiry, 1999b, para. 22). As the U nder Secretary in M AFF's A nim al H ealth G roup told his colleagues in early 1988: `... we do not know whether (BSE) can be passed to hum ans... There is no evidence that people can be infected, but we cannot say there is no (11) G illian Sheppard (BSE Inquiry transcript, 1998, 15 D ecem ber, pp. 10-11); John G um m er (BSE Inquiry transcript, 1998, 8 Decem ber, p. 50). 158 Late lessons from early warnings: the precautionary principle 1896-2000 risk... we have toface up to the possibility that the disease could cross another species g a p (em phasis added) (BSE Inquiry, 1999c, para. 59). Policy-makers had no choice but to take urgent decisions about a novel disease the im plications o f which were unknown. 15.3. Initial decisions In the early stages o f the BSE epidem ic a wide spectrum o f possible policy responses was available to policy-makers. The spectrum ranged from the m ost to the least precautionary. They could also be ranked in terms o f their likely costs, but the order was then reversed. The total eradication of the disease and its pathogen from agriculture an d foods would have required, inter alia, the slaughter and exclusion from the food chain o f all the animals which had received feed known, or suspected, to have been contam inated with the path ogen. As there were no ways o f knowing which batches o f feed were contaminated, and because almost all dairy herds had received feed containing m eat and bone meal, and because the majority o f the beef herd were bred from dairy herds, that would have entailed slaughtering alm ost the entire British herd which m ight have cost in the order o f GBP 12-15 billion (12). O n the other hand, there were num erous other options which could have been selected, which would have substantially reduced the risks without spending a great deal of public money. These m ight have included, for example, a ban on the use of animals from affected herds as hum an food, or a ban on the use in the food chain o f all bovine tissues that were suspected o f harbouring the pathogenic agent, or even ju st a ban on the use o f clinically affected animals as hum an food. In 1987, and the first half of 1988, approxim ately 1 200 clinical cases o f BSE were recorded (though at that time the disease was not notifiable and the actual incidence m ust alm ost certainly have been higher) and m ost of those were sold as hum an food. The costs of compensation for removing those clinically diseased animals would have been no m ore than GBP 1 000 p er anim al, totalling approxim ately GBP 1.5 m illion. It is n ot yet possible to estim ate the harm which eating those animals may have caused. Box 15.1. Early warnings Rendered animal slaughterhouse wastes have been recycled into animal feed since at least the beginning of the 20th century (Cooke, 1998). The known risks of that practice included the transm ission, recycling and amplification of pathogens. Those concerns prompted the 1979 Royal Commission on Environm ental Pollution to recom m end minimum processing standards in the rendering industries (RC EP, 1979). Before the Labour governm ent could follow that advice, they lost the 1979 election. It is not y et clear w hat effect such regulations might have had in diminishing the BSE epidem ic because the incoming Conservative government decided to withdraw the proposed regulations, deeming them to be unnecessary and excessively restrictive. The Thatcher government indicated that the industry should be left to decide for itself how its equipm ent should be operated (Barclay, 1996, Section II B, p. 13). A fte r 1996, minimum process standards w ere introduced in the rendering industry and deactivation experim ents have been conducted and are in progress. W e might eventually learn what the effect on the spread of TSE agents would have been if those standards had been introduced in 1979. In the m id-1970s the US D epartm ent of A g riculture decided that carcasses of sheep and goats afflicted w ith, or exposed to, scrapie should not be used in human or animal foods, partly to prevent transmission of scrapie to other flocks but also because of their concern about a possible link between scrapie and C JD (Martin, 1998). No similar action w as taken in the United Kingdom . If BSE w as indeed caused initially by scrapie jum ping from sheep to cattle then similar, relatively inexpensive, restrictions might have prevented the BSE epidem ic. The possibility that BSE might transm it to humans w as recognised by veterinary officials in M A FF as soon as the disease w as first diagnosed in 1986; however they thought that the probability that BSE might be pathogenic to humans was acceptably slight. The earliest docum ented official acknowledgem ent, of which we are aware, that the probability of transmission might be more than remote was made at a meeting at the National Institute for Biological Standards and Control in May 1988. The minutes of a meeting reported the conclusion that 'by analogy (with scrapie and C JD ) BSE may be transm issible to humans' (BSE Inquiry, 1999c, para. 186). Senior government advisers participated in that m eeting. Between 1990 and 1995, evidence gradually em erged indicating that BSE exhibited distinct transm ission characteristics from scrapie in sheep, thus indicating that BSE had an unknown and unpredictable host range. The most significant evidence was the discovery, from 1990 onwards, that BSE was transm issible, via food, to domestic cats, a sp ecies that is not susceptible to scrapie. Evidence that BSE could cause C JD did not emerge until 1995 when cases of an unusual form of C JD (later called variant C JD ) in excep tionally young people began to be discovered. The temporal and geographical association between the two diseases was circumstantial evidence of causation. In 1996 and 1997 direct evidence indicative of a causal relationship between BSE and variant C JD was produced. This included studies indicating that the pathological and clinical features of BSE and variant C JD w ere identical whilst both differed from the distinctive features of scrapie and sporadic C JD . (12) Assum ing a compensation rate of G B P 865 per slaughtered cow (the real rate of com pensation paid to farm ers in 1996), a herd of approxim ately 12 million cattle, the costs of slaughtering and incinerating cattle, and the knock-on effects on em ploym ent produces a maximum estim ated figure of G B P 15 billion ('Cash for cows', 1996). 'Mad cow disease' 1980s-2000: how reassurances undermined precaution 159 Even if the science had been massively less uncertain, scientific considerations would never by themselves have indicated where on the policy spectrum an appropriate response would have been. Policy-makers had to make political judgem ents about which actions to take, and how the costs should be distributed between public and private sources. O ne problem with taking any regulatory measures, as then seen from the perspective o f the UK government, was that any regulatory response -- indeed any admission that consum ing meat, milk or dairy products from British cattle m ight be harm ful -- would have underm ined domestic and international confidence in the safety of British b eef with adverse consequences for the m eat industry. Even the virtually cost-free option o f sharing information about the disease with those outside MAFF m ight have alerted domestic consumers and potential im porters o f UK cattle and m eat to the presence of a new potentially fatal zoonotic disease. Fear of those consequences, and a reluctance to increase public expenditure, dom inated policy-making in MAFF for the first 20 m onths o f the epidemic. For exam ple, when M AFF's C h ief Veterinary Officer first told his minister about BSE he w arned that `... the disorder could have potentially serious implications, not only domestically but for U K exports'. He advised that it was not appropriate to im pose regulatory restrictions, noting that `irresponsible or ill-inform ed publicity is likely to be unhelpful since it m ight lead to hysterical dem ands for immediate, draconian Government measures and might also lead other countries to reject U K exports o f live cattle and bovine embryos and sem en' (BSE Inquiry, 1999b, pp. 27-28). Even the option o f making the disease notifiable, an essential tool for disease surveillance, was rejected in part, because, as one official p u t it, such action `... m ight imply to the gen eral public we know som ething they d o n 't, like the m eat or m ilk is a source o f dan ger for h u m an s' (Phillips et al., 2000, Vol. 3, para. 2.130). As the epidemic rapidly began to escalate, U K policy-makers not only chose to avoid taking any regulatory action at all, but they also decided to try to keep inform ation about BSE within the confines o f the ministry. One o f M AFF's scientists later recalled: `... in Decem ber 1986 when recognition o f the disease began to crystallise, we were at the Central Veterinary Laboratory placed under strict confidentiality as to discussing it with outside people...' (BSE Inquiry, 1999b, p. 13). As the Phillips Inquiry into BSE concluded, during the first half of 1987 `... there was a policy o f restricting, even within the State Veterinary Service, the dissemination of any information about the new disease' (Phillips et al., 2000, Vol. 3, para. 2.137). Most of the scientific research community, the medical profession, and senior officials and ministers in other government departments did not learn about BSE until early 1988. In February 1988, with the m edia beginning to devote attention to the new cattle disease and increasing num bers o f affected cattle, senior MAFF officials ch an ged their views and recom m ended that their ministers introduce a slaughter and compensation policy for clinically diseased cattle which, at the time, were being sold for hum an food. Officials argued, privately, that without a slaughter policy the government would be held responsible if it later transpired that BSE was transmissible to humans. The Agriculture Minister, Jo h n MacGregor, rejected that advice. T he m inister's private secretary explain ed why: `H e (the m inister) does n ot see how we could proceed without being clear where the offsetting savings are coming from... M ore importantly... the argu m en t that slaughter com pensation policy would help to stem the spread o f the disease (advocated in these papers) is precisely the one sugar beet growers have been m aking, an d which we have strongly and publicly been rejecting. He also thinks that action along the lines recommended now would make the export position much worse, n ot better' (em phases added) (Minute, 1988). T he govern m en t's policy was not precautionary. Its prim ary objective was rather one o f trying to diminish, as far as possible, the short-term adverse im pact of BSE on the profitability o f the food industry and the level o f public expenditure. 15.4. Expert advice and regulatory controls In mid-1988, and for the first time, a small expert advisory committee was set up to provide advice on BSE. This only occurred at the insistence of the Chief Medical Officer (CMO) at the Departm ent of Health, who was first inform ed about the new disease only in March 1988 -- 17 m onths after MAFF was first alerted (BSE Inquiry, 1999c, para. 115). That only occurred because agriculture 160 Late lessons from early warnings: the precautionary principle 1896-2000 ministers were advised by their officials that they needed the support o f the CMO for the m inisters' decision not to remove clinically affected cattle from the hum an food chain (BSE Inquiry, 1999c, para. 76). The advisory committee, under the chairmanship of Sir Richard Southwood, insisted on the day o f their first m eeting (20 Ju n e 1988) that clinically affected cattle should cease to go into the hum an and anim al food chains and that farmers should be compensated. Another major change occurred when MAFF announced, on the very sam e day, that they would be introducing a ban on the use o f potentially contaminated rum inant protein in rum inant feed. The feed ban only applied to ruminants. Animals such as pigs and poultry could still be fed with the contaminated protein even though no one knew whether or not they m ight also be susceptible to BSE. MAFF officials had in fact considered, and then rejected, a ban on feeding rum inant protein m eal to all animals because that would have deprived the rendering industry o f its principal m arket (the bulk o f anim al protein was fed to pigs and poultry) (BSE Inquiry transcript, 1998, 29June, p. 35). Senior veterinary officials were nevertheless aware that their decision was a gam ble. In Ju n e 1988, the Chief Veterinary Officer, Keith M eldrum, admitted privately to a colleague that `the m ost we could say is that any rum inant protein fed to (pigs) might contain the agent of BSE or scrapie. W hether or not infection would be established in the pig and whether it might replicate is unknown' (BSE Inquiry transcript, 1998, 16 Ju n e, p. 99). One unfortunate consequence of that decision was that for the next six or so years cross contamination occurred between feed destined for cattle and feed destined for other animals, greatly prolonging the BSE ep id em ic. A lthough Southw ood's advisory com m ittee had been quick to insist that clinically affected animals be removed from the hum an food chain it did not recom m end controls on the use in food o f sub-clinically infected animals, the tissues o f which would also harbour the infectious agent. (In the absence o f an ante-mortem diagnostic test any controls would have had to have been im posed on the entire British herd.) Southwood acknowledged in March 1996, after the acute BSE crisis erupted, that a ban on the use o f all cattle brains might not have been a politically feasible option in 1988. He explained: `We felt it was a no-goer. They (MAFF) already thought our proposals were pretty revolutionary' (New Scientist, 1996). A ban on the use of cattle brains and other offal from all cattle in the hum an food supply was introduced by MAFF in November 1989, nine months after the Southwood Committee reported. That regulation was only introduced after it em erged that one o f the govern m en t's own expert advisers h ad told officials, in confidence, that he had undertaken private consultancy work for the pet food industry and had subsequently recom m ended a bovine offal ban in pet food, and after ministers decided that they did not wish to be upstaged by the m eat products industry and the pet food industry both of which had told MAFF that they would be unilaterally removing bovine offals from their products (BSE Inquiry, 1999e, paras 87-89 and 135). The rum inant feed ban, the slaughter and destruction o f affected cattle, and what becam e known as the specified bovine offal (SBO ) ban were all in place by the end o f 1989. The controls were not designed to eradicate the BSE agent, however, but only to diminish the risk. For example, the tissues selected for the SBO ban were chosen not because they alone harboured the infectious agent but because they could m ost easily be removed, and because they were o f the lowest comm ercial value. There were no experim ental data, at that time, indicating which cattle organs m ight be contaminated with the path ogen, although analogies with other species and their TSEs indicated that many other tissues would also have carried the agent. For example, lymph nodes and peripheral nerves would alm ost certainly be highly infectious but could not practicably be rem oved and organs such as the liver would, by analogy with other TSEs, also contain (lower) levels o f the infectious agent but were commercially valuable (BSE Inquiry, 1999e, para. 85). Moreover, the SBO ban excluded cattle under six m onths old. The carcasses o f calves were not normally split in abattoirs so the removal of their spinal cords would have raised abattoir costs. The exclusion o f calves would only have been sensible if one could have assum ed that vertical transmission o f BSE from cow to calf could not occur. That assum ption was im plausible because it was already known that scrapie did transmit from ewes to their lambs and because MAFF did not start funding research into maternal transmission 'Mad cow disease' 1980s-2000: how reassurances undermined precaution 161 o f BSE until 1989 (Barclay, 1996, p. 16; Nature, 1990). contaminated material from entering the food chain (Radio Times, 1992). The delay in im plem enting the principal regulations also m eant there had already been repeated hum an exposures to the pathogen. For exam ple, from mid-1988, after Southwood recom m ended that all clinically affected cattle be destroyed, to the end o f 1989, when the SBO ban was actually introduced, an estimated 30 000 infected cattle that were at least halfway through the average incubation period for BSE had been consum ed (Dealler, 1996). 15.5. Constructing a house of cards In 1987, U K policy-makers adopted the hypothesis that BSE was an innocuous version o f scrapie and they struggled to rem ain w edded to it, in the face o f accum ulating evidence to the contrary, because that narrative enabled the UK government to offer a reassuring and optimistic m essage by suggesting that the presence o f BSE in British dairy and beef herds posed no threat to hum an health. MAFF repeatedly asserted that their reassuring statem ents were fully supported by scientific evidence, expertise and advice. T hat was, however, a m isrepresentation. Policy-makers were repeatedly told, both by the scientific experts on whom they claimed to rely, and by the wider scientific community, that it was im possible to be certain that consum ing meat, milk and dairy products from anim als with BSE p osed no risk. For exam ple, in May 1990, the govern m en t's Sp on giform Encephalopathy Advisory Committee (SEAC) told policy m akers that: `in the present state o f knowledge, it would not be justified to state categorically that there was no risk to hum ans, an d it was n ot ap p rop riate to insist on a zero risk' (BSE Inquiry transcript, 1998, 24 March, p. 71). Ministers and senior policy makers insisted otherwise in public. On 7 Ju n e 1990, for exam ple, the Agriculture Minister told the House o f Com mons that there was `... clear scientific evidence that British b e e f is perfectly safe ' (Hansard, 1990, column 906). Policy-makers repeatedly claim ed a certainty which was unavailable and which they knew to be unavailable. Occasionally policy-makers acknowledged that they could n ot be certain that BSE was an inn ocuous form o f scrapie, bu t they always claimed that regulatory controls introduced in November 1989 prevented all potentially Regulatory controls were never intended, however, to elim inate exposure to the BSE agent, but only to diminish exposures, and MAFF scientists and expert advisers made sure that senior policy-makers knew that (see, for exam ple, BSE Inquiry, 1999e, para. 275). For example, in 1990 SEAC produced a draft docum ent about the safety o f beef intended for the Chief M edical Officer, in which it m ade statements such as: `some o f the edible offal... that have on rare occasions dem onstrated low titres o f infectivity are not included in the offal ban ' and `there are some who insist on nothing less than an absolute guarantee o f safety. N o scientist is in a position to do that at present for British (or Irish) b e e f' (em phasis in original) (Phillips et a/., 2000, Vol. 11, para. 4.120). T hat docum en t was circulated within the Department of Health and MAFF for suggested am endm ents. The MAFF official who subsequently forwarded the docum ent to ministers told those ministers, and other MAFF officials, that: `the m ost potentially inflammatory pieces of drafting in earlier versions (including the citations above) have now been edited o u t' (Phillips et at., 2000, Vol. 11, para. 4.118). But then as Phillips concluded, in a m em orable phrase, ministers and officials followed an approach to inform ation provision `whose object was sedation ' (Phillips et at., 2000, Vol. 1, para. 1179). T he assertion that b e e f was entirely safe was not only m isleading but it also m ade it increasingly difficult for MAFF to take a range o f other precautionary steps. Any new regulatory measure, no matter how useful or cheap, risked not only raising questions about the govern m en t's reassurances, but might also provoke serious doubts about the logic of not introducing further and more expensive controls (zero risk was always going to be unattainable without slaughtering and restocking the entire British herd and cleaning out the feed chain). In other words, partial reductions in risk were difficult from the point o f view o f presentation; the only credible options, given the rhetoric, were to draw a line, maintain that it provided total protection an d stick to it, or alternatively to try to eradicate the agen t entirely. Num erous precautionary m easures were not therefore introduced, not because o f their im m ediate costs but because o f their liability 162 Late lessons from early warnings: the precautionary principle 1896-2000 to underm ine the govern m en t's reassuring admitted that some measures of control message. For example, mechanically might be too impractical an d /o r too recovered m eat (MRM) was widely expensive, and therefore that a lower but acknowledged to contain residual pieces o f acceptable level o f risk m ight have to be potentially highly infectious nervous tissue tolerated, policy-makers m ight have found it (BSE Inquiry transcript, 1998, 6 July, pp. 104-- easier to introduce precautionary control 106 and 127). Banning MRM would have m easures without threatening the m inistry's m ad e explicit the risks associated with credibility. peripheral nervous tissue, m ost o f which could not practicably be removed from the As events and evidence eventually prom pted carcass. As the minutes o f a m eeting held at additional regulatory controls, it becam e MAFF in Septem ber 1989 indicate: increasingly hard to reconcile those `M echanically recovered m eat (MRM) -- the responses with the govern m en t's reassuring possible danger raised by several o f those narrative, particularly if those additional consulted was recognised and during the regulations m ade explicit further areas where discussion there was an expression o f the controls ought to be introduced. Policy illogicality o f what was being done and in makers therefore sometimes misrepresented particular how easy it would be to have to their reasons for introducing regulations in concede the possible dangers o f material an attem pt to prevent their narrative from other than those listed in the proposed ban. unravelling. For exam ple, MAFF insisted that It was agreed not to raise it' (BSE Inquiry, the SBO ban was not necessary on scientific 1999e, para. 263). grounds (see, for example, House of Commons Agriculture Committee, 1990, pp. In February 1990 the Institute of 9 and 71), and it was therefore difficult to Environmental Health Officers (IEHO), convince industry and other stakeholders whose m em bers were responsible for that those regulatory m easures were crucially enforcing controls in slaughterhouses, raised im portant for public health. In 1995, during concerns with MAFF abo u t existing practices un an n oun ced visits by enforcem en t officers for removing cattle brains from heads, none to U K abattoirs, som e 48 % were found to be o f which, the IEH O argued, could be failing to com ply with the SBO regulations accom plished without contaminating the (House of Commons Agriculture and Health head meat. The IEHO sensibly Select Committees, 1996, p. 10). As a recom m ended that all removal o f head meat representative o f those responsible for should take place prior to splitting the skull enforcing slaughterhouse regulations p u t it: for the removal of the brain. Although `We were bein g given the m essage that really agriculture ministers had `qualm s' about the practice o f removing brains before removing head meat, and felt that the practice ought to be ban n ed, civil servants argu ed, successfully, there probably wasn't a problem anyway so this was maybe a bit o f window dressing as opposed to serious public health m atters...' (Panorama, 1996). that no new controls should be introduced. As a senior official in M AFF's M eat Hygiene Division explained to the Minister for Food: `A m en dm en t regulations w ould fuel debate on BSE generally and, inevitably, lead to dem ands for similar action on spinal cords... A ban on splitting (spinal columns) would have grave consequences for the industry and for the export trade. N or would it end with spinal cords. Concern would then be directed at nerve trunks and lymph nodes, which cannot be removed from carcasses' (BSE Inquiry, 1999f, p. 7). It was not only new regulatory measures, however, that threatened the govern m en t's claims that risks were zero. Sustaining the governm ent's reassuring m essage also entailed that unwelcome information and evidence could underm ine the official narrative. Expert advisers were therefore carefully selected and those who did not share the m inistry's policy framework, or who m ight refuse to acquiesce to restrictions on the dissemination o f information, were excluded. As one MAFF official p u t it: `... you have to turn to external bodies to try to give Again, a cheap, simple and risk-reducing m easure was avoided in order to maintain a reassuring and misrepresentative message about safety and to counteract political pressure for ever more precautionary controls. If MAFF had publicly acknowledged some o f the uncertainties and risks, and some credibility to public pronouncem ents, you are very dependent therefore on what the Com m ittees then find... Really the key to it is setting up the Com m ittee, who is on it, and the nature of their investigations' (BSE Inquiry transcript, 1998, 29Ju n e, pp. 79-81). Some experts were also excluded because 'Mad cow disease' 1980s-2000: how reassurances undermined precaution 163 Box 15.2. The European dimension Although BSE first em erged in the United Kingdom , it spread to other countries, especially in continental Euro pe, as a consequence of trade in anim als and feedstuffs. Individual M em ber States and the European Commission have therefore had to grapple with many of the same difficulties and dilemmas as the United Kingdom . Countries such as Ireland, Portugal and France have had sufficiently high rates of BSE that, during the 1990s, their governments recognised the need to establish controls on their dom estic production system s. O ther countries, with lower incidences of BSE, such as Belgium, the Netherlands and Italy occasioned some concern and regulatory activities, but prim arily in relation to traded animals and feedstuffs. The various jurisdictio ns have, nevertheless, responded in quite different ways, regardless of their levels of industrial developm ent or the num ber of BSE cases in each country. A full account of how and why those responses differed is beyond the scope of this case study (19). N evertheless, it is clear that the location of responsibilities for BSE policy-making was fundam ental in fram ing the w ays in which the BSE problem was defined, evaluated and responded to. In general, those jurisdictions in which industrial sponsorship and responsibility for consumer protection were located within the same institutions (e.g. Ireland) adopted less precautionary approaches to protecting public health than those where sponsorship and regulation were split or shared between more than one institution (e.g. Austria) (B A SES, n.d.). For exam ple, within the European Comm ission, prior to 2000, responsibility for BSE policy was located prim arily within DG III (with responsibility for the European Union's (EU) internal market and enterprise) and DG VI (with responsibility for agriculture and fisheries). Although the United Kingdom introduced dom estic legislation banning the use of contam inated ruminant protein for use in ruminant feed in Ju n e 1988, at which point it informed the European Commission of its actions, EU-wide legislation to control the spread of BSE was not introduced for another six years. A fter June 1988, the United Kingdom continued to export contaminated feed to other Mem ber States and some of that was subsequently fed to cattle. For exam ple, exports of meat and bone meal to the EU had jum ped from 12 553 tonnes in 1988 to 25 005 tonnes in 1989 (European Parliam ent, 1997, p. 8). In the summer of 1989, the Commission asked the United Kingdom to introduce an exp o rt ban on those feedstuffs but the United Kingdom refused to do so (BSE Inquiry, 1999d, para. 257). The Commission has since claimed that it did not have the legal basis itself, prior to the Single European A ct, to ban exports of UK meat and bone meal (although the European Parliam ent has disputed that assertion); instead the Com m ission invited all M em ber States to introduce a national ban on the import from the United Kingdom of ruminant-derived meat and bone meal (European Parliam ent, 1996, p. 10). Some M em ber States then did so, or had already introduced such a ban (e.g. the N etherlands), whilst some did not ban imports of UK meat and bone meal until much later (e.g. Portugal). The European Com m ission did not insist on an EU wide ban on the feeding of ruminants with meat and bone meal until 1994. Not until 1996 did the Commission ban the exports of UK produced meat and bone meal (and all other cattle products). The European Parliament has concluded that the Commission consistently subordinated the protection of animal and public health to maintenance of the internal market (European Parliam ent, 1997). A s a result of the 1996 BSE crisis, there has been an on-going reorganisation of scientific advice and food safety policy-making in both the European Com m ission and M em ber States. In the European Comm ission, the scientific advisory system has been reformed and m anagement of scientific advice has moved to DG XXIV, now renamed DG SA N C O or the Directorate General for Health and Consumer Protection. In 2000, the regulatory and sponsorship functions of DG III and DG VI w ere sp lit and the regulatory functions were also transferred to DG SA N C O . The European Comm ission's 2000 W hite Paper on food safety has now proposed a further institutional split by proposing the creation of a European Food Authority to provide independent science-based advice to DG SA N C O (European Comm ission, 2000). Analogous reforms and proposals for reform have occurred in m any of the Member States. A s of 2000, many individual M em ber States have discovered that they have rising cases of BSE (e.g. France and Ireland) and those countries that previously thought they might be free of the disease have discovered cases amongst their dom estic cattle population (e.g. Germ any and Spain). Substantial historical d ifferences in the types of controls that have been im posed by different M em ber States, and the extent to which those controls have been enforced, will mean that some jurisdictions face considerable animal and public health challenges for many years. their institutional location m ight provide the wrong im pression. For exam ple, the UK Public Health Laboratory Service (PHLS) -- the established disease surveillance institution in the U nited Kingdom for new an d em erging diseases -- was always excluded from BSE policy. As the Welsh Chief M edical Officer has recalled: `the basis o f the consistent opposition to the involvement of the PH LS was the anxiety that their involvement would be tantam ount to admitting the possibility o f a hum an health risk (Phillips et al., 2000, Vol. 11, para. 4.28). Senior policy-makers also attem pted to ensure that the commissioning, conduct and reporting o f BSE research were tightly controlled. Many key experim ents were never started or were seriously delayed, information and evidence were sometimes withheld, and data and materials were not always shared with other researchers. For exam ple, a random post-mortem survey of anim als in abattoirs would have helped to provide estimates of the num bers of infected but asymptomatic animals entering the hum an food chain. The laboratory resources (13) National reports about the policy responses to BSE in 11 European countries, conducted as part of a European Commission sponsored research project, provide more information. These are available at: http:// www. upmf-grenobl e.fr/inra/serd/BASES/ 164 Late lessons from early warnings: the precautionary principle 1896-2000 were available and the costs would have been relatively low but only one such survey has been conducted in the United Kingdom and that was in 1999 and only in cattle that were not permitted to enter the hum an food chain. A crucially im portant experim ent to see whether cattle fed on rations deliberately infected with scrapie would get BSE was not started until 1996 (BSE Inquiry transcript, 1998, 11 M arch, p. 132). O nce BSE becam e notifiable, all infected cattle brains became the property o f MAFF and the ministry showed extraordinary reluctance to provide som e o f the m ost senior scientists in the U nited States with pathogenic m aterial (BSE Inquiry, 1999a, paras 493-505). A precautionary approach to policy-making, on the other hand, would have involved producing and disseminating far more information and evidence. 15.6. The failures and eventual collapse of the policy edifice In the years following the Southwood Report, M AFF's reassuring narrative repeatedly and progressively unravelled. It did so partly because the scientific evidence, which was never entirely reassuring, became progressively less reassuring, because the m inistry's control o f the diffusion o f inform ation was im perfect, and because actors an d agents outside the m inistry's control took independent decisions. This occurred despite M AFF's vigorous attem pts to sustain the policy in the face o f its inherent weaknesses and accumulating counter evidence. By late 1995 a lengthy series o f events and evidence had obliged MAFF progressively to tighten regulations, though in a reactive, rather than in an anticipatory, fashion. By then, a growing proportion o f the national and global scientific and research community had become increasingly concerned about the risks from BSE. M AFF's policy eventually collapsed in March 1996 after a new variant o f CJD (now, with the passage o f time, labelled variant or vCJD) had em erged in the United Kingdom, and after SEAC had concluded that consuming food contam inated with BSE was the m ost probable cause. Box 15.3. The costs of BSE BSE has had substantial, and as yet incom plete, ram ifications, som e of which can be defined in monetary term s but others of which are incalculable. In 1998, the cum ulative expenditure by UK agricultural departm ents in response to the BSE crisis, from 1996 to 2001, was forecast to be G BP 4.2 billion (House of Com m ons, 1999). The bulk of that expenditure was, and will be, for compensating commercial enterprises, especially compensation to farmers for the removal of cattle over 30 months of age from the human food chain and support to the slaughtering and rendering industries. The figures also include expenditure on research programmes and administration. Other costs since 1996 that have been borne by public expenditure include the public inquiry into BSE, at an estim ated G B P 25 million (Farm ers W eekly Interactive Service, 1999). The costs to the private sector of BSE have also been considerable. The ban on British beef exports in March 1996 led to the com plete loss of a trade worth G B P 700 million per year (DTZ Pieda C onsulting, 1998). In the first 12 months since March 1996, the total value of the market for UK produced beef fell by an estim ated 36 % in real term s (a com bination of loss of exp orts and the drop in dom estic dem and), am ounting to an estimated loss of value added to the UK economy of G BP 1.15 billion (DTZ Pieda Consulting, 1998). It would be prem ature to try to provide precise estim ates of the total costs of BSE, not least because we still cannot estim ate how many people will eventually succumb to new variant C JD ; there may be no more than another 100 cases, or there may eventually be up to a million (C ollinge, 1999). 15.7. Conclusions BSE was always goin g to be a difficult issue to apply a precautionary approach to once it had been discovered in the U K cattle herd. In November 1986, many infected cattle had already been consum ed and even at that stage it would have been hugely expensive to eradicate. Nevertheless, there was a great deal which the governm ent could have done to diminish the risks to consum ers, and in the long run to the m eat industry and to the Exchequer, especially if they had openly acknowledged what the scientific evidence did, and did not, indicate. Instead the U K government claimed to be prudently protecting public health, while in practice it covertly subordinated the protection o f public health to the support of agricultural sales, with a view also to minimising state intervention and public expenditure. The regulations which were set were, consequently, too little and too late, and even then they were not properly enforced. M oreover too little was invested in scientific research and the involvement of 'Mad cow disease' 1980s-2000: how reassurances undermined precaution 165 independent scientists was actively discouraged. If the U K governm ent had adopted a genuinely precautionary approach, it would have necessitated, firstly and m ost fundamentally, reform ing policy institutions so as to separate responsibilities for regulation from those of sponsorship. The BSE saga has provoked a reappraisal o f the ways in which risks to public and environmental health are assessed and m anaged in the United Kingdom . Since May 1997, the U K government has acknowledged that there was a fundam ental contradiction at the core o f M AFF's rem it, an d that recognition informed the decision to create the Food Standards Agency. A precautionary approach would also have necessitated acknowledging how little `sound science' was available, and would have involved open and accountable discussions of the possible costs and benefits o f taking, or failing to take, a wide range o f different possible courses of action. A Freedom of Information Act might have the effect of initiating a cultural change in the way policy making bodies represent scientific evidence. An institutional separation between those responsible for providing scientific advice, risk assessments and research and those responsible for regulatory policy-making m ight also have encouraged more open and robust discussion o f possible risks. Furtherm ore, the m ore sup p ort there is for research, conducted by a wide range of disciplines and interdisciplinary groups and in a wide range o f institutions with open access to evidence and data, the harder it would be to conceal uncertainties, and therefore the uncertainties might be more readily diminished. O ne o f the factors which, prior to March 1996, helped MAFF to sustain its optimistic narrative was the willingness o f expert scientific advisers to acquiesce to an arrangem ent under which the scientists provided advice which was based on both scientific and non-scientific considerations but which was represented to the general public as if it was purely scientific. That suited ministers because it allowed them to argue that they were doing what, and only what, their scientific advisers recom m ended, and it flattered the scientists by representing them as authoritative and influential. A more precautionary approach could have been expected if the role o f the scientific advisers Box 15.4. Conclusions from the Phillips Inquiry In Janu ary 1998, an inquiry into BSE w as set up by the Labour government to 'establish and review the history of the em ergence and identification of BSE and variant C JD and of the action taken in response to it up to 20 March 1996' and 'to reach conclusions on the adequacy of that response, taking into account the state of knowledge at the tim e'. The inquiry com m ittee, chaired by Lord Ju stice Phillips, reported in O cto b er 2000 after having taken oral and written evidence from over 600 witnesses. The 16-volume report focused prim arily on procedures rather than outcom es; indeed, the inquiry team was reluctant to comment on the extent to which policy was or was not proportionate, preferring instead to highlight issues of communication within and outside government, the use of expert advice and cooperation between government departm ents. Phillips' main criticism was that public policy was dominated by the political objective of reassuring the public. The inquiry team maintained, nevertheless, that appropriate policy decisions were taken although they were not always tim ely, properly designed, or adequately implemented or enforced. Tw o of the conclusions reached by Phillips and his colleagues were that 'the government was anxious to act in the best interests of human health' and that 'it was not (the Ministry of Agriculture, Fisheries and Food's (M AFF)) policy to lean in favour of the agricultural producers to the detrim ent of the consum er'. It is, how ever, difficult to see how the evidence available to the inquiry, som e of which is sum m arised in this case study, is consistent with those conclusions. The Phillips Report contains 160 individual lessons. Many of these are specific to animal health policy and agricultural production practices but the main generic conclusions concern: the appropriate use and role of advisory co m m ittees; the maintenance of sufficient in-house expertise; greater cooperation between animal and human health professionals; the proper implem entation, operation and enforcement of policy measures; coordination of research to diminish policy relevant uncertainties; the principle that uncertainty can justify action; the importance of establishing credibility and trust; communication of uncertainty; openness and transparency. Many of those conclusions are relevant to debates about precaution but none provided an explicit discussion of w hat precaution might im ply in practice for the conduct of policy-making and research. had been m ore strictly delineated, and open to expert and public scrutiny. Since risk assessm ents are always fram ed by socio economic considerations, the responsibility for articulating and justifying those framing assumptions should be the responsibility of democratically accountable ministers, and expert scientific advisers should be responsible for showing how they have gathered and interpreted all the appropriate evidence. Since March 1996, m em bers of 166 Late lessons from early warnings: the precautionary principle 1896-2000 Table 15.1. SEAC have becom e more independent than their predecessors. When advising on the possible risks from eating `beef on the bon e' SEAC set out the possible consequences of various courses o f action, and explicitly indicated the decisions which ministers would have to take. Policy-making on BSE has becom e more precautionary, and in part that has been because it has becom e more open and accountable, especially following the creation o f the U nited K in gdom 's new Food Standards Agency. BSE: early warnings and actions M id-1970s 1979 1986 1988 1988 1989 1995 1995 1996 1996 1998 2000 United States bans scrapie-infected sheep and goat meats from cattle food chain UK Royal Com m ission on Environm ental Pollution recognises risks of pathogens in animal feed and recom m ends minimum processing standards in rendering industries First cases of bovine spongiform encephalopathy (BSE) are officially acknowledged First documented official acknowledgem ent that BSE may be transmissible to humans Southwood C om m ittee is set up and recom m ends that clinically affected cattle should not go into human and animal food Ruminant feed ban, slaughter and destruction of affected cattle and specified bovine offal (SBO) ban Alm ost 50 % of the abattoirs checked are found to be failing to comply with the SBO ban Evidence that BSE may cause Creutzfeldt-Jakob disease (CJD) A t last, experim ents start to see whether cattle fed on rations deliberately infected with scrapie would get BSE BSE crisis, after a new variant of C JD em erged in the United Kingdom and consuming BSE contaminated food was considered the most probable cause The Phillips Inquiry takes place and its 16-volume report is published. Its conclusions do not seem sufficiently rigorous on judging governm ent actions over tim e. Th ese conclusions state that appropriate policy decisions had been taken, although not always tim ely, or adequately implemented or enforced. 15.8. References Barclay, C., 1996. `Bovine spongiform encephalopathy and agriculture', House of Commons Library Research Paper No 96/62, 15 May. BASES, n.d. General introduction, national action systems and TSEs in Europe, at h ttp :// www.upmf-grenoble.fr/in ra/serd /B A S E S / BSE Inquiry, 1999a. `The Central Veterinary Laboratory 1985-1989', Revised factual account, No 4. BSE Inquiry transcript, 1998. At h ttp :// www.bse.org.uk `Cash for cows', 1996. The Economist Vol. 338, No 7959 (30 March). Collinge, J . 1999. `Variant Creutzfeldt-Jakob disease', The Lancet Vol. 354, No 9175 (24 July), pp. 317-323. Cooke, B. C., 1998. `A b rief history o f the use of m eat products, especially m eat and bone m eal in the feeds for rum inant anim als', The BSE Inquiry, Statem ent No 27. BSE Inquiry, 1999b. `The early days', Revised factual account, N o 5. BSE Inquiry, 1999c. `Slaughter and com pensation', Revised factual account, No 6. BSE Inquiry, 1999d. `Rum inant feed b an ', Revised factual account, No 7. BSE Inquiry, 1999e. `The introduction o f the SBO b a n ', Revised factual account, N o 8. BSE Inquiry, 1999f. `Consideration o f the risk from m echanically recovered m eat (MRM) in 1989', draft factual account, No 14. Dealler, S. 1996. `Can the spread o f BSE and CJD be p red icted ?' in Ratzan, S. C. (ed.), The mad cow crisis, U C L Press, Lon don . DTZ Pieda Consulting, 1998. `Economic im pact o f BSE on the UK economy', Report to UK agricultural departments and HM Treasury. European Commission, 2000. White Paper on food safety in the European Union C O M (99)719, 12 January, at h t t p :/ / europa.eu.int/com m /food/fs/in tro/ in dex_en .h tm l 'Mad cow disease' 1980s-2000: how reassurances undermined precaution 167 European Parliament, 1996. `Replies from the Commission to questions from the Committee m em bers', Temporary Committee o f Inquiry into BSE, Doc_PE 218.980, 18 Septem ber. European Parliament, 1997. `R eport on alleged contravention or maladministration in the im plem entation o f Community law in relation to BSE, Part B, Work o f the Committee o f Inquiry and basic data', Doc_EN\RR\319\319055 A4-0020/97A, 7 February. Farm ers Weekly Interactive Service, 1999. At http://www.fwi.co.uk (accessed 26 April 1999). Hansard, 1990. 8 Ju ne. H ouse o f C om m ons, 1999. BSE: The cost of a crisis, Select Com m ittee on Public Accounts, 24th report, HC 790, HMSO, London. House of Commons Agriculture Committee, 1990. `Bovine spongiform encephalopathy (B S E )', Fifth report, 10 July. House of Commons Agriculture and Health Select Committees, 1996. `Bovine spongiform encephalopathy (BSE) and Creutzfeldt-Jakob disease (CJD): Recent developments', Minutes o f evidence, HC-331. Kim berlin, R. H., Cole, S. an d Walker, C. A., 1987. `Temporary and perm anent m odifications to a single strain o f m ouse scrapie on transmission to rats and ham sters', Journal of General Virology Vol. 68, pp. 1875 1881. Martin, W. B., 1998. `Involvem ent with scrapie as Scientific Director o f the M oredun Research Institute', The BSE Inquiry, Statem ent N o 5. Minute, 1988. Minute from the private secretary o f Mr M acGregor to the private secretary o f Mr Thom pson, dated 29 February 1988, BSE Inquiry Docum ent No YB 88/2.29/4.1. Nature, 1990. 24 May, p. 278. New Scientist, 1996. 30 M arch, p. 4. Panorama, 1996. BBC Television, broadcast on 17 Ju ne. Phillips et al., 2000. The BSE Inquiry Report. Evidence and supporting papers of the inquiry into the emergence and identification of bovine spongiform encephalopathy (BSE) and variant CreutzfeldtJakob disease (vCJD) and the action taken in response to it up to 20 March 1996, T he Stationery Office, London. Radio Times, 1992. Keith M eldrum , quoted in issue o f May 31. RCEP, 1979. `A griculture an d pollu tion ', Seventh report, Royal Commission on Environmental Pollution. 168 Late lessons from early warnings: the precautionary principle 1896-2000 16. Twelve late lessons Editorial team 16.1. Introduction The case studies provide a wealth of `late lessons' for future policy. To attem pt to sum m arise them all would effectively replicate the studies. Returning to the four broad questions that we posed to the authors and trying to distil a num ber o f specific lessons seem ed more useful. These then might be applied to future policy in order to minimise repetition o f the mistakes -- or at least the oversights -- o f the past. The first question posed was when was the first credible scientific early warning? The second was when and what were the m ain actions or inactions on risk reduction taken by regulatory and other responsible bodies? T he key p oin t here is the length o f the gap between the specific problem being identified and effective action being taken. The answer for many case studies was that the gap was long, certainly many years or decades, and in som e cases over a century. This m ight not be so surprising before the need for the precautionary principle had been explicitly identified in the 1970s and 1980s. But even after that, exam ples of unequivocal precautionary action were relatively scarce. It is also notable that, while the precautionary principle debate has sometimes been characterised as a battle between the European U nion (EU) and North America, the case histories suggest another story, one o f different degrees o f acceptance o f the need for precaution within different institutions in both North America and Europe. In many o f the case studies, adequate inform ation about potential hazards was available well before decisive regulatory advice was taken, but the inform ation was either not brought to the attention o f the appropriate decision-makers early enough, or was discounted for one reason or another. It is also true that in som e o f the case studies, early warnings -- and even `loud and late' warnings -- were effectively ignored by decision-makers because o f short-term economic and political interactions (see the case studies on asbestos, PCBs, the Great Lakes, and sulphur dioxide and acidification). The third question was what were the resulting costs and benefits of the actions or inactions? This proved to be the most difficult question for the case study authors to answer, at least in a comprehensive manner. In part this is due to the background of many o f our authors, who are generally technical experts in the field, rather than experts in assessing the economic costs and benefits, or the wider pros and cons of action. B ut this is also an intrinsically difficult and controversial area. There is no credible way o f reducin g the pros an d cons o f alternative courses o f action to a single figure, economic or otherwise, not least because of the problem of comparing incommensurables and because the pros and cons are unlikely to be spread evenly across all interest groups. There are constructive ways o f dealing with these com plications, but ultimately a general analysis lay beyond the scope o f the current publication. However, dealing with som e aspects o f costs and benefits is inevitable when addressing the fourth question posed to the authors, namely, `what lessons can be drawn that may help future decision-making?'. For the purposes o f this publication, this is a key issue. The European Scientific Technology Observatory (ESTO) project on technological risk and the m anagem ent of uncertainty (see, for example, Stirling, 1999) provided the initial fram ing o f this analysis. This sets out a comprehensive structure for the consideration o f issues relating to precaution. N ot only did this help in organising the consideration o f the lessons, it also gave an opportunity to test or elaborate many o f the points arising from the ESTO studies against the large body of historical material contained in the case studies. Most o f the key issues that em erged from the case studies could be addressed by 12 late lessons: 1. Acknowledge an d respon d to ignorance, as well as uncertainty and risk, in technology appraisal and public policy m akin g. 2. Provide adequate long-term environmental and health monitoring and research into early warnings. Twelve late lessons 169 3. Identify and work to reduce `blind spots' and gaps in scientific knowledge. 4. Identify and reduce interdisciplinary obstacles to learning. 5. Ensure that real world conditions are adequately accounted for in regulatory appraisal. 6. Systematically scrutinise the claim ed justifications and benefits alongside the potential risks. 7. Evaluate a range o f alternative options for m eeting needs alongside the option under appraisal, and promote more robust, diverse and adaptable technologies so as to minimise the costs o f surprises and maximise the benefits of innovation. 8. Ensure use o f `lay' and local knowledge, as well as relevant specialist expertise in the appraisal. 9. Take full account o f the assum ptions and values of different social groups. 10. M aintain the regulatory independence of interested parties while retaining an inclusive approach to inform ation and opinion gathering. 11. Identify and reduce institutional obstacles to learning and action. 12. Avoid `paralysis by analysis' by acting to reduce potential harm when there are reasonable grounds for concern. The distinctions between these different aspects are intended to be illustrative, rather than definitive. Many are clearly interlinked. Some might be combined together or further distinguished. However, the issues sum m arised here provide a basis for the practical im plem entation o f the precautionary principle. Many o f the lessons relate to the type, quality, processing and utilisation o f information set within the context of a m ore participative and democratic process. Such an integrated and comprehensive process of hazard and options appraisal clearly needs to be related to the likely scale o f the potential consequences (environmental, social, econom ic) o f the activity in question. In elaborating these lessons the rule was of not to introduce m aterial extraneous to the case studies. However, in a final section som e m ore general points are raised, which attem pt to set the conclusions within the context of other developments in the field. 16.2. Twelve late lessons 16.2.1. R espond to ignorance as well as uncertainty A central lesson o f this book concerns the im portance o f recognising and fully understanding the nature and limitations of our knowledge. What is often referred to as `uncertainty' actually hides im portant technical distinctions (see Box 16.1.). All the activities in the case studies were subjected to some form of (formal or informal) assessment o f risk. What rem ained neglected, however, was the virtual certainty that there would be factors that rem ained outside the scope o f the risk assessment. This is the dom ain o f ignorance -- the source of inevitable surprises, or unpredicted effects. N o m atter how sophisticated knowledge is, it will always be subject to som e degree o f ignorance. To be alert to -- and humble about -- the potential gaps in those bodies of knowledge that are included in our decision m aking is fundam ental. Surprise is inevitable. Ju s t as one basis for scientific research is the anticipation o f positive surprises -- `discoveries' -- so it will always yield the corresponding prospect of negative surprises. By their nature, complex, cumulative, synergistic or indirect effects in particular have traditionally been inadequately addressed in regulatory appraisal. Thus a key element in a precautionary approach to regulation involves a greater willingness to acknowledge the possibility of surprise. This does not m ean resorting to blanket opposition to innovation. But acknowledging the inevitable limits of knowledge leads to greater humility about the status of the available science, requiring greater care and deliberation in making the ensuing decisions. It also leads to a broadening o f appraisals to include more scientific disciplines, m ore types o f information and knowledge, and more constituencies. The consequences of ignorance can be dramatic, as dem onstrated by the case study on halocarbons. Prior to the hypothesis of a m echanism for stratospheric ozone depletion in 1974, the now well-known im pact o f chlorofluorocarbons (CFCs) was a very strong candidate for ignorance extending over many decades. Not only the likelihood, but the very possibility o f an `ozone h ole' was unappreciated. Chemicals that were relatively inert and benign under `norm al' conditions 170 Late lessons from early warnings: the precautionary principle 1896-2000 Box 16.1. Risk, uncertainty and ignorance The precautionary principle is seen principally as a w ay to deal with a lack of scientific certainty. A basic foundation for our conclusions concerns the nature of scientific certainty itself. Th ere is an urgent need for a more complete and system atic basis for thinking about the different w ays in which scientific uncertainty may pervade regulatory appraisal. First there is the fam iliar condition of risk, as form ally defined in probability theory. Th is is w here all possible outcom es are known in advance and w here their relative likelihood can be adequately expressed as probabilities. W here this condition prevails, risk assessm ent is a valid technique that can save lives, prevent damage to the environment and provide a robust basis for decision-making. Still, the jud gem ents over w hat is defined as at risk, and over the right balance to strike in decision-m aking, are necessarily laden with subjective assumptions and values. Under the condition of uncertainty, as formally defined, the adequate empirical or theoretical basis for assigning probabilities to outcomes does not exist. This may be because of the novelty of the activities concerned, or because of com p lexity or variability in their contexts. Either w ay, conventional risk assessm ent is too narrow in scope to be adequate for application under conditions of uncertainty. Although techniques such as safety factors, scenario or sensitivity analysis can be useful, they do not provide a w ay adequately to assess the impacts of different options. Here, more than ever, jud gem ents about the right balance to strike in decision-making are laden with subjective assumptions and values. M any case studies in this book involve exam ples where regulatory appraisal laboured not only under a lack of certainty as to the likelihood of different outcom es, but where some of the possibilities them selves remained unknown. Here, decision making is faced with the continual prospect of surprise. Th is is the condition form ally known as ignorance. Even more than uncertainty, this underscores the need for a healthy humility over the sufficiency of the available scientific knowledge and, crucially, for an institutional capacity for open reflection on the quality and utility of available bodies of knowledge. Regulatory appraisal must explicitly address the implications of alternative assumptions and system atically docum ent how these relate to the perspectives of different social groups and to the enviro nm ent. O nce it is acknow ledged that the likelihood of certain outcom es may not be fully quantifiable, or where certain other possibilities may remain entirely unaddressed, then uncertainty and ignorance, rather than mere risk characterise the situation. The adoption of robust, transparent and accountable approaches towards the various aspects of risk, uncertainty and ignorance can be identified as one crucial m eans of regaining public confidence in regulatory decision-making. The decision-making process needs to be more explicit and system atic about the level of proof needed to justify reducing hazards. Exam ples include 'scientifically based suspicion', 'reasonable grounds for concern', the 'balance of evidence' and 'beyond reasonable dou bt'.(see Tab le 1 6 .1 ., in the last section of this chapter, and its elaboration in the last part of this report). There is therefore a range of choices of levels of proof for determining the basis for action, each with different cost and benefit implications for different groups. These different levels of proof provide a more sophisticated basis for the weighing up of potential benefits and harm than do sim ple pronouncements of truth or falsity. (and less conventionally hazardous than the substances they replaced) turned out to behave very differently under conditions that were not considered in the risk appraisal. The effects o f the synthetic oestrogen diethylstilboestrol (DES) on the next generation o f the treated patient came as a com plete surprise, while the accum ulation of organotins in top predators, arising from tributyltin (TBT) antifoulants was simply not envisaged. According to the authors o f the antim icrobials case study: `The justification for the later dilution o f (the Swann C om m ittee's) conclusions an d com prom ises on its recom m endations was based mainly on narrow considerations o f what was precisely known rather than on taking account of what was not known, o f the ignorance within the field ... In other words science that em braces complexities, uncertainties and unknowns with m ore humility an d less hubris is n e e d e d .' The lesson seems clear. Rather than focusing only on the most straightforward and direct impacts, the process of regulatory appraisal should extend attention to as wide a range of conditions and effects as can reasonably be anticipated. Whilst accepting that even the broadest appraisal processes may still fail to foresee `surprises', there is m uch that can be done to guard against som e o f the consequences of the ubiquitous experience of ignorance and surprise. This insight lies at the heart o f the case studies and is a central elem ent o f what it m eans to be precautionary. At first sight, responding to ignorance may seem to ask the im possible. How can strategies be devised to prevent outcomes, which, by definition are n ot known? Yet the case studies suggest that it is possible to do rather better than in the past. For exam ple, although not quite as simple as it seems, account can be taken o f the potential irreversibility o f actions, even if the consequences might not be known. For halocarbons, polychlorinated biphenyls (PCBs) and methyl tert-butyl ether (M TBE), as artificial chemicals, their very novelty might be taken as a warning sign. Enough was known at the outset regarding their persistence in the environment to serve as another warning. They would also readily disperse to becom e ubiquitous throughout the physical environment -- one more warning. It could have been deduced from the outset that if these substances were Twelve late lessons 171 released into the environment, and if a problem subsequently developed, it would take many years for both them and the problem to `go away'. The Great Lakes case study illustrates long-term hazards associated with other persistent organic pollutants (POPs). For other cases, while it may not have been known in the first instance how irreversible an action was, when this did becom e ap p aren t regulators were often slow to react. It was relatively quickly established that T BT was m ore persistent than first assumed; and the perm anence o f asbestos dust has been known to be part o f the problem for many decades. In neither case did hazard reduction actions take account of the long-tem effect early enough. If persistence and bioaccumulation are used as screening for eliminating potential hazards, then the size and seriousness o f future `surprises' are likely to be smaller. O f course, irreversibility is n ot restricted to the possible consequences o f chemicals. Resistance to antimicrobials, it is now known, is long term. For fisheries, stocks can take a lon g time to recover from a crash, if ever. Consideration o f the irreversibility, or slow reversibility, o f actions is a necessary com ponent of a m ore broadly based approach to the appraisal process. The scale o f the potential hazard, particularly if global, where there is only one `exp erim en tal' m odel, is also a relevant consideration in the appraisal. The case studies also provide some confirmation that potential problem s can be anticipated. For antimicrobials in animal husbandry, confidence over the low risks o f transmission o f antibiotic resistance to hum ans was progressively qualified as new understandings developed. Yet as early as the 1960s the UK expert Swann Committee had anticipated many of the subsequent difficulties. This early exam ple o f sensitivity to possible sources o f ignorance was subsequently overwhelmed by scientific over confidence in the safety o f antimicrobials. Similarly, for PCBs, early results -- such as those obtained in anim al testing in 1937 -- might have been taken as m ore o f a warning. If a harm ful property o f a chem ical is identified, then it may be prudent to consider that this may be correlated with other potentially harmful but less obvious effects. O f the case studies, short-term acute effects that were readily identified preceded -- sometimes by many decades -- less obvious chronic problem s for sulphur dioxide emissions, ionising radiation, benzene, asbestos, TBT and PCBs. However this should not necessarily be taken as a general rule. At the very least the relationship is an asym m etrical one. While the presence o f acute effects may be correlated with chronic im pacts, chronic effects will not necessarily be preceded by acute ones -- as illustrated by bovine spongiform encephalopathy (BSE) and halocarbons. Similarly, harm ful effects seen in wildlife can be useful `sentinel events' warning us o f potential hazards for humans. This calls for integrated ecological and health hazard appraisals. It is also necessary to draw a distinction between a condition o f ignorance that is located at the point of decision-making itself and a condition of ignorance endemic throughout society as a whole. The form er predicam ent, which m ight be term ed `institutional ignorance', refers to a situation where information relevant to the decision may be extant in society, but is not available to the decision-makers. H ere, the consequent `surprises', though they can be serious, may be quite localised. This problem is exem plified in m ost o f the case studies in this book. It can be rem edied by a series o f provisions for m ore effective communication and social learning. The latter condition of `societal ignorance' is m ore intractable. This problem is also exem plified in many o f the case studies (including BSE) and requires rather different remedies, involving scientific research and the fostering of greater diversity, adaptability an d flexibility in decision-making and technological choices. These issues are returned to in the final section o f this chapter. 16.2.2. Research and monitor for `early warnings' General research and long-term monitoring can be dismissed as being too expensive and unfocused. Yet well-planned research and m onitoring are essential to the systematic identification o f areas o f uncertainty. It is, however, necessary to consider how to conduct general m onitoring to increase the prospect o f timely alerts to problem s arising out of ignorance. Awareness of uncertainty and ignorance helps the posing of appropriate research questions for scientific evaluation. It follows that the adequate funding of research and monitoring intended to pick up early warnings is central 172 Late lessons from early warnings: the precautionary principle 1896-2000 to a robust approach to regulatory appraisal o f potential hazards. The case study on halocarbons and the ozone hole carries a m ixed m essage. It was `curiosity driven' general research for another purpose that resulted in the discovery of stratospheric ozone depletion in the Antarctic. The discovery was essentially serendipitous (see next section). While em phasising the value of purely academ ic scientific enquiry, it is hardly a reassuring reflection on the status of deliberate monitoring. Many case studies indicate the value of thorough, long-term, m onitoring. While for asbestos, benzene and PCBs evidence was accumulating of the adverse health effects as early as the 19th century, no role was then played by systematic m onitoring. Data were either not collated (benzene), or became available only in a slow an d rather ad hoc fashion over a period o f many decades. Presumably the assum ption was m ade that if there were harmful effects, evidence would em erge o f its own accord and in good time for corrective action. A different attitude could have prevented harmful effects at an earlier stage. As for the current relevance, if the trend is for h um an actions to becom e geographically m ore widespread and less reversible, then the use o f the `world as a laboratory' becom es increasingly problem atic. It is probably still true that in many cases the underlying assumption rem ains that any m ajor problem s will em erge in good time for corrective action. However, there ought to be m ore ecological and biological surveillance o f the only biosphere we have. It is also a feature o f the case studies that even `critical path' issues, identified at an early stage, were not necessarily followed up in a timely or effective fashion. For BSE, research into a num ber o f crucial issues was not undertaken in the United Kingdom until late in the story. BSE was first identified as a new disease in cattle in 1986, but research to verify its supposed absence o f m aternal transmission in cattle -- im portant to the early position o f the UK Ministry of Agriculture, Fisheries and Food (MAFF) -- was n ot initiated until 1989. Ultim ately it showed that m aternal transmission did occur. Similarly, experim ents concerning the transmissibility o f sheep scrapie to cattle (a favoured hypothesis o f the source o f the disease) were not begun until 1996. No surveys o f the num ber of infectious but asymptomatic cattle entering the food chain have ever been conducted. Yet, throughout this period, reassurances on the part o f the UK governm ent continued prom inently to cite the absence o f evidence, when no evidence was actually being sought. This was a classic exam ple o f `no evidence o f h arm ' being m isinterpreted as `evidence o f no harm '. Similar delays in the conduct o f relevant scientific research are docum ented in other cases. Regarding the routine use of antimicrobials in livestock m anagem ent, concerns over the implications o f the developm ent o f resistance, identified by the 1967 Swann Committee, were not followed up until the 1990s. This was despite longstanding knowledge that widespread antibiotic usage could lead to the rapid development o f resistance. Likewise for asbestos, there was no systematic m onitoring o f health impacts, despite clear warnings and recom m endations on mortality studies between 1898 and the 1920s, and despite the fact that techniques for workforce health m onitoring were well within the capabilities o f the day. T he authors o f the T B T case study concluded that: `T hough frequently under rated, baseline studies play a vital role in the early detection o f adverse trends and may consequently serve the application of precaution'. The case studies on MTBE and PCBs also com m ent on the relative lack o f research on identified causes o f concern. M onitoring alone is n ot enough. A dequate reporting, dissemination and utilisation of research and m onitoring results are also essential. However, neither long-term m onitoring specifically, nor environmental science in general, offers a panacea. They may answer som e questions, but they raise others and the science progresses from relatively simple and often linear proposition to m ore dynamic and com plex `system' science. Research may convert som e aspects o f our ignorance into uncertainty -- and even uncertainties into risks -- bu t this will n ot necessarily be the case. There are examples where research can com pound uncertainty and reveal new sources of ignorance. For fisheries, a Canadian mathematical m odel of the interactions between various fish species suggested that these became more unpredictable as progressively more biological data were incorporated into the m odel. For the Great Lakes, intensive research am plified the uncertainties. It Twelve late lessons 173 progressively raised m ore questions about the possible causes o f the observed bird population crashes. At face value this presents a serious challenge to the notion that further scientific research is necessarily a precautionary measure, or indeed whether broadening out the appraisal process to different disciplines is necessarily precautionary (see Identify and reduce interdisciplinary obstacles to learning, below). O f course, where thorough research genuinely reveals that concerns over particular agents are unfounded -- perhaps by demonstrating a convincing alternative m echanism for an observed effect -- then it is in no way precautionary to persist in restricting the original agent o f concern. Other case studies, such as those on antimicrobials, sulphur dioxide and PCBs, illustrate the opening up o f the research domain and of sources o f ignorance as increasing complexities are discovered. Finally, as the antim icrobials case study illustrates, hazard appraisals should be as specific as possible about the nature o f the scientific question that further research needs to address; the time such research may take and the source o f the funding needed; and the independence o f the relevant organisation carrying it out. The appraisal should also say, as the Swann R ep ort did, whether actions to reduce hazards should h ap p en before or after the research is com pleted. 16.2.3. Search out and address `blind sp ots' and gaps in scientific knowledge Several o f the case studies illustrate blind spots existing within the dom inant discipline involved in the regulatory appraisal process. For halocarbons the chemical m echanism for depletion o f stratospheric ozone was identified in the prestigious jo u rn al Nature in 1974. Nevertheless, this did not prevent regulatory neglect until firm empirical evidence o f actual effects became available. Even then, as has already been noted, the m anner of the emergence of that empirical evidence presents a salutary lesson. The confirm ation o f an Antarctic `ozone h ole' in 1985 was essentially by accident -- a by product of an experim ent conducted for other purposes. A dedicated satellite observation program m e to monitor stratospheric ozone had earlier detected m ajor depletion, but the results were considered suspect and set aside. This provides a pertinent exam ple of where assum ptions adopted in analysis can prejudge the results, leading to serious blind spots for policy-making. In this case these were not only at the centre o f the m ain scientific disciplines involved in regulatory appraisal, but they involved both the theoretical mechanism and the empirical evidence. A blind spot within a scientific discipline was also apparent in the case of agricultural antimicrobials. Evidence available in 1968 to the U K Swann Committee allowed explicit anticipation of a series of potential environmental, anim al welfare and hum an health problem s that were to becom e of pressing concern. Although initially influential, the Swann recom m endations became m arginalised over the ensuing decades. H ad they been im plem ented and sustained, they m ight at least have mitigated the difficulties now recognised. Similarly, for MTBE the key problem atic property of persistence was apparent at the outset, and this m ight reasonably have been expected to lead to m ore questions over the potential environmental problem s of widespread use o f this chemical than were actually raised in the formal regulatory process. In the case of TBT, rates of degradation were based on assum ptions about the nature o f the marine environment that were evidently incorrect for many areas. For horm ones as growth prom oters there was a failure to appreciate that young children with low natural levels o f oestrogens were the likely `at risk' group. Radiation risk estimates o f typical doses long overlooked the uncertainties resulting from their derivation largely from the health records o f the survivors o f the atomic bom bings in Jap an at an atypical high dose and dose rate. Another blind spot can occur where the adoption o f a new practice is held, o f itself, to have solved historic problems. For asbestos successive claims were m ade that past health impacts were due to conditions that had been superseded. The long lineage of claims that `the disease is n ot so likely to occur (in fu tu re )' can be traced back to 1906. With each successive minor incremental improvement in conditions, the persistent risks associated with the new conditions would then in turn take further decades to becom e evident. Remarkably, when the recognised magnitude of the health effects eventually prom pted asbestos substitution, this was initially attem pted with fibrous minerals that shared some of the properties that had led to the effects o f asbestos (such as fibre size). In their turn, these substitutes also 174 Late lessons from early warnings: the precautionary principle 1896-2000 came -- eventually -- to be recognised as sources o f essentially similar, if m uch lower, risks. For sulphur dioxide, the building of taller smokestacks, while helping to am eliorate local effects, did not address the wider problem arising from cumulative emissions and long-range transport. A more precautionary approach therefore m eans systematically searching out blind spots at the heart o f the disciplines historically involved in the regulatory process. This will be easier if m ultiple relevant disciplinary and other sources of knowledge are engaged, helping to stimulate the sometimes uncom fortable interactions that will be m ore likely to expose false assumptions and other questions. This leads directly to the next lesson. 16.2.4. Identify and reduce interdisciplinary obstacles to learning Where effects that are the domain of a particular specialist field may initially be m ore pronounced, or discovered at an earlier stage, this can lead to a situation where regulatory appraisal becom es unduly dom inated by, even `captive' to, a particular discipline. This can lead to a form of `institutional' ignorance, as opposed to the society-wide ignorance discussed above. For both asbestos and ionising radiation the setting o f standards was strongly influenced by the preoccupation o f medical clinicians with im m ediate acute effects. In both cases, the toxicology and epidemiology of long term chronic effects rem ained relatively neglected. The introduction o f M TBE was based on bodies of knowledge concerning engines, combustion and air pollution. The water pollution aspects associated with persistence and significant taste and odour problem s were essentially disregarded, though the inform ation was available. For sulphur emissions the regulatory appraisal was initially focused on hum an health concerns. When ecological effects became apparent, a regulatory process constructed to address health issues experienced problem s assim ilating and reacting to these. Similarly, for the use o f horm onal growth prom oters in livestock, the initial emphasis was also on hum an health effects. Although concerns over the im pact on wildlife were raised, they initially appeared to attract little attention from the regulators. In contrast, for both livestock antimicrobials and BSE, considerations of the human impacts were initially m arginalised by the regulatory focus on veterinary science. Indeed, for antimicrobials this concern underlay the early 1968 recom m endation, by the Swan Committee, that a single advisory process be established which `should have overall responsibility for the whole field of use of antibiotics and related substances whether in man, animals, food preservation, or for other purposes'. This recom m endation was not taken up, in the United Kingdom or elsewhere, for many years. In the case o f BSE, U K veterinary officials considered the possibility of transmissibility to hum ans as acceptably slight. This contrasts with the attitude in the United States, where the possible link between sheep scrapie and human Creutzfeldt-Jakob disease (CJD) had been regarded as a possibility since the 1970s, when the entry o f infected anim als into the food chain was banned. 16.2.5. Ensure that real world conditions are fully accounted for `Real w orld' conditions can be very different from theoretical assumptions, and these differences can have serious consequences. In principle this problem is well recognised, and it is possible to considerably reduce exposure to oversights o f this type. Yet in practice the case studies reveal a variety of incom plete assessments, resulting in erroneous regulatory appraisals and decisions. In the hum an sphere it is often assum ed that technologies will perform to the specified standards. Yet real life practices can be far from ideal; and it may be a long time before we realise that this is the case. Som etim es actions are taken that appear to be in defiance of prior experience. The leakage from petrol station storage tanks was underestimated in the U S regulatory appraisal o f M TBE and so led to an underestimation of the resulting exposures. Although storage tanks can be redesigned to reduce the chance o f leakage, this benefit can be lost by incorrect installation. For PCBs it was assum ed that these could be constrained within `closed' operating systems. This proved im possible, resulting in accidents such as Yusho and Yucheng, losses from poorly maintained equipm ent and even illegal disposal into the hum an food chain. Similarly optimistic assum ptions as to the perform ance of engineered containment equipment, or the efficiency of decom m issioning, also played a role in reducing the effectiveness o f halocarbon Twelve late lessons 175 control m easures. While som e aspects o f public exposure to benzene have received rigorous control, these are not necessarily the routes o f greatest exposure. The failure to develop mitigation practices, or even warnings, regarding benzene exposure via petrol is one notable exam ple. For the clinical use o f radiation, the im portance o f establishing, and using, the optim al dose for any exam ination can be traced back to 1949. Yet the dosage for the same exam ination in different hospitals can still vary by a factor o f 100. And for growth prom oters scientific advisory committees such as the WHO / FAO Join t Expert Committee on Food considered only a restricted range o f options, including ju st those circum stances relating to authorised use, and assessments of individual growth prom oters, rather than in combination. They also gave little attention to the misuse o f growth horm ones, such as higher doses than recommended, inappropriate injection sites, failure to withdraw horm one implants from slaughtered animals and shortened withdrawal periods. A gap can also open between assumptions and the real world when applications change without corresponding adjustments to their regulation, such as the expansion o f the use o f asbestos in consumer products and housing as well as boiler insulation. Asbestos is also notable for the long delay before it was acknowledged that real world conditions m eant that users (or even local residents around a factory), as well as workers, could be at risk from exposure. Significantly, the 2001 World Trade Organization Appellate Body ruling on asbestos concluded that `controlled use' risk m anagem ent could not be relied on to protect workers' health in real world conditions. In som e cases there will be deliberate non compliance. Apart from the illegal disposal of PCBs already noted, this included poor practice in the agricultural use of antimicrobials that increased the spread of resistance, and the circumvention of attem pted regulation of fisheries that helped deplete stocks. Similarly, unrealistic assum ptions were m ade about the im plem entation o f changes to UK slaughterhouse practices as a crucial part of the response to the BSE crisis. Acute hum an effects observed in Puerto Rico and Italy were blam ed on the illegal or incom petent use of anim al growth prom oters. Similarly, the sm uggling o f h alocarbons is threatening the effectiveness o f global controls. False assum ptions about the real world also affect our interpretation o f events in the natural environment. The PCBs case study describes the surprising potential for these chemicals (along with certain other POPs) to concentrate disproportionately on a global scale in high-latitude regions. Also for PCBs, in the real world bioaccum ulated PCBs were found to be disproportionately m ore toxic, with the result that the effects were greater than those indicated by experiments using the original commercial formulations. For fisheries, stock assessm ent m odels tend to concentrate on single stocks, downplaying the interactions between stocks or other marine species. There are enormous, perhaps insoluble, problem s in broadening the scope using the stock m odelling approach. Nevertheless, the single stock approach does not reflect actual conditions. The unexpected consequences of halocarbon degradation in the stratosphere is an oth er exam ple o f the unanticipated complexity of the real world. In these, and many other cases, the behaviour o f real natural systems can prove far removed from the standard assum ptions m ade in regulatory appraisals and consequent decisions. The issue o f horm one growth prom oters illustrates how the real world also contains groups o f particularly sensitive individuals (young boys, for exam ple) who may react differently to exposures com pared with assum ptions based on the `average' response. 16.2.6. Systematically scrutinise and justify the claim ed `p ro s' and `cons' One feature o f debates about environmental risk is how all protagonists typically call for greater and m ore systematic attention to be paid to the claimed benefits o f the technologies in question. Some are concerned that undue attention on, or inaccurate appreciation of, the risks might diminish recognition o f the anticipated benefits of their favoured technologies. Others are concerned that the justification for new technologies, and claims as to their efficacy, are insufficiently critically scrutinised. Either way, both positions effectively call for the regulatory appraisal process to m ore deliberately and systematically examine the claims m ade about the benefits o f a technology or product, including an identification and assessment o f the conditions under which the 176 Late lessons from early warnings: the precautionary principle 1896-2000 claimed benefits might, or might not, m aterialise. accrue from allowing depleted stocks to recover. Problems can arise from the partial assessment of environmental protection m easures. Raising the height of smokestacks and shifting to smokeless fuel were an effective response to the serious episodes of respiratory illness from European sulphur emissions and urban air pollution in the 1950s. Yet this very effectiveness may have inhibited attention to the problem s o f long distance transport o f acid gases and the consequent acceleration of environmental problem s in vulnerable areas. This was a classic case o f an end-of-pipe `solution' only creating other, less visible hazards. A more integrated approach eventually reduced all the problem s. Similarly, because M TBE apparently prom ised a simple solution to the serious environm ental problem s caused by lead emissions from m otor vehicles, the environm ental problem s associated with taste and odour in com bination with persistency in groundwater were perhaps more readily overlooked than might otherwise have been the case. The introduction of horm onal growth prom oters can also be considered to be an example of narrow and partial assessments of the pros and cons. Hum an health issues, possible environmental impacts and farm anim al welfare were `not given significant attention', according to the case study authors. Another dimension of the sulphur dioxide case study was that the distribution o f the pros and cons between the United Kingdom and the Scandinavian countries was very different, or at least was perceived to be very different. In relation to the effects on the natural environment, Sweden seem ed to be suffering the most. It was only with the realisation o f m ajor dam age to buildings in the United Kingdom -- caused by air pollution -- that both countries realised that they were suffering from the im pact o f acid emissions. In two cases, BSE an d fisheries (the C anadian cod collapse), the pros and cons of taking early action were identified prior to the event. But because the uncertainties were significant, and the financial costs of action high, limited action was taken. This resulted in the far greater costs (many o f which were anticipated) when the m easures proved inadequate. In the case o f fisheries, even in the absence of stock collapses, major economic, social and environmental benefits One might think m edicines would be validated for efficacy. However, for DES, the data from 1953 trials showed that D ES was ineffective as a means o f reducing risks of spontaneous abortion in certain groups o f m others and that it was positively harmful. This seem s not to have been appreciated at the time. As a result the rate o f reduction in use o f the drug was m ore gradual than m ight have been the case; there was no regulatory action; and marketing continued unabated. It was not for another 20-30 years that use of this drug was actually banned in different countries, in response to the discovery o f an increase in a rare cancer o f the vagina in daughters of treated women. Had greater critical attention been paid at the outset to the claims o f efficacy, then som e o f these second-generation cancers might have been avoided. A step beyond passive assessment of the pros and cons is actively directed prior to justification. Ionising radiation is a rare exam ple o f such a `ju stification prin ciple', developed by the International Committee on Radiological Protection in the 1950s. This was a response to the burgeoning o f a variety of dubious or ineffective uses of radioactive materials (for instance in countering ringworm, fitting shoes for children, the cosmetic removal of hair and the treatment o f m ental disorders). Yet even with this criterion, the degree to which exposure to ionising radiation is ju stified by the benefits remains open to question. Surveys of radiography practices over the past decade or so conclude that, while doses have reduced drastically, a large proportion o f m edical Xrays are still o f doubtful clinical use. The reduction o f the use o f antimicrobials, initially in Sweden, can also be considered to be an exam ple o f the wider assessm ent o f the pros and cons. The appropriate allocation o f costs and benefits is an essential pre-condition for optimising resources between technological options. A host o f risk m anagem ent measures are available under the broad remit o f the `polluter pays principle' -- including taxes, subsidies and liability regimes, which offer a way to ensure m ore equitable social distributions of the costs and benefits associated with risk governance decisions. The failure, in cases like asbestos, halocarbons and PCBs, to reflect full Twelve late lessons 177 environmental and health costs in m arket prices gave these products an unjustifiable advantage in the m arketplace. This in turn helped to keep technically superior substitutes off the m arket for longer than was optim al from society's p oin t o f view. Although the mechanisms for the internalisation o f external environmental costs and the practical im plem entation of liability regim es are controversial, such m easures are essential if both efficiency and equity objectives are to be addressed effectively. 16.2.7. Evaluate alternatives and promote robust, diverse and adaptable solutions Even where the pros are scrutinised alongside the cons, if attention is restricted simply to isolated technologies or products then im portant practical insights may be m issed. O ne concern is that, once a technological com m itm ent is m ade, a h ost o f institutional and m arket processes act to reinforce its position, even if markedly inferior to potential alternatives. So while in principle the function o f MTBE m ight be substituted by alternative oxygenates such as bioethanol, improved engine technology or an increase in the octane rating o f the fuels themselves, little formal scrutiny appeared to have been undertaken at the time o f adoption o f MTBE. In California, the search for a successor to MTBE has recognised that all the proposed alternatives m ust be fully evaluated. The authors o f the TBT case study note that `B roader consideration o f problem s may give rise to m ore beneficial solutions than simple `chemical for chemical' substitution', and give exam ples o f such alternatives. For ionising radiation, substitutes for the diagnostic use o f X-rays rem ain underutilised. The ozone-depleting properties o f second-generation CFC substitutes were perhaps also unduly tolerated, simply because o f their relatively low im pacts when com pared with the original substances, and the existence o f more benign substitutes or alternative approaches was not properly looked at. And while effective husbandry practices not involving the routine large-scale administering of antimicrobials are increasingly used in a num ber of European countries, these more benign alternatives are not actively prom oted. This raises some challenging issues about the relationship between regulatory processes and product developm ent by private companies. The prom otion and production o f alternatives needs to take place within a culture o f `eco-efficiency', `clean production' and closed-loop m aterial flows so as to minimise the size o f any future `surprises' in the use and im pact of technologies. These issues are returned to in the concluding section o f this chapter. 16.2.8. U se `lay' and local knowledge as well as all relevant specialist expertise The importance of ensuring that regulatory appraisal includes the full range o f relevant disciplines has already been covered. A related but distinct lesson concerns drawing upon knowledge held by lay people. These may include industry workers, users o f the technology an d people who live in the locality or, because o f their lifestyle or consum ption habits, stand to be most strongly affected. The point is not that lay people are necessarily more knowledgeable or environmentally committed. Rather the benefit o f attending to lay knowledge rests in its com plem entary character, its som etim es firmer grounding in real world operational conditions -- as already discussed -- and the associated independence from the narrow professional perspectives that can be a downside of specialist expertise. Often too, lay knowledge o f a technology or risk may be based on different assum ptions abo u t what is salient, or what degree o f control is reasonable to expect or require, whereas technical specialists may simply respond to granted authority without further reflection. O ne prom inent contribution from lay knowledge relevant to the regulatory process concerns workplace awareness of emerging patterns o f ill health. T he histories o f usage of asbestos and PCBs provide examples where workers were aware o f what regulators subsequently recognised to be a serious problem . Similarly, local com m unities may becom e aware of unusual concentrations of ill health before the authorities, such as the Love Canal exam ple cited in the Great Lakes case study. Another form o f lay knowledge concerns remedial measures. Fisheries highlight several aspects and, although fishers can be less precautionary about stock depletion than others, there are many exam ples where fishers wish to act in a precautionary m anner but are prevented from doing so because of a 178 Late lessons from early warnings: the precautionary principle 1896-2000 systems failure. T here is an increasing emphasis in Canada and elsewhere on the need to involve fishers in m anagem ent, and take full account o f their knowledge and perspectives. Similarly for livestock antimicrobials, Swedish farm ers' knowledge o f alternative anim al husbandry techniques allowed them to prom ote anim al health and growth without the large-scale use o f antimicrobials. Not only did they bring valuable insights to the regulatory debate, but they were able to undertake voluntary controls in advance o f regulatory requirem ents. Another aspect to emerge from this case study is the need for knowledge to be widely based in order for less harm ful alternatives to be effectively im plem ented -- indeed it is fundam ental to ensuring that the existence of possible alternatives is actually recognised. 16.2.9. Take account o f wider social interests and values G athering available knowledge is n ot the only reason for opening up the appraisal process. Historically there is little doubt that social and political conflicts can be aggravated by a regulatory preoccu pation with expert judgem ents and a lack o f attention to public perspectives and values. In part this relates to a wider assessm ent o f the pros and cons, as already discussed. The m eeting o f specialists and interest groups with different viewpoints can be productive, helping ensure a reflective approach, and allowing the implicit assumptions o f all parties to be tested. The Swedish farm ers in the antimicrobials case study show how lay views can help ensure that the regulatory process remains (or becomes) attached to prevailing ethical and socio cultural values. Another aspect o f lay knowledge is where workers know that real practices do not m atch the theoretical assum ptions o f risk assessors. The UK slaughterhouse rules over BSE separation o f `specified bovine offal' from m eat were widely flaunted, yet failures in the inspection regim e m eant that this was not drawn to the attention o f the regulators. In this case, workers in the industry in question were apparently better informed about the operational realities than were high-level regulatory advisers and officials. O f course lay knowledge should be subject to the same intensity o f critical scrutiny as specialist expertise. Lay perspectives are not im m une to the pitfalls and difficulties noted in these conclusions -- and may be more vulnerable. One exam ple is the `pensioners' party fallacy' am ongst asbestos workers who pointed to the presence o f healthy p ension ers at the firm 's Christm as party as evidence of the apparent harm lessness of asbestos. Nevertheless, workers, users and neighbours evidently can bring im portant inform ation to the regulatory appraisal process, requiring greater attention to be devoted to the development of methods to enable those groups with potentially valuable knowledge to provide this, and for this to be fully taken into account. Such broadening o f the knowledge base can strengthen the appraisal, improve governance and democracy, and enhance the acceptability and legitimacy o f the process. It is im plicit that the intuitions em bodied in public values may sometimes prove quite robust in relation to the fram ing o f the regulatory science itself. An aversion to situations far outside the bounds of normal experience, or at least a desire to proceed with caution, can certainly be defen ded as a rational response to uncertainty. A key feature o f the public reaction to the em erging evidence o f BSE in the late 1980s was the surprised revulsion that rum inants were being fed on offal and bodily wastes. It seem s likely that avoiding offal in rum inant feed would have at least significantly limited the scale of the subsequent BSE and CJD problem s. Similarly, had w idespread public misgivings over the use o f antim icrobials in anim al husbandry been heeded, the development o f resistance would have been held in check. The fisheries study illustrates how the objectives of a precautionary approach -- to prevent stock collapses, to m aintain m axim um sustainable yields or to ensure the protection o f other species -- depend on the value judgem ents o f the interest groups. 16.2.10. Maintain regulatory independence from economic and political special interests A m ajor elem ent o f a broader approach to regulatory appraisal is to ensure that an appropriate distance is m aintained between those responsible for regulatory appraisal and various contending interest groups seeking to influence their decisions. It is a necessary part of the regulatory process that claims over the pros and cons should be actively advanced and argued by interested Twelve late lessons 179 parties. It then becomes a matter for independent, accountable institutions to adjudicate between the contending claims. There is evidence in the case studies that interested parties are often able to unduly influence regulators. As a result decisions that might reasonably have been m ade on the basis o f the available evidence were not taken. Benzene was dem onstrated to be a powerful bone marrow poison in 1897; the potential for acute respiratory effects of asbestos was first identified in 1898; an d the first cases o f PCB-induced chloracne were docum ented in 1899, with effects on workers known by the late 1930s. Yet it was not until the 1960s and 1970s that significant progress began to be m ade in restricting the damage caused by these agents. One factor in the slow U K response to BSE was that the governm ental regulatory body was responsible first to the industry and only second to consum ers. Similarly, the temporary lifting o f the ban on DES as a growth prom oter in the United States in 1974 followed strong pressure from the farming lobby, and occurred despite the availability o f alternatives. It is also notable how poorly substantiated some o f the sustained `refutations' o f critical findings have been. This was true o f the identification o f PCBs as widespread environmental pollutants in the 1960s and of the U K response to the im pact o f acid deposition in the 1970s and early 1980s. Asbestos provides a clear example of persistent obstruction and m isinform ation by vested interests, and of drastic miscalculation in the wider regulatory process. The accumulation of medical and pathological evidence, while having very little im pact on regulation, was sufficient to m ake parts o f the U S an d C anadian insurance industry wary about providing cover by 1918; a precautionary attitude that ironically was not maintained as evidence accumulated -- an error that was to cost the insurance industry billions of dollars. Although benzene had been unequivocally identified as a human carcinogen, human health protection m easures continued to be im peded by claims over a lack of evidence for animal carcinogenicity. Similarly demonstrably erroneous claims (based, for example, on basic statistical errors) were later repeated regarding evidence o f effects arising from very low benzene exposures. Even where the evidence is essentially not contested, data and reports are sometimes suppressed, or publishers intimidated, as reported in the G reat Lakes case study. For antimicrobials, research that m ight have been expected to reinforce a critical position was delayed. Efforts were m ade in the case o f BSE and asbestos unjustifiably to discredit independent critics. In the Californian sardine fishery in the 1930s, critical agency scientists were dismissed. `Shooting the m essenger' has been a typical response to those bearing disturbing news ever since Galileo, but it rarely, if ever, prom otes societal welfare. BSE provides an explicit exam ple of independent expert advisory committees who advise the regulators subjecting themselves to self-censorship, based on a ju d gem en t o f what was `realistic' or `achievable'. The Southwood BSE Committee considered in 1988 that a ban on the use o f all cattle brains in the hum an food chain m ight bejustified on scientific grounds, but it was considered not to be a politically feasible option. Regulatory appraisal frequently fails due to the dependence of risk assessment on information produced and owned by the very actors whose products are being assessed. Independent sources of risk information are a necessary, if not sufficient, condition of independent, rigorous and trustworthy regulatory appraisal. Independent inform ation about risks and possible risks was often lacking in the cases studied. In som e of the exam ples cited above, for example benzene, PCBs, asbestos, halocarbons and DES, knowledge concerning hazards was available long before any regulatory action was decided. N ot all o f these cases demonstrate the delaying or distorting effect of non-independent sources of such knowledge. Nevertheless, the em ergence of hazard evidence can be accom panied by vigorous though often low public profile interpretative jockeying to try to justify inaction. W hat is very difficult to dispute is that had such information -- from wherever it em erged -- been taken to be the rightful subject of control and disseminated by an in depen den t public body whose very raison d 'trewas the provision o f salient public policy information, then the corresponding interpretative debate about policy intervention would have been m ore open, and m ore pluralistic. Diverse interests would have been m ore equitably, thoroughly and probably m ore rationally represented. Independent information institutions, allied 180 Late lessons from early warnings: the precautionary principle 1896-2000 to corresponding rights, resources and responsibilities, are thus a key elem ent of authentic regulatory independence and robust governance an d appraisal. This is increasingly being recognised, for example by the shifting of advisory committees from `producer' directorates in the European Com m ission (for exam ple, agriculture) to the Health and Consumer Directorate. The setting up o f independent food agencies in som e M ember States and at the EU level also reflects this concern for m ore independent hazard appraisal institutions. 16.2.11. Identify and reduce institutional obstacles to learning and action The progressive unfolding o f the asbestos, benzene and PCB episodes from the late 19th century provide various exam ples of how short-term horizons, notably governm ent and business cycles, can militate against social welfare in the m edium and long term. However, institutional obstacles against timely protection o f health and the environm ent can take other form s as well. The case studies illustrate three other areas which can present difficulties: those resulting from periods of transition (for example between succeeding elected administrations) or from tensions between different departm ents or levels o f governm ent and `their' agencies, and the issues that can arise from differing national approaches. One instance where a change in political administration may have contributed to poor im plem entation o f prior knowledge is provided by the case o f BSE. An official UK commission in 1979 recom m ended the setting o f minim um processing standards in the rendering industries. A new adm inistration later that year decided to withdraw the resulting proposed regulations, deem ing them to be an unnecessary burden on industry. It is not clear to what extent such tighter standards might actually have inhibited the later BSE outbreak, bu t it is notable that the im plem entation o f standards o f this sort featured prominently am ong that sam e govern m en t's later responses to the BSE crisis in 1996. Similarly, in the Californian sardine fishery, what was in effect a precautionary program m e o f stock conservation was reversed with a change in governm ent. T he Californian fishery also provides a clear exam ple o f tension between different levels o f government, albeit dating back to the 1930s, where the state agency's recom m ended precautionary action was strongly opposed by the U S Federal Bureau o f Fisheries on the grounds that it would unduly inhibit com m ercial activity. An exam ple of similarly fraught communication between different governmental departments may be found in the case o f BSE in the United Kingdom. Here, the Departm ent of Health was not inform ed by MAFF about the em ergen ce o f the new disease until som e 17 m onths after MAFF was first alerted, and only then because its assent was required for a decision not to remove clinically affected cattle from the hum an food chain. Regulatory agencies can face a difficult relationship with governm ent, who can exercise m ore or less subtle influence, even where the agency has explicit independence. Difficulties can also arise where the regulatory agency becom es part o f the issue through its past decisions. Both situations are illustrated by BSE, where identification and recognition o f the problem was influenced and delayed as a consequence of wider policy considerations (the economic im pact on farming) and the concern that lack of consistency would underm ine the credibility of government and agencies alike. Scepticism over the scientific results obtained in other countries was a key feature o f the attitude o f U K regulatory authorities to the sulphur emissions in the mid-1980s. Although research results suggesting both the seriousness o f the environmental effects, and the identity o f the responsible agent, were accepted in Norway in 1976, it was not until 1985 that the scientific case for a causal link was acknowledged in the United Kingdom . As already noted, during this period the distribution o f pros and cons was perceived to be unevenly balanced between the contesting parties, and it would not have been surprising if this influenced the level of scepticism on the part of the United K in gd om . Similar tensions have existed between Sweden and the EU in the case of antimicrobials, where different regulatory systems were based on different presum ptions. W hatever one may consider to be the pros and cons o f the argument, EU m embership requirem ents were an institutional obstacle to the continued im plem entation o f a national policy. At a different level, global problem s clearly require a global response. This is brought into focus by the circumstances surrounding the T B T issue. Here, effective action to Twelve late lessons 181 regulate the use of antifouling biocides on the hulls o f wide-ranging comm ercial vessels is dependent on agreem ent within global institutions such as the International Maritime Organization. For BSE it is notable that the sam e scientific evidence was available to the United States and the United Kingdom in the mid-1970s concerning the transmission o f scrapie and possible links between scrapie and CJD. However, it led the U S Departm ent of A griculture, bu t n ot the U K MAFF, to decide that scrapie-affected animals should not be used in hum an or anim al foods. The different timing o f national decisions on the use of DES both as a pharmaceutical and as a growth prom oter in animals also illustrates how widely decisions can vary even though based on the same information. In these, and other examples, institutional obstacles appear to have played a role. 16.2.12.Avoid paralysis by analysis The general tenor o f the lessons so far is to `know m ore', for exam ple by searching out blind spots within disciplines, reach ing out to other disciplines, accessing lay and local knowledge, and taking account o f wider social perspectives. O ne respon se is to ask how m uch inform ation is enough to justify action to reduce potential hazards. An obvious concern is that o f the danger o f paralysis by analysis where either information overload or lack o f political will leads to a failure o f timely hazard reduction measures. One exam ple is the evidently anti precautionary `straightjacket' im posed upon U S benzene regulation by a Suprem e Court decision, which required layer upon layer o f additional information before regulatory action to reduce risks was possible. Experts have often argued at an early stage that `we know en o u g h ' to take protective action. For antimicrobials the UK Swann Committee in 1969 concluded: `despite the gaps in our knowledge... we believe... on the basis o f evidence presented to us, that this assessm ent is a sufficiently sound basis for action... the cry for m ore research should not be allowed to hold up our recom m endations'. Other case studies, such as asbestos and BSE, suggest that more, or better-targeted, research, at an earlier stage, would have helped minimise future costs. Similarly, for fisheries, the Ecosystems Principles Advisory Panel to the US Congress concluded: `T here will always be unm easured entities, random effects, and substantial uncertainties, but these are not acceptable excuses to delay im plem enting an ecosystembased m anagem ent strategy.' On the other hand, the Great Lakes study argued that m ore uncertainty was generated as the field was opened up to new disciplines, and that this indeed could lead to paralysis by analysis if a precautionary approach was applied. The Great Lakes case raises a very im portant issue that is best understood by appreciating the significant distinction between this and the other case studies. Most deal with concerns raised in relation to particular identified agents. The issue in such cases is the `prospective' assessm ent o f risks, starting with an agent and seeking possible effects. The Great Lakes case, however, is an exam ple o f a rather different, `retrospective' process. This starts with the docum enting o f a series of manifest health or environmental effects and seeks the identification of possible agents. The broadening out of prospective appraisal is precautionary because it has the effect of focusing attention on a greater num ber o f possible effects. The broadening out of retrospective appraisal, however, may have the effect o f raising uncertainties over the basis for regulating individual agents. As has been pointed out in discussing m onitoring, it is in no way precautionary to persist in restrictions of the wrong agent. However, the precautionary principle applies as m uch to uncertainties over agents as to those over effects. If a broad-based retrospective process raises scientific uncertainties or ambiguities about the grounds for targeting a particular agent, the precautionary principle may nonetheless be invoked, entirely legitimately, to defend continued action on this agent, until such uncertainties are resolved. The fact that the precautionary principle was not invoked in this fashion in the Great Lakes case is m ore a reflection o f the value judgem ents in the prevailing legal and socio political context than it is o f intrinsic inconsistencies in the concept o f precaution itself. The prospects for successful appeal to the precautionary principle will depen d on the culture within which the appeal is m ade. If the culture is n ot prepared to act even when there is dem onstrable evidence of cause and effect, then appeals to precaution are unlikely to succeed. In reality the regulatory and wider culture o f a society may vary between these extremes, and even 182 Late lessons from early warnings: the precautionary principle 1896-2000 between different regulatory sub-cultures (cf. the differing US approach in the case studies on BSE, fisheries, MTBE, benzene and the Great Lakes). W hether or not the need for more inform ation risks `paralysis by analysis' or is merely part of a `prudent and careful evaluation' o f the situation, will be influenced by the individual, social or interest g ro u p 's assessm ent o f the likely pros and cons as they im pinge upon them. If the benefits o f taking early precautionary action are large, and the adverse consequences deem ed relatively small, and if these are evenly spread across interest groups, then early action is likely to be considered. If the advantages are less certain and, like the costs, are very unevenly spread across different interest groups, or time periods, then reaching consensus about the appropriate am ount o f research, or about actions to reduce hazards, will be m ore difficult. O f course, it may be that in som e cases it will be necessary to severely curtail or end innovation in a particular field or technological direction where societyju d ges the risks to be unacceptable. B ut there is an enorm ous difference between the discouraging o f a particular innovatory pathway, and the channelling o f innovation into alternative routes. As illustrated by asbestos, halocarbons, PCBs and antimicrobials, the curtailment o f a particular option may actually serve to foster and intensify innovation in other areas. It may also provide a competitive edge to the economies o f the countries leading such innovations. The intelligent use o f foresight and the precautionary principle may therefore not only reduce overall costs to society o f some economic developments, but can stimulate innovation, encourage better and m ore systems-based science, and improve public decision-making. Som e o f these wider implications of a more precautionary approach to potential hazards and innovation are further elaborated in the next section. 16.3. The wider implications of precaution T he 12 broad lessons for policy-making on risk ju st described have been developed under a num ber o f criteria. First, they are well grounded in the empirical detail of particular case studies. Second, they are sufficiently general in nature to be found relevant to virtually any risk m anagem ent problem . Third, taken together, they address a balanced and fairly comprehensive range of considerations, spanning a large part o f the current debate on the m anagem ent of hazards and the im plem entation o f the precautionary principle. Finally, although necessarily general in nature, they are sufficiently concrete to inform practical policy m easures and institutional procedures, even though the precise nature o f these in a given case will necessarily be defined by variable local circumstances. T hat said, it is n ot possible to anticipate in one short discussion, or one set o f `lessons', the full range and diversity o f detailed practical precautionary measures and procedures, still less to explore the specific contextual considerations bearing on their application. Such matters are the subject of a wide and burgeoning literature (see, for exam ple: O 'Riordan and Cam eron, 1994; H arding and Fisher, 1999; Raffensperger and Tickner 1999; Stirling, 1999; O 'R iordan et al., 2001). A lthough n ot always labelled as `precau tion', many o f these lessons have been strongly developed and elaborated by a variety o f influential policy studies in industrialised countries over recent years. In the United States, for instance, the seminal study by the National Research Council (NRC), `Understanding risk' (NRC, 1996) and the subsequent report by the presidential com m ission (O m en et al., 1997) docum ented the limitations of conventional narrow risk assessment and highlighted the im portance o f interdisciplinary, lay knowledge and divergent stakeholder viewpoints in the characterisation o f risk issues and o f appropriate assessment approaches. The 1998 report of the UK Royal Commission on Environmental Pollution developed this theme (RCEP, 1998), underscoring the potential significance of uncertainty and different `fram ing assum ptions' in the shaping and interpretation of formal appraisal. In France (Kourilsky an d Viney, 1999) recom m endations on im plem entation o f the precautionary principle stressed the need to organise systematically national expertise capacities, including both scientific and technical expertise, alongside economic and social expertise. In Germany, the importance o f m ore broad-based discursive procedu res is recognised in the m ajor report o f the Germ an Advisory Council on Global Change Twelve late lessons 183 -- WBGU (WBGU, 2000). The development o f the Swedish chemicals policy is based on recognition o f many of the lessons noted here concerning the fundam ental limitations o f risk assessment, particularly the use of persistence and bioaccum ulation as `proxies' for unknown but possible impacts. Various specialists have also detailed the general points o f these lessons in respect of the structural limitations of EU regulatory risk assessment of genetically modified organisms (GMOs) (van Dommeln, 1997). In the light o f these cogent critiques o f the overly reductionist and narrow character of European scientific risk assessment o f GMOs, it is worth rem em berin g that this sam e EU risk assessm ent is regarded by U S counterparts as irrationally exaggerated when m easured against what they define as `soun d scien ce'. At any rate, it is abundantly clear that in m ost of these issues, a scientific fram ing o f the questions as if they were all resolvable by existing or available knowledge o f risks is radically incom plete -- and not only in respect o f the answers, but m ore particularly in respect o f the questions which are deem ed salient to address even if answers are difficult or impossible to achieve. In the light o f recognition o f these wider dimensions, some practical yet m ore broadly based institutional procedures such as consensus conferences and scenario workshops have been developed in Denm ark an d the N etherlands to try to articulate public questions and values with respect to scientific presum ptions about the answers, and these have been exported widely over recen t years (Renn et al., 1996). In the U nited Kingdom , the advent o f new `strategic comm issions', on food, hum an genetics, and agricultural genetics and environment, is a recent innovation that opens up the risk policy process in the way suggested by som e o f these lessons. Detailed policy appraisals in areas such as BSE (Phillips et al., 2000) an d m obile phones (IEGMP, 2000) have seen various of these lessons explored in some detail, with specific recom m endations on how to handle issues such as institutional conflicts o f interest and unrealistic expectations o f the role o f science as a touchstone, or arbiter, o f ultimate truth. The twelve lessons should be useful aids to policy-making. However, one should avoid caution against over-reliance on any single set of prescriptions for what constitutes a `precautionary approach ' in any given case. The im plem entation o f the precautionary principle involves drawing on an entire spectrum of methods, procedures and instruments, many falling on the same continuum as orthodox risk m anagem ent approaches. The business of deciding upon one particular set of responses rather than another must necessarily remain, at some level, an essentially political business -- subject to all the norm al processes of rational policy deliberation, professional review and democratic debate and accountability. Finally, it is worth rem em berin g that the precautionary principle began life as an environmental policy approach but it has been rightly recognised that constructive and effective environmental policy dem ands integration of environmental objectives into all areas o f decision-making and technological, as well as public policy, com m itm ent. T h erefore it is axiom atic that a proper framework for m easured and effective precautionary policy-making must encom pass these wider domains, even if only indirectly. Beyond the twelve lessons, som e general principles and specific practical m essages are proposed, covering the relationships between precaution and science, precaution and innovation, and precaution and governance. 16.3.1. Precaution and science The precautionary principle raises im portant issues for science. Som e are to do with what many might perceive as the mechanics o f science, such as issues of statistical proof and the framing of hypotheses. But it also raises some very fundam ental interdisciplinary issues regarding the very nature of knowledge-gathering. `Statistical proof' and thefram ing of hypotheses For the environmental sciences, issues of proof and precaution often come up via the interpretation o f statistics. Given that it is impossible to prove a hypothesis (such as `all swans are white'), only to disprove it (the discovery o f Australia, an d with it the Australian black swan), the statistical workaround is to attem pt to falsify a `null hypothesis'. This is the opposite o f the hypothesis of interest, for exam ple that chemical concentrations have increased. If an increase exceeds, by som e arbitrary threshold, that which m ight be expected by chance fluctuation, the null hypothesis is rejected and it is assum ed that an increase has occurred. However it is still possible that the result is a freak an d that the increase is 184 Late lessons from early warnings: the precautionary principle 1896-2000 Table 16.1. due to chance. This is known as a `Type I ' statistical error, or a `false positive'. Traditionally there has been a strong emphasis on avoiding Type I errors. In the exam ple here, a recorded increase in the chemical would be assum ed to reflect a real underlying change only if there is a very small probability o f the change occurring by chance -- typically less that 1 in 20 or 1 in 100 (a result that is significant at the 95 % or 99 % level). In effect n ot being w rong is more im portant than being safe. However this issue is being increasingly recognised in policy. Table 16.1. gives som e exam ples of policy action taken at different levels o f proof. This is further elaborated in the `Im plications' part o f this report. Different levels of proof for different purposes: some illustrations Verbal description 'Beyond all reasonable doubt' 'Balance of evidence' 'Reasonable grounds for concern' 'Scientific suspicion of risk' Exam p les Criminal law; Swedish chemical law, 1973 (for evidence of 'safety' from manufacturers) Intergovernmental Panel on Climate Change, 1995 and 2001 European Commission communication on the precautionary principle Swedish chemical law, 1973, for evidence required for regulators to take precautionary action on potential harm from substances But there is another fundam ental statistical problem . Because environmental monitoring is expensive it is usually lim ited in scope. Yet the smaller the sam ple size a n d /o r the greater the natural variation, the less likely it is that a real increase could be identified through the statistical noise. The possibility o f calling a real effect false is known as a `Type II' error, or a `false negative'. U nderw ood (1999) concluded `Typically there has been little concern abo u t Type II error. The chances o f erring in `favour' o f the environment (a Type I error) is deliberately kept small, whereas the chances of erring `unfavourably' to environmental issues is not! ' A great deal of money can be wasted, for example on expensive ship-borne marine m onitoring, if the actual am ount of sampling that can be carried out on a few trips has no prospect of separating adverse effects from statistical noise, that is, it has low statistical power (HELCOM , 1996). On top o f this come complications arising from the form ulation o f the hypothesis to be tested, and of experimental design: for example, asking the wrong question; dismissing a factor from assessm ent due to an erroneous prior assum ption that it is unimportant; deciding when, what and how frequently to monitor; the im portance to be accorded to rare events; or deciding how to deal with com plex interactions, or non-linear responses such as chaos. Similarly, even small misclassifications of exposure in epidem iological studies can result in m ajor reductions in relative risks, with associations being m ore likely to be m issed than falsely im plicated (C opeland et al., 1977). In general, the power o f epidem iological studies to detect relevant risks is critical but often overlooked, leading to a false sense o f security from so-called `negative' studies that fail to find a risk. The bias in science towards avoiding false positives inevitably involves generating false negatives, which, if they are hum an and or environmental disasters, as in m ost o f these case studies, is not sound public policy. Clearly, such a bias in favour o f generating false negatives does not conform to the precautionary principle, and is an issue that is taken up briefly in the `Im plication s' part o f this report. While such issues are often seen as simply a matter o f scientific judgem ent, in fact they lead on to som e rather deeper points. Fundamental issues The commonly used word `uncertainty' needs to be differentiated at least into risk, uncertainty and ignorance. A wider examination of just how agreem ent can be reached on what are the `facts' would have to take account o f other distinct dimensions such as complexity, indeterminacy, ambiguity, and the nature o f disagreem ent (Wynne, 2001; Stirling, 1999). The concept of precaution has resulted in new thinking in the fields o f sociology o f science and in philosophy, and with it the recognition o f new kinds o f ignorance underlying the very processes of knowledge procurement. Authors such as Krohn and Weyer (1994) have explained how full knowledge o f the consequences of innovations can only be Twelve late lessons 185 gained by treating society and the larger environment as themselves the experimental laboratory. This point has profound though as yet unrealised implications for democratic policy-making abo u t new, an d in deed existing, technologies. Society's growing com m itm ent to the precautionary principle is essentially a respon se to a growing tension between two aspects o f science: its growing innovative powers were increasingly outrunning its capacity to anticipate the consequences. Moreover, too often from within the scientific community there was a denial o f the waning ability to predict those consequences. This encouraged the reasonable democratic response of demanding more circum spection. This circum spection was not about innovation and risk themselves, but about our ability to know. In other words, it was about science and its presum ed powers. It is not at all anti-scientific to raise such questions. Indeed it could be said that it is anti-scientific to deny them. There is nothing scientific about the `pretence at know ledge' (von Hayek, 1978). Such pretence has the consequence o f underm ining the authority and credibility of the institutions o f science -- society's m ost pow erful intellectual resource. With this in m ind som e further observations are relevant. For exam ple, in discussions between peers, it is accepted that the nature o f scientific p ro o f is essentially complex, open, an d always provisional. Yet externally it has in som e quarters becom e a requirem ent that science provide the policy process both with sim ple answers and certainty. This dual identity for `science' causes considerable tensions, not least because the contradiction between intrinsic provisionality and pretended certainty often goes unacknowledged -- but not, it appears, unnoticed. This brings us to a deep dislocation between policy institutions and the public about understandings and representations of scientific uncertainty and ignorance. Public surveys in relation to GMOs on both sides of the Atlantic (Levy and Derby, 2000; Wynne et al., 2000), indicate that non-experts do make a basically correct distinction between uncertainty and ignorance. Whilst scientific risk assessment focuses on (known) uncertainties, public concerns instead centre on unacknowledged ignorance lying behind even the best science. Especially with the rapid expansion of very novel innovations, a m ajor public concern is the possible consequences of ignorance. The reaction can be sum m ed up as `if we can never fully know the consequences, then we h ad better at least ensure that the purposes driving the enterprise, and the interests which control the responses to the resultant surprises, are good on es' (Wynne, 1992 and 2001). In other words the issues of what are the driving purposes and who benefits are forem ost in p e o p le 's m inds. Yet the policy response, in order to reassure the public, has often been to intensify research on identified uncertainties, with the intention o f demonstrating intellectual mastery o f the issue, and to show that concern about (known) risks is unfounded. These policy responses to what are believed to be misconceived public dem ands for zero risk and zero uncertainty are futile, because they presum e the problem s o f public mistrust lie with the p u b lic's erron eous expectation o f certainty, an d the p u b lic's supposed m isunderstandings of science, risk and uncertainty. This institutional approach fundamentally m isunderstands typical public attitudes and expectations; and it only feeds public mistrust by inadvertently dem onstrating its own denial o f ignorance and lack o f intellectual mastery -- which the public appears intuitively to understand rather better than does institutionalised science itself. Thus whereas som e views critical o f precaution see it as pandering to populist anti-science sentiment, in the form of supposed dem ands for certainty before sanctioning any innovative commitment, there is am ple evidence that people are typically quite ready to accept a m uch more radical kind o f uncertainty than institutional science is able to acknowledge -- namely ignorance, and corresponding lack of control. As Stirling (1999) and colleagues have elaborated in detail elsewhere, the precautionary principle has nothing to do with anti-science, an d everything to do with the rejection o f reductionist, closed and arbitrarily narrow science in favour of sounder, more rigorous and more robust science. Perhaps the most fundam ental general insight to em erge is that scientific uncertainty, like scientific knowledge itself when deployed to provide authority to policy, is em phatically n ot just a private m atter for 186 Late lessons from early warnings: the precautionary principle 1896-2000 scientific bodies to autonom ously resolve, define, or otherwise interpret on behalf of the public policy dom ain, before it is rendered visible to the latter. As the NRC (1996) in the United States and the RCEP (1998) in the United Kingdom independently concluded, prior questions need public deliberation. These include: what questions should the science be addressing, what are the salient factors and what general principles should define good science (for example, the balance between comprehensiveness and precision) for environm ental policy, as distinct from other more confined domains such as engineering risks. These questions lead on to the im portance of distinguishing between facts an d values. It is som etim es assum ed that, having agreed to adopt a precautionary approach, this should automatically lead to one `correct' outcome. This is an oversim plification. W hat is at issue itself will vary. Preferred outcom es will also vary d epen d in g on p e o p le 's interests, objectives and values. As Popper pointed out long ago, it is rationally impossible to derive a proposal for a policy from facts alone (Popper, 1962). Policies that unduly emphasise the factual basis of decisions, without explicitly acknowledging and en gagin g with the value ju d ge m en ts that are also part o f such decisions, are unlikely to achieve consensus, or at least acceptance, where substantial divisions o f opinion exist. (R M N O ,2000) It is for these reasons that the involvement of stakeholders in regulatory appraisal needs to begin at the beginning rather than being artificially confined to the later `risk m anagem ent' stages of the conventional approach. The stages o f hazard and risk appraisal, m anagem ent and communication are not sequential, as in the traditional model, but require stakeholder involvement at the earliest stage. This has been recognised by the NRC, the RCEP and the Nice Council o f Ministers conclusions on the precautionary principle in 2000. Again, it is em phasised that the raising o f such issues should not be seen as creating a new form of `paralysis by analysis'. The case studies, and the lessons that em erge, do provide a positive and robust way forward. But if these complexities are not explicitly addressed, progress will be slower an d the m istakes m ade will be m ore serious. 16.3.2. Precaution and innovation In the Introduction it was n oted that, the G erm an Vorsorgeprinzip ( `foresigh t' or `precautionary' principle) o f the 1970s considered the stimulation of innovation, employment and forward planning to be integral components of the precautionary principle. An overarching principle arising from all the case studies, as well as from m uch wider analysis and experience, is that the polarised processes of technological innovation and risk regulation need to become less separate and antagonistic. Many o f the lessons are as applicable to the innovation process itself as to the regulation of the resulting products and technologies. Indeed, accom m odating them within the innovation process could overcome the adversarial relationship between innovation and regulation. For example, traditional risk assessments of chemicals concentrate on the risks associated with a particular chem ical. T he use o f the chem ical is identified, bu t this identification o f use is not then used in a m ore integrated way to assess the potential o f alternatives. It is not the first time that the need for the integration of innovation and risk regulation, and for the integration o f technology appraisal approach es m ore generally is identified. It is addressed, for instance (at least in part), by the techniques of constructive technology assessment (CTA) developed in the N etherlands (Rip et al., 1996; Wynne et al., 2001). CTA starts from recognition that technologies are more than hardware. Their prom otion and adoption depends upon the goals, relationships, understandings and skills o f different sectors o f society. The greater the scope, power, complexity o r interconnectedness o f the technological systems concerned, the more im portant a consideration of these social and institutional aspects becomes. In essence, CTA attem pts to foster the recognition o f this neglected but fundamental dimension of technology and innovation. By a variety of m eans it aims to improve integration, from the outset, o f the perspectives of innovators, regulators, users and other stakeholders. In this way, innovation benefits from creative inputs at a stage when these may realistically be harnessed, rather than forcing such wider social interaction simply to take the form of an (often adverse) downstream reaction to a technology already developed through a m ore closed process. The analysis underlying CTA recognises how technological systems otherwise have a tendency to `lock-in' to Twelve late lessons 187 particular configurations at a relatively early stage in their development, thus foreclosing other options and raising the costs o f shifting to alternatives. The particular technologies that gain ascendancy in this way may do so for arbitrary reasons which have nothing to do with intrinsic qualities, an d everything to do with chance and first-leader advantage. CTA attempts to highlight such questions and to provide the intellectual resources for resolving them. The potentially positive relationship between regulation and innovation is also em phasised in the technological options analysis (TOA) approach, developed in the United States (Ashford, 1981 and 1994; Tickner, 2000). H ere, a variety o f practical procedures are routinely employed in order to include consideration o f `off the shelf' and `on the horizon' alternatives alongside the technology or process in question. Again, like CTA, TOA can be perform ed by regulatory agencies or private companies, depending on the context. Together, they form part o f a broader culture o f `alternatives assessm ent' (O 'Brien, 2000), which highlights the importance o f innovation and understands this as an open-ended process, subject to deliberate choices and commitments. These innovative approaches represent a more reflective and intelligent approach to the design, assessment, choice and im plem entation o f technologies. By addressing innovation at the earliest stages, they offer a m eans to im plem ent the lessons developed here in a fashion that minimises economic inefficiencies and social tensions, and actively fosters innovation pathways that are more sustainable over the longer term. The prom otion of robust, diverse and adaptable technologies not only helps to stimulate innovation but it can also provide `insurance' against surprises, such as the case study exam ples of the asbestos cancer, mesothelioma, and the halocarbon damage to the ozone layer. This is because the size o f any future surprises will be sm aller if there are several competing technologies that are being used to m eet hum an needs, rather than ju st one, global, near monopoly, as was the case with asbestos, halocarbon s and PCBs. Diverse technologies and other ways o f m eeting n eeds can help deal with the seem ingly intractable problem o f `societal ignorance' and attendant surprises. 16.3.3. Precaution and governance `Governance' is about the m anner in which som ething is governed by m ethods of m anagem ent and systems o f regulations, be they formal or informal. M ore broadly, it refers to the conduct o f life or business in general and the m ode o f living, or behaviour, in society. The challenges which the precautionary principle heralds involve m ore than simply new decision rules and technical instruments, and this implies a learning process in public policy, industry, science and civil society at large. H ere three aspects o f governance relevant to precaution are considered, namely: how current institutions associated with the appraisal o f risk m ight evolve to take account o f precaution; the relevance of participatory approaches and subsidiarity; and the im portance o f greater awareness within civil society, including ethical awareness, so as to better exercise both rights and responsibilities. Evolution, not revolution Beyond the disciplines of CTA, there are a variety o f perspectives to be found am ong established approaches to the regulatory appraisal o f risk which could be constructively enlarged and developed. Suitably am ended and interpreted, many of these offer ways to respond to the lessons discussed here. Multi-criteria m apping (M CM ), for instance, com bines the flexibility and scope o f qualitative approaches with the transparency and specificity o f quantitative disciplines. It offers one way to accom m odate a diverse array of stakeholder perspectives with different technical an d scientific factors, including uncertainties, without placing undue constraints on the divergent framings o f the issues in question (Stirling an d Mayer, 1999). Likewise, life cycle analysis (LCA) has developed an array of m ethods to ensure that attention extends to encom pass the full technical life cycles and resource chains associated with different options (van den Berg, 1995). Cost-benefit analysis, for all its idiosyncrasies an d serious lim itations, is unusual in upholding the importance, emphasised in these lessons, o f considering the `pros' (justifications and benefits) alongside the `cons' (risks and costs) (Hanley and Spash, 1993). Although rather infrequently used, sensitivity and scenario analysis techniques offer ways to explore the implications of different assumptions and perspectives and so be m ore `h um ble' about the status of any particular understanding of a risk issue (Godet, 1992). 188 Late lessons from early warnings: the precautionary principle 1896-2000 One broad approach that has em erged over the past decade or so, integrated environmental assessment (IEA), offers an architecture for constructing a new synthesis o f these kinds o f positive attributes in existing approaches to the appraisal o f the pros and cons o f contending alternatives (EFIEA, 2000; Dowlatabadi and Rotmans, 2000). Specifically developed to address large, complex environmental problems, IEA provides a m eans to integrate such m ethods into the kind of broad-based, multi perspective and more humble and openended approach highlighted in these lessons. IEA in its present form addresses the interface between science-engineering and policy with em phasis on the n eed for interdisciplinarity; integration across the environm ental m edia: water, air and soil; mass balance accounting in society and sectors; an d analysis o f alternatives. In view o f the lessons identified in the case studies, equal emphasis on the following issues would improve the perspectives o f IEA: m ore openness and deliberation about framing agendas; distinguishing between risk, uncertainty and ignorance; taking better account o f the risk and consequences of being wrong; accounting for all values; expansion o f cost-benefit analysis to the appraisal o f wider pros and cons (as defined in this book); and m uch earlier involvement o f these broader concerns in the relevant decision-making processes. Here, the use of `what-if' scenarios an d participatory scenario developm ent techniques can assist greatly in the m anagem ent o f surprises and uncertainties. Participatory approaches and subsidiarity The historic trend o f legal and economic regulation has been to centralise at global, regional or national governm ent level, and to disenfranchise many interest groups. While there are signs of change, attempts to create participatory approaches are as vulnerable as any other initiative to centralising pressures. For example, the consensus conference, without safeguards, can potentially degenerate into little m ore than a form of consultation driven by the sp o n so r's agen da. A ppeals procedures may also involve participatory approaches, but at far too late a stage in the process o f regulation to deal with anything other than m arginal issues. It follows from the `late lessons' that participation o f interest groups should be at an early stage, broadly drawn, and carried down to the appropriate local level. O f course, local involvement is only m eaningful in the context of democratically legitimate strategic frameworks o f innovation and overall policy. The basic point is that often experts involved in technological developments -- and officials in government -- need better sensitivity to (often under articulated) public values, priorities and concerns, at a variety o f levels, from local to international. This can be achieved without the need for what would be paralysing indiscriminate full public participation in every single decision. In addition to European examples already mentioned, there are also developments outside the EU worthy o f evaluation. These include the environm ental laws o f New Zealand, where the principle of subsidiarity has been introduced as an alternative approach to centralised regulation (Ministry for the Environment, 2001), and where the stated aim is for procedu res for local participation and consensus-building to take prom inence over central regulation. Both the development and implementation of A ustralia's ocean policy include significant attem pts to involve stakeholders (National Oceans Office, 2000). Among many other findings, the research underlying the ocean policy m ade the relevant poin t that the views of local interest groups are not necessarily identical to their national or international equivalents, be these, for example, environmental non-governmental organisations or industry. The tools for participatory approaches are in various stages o f development, and the challenges are far from trivial (Brookes, 2001). But this has to be set against traditional approaches, where the costs of failure can also be high, as illustrated by the public rejection of irradiated foods, the abandoned attem pt to dum p the North Sea Brent Spar oil installation and the response to GMOs. Awareness and ethics Participatory approaches cannot work without heightened awareness, interest and engagem ent on the part o f stakeholders and the public at large. Broadening the scope o f environm ental assessm ent will have a lower chance of success without a development of society's attitudes, responsibilities an d ethics in relation to the environment. Parallel efforts are n eeded to increase society's awareness, foster discriminating involvement, and increase the educational basis for consensus building, by increased knowledge Twelve late lessons 189 o f all aspects o f the environment over the whole range o f the disciplinary spectrum. This process starts at hom e and school and reaches its full potential by engagem ent in participatory procedures as a matter of interest and comm itm ent. In stating this indoctrination to any one point o f view is not im plied, but rather the developm ent o f skills to critically assess argum ents, express a point o f view and engage in the dem ocratic process. A related area is the professional com m unication o f the issues surrounding precaution. It is evident that at least som e com m unication o f the complexities involved is necessary to move forward, for the sake all interest groups. It is therefore ironic that the m edia and other professional com m unicators have been moving in the opposite direction, em phasising only `positive knowledge' and the `clear and simple m essage' in a sound-bite culture. This tends to exclude the comm unication of ignorance, of complexity and of responsibility in face o f the essential limits o f all knowledge. involvement in policy-making (including science policy and technological innovation). This in turn will dem an d that the expert-led institutions o f science, industry and policy learn to trust, to challenge and to build the opportunities an d the fram eworks for civil society to take on those responsibilities. From the case studies, the 12 late lessons and these wider considerations o f science, innovation and governance, it is evident that the precautionary principle -- the need to exercise foresight -- once accepted, leads far beyond the sim ple definition. R ather it is playing its p art in the developm ent o f civil society and policy-making during the early 21st century which, it appears, will have its own distinctive character, as great in its differences as those which set apart previous centuries. 16.4. References A shford, N., 1984. `Alternatives to costbenefit analysis in regulatory decisions', Annals of the New York Academy of Sciences Vol. 363, April. There is also an ethical and cultural issue raised by the im portant need for institutional recognition o f ignorance as well as uncertainty. This arises because o f the way in which the ethical boundaries o f acknowledged responsibility for uncertainties about the consequences o f hum an innovative commitments have been drawn by scientific knowledge. Any possible future consequence which lies beyond existing scientific knowledge and predictability is deem ed by definition to be beyond responsibility. This is defined as such even though it is known that such surprises will occur as a result o f choices and commitments. The precautionary principle implies the need, as a m atter of cultural change, for society's institutions to enlarge existing notions o f ethical responsibility to encom pass these unknowns, which are predictable in principle even though not in specifics. Suggestions have been m ade as to how this process of intellectual enlargem ent might begin, starting with the twelve late lessons. This report expresses the honest conviction that to achieve sustainable environmental policies and a properly balanced practice of the precautionary principle requires the achievement o f a cultural shift towards a greater civil sense o f responsibility and A shford, N., 1991. `An innovation-based strategy for the environment and for the workplace' in Finkel, A. and Golding, D. ( eds), Worst things first: the debate over risk-based national environmental priorities, R esources for the Future, Washington, DC. C opelan d et al., 1977. `Bias due to misclassification in the estimate o f relative risk', AmericanJournal ofEpidemiology Vol. 105, pp. 488-495. Dowlatabadi, H. and Rotmans, J., 2000. `Integrated assessm ent', Special issue of Integrated Assessment Vol. 1, N o 3, Battzer Science Publishers. EFIEA (European Forum for Integrated Environmental Assessment), 2000. `Integrated environmental assessm ent in European Environment Agency reporting', Report of the EEA /EFIEA special session, C open h agen /A m sterdam . Hanley, N. and Spash, C., 1993. Cost benefit analysis and the environment, Edward Elgar, London. H arding, R. an d Fisher, E. ( eds), 1999. Perspectives on theprecautionary principle, T he Federation Press, Sydney. 190 Late lessons from early warnings: the precautionary principle 1896-2000 H ELCO M , 1996. `Third periodic assessm ent o f the state o f the m arine environment o f the Baltic Sea 1989-1993, Executive sum mary', Baltic Sea Environment Proceedings 64 A. IEGMP (Independent Expert Group on M obile Phones), 2000. Mobile phones and health (T h e Stewart R ep ort), Chilton, U nited Kingdom, at http://www.iegm p.org.uk/ IEGM Ptxt.htm (accessed 16 May 2001). report, Vol. 1, Environm ental Protection Agency, Washington. O 'R iordan an d Cam eron, 1994. Interpreting the precautionary principle, C am eron & May, London. O 'Riordan, Cam eron and Jo rd an , 2001. Reinterpreting the precautionary principle, C am eron & May, Lon don . Kourilsky, P. an d Viney, G., 1999. `Le principe de prcaution', Rapport au Premier ministre, France. Krohn, W. an d Weyer, J., 1994. `Society as a laboratory: The social risks o f experim ental research ', Science and Public Policy Vol. 21, pp. 173-183. Levy, A. an d Derby, B., 2000. `R eport on consumer focus groups on biotechnology', U S Food and Drug Administration, Center for Food Safety and Applied Nutrition, Office o f Scientific Analysis and Support, Division of Market Studies, Washington, DC and Paper presented to session on `Public attitudes to agricultural biotechnologies', Annual M eeting o f the Society for the Social Studies o f Science (4S), University o f Vienna, 30 September. Ministry for the Environment (New Zealand), 2001. Resource M anagem ent Act, text and background discussion, at h ttp :// www.m fe.govt.nz/m anagement/rm a/ rm a11.htm (accessed 17 May 2001). Phillips et al., 2000. Return to an order of the Honourable the House ofCommons dated October 2000for the Report, evidence and supporting papers of the Inquiry into the emergence and identification of bovine spongiform encephalopathy (BSE) and variant Creutzfeldt-Jakob disease (vCJD) and the action taken in response to it up to 20 March 1996, The Stationery Office, London, at http://www.bse.org.uk/ index.htm (accessed 16 May 2001). Popper, K. R., 1962. The open society and its enemies, Vol. 1, C hapter 5, fourth edition, Routledge & Kegan Paul, London. Raffensberger, C. and Tickner, J. (eds), 1999. Protecting public health and the environment: Implementing the precautionary principle, Island Press, W ashington, DC. Renn O., Webler, T. an d Kastenholz, H., 1996. `Procedural and substantive fairness in landfill siting: A Swiss case study', Risk, Health, Safety and Environment, Vol. 7, N o 2 (Spring 1996), pp. 95-98, Franklin Pierce Law Center, Concord, NH. N ational O ceans Office, 2001. A ustralia's oceans policy hom e page, at h ttp :// www.oceans.gov.au/aop/m ain.htm (accessed 26 April 2001). NRC (US National Research Council), 1996. `Understanding risk', Report of an ad-hoc working party chaired by H. Feinberg, summary at http://www.riskworld.com/ N reports/1996/risk_rpt/htm l/ nr6aa045.htm (accessed 17 May 2001). O 'Brien, M., 2000. M aking better environmental decisions: an alternative to risk assessment, MIT Press, Cam bridge, MA. O m en, G. S., Kessler, A. C., A nderson, N. T., et al., 1997. Frameworkfor environmental health risk management, U S Presidential / Congressional Commission on Risk Assessment and Risk M anagement, Final Rip, A., Misa, T. an d Schot, J., 1996. M anaging technology in society, Pinter, L on don . RCEP, 1998. Setting environmental standards, 21st Report o f the Royal Commission on Environmental Pollution, Cm 4053, HMSO, London, summary at h ttp :// www.rcep.org.uk/reports2.html#21 (accessed 16 May 2001). RMNO, 2000, `Willingly an d Knowingly. The roles o f knowledge about nature and the environment in policy processes', Advisory Council for Research on Spatial Planning, Nature and the Environment, Lem m a Publishers, Utrecht. Stirling, A., 1999. `On science, and precaution in the m anagem ent o f technological risk', Final summary report, Technological Risk and the M anagem ent of Uncertainty project, European Scientific Twelve late lessons 191 Technology Observatory, EC Forward Studies Unit, Brussels, cited in European Commission, 2000. Com m unication from the Commission on the precautionary principle, CO M (2000)1, Brussels. Stirling, A. and Mayer, S., 1999. `Rethinking risk: A pilot multi-criteria m apping o f a genetically m odified crop in agricultural systems in the U K ', SPRU, University of Sussex. Tickner, J . A., 2000. `Precaution in practice: A framework for im plem enting the precautionary principle', Dissertation prepared for the D epartm ent o f Work Environment, University o f Massachusetts, Lowell, MA. U nderw ood, T., 1999. `Precautionary principles require changes in thinking about and planning environm ental sam pling' in H arding, R. and Fisher, E. (eds), 1999. Perspectives on the precautionary principle, pp. 254-266, The Federation Press, Sydney. van den Berg, N., Dutilh, C. and H uppes, G., 1995. Beginning LCA: A guide into environmental life cycle assessment, CM L / Unilever / Novem / RIVM. van D om m eln, 1997. Coping with deliberate release: The limits of risk assessment, International Centre for Hum an and Public Affairs, T ilburg/B uen os Aires. von Hayek, F., 1978. New studies in philosophy, politics, economics and the history of ideas, Chicago University Press. W BGU, 2000. Welt im Wandel. Handlungsstrategien zur Bewaltigung globaler Umweltrisiken (World in transition: strategies for m anaging global environmental risks), Jahresgutachten 1998 (Annual report 1998), Springer, Berlin. Wynne, B., 1992. `Uncertainty and environmental learning: Reconceiving science and policy in the preventive p arad igm ', GlobalEnvironmental ChangeVol. 6, pp. 111-127. Wynne, B., 2001. `M anaging and comm unicating scientific uncertainty in public policy', Harvard University Conference on Biotechnology and Global Governance: Crisis and Opportunity, Kennedy School o f Government, April. Wynne, B., Marris, C., Sim m ons, P., De Marchi, B., Pellizoni, L., Renn, O., Klinke, A., Lemkow, L., Sentmarti, R. and Carceras, J., 2000. `Public attitudes to agricultural biotechnologies in Europe', Final report of project PABE, 1997-2000, DG Research, European Commission, Brussels. Wynne, B. et al., 2001. Wising up, Lan caster University. 192 Late lessons from early warnings: the precautionary principle 1896-2000 17. Conclusions Table 17.1. The case studies reviewed in this book show that there is m uch that can be learnt from history. Such learning begins with two basic observations: Regulatory appraisal and control of technologies and economic development involves balancing the costs o f being too restrictive on innovation with the hazards and costs o f being too permissive, in situations of scientific uncertainty and ignorance. The case studies provide many exam ples where regulatory inaction led to costly consequences that were not -- and sometimes could not have been -- foreseen. The case studies also provide many exam ples where `early warnings', and even `loud and late' warnings, were clearly ignored; where the scope of hazard appraisal was too narrow; and where regulatory actions were taken without sufficient consideration o f alternatives, or o f the conditions necessary for their successful im plem entation in the real world. If m ore account, scientifically, politically and economically, is taken o f a richer body of information from more diverse sources, then society may do substantially better in the future at achieving a better balance between innovations and their hazards. Discussion of the case studies led to the distillation of twelve late lessons which, if applied to future decision-making, could help achieve this better balance. The precautionary principle is an overarching framework o f thinking that governs the use of foresight in situations characterised by uncertainty and ignorance and where there are potentially large costs to both regulatory action and inaction. However, the intrinsic difficulties o f applying the precautionary principle to issues of complexity, uncertainty and controversy are com pounded by a lack o f agreem ent on the definition and m eanings o f key terms. The table below is a contribution to clarifying the m eaning o f six basic concepts that lie at the heart o f this debate. What is sometimes loosely referred to as `uncertainty' mixes up the analytically distinct concepts o f `risk', `uncertainty' and `ignorance'. The public action concepts of `prevention', `precautionary prevention', and `precau tion ' can then be usefully related to these three states o f knowledge, as in Table 17.1. Uncertainty and precaution -- towards a clarification of terms Situation Risk Uncertainty Ignorance State and dates of knowledge 'Known' impacts; 'known' probabilities e.g. asbestos causing respiratory disease, lung and mesothelioma cancer, 1965-present 'Known' impacts; 'unknown' probabilities e.g. antibiotics in animal feed and associated human resistance to those antibiotics, 1969present 'Unknown' impacts and therefore 'unknown' probabilities e.g. the 'surprises' of chlorofluorocarbons (CFCs) and ozone layer damage prior to 1974; asbestos mesothelioma cancer prior to 1959 Exam ples of action Prevention: action taken to reduce known risks e.g . eliminate exposure to asbestos dust Precautionary prevention: action taken to reduce potential hazards e.g. reduce/ elim inate human exposure to antibiotics in animal feed Precaution: action taken to anticipate, identify and reduce the impact of 'surprises' e .g . use of properties of chem icals such as persistence or bioaccum ulation as 'predictors' of potential harm; use of the broadest possible sources of information, including long term monitoring; promotion of robust, diverse and adaptable technologies and social arrangements to meet needs, with few er technological 'm onopolies' such as asbestos and CFCs T he procedu res for dealin g with the situations o f risk, uncertainty and ignorance n eed to be fair, transparent an d accountable, key elem ents o f the `goo d governance' which is needed to regain public confidence in Conclusions 193 policy-making on technologies, their benefits and potential hazards. M ost o f the cases in this book involved costly impacts on both public health and the environm ent, two fields o f science an d policy making that have become specialised and somewhat polarised during the last 100 years. Individuals experience their health and their environm ent as one, interconnected reality: science, regulatory appraisal and policy m aking need to be similarly integrated. Late lessons should m ake som e contribution to this integration of health and environment. The scope o f regulatory appraisal needs to be broadened to include adequate consideration of relevant social issues alongside the physical, chemical, biological and m edical aspects o f technologies. Involving a wide range o f stakeholders, and taking account of their values and interests at the earliest stage o f the appraisal and choice o f technological and social options for m eeting h um an needs, brings two key benefits. It not only augm ents the inform ation available to policy-making, but may also im prove public trust in society's capacity to control hazards, without necessarily stifling innovation or compromising science. Recent controversies over emerging technologies such as genetically modified organism s and oil-rig disposal have m uch to do with public values an d scientific uncertainties, in contrast to the previously low emphasis on values and to the dem and for unequivocal scientific p roof before action to reduce hazards. An open recognition of this changing context for such controversies is a first step towards their improved governance. However, the inclusion o f different socio political perspectives in regulatory appraisal becom es notju st a matter of better policy making: it can make for better science as well. For example, reductionist science and linear causality are useful approaches, but they are lim ited. They do n ot cope well with the dynamics o f com plex and som etim es chaotic systems, characterised by feed-back loops, synergisms, thresholds, and equilibria/ instability issues, and linked by multi-factoral and interdependent causal chains. Such com plex reality dem ands better science, characterised by m ore humility and less hubris, with a focus on `what we d o n 't know' as well as on `what we do know'. One im portant consequence of acknowledging both scientific uncertainties (including ignorance) and the urgency of hazard reduction in situations of high stakes is the n eed for agreem en t on the sufficiency o f evidence o f harm ful effects that is required to justify action. Such `levels o f proof' can vary from the `reasonable grounds for co n cern ' o f the E u ro pean C om m ission 's Com munication on the Precautionary Principle, to the `beyond reasonable dou bt' o f crim inal law. C hoosing which level to use in particular situations involves a decision that can radically shift the size, nature and distribution o f the costs o f being wrong. This is a key political decision with profoun d ethical im plications. T he level o f p ro o f that is appropriate for particular issues depends upon the size and nature o f the potential harm, the claimed benefits, the available alternatives, and the potential costs o f being wrong in both directions, ie o f acting or not acting in the context o f uncertainty, ignorance and high stakes. This type o f public decision-making is not unknown: military intelligence has long adopted similarly precautionary approaches to uncertainty and high stakes, where the costs of being wrong can be catastrophic. 17.1. Late lessons from early warnings The case studies in this book both support an d illustrate the n eed for the twelve late lessons derived from the century o f history reviewed. 1. Acknowledge an d respon d to ignorance, as well as uncertainty and risk, in technology appraisal and public policy m aking. 2. Provide adequate long-term environmental and health monitoring and research into early warnings. 3. Identify and work to reduce `blind spots' and gaps in scientific knowledge. 4. Identify and reduce interdisciplinary obstacles to learning. 5. Ensure that real world conditions are adequately accounted for in regulatory appraisal. 6. Systematically scrutinise the claim ed justifications and benefits alongside the potential risks. 7. Evaluate a range o f alternative options for m eeting needs alongside the option 194 Late lessons from early warnings: the precautionary principle 1896-2000 under appraisal, and prom ote more robust, diverse and adaptable technologies so as to minimise the costs o f surprises and maximise the benefits of innovation. 8. Ensure use o f `lay' an d local knowledge, as well as relevant specialist expertise in the appraisal. 9. Take full account o f the assum ptions and values of different social groups. 10. M aintain the regulatory independence o f interested parties while retaining an inclusive approach to information and opinion gathering. 11. Identify and reduce institutional obstacles to learning and action. 12. Avoid `paralysis by analysis' by acting to reduce potential harm when there are reasonable grounds for concern. uncertainties or avoid the consequences of ignorance. However, they would at least increase the chances o f anticipating costly impacts, of achieving a better balance between the pros and cons of technological innovations, and o f minimising the costs of unpleasant surprises. The use o f the precautionary principle can also bring benefits beyond the reduction o f health and environmental impacts, stimulating both m ore innovation, via technological diversity and flexibility, and better science. The `late lessons' may also help to achieve a better balance between proportionate and precautionary public policies, recognising that over-precaution can also be expensive, in terms of lost opportunities for innovation and lost lines o f scientific inquiry. Most o f these lessons involve im provem ents in the quality, availability, utilisation and processing o f information in public policy making on environment and health. However, none o f the lessons would themselves remove the dilemmas of decision making under situations of uncertainty and high stakes. They cannot eradicate It is the central conclusion o f this report that the very difficult task o f maximising innovation whilst m inimising hazards to people an d their environm ents, which is ultimately a matter of political discourse, could be m ore successful if it em braced the twelve late lessons from the histories o f hazards reviewed. A u t h o r b i o g r a p h i e s ' 14' Author biographies 195 Jim W. B ridges Dean for International Strategy Vice C h an cellor's Offices University o f Surrey Guildford GU2 5XH United Kingdom Tel: (44) (0)1483 873802 e-mail: j.bridges@ surrey.ac.uk Lars-Erik Edqvist Swedish National Veterinary Institute Ulls vag 2A-C Box 7073 750 07 Uppsala Sweden Tel: (46) 018 67 40 00 e-mail: Lars-Erik.Edqvist@sva.se Jim Bridges has been Professor o f Toxicology and Environmental H ealth at the University o f Surrey since 1979. H e is a m em ber o f the EU Scientific Steering Committee (since 1997) and Chairm an o f the UK Veterinary Residues Committee (since 2001). H e was previously a m em ber o f the EU Scientific Committee on Animal Nutrition (1991-97) and the UK Veterinary Products Committee (1982-97). He acted as a scientific adviser to the European U nion (World Trade Organization hearing in Geneva) on the legitimacy o f the EU ban on the use of horm ones as growth prom oters. H e is author/joint author of more than 300 research p ap ers an d reviews. Olga Bridges European Institute of Health and Medical Sciences University o f Surrey Guildford GU2 5XH United Kingdom Tel: (44) (0)1483 873802 e-mail: o.bridges@ surrey.ac.uk Dr. O lga B ridges is a lecturer in Environmental and Public H ealth at the University o f Surrey. H er research interests are the im pact o f chemicals in the environment and the diet, public perception o f environmental risks and application of HACCP (hazard analysis critical control point). She has participated in a num ber of public enquiries, and has recently been involved in the training program m es of environmental health officers abroad. Professor Lars-Erik Edqvist DVM, Ph.D. is Director General of the National Veterinary Institute (SVA), Uppsala, Sweden. The antibiotics departm en t at SVA has been involved in research for many years and has advocated prudent use o f antim icrobials in veterinary m edicine. In 1995 the Swedish government appointed Professor Edqvist chairman of the Commission on Antimicrobial Feed Additives. The commission gathered and evaluated scientific information on the use o f antimicrobials as feed additives in relation to the temporary derogation from Community legislation that Sweden had when it becam e a m em ber o f the EU. The com m ission com pleted its task in 1997. Joe Farman European Ozone Research Coordinating Unit C am bridge United Kingdom e-mail: general@ ozone-sec.ch.cam .ac.uk Dr. J . C. Farm an, CBE and Polar M edal, is currently a consultant to the European Ozone Research Coordinating U nit in Cam bridge, and is a m em ber o f the Science Panel on Stratospheric Ozone Research of the European Commission. From 1956 to 1990 he worked for the British Antarctic Survey. He is on the UNEP Global 500 Roll o f Honour, and has received the UNEP Global Ozone Award, the Environm ent M edal o f the Society of Chemical Industry and the Charles Chree M edal o f the Institute o f Physics. H e is an Honorary Fellow o f Corpus Christi College, Cambridge. (14) The view s expressed in the chapters are those of the authors and do not necessarily reflect those of the institutions where they work. 196 Late lessons from early warnings: the precautionary principle 1896-2000 David Gee Inform ation N eeds Analysis and Scientific Liaison European Environment Agency Kongens Nytorv 6 1050 Copenhagen K D enm ark Tel: (45) 33 36 71 42 e-mail: david.gee@ eea.eu.int David Gee graduated in economics and politics but has been working at the interface o f science, economics, production and policy-making, within occupational and environmental health, since 1974, for trade unions and non-governmental organisations (he is a form er D irector o f Friends o f the Earth, England, Wales and Northern Ireland). Since 1995 he has worked with the European Environment Agency on emerging issues and scientific liaison. Michael Gilbertson International Join t Commission 100 O uellette Avenue, 8th Floor W indsor Ontario N9A 6T3 Canada Tel: (1) 519 257 6706 e-mail: GilbertsonM@W indsor.IJC.Org Dr. M ichael Gilbertson has worked with the Canadian federal governm ent since 1969, when he initiated research on the outbreaks of reproductive and developmental disorders in fish-eating birds in the lower Great Lakes. He has been at the interface between forensic environmental science and regulatory action since 1975. His interest is in the social processes involved in moving from use of inappropriate technologies that require causal demonstrations of injury for their control to appropriate technologies based on sound ecological principles. Morris Greenberg 74 End Road London NW11 7SY United Kingdom Tel: (44) (0)20 8458 2376 e-mail: GillM orris.Greenberg@ talk21.com Dr M orris G reenberg is a form er M edical Inspector of Factories, Health and Safety Executive, United Kingdom, and a former m em ber o f the Environmental Toxicology Unit, U K Departm ent of Health. He helped in setting up the first asbestos m esotheliom a register in the United Kingdom and is an expert in the medical aspects o f inhaled fibres. Poul Harremoes Department of Environment and Resources Technical University o f Denm ark e-mail: ph@ er.dtu.dk Poul Harrem oes has been a full Professor at the Technical University o f D enm ark in Environmental Science and Engineering since 1972. He has carried out research in urban water issues, water pollution and purification, and environmental m anagem ent. H e is a m em ber o f the Scientific Committee o f the European Environment Agency and former president of the International Water Association. He has the following international awards: the Stockholm Water Prize (1992) and the Heineken prize for environmental science (2000). H e has publish ed five textbooks and some 350 further publications (see www.er.dtu.dk). Dolores Ibarreta IPTS Institute for Prospective Technological Studies (Joint Research Centre) European Commission Edificio Expo-WTC C / Inca Garcilaso, s/n E-41092 Seville Spain Tel: (34) 95 448 84 45 e-mail: dolores.ibarreta@ jrc.es Dr. D olores Ibarreta is a biologist with a Ph.D. in genetics from the Universidad Complutense, Madrid, Spain. She has worked as a researcher at the Centro de Investigaciones Biologicas (CIB-CSIC) in M adrid and Georgetown University M edical Center in the U n ited States. She is currently a scientific officer at IPTS (Joint Research Centre, European Commission), where her main focus has been the risk assessment of endocrine disrupters (DES being one the main references), specifically hormonally active substances in the diet. Author biographies 197 Peter Infante Health Standards Programs Occupational Safety and Health A dm in istration US Department of Labor Washington DC United States e-mail: Peter.Infante@ osha.gov For the past 25 years Dr. Peter F. Infante D .D .S., Dr.P.H. has evaluated an d regulated toxic substances found in U S workplaces for the D epartm ent o f Labor. H e is the author o f m ore than 100 scientific publications. He has a doctorate in public health from the D epartm ent o f Epidemiology, University o f M ichigan and is a Fellow o f the American College o f Epidemiology. He has served on num erous panels and advisory committees related to the U S National Cancer Institute, the P residen t's Cancer Panel, the O ffice o f Technology Assessment of the US Congress and the National Academy o f Sciences, and m ost recently as an expert consultant to the World Trade O rganization in its deliberations on the banning of products containing asbestos in the European Union countries. semi-natural ancient woodlands and urban street trees. Janna G. Koppe Emeritus Professor of Neonatology University o f Amsterdam Hollandstraat 6 3634 AT Loenersloot The Netherlands Tel: (31) 294 291589 e-mail: Janna.K oppe@ inter.N L.net After studying medicine in Amsterdam Professor Koppe specialised in paediatrics and neonatology, also at the University of Amsterdam. She becam e an assistant professor in neonatology in 1969 and a full professor in 1986, becom ing head o f the neonatology departm ent in 1977, and retiring in 1998. H er contribution to this book was written as part o f her role as Emeritus Professor of Neonatology, University o f Amsterdam and Chair o f the Ecobaby society. Studies on the effects of dioxins and PCBs on the unborn and newborn hum an baby were started in 1985 and are still ongoing. Paul Johnston Greenpeace Research Laboratories Departm ent o f Biological Sciences University o f Exeter Exeter EX4 4PS United Kingdom e-mail: p.johnston@ exeter.ac.uk Dr. Paul Jo h n sto n is an aquatic toxicologist who established the Greenpeace Research Lab oratories in 1987 and continues as its principal scientist. H e has m ore than 15 years research experience in the field o f marine pollution. Jane Keys 22 Brook End Potton, Bedfordshire SG19 2QS United Kingdom Tel: (44) (0)1767 261834 e-mail: Jane.Keys@ btinternet.com Jan e Keys is a freelance researcher. Since the late 1980s she has researched environmental issues for industry, environmental non governm ental organisations and the European Environment Agency. Recent work includes, on behalf o f the UK Forestry Commission and others, the identification of feasible and sustainable econom ic activities to fund an d sup p ort the U n ited K in gdom 's Barrie Lambert St Bartholom ew 's an d the Royal L o n d on School of Medicine Charterhouse Square London EC1M 6BQ United Kingdom Tel: (44) (0)1273 471973 e-mail: barrie.lambert@ which.net Dr. Barrie L am bert is an in depen den t radiation biologist. He has researched and taught the subject for m ore than 30 years, with particular reference to the com parative long-term risk o f radiation exposure. He is a consultant to environmental groups, the nuclear industry and government. William J . Langston Plymouth Marine Laboratory Citadel Hill Plymouth PL1 2PB United Kingdom Dr. Bill L an gston is a research scientist at the Plymouth Marine Laboratory and has been engaged in studies o f tributyltin behaviour and im pact in the marine environment since the early 1980s. Malcolm MacGarvin m odus vivendi Ballantruan, Glenlivet Ballindalloch AB37 9AQ 198 Late lessons from early warnings: the precautionary principle 1896-2000 Scotland Tel: (44) (0)1807 590396 e-mail: macgarvin@modus-vivendi.co.uk Dr. M alcolm M acGarvin has a Ph.D. in ecology. Since the mid-1980s he has worked as a consultant on environmental issues for industry, environmental non-governmental organisations, the European Commission and the European Environment Agency. In addition to the precautionary principle, recent work includes that on marine eutrophication and on marine fisheries. Erik Millstone SPRU - Science and Technology Policy Research University o f Sussex Brighton BN 1 9RF United Kingdom Tel: (44) (0)1273 877380 e-mail: e.p.m illstone@ sussex.ac.uk Dr. Erik M illstone trained initially in physics and philosophy. Since 1974 he has been researching the causes, consequences and regulation o f technological change in the food industry. Since 1998 he has been investigating the links between the science and politics o f bovine spongiform encephalopathy for projects funded by the European Commission. Knud Bprge Pedersen Danish Veterinary Laboratory 27 Blowsvej DK-1790 Copenhagen V D enm ark Tel: (45) 35 30 01 23 e-mail: Kbp@svs.dk Dr. K nud B 0rge Pedersen DVM, Ph.D., DVSc. has been Director o f the Danish Veterinary Laboratory (DVL) since 1985, and o f the Danish Veterinary Institute for Virus Research since October 2000. His doctoral thesis was on infectious keratoconjunctivitis in cattle, and he is auth or/join t author o f a num ber o f scientific articles on microbiology and infection pathology. During recent years, DVL has established research groups on antimicrobial resistance mechanisms and epidemiology, particularly concerning antimicrobial growth prom oters. Dr Pedersen has initiated the formulation o f a veterinary antibiotic policy and guidelines for prudent use o f antimicrobials in veterinary m ed icin e. David Santillo Greenpeace Research Laboratories Department o f Biological Sciences University o f Exeter Exeter EX4 4PS United Kingdom e-mail: D.Santillo@ exeter.ac.uk Dr. David Santillo is a m arine an d freshwater biologist and a senior scientist with the Greenpeace Research Laboratories. He represents Greenpeace International at various international conventions addressing marine environmental protection. Arne Semb Norwegian Institute for Air Research (NILU) P.O. Box 100 N-2027 Kjeller Norway Tel: (47) 63 89 80 00 e-mail: arne.sem b@ nilu.no Dr. A rne Sem b gradu ated in chemistry from the University o f Oslo and has worked on a num ber of different air pollution problem s at NILU since 1971. His major involvements include the OECD long-range transport of air pollutants program m e 1972-77, the Norwegian SNSF project on acid precipitation and its effects on forests and fish 1972-80, and the EMEP program m e for monitoring and evaluation o f long-range transmission o f airborne pollutants in Europe from 1977 to the present. Andrew Stirling SPRU - Science and Technology Policy Research University o f Sussex Brighton BN1 9RF United Kingdom Tel: (44) (0)1273 877118 e-mail: A.C.Stirling@ sussex.ac.uk Dr. Andrew Stirling is a Senior Lecturer at SPRU. He has researched and published widely on issues relating to technology assessment, environmental appraisal and risk analysis. H e has worked with a variety o f academic, public interest, industry and government organisations and served on a num ber of Advisory Committees. Author biographies 199 Shanna H. Swan Research Professor University o f Missouri, Columbia D epartm ent o f Family and Community M edicine Medical Sciences Building MA306L Columbia MO 65212 United States Tel: (1) 573 884 4534 e-mail: swans@ health.m issouri.edu Professor Sh an n a H . Swan Ph.D. is an epidem iologist and statistician who studies the reproductive risks o f DES, other horm ones and environmental exposures. She is currently Research Professor at the University o f Missouri, Colum bia and directed the Reproductive Epidemiology Section o f the California Departm ent of H ealth Services 1981-98. H er approach to public health, particularly in relation to risks to the unborn foetus, has always been precautionary. S o fia G u ed es Vaz Integrated Assessment and Reporting European Environment Agency Kongens Nytorv 6 1050 Copenhagen K Denmark Tel: (45) 33 36 72 02 e-mail: Sofia.Vaz@eea.eu.int Sofia G uedes Vaz is an environm ental engineer who has worked at the European Environment Agency since 1997, specialising in data, reporting, targets and environmental em erging issues. She has an MSc in Environmental Technology from the Imperial College and before working at EEA did environmental consultancy. Martin Krayer von Krauss Ph.D. student Department of Environment and Resources Technical University o f Denm ark e-mail: m kvk9@yahoo.com Patrick van Zwanenberg SPRU - Science and Technology Policy Research University o f Sussex Brighton BN1 9RF United Kingdom Tel: (44) (0)1273 877141 e-mail: p.f.van-zwanenberg@sussex.ac.uk Martin Krayer von Krauss has a BEng from the Royal Military College o f C anada and has worked as project engineer on petrolcontaminated site investigations. He has an MSc in environmental engineering from the Technical University o f Denmark. His Ph.D. study is on `Incertitude and its im plications for environmental engineering'. Dr. Paddy van Zw anenberg trained initially in environmental science before taking postgraduate degrees in science and technology policy. His research focuses on issues o f science an d governance, with a particular emphasis on the role o f scientific expertise in public and environmental health policy-making. H e has worked on two European Commission funded projects on the science and politics of bovine spongiform encephalopathy. Brian Wynne Centre for the Study o f Environmental Change Institute for Environment, Philosophy and Public Policy Furness College Lancaster University Lancaster LA1 4YG United Kingdom Tel: (44) (0)1524 592653 e-mail: b.wynne@lancaster.ac.uk Brian Wynne is Professor o f Science Studies at the Institute for Environment, Philosophy and Public Policy. H e was a m em ber o f the E u ropean Environm ent A gency's m anagem ent board and Scientific Committee from 1995 to 2000. He has completed extensive research and publications on risk, environmental and science studies issues and on public understanding o f science and risk. 200 Late lessons from early warnings: the precautionary principle 1896-2000 Index: Late lessons from early warnings: the precautionary principle 1896-2000 Note: Page references in bo ld refer to figures; those in italics to tables or boxed m aterial A abortion, spontaneous 84, 86-7, 176 acid rain 102-8 aerosols 12, 79, 80 A griculture, Fisheries an d Foods, Ministry, U K (MAFF) 157, 161-2, 180, 181 air quality standards 101-2 Airborne Antarctic Ozone Experiment, US (AAOE) 80 aircraft, supersonic stratospheric 79 Aksoy, M 39 Albers-Schonberg, HE 32 aldrin 127 Alice, a Fight for Life 57 Alkali Act 101 alternatives, evaluation 177 Alverson Report 21, 27 American Conference o f Governmental Industrial Hygienists (ACGIH) 39, 43, 47 American mud-snail 136 American Petroleum Institute 39, 47 amosite 52, 56 androgens 149 animal feeds PCB contam ination 71 slaughterhouse wastes 157, 158, 160 Antarctica, ozone hole 76, 80, 82, 172, 173 antibiotics 93 antifouling paints, see T B T (tributyltin) antifoulants antimicrobial growth prom oters 173, 177 advantages/disadvantages of use 97-8, 99 alternatives 177, 178 ban of avoparcin 96 delays in research 172 early warnings and actions 99 introduction o f 93 lessons from 98 research and uncertainty 204 scientific reports and recom m endations on 97 secondary benefits o f precaution 202 Swann R ep ort 94-5, 98, 99, 170, 171, 172, 173, 174, 181 Swedish ban 95-6 antimicrobial resistance 93-4, 98 antimicrobials 93 aplastic anaem ia 38-9, 42 A rcachon Bay, France 136-7, 138, 139 Archives ofEnvironmental Health 112 Arvin, Professor Erik 114 asbestos substitution 57, 58, 61, 173-4 types o f 52, 56-7 uses 58 asbestos exposure 3, 11, 15 actions an d inactions 56-8, 59-60, 61, 172, 173, 179 and asbestosis 53-4 com pensation claims 52, 57, 58, 60-1 cost-benefits o f action/inaction 58-9 econom ic factors 59 `healthy survivors' fallacy 60, 178 and insurance companies 54, 58, 179 latency of disease appearance 55, 60 lessons from 59-61, 179 and lung cancer 54-5 and mesothelioma 52, 55-6 asbestosis 52, 53-4 assumptions 174-5 asthma 112-13 atmosphere, structure of 78-9 Atomic Energy Agency, International (IAEA) 35 avilamycin 97 avoparcin 94, 96, 97 B bacitracin zinc 96 bald eagle 126 Baltic Sea fauna 65, 72 Barcelona Convention for the Protection of the M editerranean Sea against Pollution 69 BBC (British Broadcasting Corporation) 56 Beaver R eport 101 Becquerel, H enri 32 Beijing Declaration 81, 83 Belgium 71 Benedick, Richard 80 benzene actions and inactions 39-42 causes o f inaction 42-5 costs o f inaction 41 in gasoline 45, 46, 175 industrial uses 38 risk assessment debate 40-1 substitutes 42-3 toxicity and carcinogenicity 38-9, 39-40, 41-2, 42-3, 179 US controls 40-1 Benzene Decision (US) 40-1 Bertram , Ja m e s 17, 18 bioaccum ulation 170-1, 175 Index: Late lessons from early warnings: the precautionary principle 1896-2000 201 organochlorines 64-5, 66, 67, 70, 126-7 organotins 140 birds 64, 126, 127, 152 `blind spots' 173-4 Blue Ribbon Panel, Oxygenates in Gasoline 114-15 Blum, Theodore 32-3, 36 bone sarcoma 32-3 bovine spon giform encephalopathy, see BSE breast cancer 151, 153 breast milk 67, 70 Bridges, Professor Jim W 149, 195 Bridges, Dr O lga 149, 195 brom ine, tropospheric 77, 78 brom och lorom eth an e 81 Brown, Sandford 64 BSE (bovine spongiform encephalopathy) costs o f 164 early warnings and actions 166 European response 163 first cases 157 government policy options 158-9 government reassurances and inactions 161-4, 176 Phillips enquiry 159, 165 potential o f precautionary approach 164-6, 196 public reaction 178 rendering o f anim al wastes 157, 158, 180 research 163-4 slaughterhouse controls 160-1, 162, 178 Southwood Committee 159-61, 179 transmission 157, 158, 160-1, 172, 174 U S precautions 12, 158, 202 Buccinum undatum 139 burden o f p roof 107, 199-200 Bureau o f Fisheries (US) 20, 180 Byers, Elsen 33 C California MTBE pollution 114 sardine fisheries 19-20, 27, 180 Canada appeal against French asbestos ban 53, 57-8 CFCs 80 cod fisheries 20-3, 24-5, 27 Department of Fisheries and Oceans (DFO) 20-2 cancer and asbestos exposure 52, 54-6, 58 and benzene exposure 38-40, 41-2, 42-3, 179 and ionising radiation 36 latency periods 55, 56, 60 and oestrogenic com pounds 84, 85-8, 151, 153 and UV radiation 76 Cape Asbestos Company 54 C a p e 's Acre Mill asbestos plan t 56, 57 carbon m on oxide 111 carbon tetrachloride 78 Carson, Rachel 116, 126, 127-8 Cartagena Protocol on Biosafety 14 Carter, President Jim m y 127 case studies authors 11, 12, 195-199 structure o f 11 catfish 152 cattle offal 160-1, 162, 178 causal relationships 129-30, 131 Center for Health, Environment and Justice (US) 127 Central Veterinary Laboratory (UK) 159 cetaceans 140 CFC-11 77, 78, 79, 80 CFC-12 77, 78, 79, 80 CFCs (chlorofluorocarbons) 3 actions on releases 79-81 atmospheric persistence 77, 170-1 essential uses 77 late lessons 82-3 Montreal Protocol 80-1 production and release 76, 77, 79, 82 stratospheric ozone depletion 76, 78, 79, 82, 173 substitutes 81, 177 use of precaution 12 Chant, Professor Donald 128 Chemical M anufacturers Association 80 chemotherapeuticals 93 chickens antimicrobial growth prom oters 93 PCB contam ination o f feeds 71 child development m aternal oestrogenic com pounds 86, 88, 151,152-3 m aternal PCB exposure 67, 69, 70, 72, 127 childhood asthma 112-13 childhood leukaemia 34 China 81 Chinese Academy o f Preventative Medicine 44 chloracne 64, 66 chlorine atm ospheric 76, 77, 78 industrial use 79 chlorofluorocarbons, see CFCs cholera 14-15 chrysotile 52, 56 cigarette sm oking 55 cities, air pollution 101, 106, 111, 112 CJD (Creutzfeldt-Jakob disease) 157, 158, 174 Clean Air Act (U S) 80, 102, 110-11 clear-cell carcinoma, vaginal 84, 85-8 CLRTAP (Convention on Long-range Transboundary Air Pollution) 104, 105-6 coal burning 101, 104, 105 cod fisheries 20-3, 24-5, 27 cognitive developm ent 69, 127 202 Late lessons from early warnings: the precautionary principle 1896-2000 Collins R eport (1989) 152 Com m on Fisheries Policy (EC) 24, 27 com m unication, professional 189 compensation 60-1 asbestos dam age 52, 57, 58 consensus organisations 43 Conservative Government (UK) 158 constructive technology assessment (CTA) 186-7 contraceptives 89, 151 Convention on Long-range Transboundary Air Pollution (CLRTAP) 104, 105-6 Convention for the Protection o f the Marine Environment o f the Northeast Atlantic (OSPAR) 140, 143 Cooke, Dr W 54 cost-benefit analyses 175-7, 187 Creutzfeldt-Jakob disease (CJD) 157, 158, 174 crim inal law 184 critical load concept 106, 107 `critical path' issues 172 crocidolite 52, 56 Curie, Marie 32, 33 Curie, Pierre 32 Czechoslovakia 104, 105 D Dally, Clarence 31 Daly, Dr H elen 129 DDT (dichlorodiphenyl trichloroethane) 126,152 de Bort, Teisserene 78 D eane, Lucy 11, 53, 60 Delaney Clause 12, 149 Denm ark 96, 114, 115, 118 Dennis, Jo h n 31, 36 D epartm ent o f Agriculture (US) 128, 158, 181 Department of Fisheries and Oceans (Canada) 20-3 Departm ent o f Labor (US) 40 Der Spiegel 104 DES (diethylstilboestrol) as livestock growth prom oter 149-53 EU actions 149, 150, 153 human health concerns 149-51, 152-3 illegal use 151 U S actions 12, 149-50, 153 therapeutic use 84-90 carcinogenicity and teratogenicity 84, 85-6, 88 economic interests 88-9 extent o f 87-8, 89 inefficacy in m iscarriage 86-7, 88, 176 lessons from 88-90 detergents 116 Detroit Wastewater Treatm ent Plant 130 dibenzo-p-dioxin 68 dibenzofuran 68 dichlorodiphenyl trichloroethane, see D D T dieldrin 126, 127, 128 diethylstilboestrol, see D ES dioxins 68, 127 Dodds, Charles 84-5 dogwhelks 136, 137, 139 Doll, Sir Richard 54, 57 dolphins 140 Drinker, Cecil K 64 drinking water 112, 114 drug assessment 89-90 Du Pont 79, 80 E eagle, bald 126 Eastern Europe 104, 105, 107 Economic Commision for Europe (ECE) 104, 107 economic interest groups 59, 88-9, 178-9 ecosystem approach 25-6 Ecosystem Principles Advisory Panel (US) 25 Edison, T hom as 31 Edqvist, Professor Lars-Erik 93, 195 eels 66 Egner, H ans 102 Eli Lilly 89 `em antoria' 33 endocrine disruption 67, 88, 113, 152 endrin 128 An Enemy of the People 55-6 energy consum ption 102 engine technology 118 Enhydra lutris nereis 140 Environmental Defense Fund 128 Environmental Health Officers, Institute (UK) 162 Environmental Protection Agency (Denm ark) 114, 115 Environmental Protection Agency (Finland) 113, 115 Environm ental Protection Agency (U S) 113, 114-15, 128, 130 Erie, Lake 126, 128 ethanol 118 ethers 115 ethics 188-9 European Chem icals Bureau 112 European Commission Com m on Fisheries Policy 24, 27 Communication on theprecautionary principle 13, 24 scientific advisory system 163, 180 Scientific Steering Committee 97 Scientific, Technical and Econom ic Committee 24 European Court o fJustice 96 E u ropean Environm ent Agency (EEA) 3, 11 European Scientific Technology Observatory (ESTO ), Technological Risk and the M anagem ent of Uncertainty project 168-9 Index: Late lessons from early warnings: the precautionary principle 1896-2000 203 European Union 3 action on MTBE 111, 112, 116 action on PCBs 69 antimicrobial growth prom oters 97 fisheries policies 23-4 horm one growth prom oters 149, 153 ionising radiation directive 35, 36 M aastricht Treaty 14 PCBs/PCTs directive 69 response to BSE crisis 163 Scientific Committee on Animal Nutrition 96 Scientific Committee on Veterinary M easures Relating to Public Health 152 everninomycin 97 `evidence of no harm ' 172 `exp ert' opinions 61 F Factory D epartm ent (UK) 54, 56 Facts versus fears 12 `false negatives' 12, 53, 60, 184 `false positives' 12-13, 60, 184 Farm an, Dr Jo e 12, 207 FDA, see Food an d D rug A dm inistration Federation o f Swedish Farm ers 95 Feedingstuffs Act (Sweden) 95 fetotoxicity 67, 69, 70, 84, 86, 88 Finnish Environm ental Institute 116 Finnish Environmental Protection Agency 113 fish and acid deposition 102-4 organochlorine contamination 126-9, 131-2 fisheries 176 Californian sardine 19-20, 27 cost-benefits o f actions 19, 20, 176 ecosystem approach 25-6 FAO Code o f Conduct 23, 24, 25 international agreem ents 23 late lessons 26 lay knowledge 17-18, 177-8 Newfoundland cod 20-3, 24-5, 27 precautionary action 23-5 research and m odels 172, 175, 204 Scottish herring 17-18 steam trawling 18-19 UN Agreem ent on Straddling Fish Stocks and Highly Migratory Fish Stocks 23 fishing mortality 24-5 fishing nets 17 Fishingfor truth 22 flame shell 138 Fleming, Alexander 93 flue gas desulphurisation 105 Food and Agriculture Organization (FAO) 175 Code o f Conduct for Responsible Fisheries 23, 24, 25 Jo in t Expert Committee on Food (JECFA) 150,152, 153 Food and Drug Administration, U S (FDA) 84, 86, 89, 149-50, 154 Food Standards Agency (UK) 165, 166 forest decline 104-6 Framework Convention on Climate Change 14 France asbestos ban 53, 57 marine antifoulants 136-7 freshwaters acidification 102, 106 detergents in 116 PCB pollution 66, 70 see also groundw ater fuel desulphurisation 104 furans 68, 69 G gasoline benzene 45, 46 lead 110 MTBE 111, 112 gasoline direct injection (GDI) 118 Gee, David 208 genetically m odified organism s (GM Os) 185 G eological Survey (U S) 111 Germany 96 Advisory Council on Global Change 182-3 Clean Air Act (1974) 13 Council o f Environmental Advisers 105, 107 Green Party 107 sulphur dioxide emissions 104-5 Gibbs, Lois 126-7 Gilbertson, Dr M ichael 12, 126, 196 glass fibre 57, 61 GMOs (genetically m odified organism s) 185 governance 187-9 governments 180-1 Great Lakes contamination actions/inactions 127-8 causal relationships 129-30, 131 costs and benefits of actions 130-1 early warnings and actions 132 effects on wildlife 67, 126-7 hum an health 127, 129, 131-2 research 172-3, 181 water quality improvements 129-30 Great Lakes Science Advisory Board 130-1 Great Lakes Water Quality Agreem ent 128,129 Great Lakes Water Quality Guidance 130 G reen Party, Germ any 107 Greenberg, Dr Morris 12, 208 G reenpeace 25 groundwater pollution 111-12 204 Late lessons from early warnings: the precautionary principle 1896-2000 growth prom oters, see antim icrobial growth prom oters; horm one growth prom oters H haemopathy 38 halocarbons actions on releases 79-81, 83 atm ospheric persistence 77, 170-1 early warnings and actions 83 essential uses 77 late lessons 82-3 Montreal Protocol 80-1 production and release 76, 77, 79, 82 stratospheric ozone depletion 76, 78, 79, 82, 173, 195 Halowax Corporation 64 H am ilton H arbour 131 Harremoes, Professor Poul 208 Harris Report 21-2 hazardous waste sites 130-1 H C FCs (hydrochlorofluorocarbons) 76, 81 health m onitoring 203 Health and Safety Commission (UK) 58 H ealth an d Safety Executive (UK) 61 healthy survivors fallacy 60, 178 heavy fuel oil 101 heptachlor 127 herring fisheries 17-19 Hickey, Dr Jo sep h 126 H ooker Chem ical Com pany 127 Horizon 56 horm one growth prom oters early warnings and actions 154 effects on wildlife 151-2 EU actions 149, 150 hum an health concerns 149-51, 152-3 illegal use 151 U S actions 149-50, 153 Huxley, T hom as 18 h ydrochlorofluorocarbons (H CFCs) 76, 81 hydrogen fuel cell 118 hyperactivity 70 hypersusceptibility 38 I IAEA (International Atomic Energy Agency) 35 Ibarreta, Dr Dolores 84, 196 Ibsen, HJ 55-6 ICES (International Council for the Exploration o f the Sea) 24, 25 ICNAF (International Commission for the Northwest Atlantic Fishery) 20 ICRP (International Committee on Radiological Protection) 33, 34, 35, 36 ignorance 3 defined 192 responding to 169-71 and uncertainty 185 IJC (International Jo in t Com mission) 128, 130-1 IMO (International Maritime Organization) 138, 181 im posex 135, 136, 137, 138-40 India 81 Industrial Maladies 52 Infante, Dr Peter 12, 209 innovation 182, 186-7 Institute o f Environmental Health Officers (UK) 162 institutional obstacles 180-1 insurance com panies 54, 58, 179 integrated environmental assessm ent 188 Intergovernmental Panel on Climate Change (IPCC) 184, 199 International X-ray and Radium Protection Com m ittee (IXRPC) 33 International Atomic Energy Agency (IAEA) 35 International Commission for the Northwest Atlantic Fishery (ICNAF) 20 International Committee on Radiological Protection (ICRP) 33, 34, 35, 36 International Council for the Exploration of the Sea (ICES) 24, 25 International Jo in t Com mission (IJC) 128, 130-1 International Maritime Organization (IMO) 138,181 International Programme on Chemical Safety (IPCS) 113 Invitational EU Conference on The Microbial Threat 97 ionising radiation, see radiation, ionising Ionising Radiation Regulations (UK) 34, 35 IPCC (Intergovernmental Panel on Climate Change) 184 IQ scores 69, 127 Irgarol 1051 141 irreversibility 170-1 Italy 39 IXRPC (International X-ray and Radium Protection Committee) 33 J Jap an 65, 66, 69 Jefferson, Alice 57 Jen sen , S0ren 64, 68, 71 Joh n s Manville company 56 Johnston, Dr Paul 135, 197 Jo in t Expert Committee on Food (JECFA) 150, 152, 153 K Karnaky, KJ 84 Keats Report 21, 27 Kershaw, Nellie 54 Keys, Jan e 64, 197 knowledge 4-5 Index: Late lessons from early warnings: the precautionary principle 1896-2000 205 `blind spots' 173-4 lay/local 17-18, 177-8 paralysis by analysis 181-2 Koppe, Professor Jan n a G. 64, 197 Kyoto Protocol on Climate Change 82 L Lam bert, Dr Barrie 31, 197 Lam m ing Committee 150, 154 Lancet 58 landfill sites 130-1 Langston, Dr William J 135, 197 Laplan, Bernard 88 `late haem orrhagic disease o f the new born' 70 late lessons 16, 168-9, 193-4 alternatives, evaluation 177 gaps in scientific knowledge 173-4 ignorance and uncertainty 169-71 institutional obstacles 180-1 interdisciplinary obstacles to learning 174 lay knowledge 177-8 paralysis by analysis 181-2 pros and cons 175-7 `real world' conditions 174-5 regulatory independence 178-80 research and monitoring 171-3 social interests and values 178 `latency lacuna' 55, 56, 60 latency periods 55, 56, 60, 195-6 Lawther, Professor 102 lay knowledge 17-18, 177-8 lead 110 Lear, Linda 127 learning institutional obstacles to 180-1 interdisciplinary obstacles to 174 Legge, Dr Thom as 52, 59 leu kaem ia and benzene exposure 38-44 childhood 34 defined 38 life cycle analysis (LCA) 187 Lima hiams 138 lives, valuation o f 58 Lloyd's o f L o n d on 58 London cholera outbreak 14-15 sm ogs 101 loud and late warnings 205 Love Canal, Niagara Falls 126-7 luminous paint 32-3 lung cancer 52, 54-5 Lykke, Erik 103 M M aastricht Treaty 14 M acGarvin, Dr M alcolm 17, 197-198 MacGregor, Jo h n 159 macrolide antibiotics 94 `m ad cow disease', see BSE M adden, M ax 56 MAFF (Ministry o f Agriculture, Fisheries and Foods) 157, 161-2, 180, 181 Marine Environmental Protection Committee (MEPC) 140 marine gastropods 135, 136, 137, 138-40 marine m am m als 65, 140 Marine Pollution Bulletin 13, 137 Martland, Harris 33, 36 Massachusetts General Hospital 86 meat horm one residues 152 risk o f BSE 159, 160-2 m eat industry 150, 152, 159, 163, 164 mechanically recovered m eat 162 m edia 56, 57, 189 Meldrum, Keith 160 m elengestrol acetate 150 Merewether, Dr E. R. A. 54 mesothelioma, compensation payments 52, 55-6, 58 methyl chloroform 77, 78 methyl tert-butyl ether, see M TBE Michigan, Lake 69, 126, 127, 129 Midgely, Thom as 79 Millstone, Dr Erik 157, 198 m ineral wool 61 mink 67, 126 m irex 126, 128 m iscarriage 84, 86-7, 88, 176 Mississippi River 128 m onitoring 181 long-term 171-3 Monsanto 64, 65, 66, 71, 128 M ontreal Protocol 14, 77, 80-1, 83 Beijing m eeting 81, 83 Multilateral Fund for Implem entation (MFMP) 81 M TBE (methyl tert-butyl ether) alternatives 117-18, 119, 177 and asthm a 112-13, 119 benefits o f 110, 111 commercial production 110-11 cost-benefit analysis 118-19 in drinking water 114 early warnings and actions 121 endocrine disruption 113 groundwater contamination 112, 119, 174, 176 national and government recommendations on 114-15 persistency 111-12, 116-17, 120, 170-1 potential carcinogenicity 112, 113 potential use o f precaution 115-20 present trends in use 115 multi-criteria m apping 187 Multilateral Fund for the Implem entation of the M ontreal Protocol (MFMP) 81 multiple myeloma 38, 41, 42, 44 206 Late lessons from early warnings: the precautionary principle 1896-2000 Murray, Dr M ontague 53 mussels, blue 139 Mutscheller, Arthur 33 myelodysplastic syndrome 42 myelofibrosis 45 myeloproliferative disorders 42 Mytilus edulis 139 N NAFO Northwest Atlantic Fisheries Organization) 20 NASA (North American Space Agency) 82 Nassarius obsoletus 136 National Agricultural Chemicals Association (US) 127-8 National Cancer Institute (US) 44, 87 National Institute for Occupational Safety and Health (NIOSH) 41, 43, 44 National Institutes of Health (US) 86 National Oceanic and Atmospheric Administration, US (NOAA) 138 National Ozone Expedition, U S (NOZE) 80 National Radiation Protection Board, UK (NRPB) 32, 36 National Research Council (US) 182, 186 National Toxicology Program (US) 113 Nature 80, 103, 173 Netherlands 58, 67, 88 neutropenia 45 New York Times 80, 93 New Yorker 127 New Zealand 188 Newhouse, Dr M 56 N iagara river 130-1 Nice D ecision 13 NIOSH (National Institute for Occupational Safety and Health) 41, 43, 44 nitrogen oxides 111 Nixon, President R 128 NOAA (National Oceanic and Atmospheric Adm inistration) U S 138 `no evidence o f h arm ' 172 non-governmental organisations 79-80 non-H odgkin's lym phom a 41, 42, 44, 45 N orth A m erican natives 17 North Sea fisheries 17-19, 24-5 and PCBs 69, 72 TBT antifoulants 139-40 North Sea Conference 14 Northwest Atlantic Fisheries Organization (NAFO) 20 Norway 80, 96, 103, 180 Norwegian Institute for Air Research 103 NOZE (National Ozone Expedition) US 80 NRPB (National Radiation Protection Board) U K 32, 36 Nucella lapillus 136 nuclear industry 34, 61 null hypothesis 183-4 O O ccidental Chem ical C orporation 131 Occidental Petroleum 127 Occupational Health and Safety Administration (US) 40 Ocenebra erinacea 136 Odn, Svante 102 oestradiol-17B 149, 150, 152, 153 oestrogens biological role 84, 149 synthetic, see D ES (diethylstilboestrol) Ontario, Lake 126, 128, 129 Organisation for Econom ic Co-operation and Development (OECD) 66, 72, 116 Air M anagem ent Sector Group 101, 103 organochlorines bioaccum ulation 126, 127 Great Lakes contamination 126-32 and human health 126-7 see also named compounds OSPAR Convention (Protection of the Marine Environment o f the Northeast Atlantic) 140, 143 O utboard M arine C orporation 131 overfishing Californian sardines 19-20, 27, 180 early recognition o f 17 fishing mortalities 24-5 Newfoundland cod 20-3, 24-5, 27 Scottish herring 17-18 oyster drill 136 oysters fisheries 136-7, 138 ozone, industrial uses 78 ozone layer depletion 76, 79-80, 82, 172, 173 discovery 78-9 P `paralysis by analysis' 181-2 Paris Commission (PARCOM) 138 participatory approaches 186, 188-9 PCBs (polychlorinated biphenyls) 64, 72, 152 chemistry an d toxicity 68, 68 chicken feed contam ination 71 congeners o f 66, 70 early w arnings an d actions 73 environmental distribution and persistence 64-5, 66, 72, 170-1, 174 governm ent and industry actions 66, 69, 71-2 Great Lakes contamination 126-32 h um an health effects 64, 65, 67, 69, 70, 127, 129 rice oil contam ination 65, 66, 67, 69 routes of environmental exposure 70-1 uses o f 64, 70 pedascopes 34 Index: Late lessons from early warnings: the precautionary principle 1896-2000 207 Pedersen, Dr Knud B 0rge 93, 198 pelvimetry 32, 34 penicillin 93, 95 `pensioners' party' fallacy 60, 178 peregrine falcon 127 persistency 111-12, 116-17, 120, 170-1 organochlorines 131-2 or organotins 142 Peto, Ju lian 56 petrol, see gasoline pharmaceutical companies 88-9 phenobarbital 70 Phillips Inquiry 159, 165 pigs 160 Pim enta R eport 151 Plymouth Sound 137 Poland 104, 105 policy-making 4-5, 195 integration of science 15-16 levels o f proof required 197-8 public participation 188-9 political adm inistrations 180 political interests 178-9 politicians 15, 16 `polluter pays' principle 59, 176-7 Pollution Probe 128 polychlorinated dibenzofurans 68 poultry 71, 93, 160 power generation plants 101, 104, 105 precaution cost-benefit analyses 175-7, 187 and governance 187-9 and innovation 186-7 justification 197-8 and science 183-6 wider implications 182-3 precautionary prevention, defined 192 precautionary principle 3 clarification of terms 192 exam ple o f early use 14-15 examples from U S 12 in international agreem ents 13, 14 m ean in g o f 12, 13 proportionality and time 60 U S /E U disputes 12 pregnancy X-rays 32 oestrogenic com pounds 84, 85-8 PCB exposure 67, 69, 70 prevention, defined 192 Price, C. W. 54 progesterone 150 proof levels required for precaution 184, 193 proportionality principle 60, 196 `p ro s' and `cons' 175-7 prostate cancer 153 public trust in politicians an d scientists 16 understanding o f scientific uncertainty 185-6 public awareness 188-9 Public Health Laboratory Service, UK (PHLS) 163 public interests 178, 181-2 public values 178, 181-2, 193 R radiation, ionising contemporary risk estimates 34-5, 36 early warnings and actions 36 first controls on exposure 33 first reported injuries 31-3 liability for injury 35-6 m edical exam inations 35, 36, 175, 176 medical treatments 31, 33 misuse 34 radiation protection first m easures 32, 33 post-war developm ent 34-5 radiation, ultraviolet 76 Radiation Workers Com pensation Schem e 36 radioderm atitis 31 Radiothor 33 radium, discovery 32-3 `radium jaw ' 32-3, 36 radium standard 33 radon 33 raincoat manufacture 42, 46 RCEP (Royal Commission on Environmental Pollution) 158, 182, 186 `real world' conditions 174-5 reform ulated gasoline (RFG) 111 Registry o f Clear-Cell Adenocarcinom a o f the Genital Tract in Young Fem ales 87 rendered anim al wastes 157, 158, 160 reproductive tract abnorm alities 86 research lack o f baseline 171-3 paralysis by analysis 181-2 Rhine, River 66 rice oil contamination 65, 66, 67, 69 Rio Declaration on Environment and Developm ent 13, 14 risk acceptability of 72 clarification o f term 192 significant 40-1 risk appraisal approaches 187-8 interested parties 179-80 Rochdale, asbestos-related disease 54 Roentgen Society 33 R oentgen, W ilhelm C on rad 31 Rollins, William 32, 36 Rossby, C. G. 102 Rotheim, Eric 79 Royal Commission on Environmental Pollution (UK) 158, 182, 186 208 Late lessons from early warnings: the precautionary principle 1896-2000 rubber industry 38, 42 rubella infection 67 S St Bartholom ew 's H ospital 101, 102 salm on farm ing cages 137, 138 Santa Monica, California 114 Santillo, Dr David 135, 198 sardine fisheries, California 19-20, 27, 180 Scandanavia, acid deposition 102-4 Schweitzer, A lbert 13 science and precautionary principle 183-6 public trust in 16 sociology o f 184-6 refutation o f evidence 55-6, 179 scientific uncertainty 4, 15 benzene toxicity 43-5 and ignorance 185 public understanding of 185-6 and research 203-5 Scottish herring fisheries 17-19 scrapie 157, 174, 181 sea birds 64 sea otter 140 SEAC (Spongiform Encephalopathy Advisory Committee) 161, 165-6 seals 65, 140 Selikoff, Dr I. 55-6 Semb, Dr Arne 101, 198 sensitivity and scenario analysis 187-8 sex horm ones 84 Shell Chem ical Com pany 127, 128 shoe manufacture 39, 43, 46 `shooting the m essenger' 55-6, 179 `significant risk' 40-1 Silent spring 116, 126, 127-8 Simpson Committee 56-7 skin cancer 76 slaughterhouses, specified bovine offal 160-1, 178 Sleggs, Dr C. A. 55 sm og 101 smokestacks 101, 105, 107, 174, 176 Snow, Dr Jo h n 14-15 social interests 178, 181-2 Socrates 4 soil acidification 106 Southwood (BSE) Committee 159-61, 179 specified bovine offal (SBO) ban 160-1, 162,178 spiramycin 94, 96 Spongiform Encephalopathy Advisory Committee (SEAC) 161, 165-6 stakeholder involvement 186, 193 statistical p roof 183-4 steam trawlers 18-19 Stewart, Alice 32, 34, 36 stilboestrol, see D ES (diethylstilboestrol) Stirling, Dr Andrew 198 Stockholm Convention on Persistent Organic Pollutants 14 stratospheric ozone, see ozone layer subsidiarity 188 substitutes, evaluation 177 Sullom Voe 138, 139 sulphur dioxide emissions CLRTAP Protocol 104, 105-6 cost/benefits o f control 104-5, 176, 202 critical load concept 106, 107 European trends 1880-1990 103 forest decline 104-5, 105-6 freshwater acidification 102-4 from fuel oils 101 role o f precaution 106-7, 180 tall smokestack solution 101, 105, 107, 174, 176 urban air quality 101-2, 106, 111, 174, 176 supersonic stratospheric transport (SST) 79 surprises 169-71 survivors, fallacy o f healthy 60, 178 Sutherland, Ian 17 Swan, Professor Shanna H 84, 199 Swann Com m ittee 94-5, 98, 99, 170, 171, 172, 173, 174, 181 Sweden acid deposition 102-4 action on PCBs 66 actions on antimicrobial growth prom oters 95-6, 99, 180 benzene poisoning 38, 42, 46, 47 Swedish chem ical law 183, 184 Swedish Farmers, Federation 95 T Tage, Ronald 54, 55 Tait, Nancy 57 Taiwan, rice oil contam ination 67 TBT (tributyltin) antifoulants alternatives 141, 142, 177 bioaccum ulation 140, 141 control in small vessels 138-9 development and use 135-6 early warnings and actions 143 effects on marine molluscs 136, 139-40 environmental quality standards 137, 138,141 hum an intake 140 im pacts in A rcachon Bay, France 136-7, 138, 139 impacts in United Kingdom 137-8 lessons from 141-2 prohibition o f 138, 140-1 use in seagoin g vessels 139-40, 181 TCDD (tetrachlorodibenzoparadioxin) 68, 70 technological options analysis (TOA) 187 technology assessment 175-7, 186-7 testosterone 150 Index: Late lessons from early warnings: the precautionary principle 1896-2000 209 tetrachlorodibenzoparadioxin (TCDD) 68, 70 tetracycline 93, 95 T hom son, Elihu 31 trade disputes 57-8, 153, 175 transport systems 118 trawlers 18-19 trenbelone acetate 150 triazine herbicide 141 trout 152 Trudeau, Prime Minister 128 Turkey, benzene poisoning 39, 43, 46 Turner Brothers' asbestos plant 54, 56, 57, 58 tylosin 94, 96 U uncertainty 169, 170, 192 see also scientific uncertainty United Kingdom asbestos and m esotheliom a 52, 58 fisheries 17-19, 27 Ionising R adiation R egulations 34, 35 sulphur dioxide emissions 105, 106 TBT antifoulants 137-9 United Nations Agreem ent on Straddling Fish Stocks and Highly Migratory Fish Stocks 23 Convention on the Law o f the Sea 20 Fo od an d Agriculture O rganization, see Food and Agriculture Organization (FAO) Rio Declaration on Environment and D evelopm ent 13, 14 United Nations Environment Programme (UNEP) Coordinating Committee on the Ozone Layer 79 Vienna Convention 80, 83 Washington Declaration 69, 72 United States antimicrobial feed additives 95 benzene 40-1 CFCs 79-80 horm one growth prom oters 149-50, 153 scrapie precautions 12, 158 uses o f precautionary principle 12 U n ited States Suprem e C ourt 40, 181 University o f California, Davis 112-13, 114 urban air pollution 101, 106, 111, 174, 176 UV (ultraviolet) radiation 76 van Zwanenberg, Dr Patrick 157, 199 vancomycin 94 vancomycin-resistant enterococci 96, 97 Vaz, Sofia Guedes 199 Velsicol Chem ical Com pany 127, 128 Vienna Convention for the Protection o f the Ozone Layer 80, 83 virginiamycin 94, 96 VOCs (volatile organic com poun ds) 111 von Krauss, Martin Krayer 211 Vorsorgeprinzip 13, 186 W Wagner, Dr J. C. 55 Walton, J. R. 94 W ashington D eclaration 69, 72 Washington Post 80 whelk, com m on 139 White H ouse National Science and Technology Council 113 Women Inspectors o f Factories (UK) 53 World Health Organization (WHO) air quality 102 antimicrobial growth prom oters 97, 99 dioxins 71 FAO Jo in t Expert Com mittee 175 International Agency for Research on Cancer (IARC) 57, 113 Jo in t Expert Committee on Food 150 World M eterological Organization (WMO) 76 World Trade Organization (WTO) French asbestos ban 53, 57-8, 175 horm one growth prom oters 153 Sanitary and Phytosanitary Agreement 199 Wynne, Professor Brian 199 X X-rays discovery 31 injuries from 31-2 m edical exam ination 35, 67, 176 see also radiation, ionising Y Yell Sound 138-9 Yorkshire TV 57 Yucheng accident 67 Yusho accident, Jap an 65, 66, 69 V vaginal clear-cell adenocarcinom a 84, 85-8 values 178, 181-2, 186, 193 Z zeranol 150 210 Late lessons from early warnings: the precautionary principle 1896-2000 European Environment Agency Late lessons from early warnings: the precautionary principle 1896-2000 Environmental issue report No 22 Luxembourg: Office for Official Publications of the European Communities 2001 - 210 pp. - 21 x 29.7 cm ISBN 92-9167-323-4