Document e1QyvzG6YZrp0ERKXZqoYXyk9

Atmiphme finwumf Vot. 21, No. 2, pp 315--330. 131? Pnaud m Gmt Imam. V*1TWVU0>J CHw*V7>E -- CPSC. | METHYLENE CHLORIDE w TASK FORCE T0: WEALTH AND SCIENCI FROM: COMMITTEE TOM CORTINA DATE: 9/14/87 CONSIDERATIONS IN EVALUATING EMISSIONS FROM CONSUMER PRODUCTS John R. Gikman and Alfred T. Hodgson Building Ventilation and Indoor Air Quality Program. Lawrence Berkeley Laboratory. University of California. Berkeley, CA 94720, UJ5A- and SEP ' 8 '9S? Marilyn L. Wind Division of Health Effects, U.S. Consumer Product Safety Commission, Bethesda, MD 202)7, U.S.A. (First received 22 July 1985, in final form 17 February 1986 and received for publication 30 July 1986) Abstract--While several indoor air quality studies suggest consumer products (eg. aerosol sprays, paint removers, etc.) can be significant sources of volatile organic compounds, until recently characterizing emissions from consumer products has received relatively little attention. Many considerations that must be addressed in designing studies ofconsumer product emissions are similar to those addressed in studies of the emissions from building materials and combustion appliances. These stmsiahtics are discussed and. in addition, the considerations unique to studies ofconsumer product emissions an dirmmrf with reference to an ongoing study of consumer products that contain methylene chloride. These unique considerations include bulk chemical analysis, the form ofthe product (e.g. liquid, aerosol or paste)and the related consumer usage pattern. The issue of personal exposure of the product user vs the average area concentration resulting from product use must be considered, as well as the challenge of incorporating results into predictive models that adequately treat the effect of different usage patterns. Finally, post-study considerations, such as exploring new issues discovered in the study, studying similar products, and validating predictive models through extension into field studies are summarized. Key word index: Consumer product, emissions, indoor air quality, methylene chloride, personal exposure, source characterization, ventilation. INTRODUCTION While several indoor air quality studies suggest that consumer products can be significant sources of vol atile organic compounds (De Bonoli et al., 1985; Lebret et a/.. 1984), characterization of emissions from consumer products has received relatively little atten tion. This is in strong contrast to the characterization of emissions from combustion appliances (Johnson er ai7 1984; Leaderer, 1982; Traynor er al., 1982) and building materials (Girman et al, 1987; Molhave, 1982), where numerous chamber studies have been conducted. In this paper, considerations necessary for evaluating emissions from consumer products in lab oratory studies are presented. Although a broad definition of consumer products could include many combustion appliances and some building materials, a narrower definition as used in the National Academy of Sciences monograph. Indoor Pollutants (National Research Council, 1981) is em ployed in this paper. As used here, it is limited to expendable products in small containers which are readily available in retail outlets. In this context, consumer products are items such as cleaning/poUshing products, insecticides, painting/fimshing/refimshJpg products, personal grooming products, hobbyists' 'products, and deodorizers/disinfectants. Many considerations that must be addressed when designing studies of emissions from consumer prod ucts are similar to those addressed in other types of indoor air quality studies. These similarities include: defining the potential health hazard; establishing study goals; assessing analytical instrument and chamber capabilities; pretesting and establishing the experimen tal procedures; and establishing quality assurance/con trol objectives. There are, however, additional con siderations that are unique to studies of consumer product emissions. These unique considerations in protocol develop ment are discussed with reference to an ongoing study of consumer products that contain methylene chloride (CHjC12). This study, supported by the U.S. Consumer Product Safety Commission (CPSCX was undertaken by Lawrence Berkeley Laboratory (LBL) to resolve uncertainties about the exposure consumers can receive by using consumer products containing CH2C1;. LBL has conducted a controlled study in an environmental chamber to measure both the airborne concentrations produced by the use of these products and the exposures of individuals using them. Development and testing of models for both average area concentration and personal exposure are import ant goals of this study. In discussing considerations for studies ofconsumer 315 SL 037195 '!!! 316 John R. Girman k al. produces, eight major tasks are defined. While the order of presentation ofthe tasks is reasonable, it is not a rigid prescription. Certain tasks should be ac complished before starting the next task: however many tasks can be accomplished in parallel or, as is often required, in an iterative fashion. DEFINING THE POTENTIAL HEALTH HAZARD Many interrelated factors must be considered in defining the potential health hazard of a particular product or class ofproducts: the chemical composition of the product; the toxicity of the compounds con tained in the product; an estimate of airborne concen trations and exposures produced by the use of the product; the market penetration (either at the present time or expected in the future due to changing market conditions): and an estimate of the population poten tially at risk. This task should not be confused with actually developing an estimate of a health risk due to use of a product. This would be premature since too many parameters are not sufficiently well known. Rather, the goal at this stage is to judge whether the potential risk is large enough to warrant a detailed study. This initial task is similar for most indoor air pollutant source characterization studies. In some respects, this task is more asily completed for con sumer products than for other indoor air pollutant sources such as combustion appliances, building ma terials or smoking, since consumer products generally have labels describing their chemical contenL However, these labels, while useful as a guide, are not necessarily comprehensive and chemical analyses are usually required. A good example of defining a potential health hazard is provided by CH1G2, a chemical widely used by consumers. More than half of a billion pounds of CH2CI2 are produced annually in the U.S,, much of it for use in paint removers and aerosol finishes (Anonymous, 1985). The chemical composition of these products is readily obtained. For example, semi paste paint removers contain approximately 85% CH2CI2 and aerosol finishes with CH2G1 generally contain 20-40% of this chemical. Market data also exist for these products, although these data are not discussed here. There is concern about consumer exposure to CH2G2 from these products, since CHjQj retained in inhalation is metabolized to carbon monoxide, which can lead to anoxic stress from elevated levels of carboxyhemogiobin (Rainey n aL, 1974; Stewart al,, 1972). In addition, a recent animai-inhalauon labora tory study has associated increased incidence of car cinoma with exposure to CH2G2 (Peer Review Panel, National Toxicology Program Board of Scientific Councilors, L98S). These facts strongly suggest that there is a potential health risk that is worth investigation. ACQUIRING APPLICATION TECHNIQUE AND USAGE PATTERN INFORMATION In completing this task, the goal is to acquire sufficient infonmtiou about product application and use to assure that the product is used realistically in test situations. The term "application technique" is used to denote the detailed method employed by the consumer when using the product. e.g. for a paint remover, "Use a single brush stroke, applying in one direction only. Wait at least 10 min before scraping." Since consumer products are found in a number of different forms including aerosols, liquids, semi-pastes or gels, pastes or waxes, and solids, a broad range of application techniques must be anticipated. In contrast to this, usage pattern information is broader in scope, it would, ideally, include such data as amount used in various applicauons. frequency of use. typical room volume and air exchange rates where the products are used, the frequency with which con sumers increase ventilation by opening windows and/or doors, and the size and types ofobjects to which the products ate applied. There are several ways to acquire information about application techniques and usage patterns. Manufac turing trade associations for the product type may share their market research data about techniques for application and bow consumers actually use the product User surveys are another, probably better, method of acquiring information. These range from structured surveys with questionnaires to informally questioning collogues and acquaintances about their personal use of specific products. However, one of the best and most readfly available sources of information is probably manufacturers' information sheets and the instructions printed on product containers since these are the source of information most often used by consumers. Forthe methylene chloride study two major product types with very different forms were invesugated: paint removers, whichare primarily semi-pastes, and aerosol finishes. Information about application techniques and product usage was acquired through informal surveys (including personal experience on the part of the researchers and program managers) and manu facturers' information sheets and instructions on pro duct containers. ESTABLISHINC STUDY GOALS This task allows the greatest latitude and, accord ingly, the greatest activity for researchers and pro gram managers, since a wide range of choices is available to shape the direction and goals of the study. Initially, it must be decided if the study should be biased toward a "worst case' use or toward a more typical product use or both. A wont case scenario may be appropriate wbea acute health effects are of interest and are likely to occur within the population with some SL 037196 freque prune typica Ver and if time ( metho tempe impor perhai and u: param Anc ing an both, exposi area s tratiot are be strong the im of a concei sonal Of. techm produ this sc Fin: limitei restric which time, 1 In t focus 1 It war with r> ant an would the ei attenn toluep concei this pc attemi used w 3.0 air corns action indooi (A cot object ceases condit cases? kitche be pai painte twocc remov I ;ace and i test -d to uner Use inly, inter trnu istes tion n is taas use, ethe conlows hich bout ifacmay > for the :tter, rom tioa 1 the hese l by duct aint osol ques nul rt of tnupro* sid?ros is idy. 1 be tore may nest ame* Considerations in evaluating emissions from consumer products 317 frequency. A typical use scenario may be more appro priate for chronic health effects or to estimate the typical exposure of larger populations. Very often, source characterization is a primary goal and this characterization can be done with respect to time (in which case the form of the product and the method of application are important!: with respect to temperature (where the form of the product may be important); with respect to the ventilation rate (or perhaps with respect to local air Sows near the product and user); and. or with respect to other environmental parameters. Another decision must be made regarding measur ing area concentrations vs personal concentrations or both. The person using the product can receive an exposure far different than would be predicted from an area sample taken in the same room. The concen tration gradients in rooms where consumer products are being used may be large and. therefore, may have a strong effect on personal exposures. This may lessen the importance of area concentrations in the absence of a model correlating personal exposure to area concentrations and increase the importance of per sonal sampling. Of course, as discussed previously, the application technique can affect emissions from a consumer product Therefore, it must be decided if the effect of this source of variability is to be measured. Finally, the goals must be prioritized. Because of limited resources, the scope of a study must usually be restricted in some way. The actual derisions as to which goals will be pursued need not be made at this time, but they should be prioritized. In the methylene choride study, for example, the focus of the study was to be typical use, not worst case. It was further decided that source characterization with respect to time and ventilation was most import ant and that the effects of temperature and humidity would be studied later, if at all While characterizing the emission of CHjCl* would receive the most attention, the emission of other major solvents such as toluene would also be monitored. The average area concentration and personal exposure were judged, at this point, to be equally important and the study would attempt to measure both. The ventilation rates to be used were both low, OJ air changes per hour, and high, 3.0 air changes per hour. These ventilation rates may correspond to the situation when a consumer takes no action to increase ventilation while using the product indoors and the situation when windows are opened. (A consumer could also, in some situations, take the object to be painted or stripped outside. However it ceases to be an indoor air quality problem under these conditions and, more importantly, is not an option in cases when the object is a permanent fixture, tg. a floor, kitchen cabinets or wall panels.) While the panels to be paint stripped in this study were to be primed and painted several months before the experiments with two coats ofenamel paint, which tends to be difficult to remove, they were to be modest in sire (compared to wall panels, kitchen cabinets or a floor) and they were to be relatively smooth (compared to lathe-turned legs or a carved piece of furniture) and therefore easily stripped. Overall, the choices made tended to cor respond more to typical not worst case, use of the product. ASSESSING INSTRUMENTAL AND FACILITY capabilities Based upon the prioritized list of goals, the re searchers must assess the resources available to con duct the study, in terms of both instruments and a chamber or other specialized facility. This is a straightforward task. For the methylene chloride study, a continuous i.r. analyzer was to be used to monitor both the personal and areaconcentrations on an alternate basis by means of sample line switching. Sampling with charcoal tubes was originally considered for personal sampling, but was rejected for this phase of the study because of insufficient precision and time resolution. A gas chro matograph (CO was also available. A chamber de signed for studies of organic emissions was available. Its ventilation system was adequate in terms ofair flow and size, but at the time, it lacked temperature and humidity control However, since strict requirements were not established for these parameters, control could be accomplished simply by controlling the temperature and humidity of the laboratory housing the chamber. Data acquisition systems and a host computer were available for logging ofanalog and digital input signals from the Lr. analyzer, the GC and the environmental instrumentation, but some software had to be written. WRITING THE PROTOCOL In writing the protocol all of the information collected regarding product types, application tech niques. usage patterns, study goals and the analytical instrument and facility specifications are considered and brought together to construct a unified plan. The detailed experimental procedure is written, incorporat ing the specific instruments and the chamber capabi lities. Calibration procedures are established in detail Consideration should also be given to data acceptance standards. However, in research, as opposed to moni toring, it is inappropriate to set rigid standards for all types of data. Standards can be set for the limits of acceptable data for enviromental parameters, for in strumental drift and for the precision of calibration data. However, setting standards for other data can be difficult due to lack of knowledge about the behavior ofthe source and the effect ofactivity on the pan of the person using the product Audit procedures should be set and written as pan of the protocol The audit can take two forms: internal where checks of instruments, procedures and data are SL 037197 318 John R. Giuman al. conducted and recorded by the researchers themselves according to a formal plan: and external where an outside agency or individuals check the instruments, procedures and data. A detailed checklist ofspecific actions to be executed before, during and after an experiment should be constructed from the key elements of the protocol. Writing the checklist at (his time serves to insure that the protocol is practical and reasonable. However, the checklist should not be viewed as a static endpoint but. rather, as the first draft of a document that will evolve as the researchers gain experience. In the current example, the protocol for the meth ylene chloride study was written by the LBL staffand sent to the CPSC for review. The protocol specified the instruments and the calibration procedures to be used, It established internal audit procedures and named a quality control officer. It described the sampling system, the ventilation rates, the environmental par ameters. the substrate to be finished or stripped and incorporated manufacturers* instructions for product use. Incorporating the instructions for the use of puint removers, in particular, required careful consideration. Paint remover was to be applied sequentially to small sections (0.37 m;) of a 1.5-m: panel. Paint remover would set for a minimum of 10 min prior to scraping. Tools, remover containers and paint scrapings would all be weighted so that the weight data coukl be used as a check of the emission rates developed through the use of a mass-balance, ventilation model CPSC chose to obtain outside review of the protocol Comments and suggestions from CPSC and the outside reviewen were incorporated into the protocol by LBL PRETESTING The pretest period is one of the most interesting periods of a study because the learning curve is so steep. The products are selected for screening and subjected to bulk chemical analysis, generally by GC and/or GC/mass spectrometry. Based upon the remits of the bulk analysis and market considerations, the specific products to be studied are selected. The performance of the analytical instruments is evaluated using the calibration procedures. During this period, more than one calibration system may be used as an overall check of one system against the other. Sampling systems are fabricated, if necessary, and/or tested. Data acquisition software is debugged, and the acquisition system is tested. If at all possible, trial experiments should be con ducted and concentrations measured. During trial experiments, the product application technique is evaluated and standardized. This is especially import ant if more than one person will be applying the product during the experiments and measurement of the effect in variability of application technique is not a study goal The overall experimental procedures are evaluated, as well as the adequacy of the sampling system and the analytical instruments employed. These trial experiments also provide data for testing data reduction and analysis schemes. Often it is discovered that the protocol and checklist should be modified based upon the knowledge and experience gained in the pretest period. In the example study, the bulk chemical analysis of paint removers confirmed a high percentage of CHjGj and lesser amounts of toluene and aliphatic alcohols. The bulk analysis of selected aerosol paints and other aerosol finishes showed lesser percentages of CHjGj but larger amounts of toluene and, in some cases, minor amounts of other solvents. A sampling system was fabricated that switched alternately between a personal sample taken near the breathing zone of the person using the product and an average area sample consisting of the sum of 13 sampling locations in the chamber. These 13 locations are located throughout the chamber at three different heights. The l.r. analyzer was calibrated for CHjG, using both the manufacturer's closed-loop injection system and a mass-flow controlled dynamic gas dilution system with certified gas standard mixtures. Agree ment between techniques was excellent, but the dynamic technique was chosen for use in the exper iments because of ease and speed of use. Application techniques for paint removers were standardized. Preliminary experiments demonstrated that CHjGj emissions could be well characterized and that even short-term variations could probably be characterized. However, personal exposure concen trations appeared to be highly variable and could not be tracked adequately when sampling was alternated between personal and average area locations. Therefore, it was decided to obtain a second i.r. analyzer so that both personal and average area concentrations could be measured continuously. The protocol was modified to reflect this change. The preliminary experiments also graphically de monstrated that the products must be used realistically to obtain valid emission rates. If paint remover was applied to a panel and allowed to set undisturbed, the remover didn't appear to evaporate appreciably and emissions were relatively low. However, if the paint remover was agitated by scraping, the remover would evaporate to near dryness within 5-10 min and emis sions were large. Similarly, vertical stratification of concentrations was also more evident if paint remover was undisturbed. CONDUCTING THE STUDY Once the necessary preparations have been made, completing this task is relatively straightforward and. again, is similar to most indoor air pollutant source characterization studies. If at all possible, the first experiments should be replicate experiments to quan tify the reproducibility of the experimental procedure. For these experiments, the same person should use the SL 037198 product and cate events ever of variations is sufficient, stall termine whethi detect changes The remainm replicate expert reduced. The at experiments an ventilation mod appliance studie temporal variati products, it may model than can ai., 1985). The m to adequately tn The data fron calculate personi pared to exposur centrations. Stat conducted to as: parameters, e.g. b ventilation rate o exposure derived, The chamber chloride study h Hodgson. 19851 product types var the rates were con used (as determine of product applic the average ch breathing-zone co by the use of a paii in the 20-mJ chi concentration Duct use of the paint ret Fig. 1. Temporal pro concentrations of Ch paint remover m a 2& 3.23 h ' *. Durauon of At end of work penoi switched to laboi I :overed lodified ined in lysis of age of iphatie paints ages of 1 some itched ar the and an of 15 rations iffereflt . using system lution Agreeit the :xper- were trated si >ly ncend not nated tions. d lr. area .The y detcally was i, the ' and paint ould snis* n of tover Considerations in evaluating emissions from consumer products 319 product and every attempt should be made to dupli cate events even ifa later goal is to determine the effect of variations in application technique. If the data are sufficient, statistical tests can be employed to de termine whether the reproductability is adequate to detect changes due to experimental variables. The remaining experiments are conducted after the replicate experiments are completed and the data are reduced. The average concentrations obtained in the experiments are incorporated into a mass-balance ventilation model such as has been used in combustion appliance studies (Traynor tt ai^ 1982). Because of the temporal variation of emissions from some consumer products, it may be necessary to use a version of this model than can address these variations (Traynor aL 1985). The model may require some modifications to adequately treat a unique data set. The data from personal sampling can be used to calculate personal exposures and these can be com pared to exposures estimated from average area con centrations. Statistical analyses of the data can be conducted to assess results with respect to selected parameters, e-g. by product type, by product brand, by ventilation rate or by exposure (personal exposure vs exposure derived from the average area concentration). The chamber experiments for the methylene chloride study have been completed (Giiman and Hodgson. 1985). The emission rates for the two product types varied, of course, but the differences in the rates were consistent with the amount of CH:Q2 used (as determined by bulk analysis) and the duration of product application and use. Temporal profiles of the average chamber concentrations and the breathing-zone concentrations of CHjGj produced by the use of a paint remover at a high ventilation rate in the 20-mJ chamber are shown in Fig. 1. The concentration fluctuations produced by the sequential use of the paint remover are clearly evident Temporal profiles of CHjCl- concentrations were calculated from source strengths, the ventilation rates and the chamber volume using single-equation, mass-balance models. For paint removers, two types of source strengths are used: the first, the time-averaged source strength, assumed that the product was used uniformly over time for the duration of the entire work period: the second, the time-dependent source strength, ac counted for the sequential nature of product appli cation. As illustrated in Fig. 2 which contains the modeled profiles from the same experiment illustrated in Fig. 1, these theoretical concentrations were in good agreement with measured concentrations. Exposure models based upon the concentration models were also developed and then evaluated by comparing theor etical and measured exposures for the experiments. When measured personal exposures were compared to chamber concentrations ofCH.Gi integrated over the work periods, agreement was good at the low venti lation rate but averaged 21% higher at the high ventilation rate. For the experiment illustrated in Figs 1 and 2, the exposures were 1120 ppm-h (personal exposure), 921 ppm-h (exposure based upon average chamber concentrations) and 1180 ppm-h (theoretical exposure). For the same paint remover used in an experiment at the low ventilation rate, the exposures were 2400 ppm-h, 2350 ppm-h and 2530 ppm-h, re spectively. The exposure models appeared to have sufficient accuracy and precision for use in assessment of health risk from the use of consumer products containing CHjGj. POST-STUDY CONSIDERATIONS If funding is available, the research may be extended to explore issues ignored because of limited resources; to explore new issues discovered in the study; to study ade, and, orct ure.^^ the Fig. 1. Temporal profiles of chamber and breathing-tone concentration! or CHjOj during an experiment with a paint remover m a 20-mJ chamber at a venulauon rate of 3.23 h ` *. Duration of work period is shown above curves. At end of work period, breathing-zone sampling line was switched to laboratory air external to chamber. S*p*ad wn* Imwi Fig. 2. Temporal profiles of theoretical chamber concen trations of CHjGj for an experiment with a paint remover in a 20-m *1 chamber at a ventilation rate of 3213 h`\ Concentrations were calculated using both rime-dependent and rime-averaged souice strengths. Dunoon of work period is shown above curves. SL 037199 320 John R. Gi*man r at. similar products; or perhaps most importantly, to validate the model developed by conducting a field study. For the methylene chloride study, the remaining issue is the validation of the exposure models de veloped in a field study which will examine the effects of variations in ventilation patterns and rates, in volumes and in product use patterns. Office .of Building Energy Research and Development. Buildings Systems Division of the U.S. Department of Energy under Contract No. DE-ACQ3-76SFQ009S and by the Directorate of Health Sciences of the U.S. Consumer Product Safety Commission under Contract No. CPSC-IAG-84.1171. The ideas and opinions expressed are those of the authors and not necessarily those of the U.S. Department of Energy or the U.S. Consumer Product Safety Commission. This material is in the public domain and in accordance with 17 USC 103 may be fully copied or reprinted. PERSONAL protection Personal protection for the researchers deserves special mention. Unlike studies of combustion appli ances and building materials where it is not necessary for researchers to have prolonged exposures to emis sions, using consumer products can often result in extended periods of exposure. Researchers should be protected even if some accuracy is sacrificed. In the methylene chloride study, a pressure-demand breathing apparatus supplied by a cylinder of air outside the chamber was worn by the person using the product during the experiments. Because the mask exhausted air under the chin of the wearer, it may have diluted the personal sample. Nonetheless, personal protection took precedence. Disposable coveralls were worn during the experiment by the person using the product to prevent dermal contact. Viton gloves were also worn during the use of paint removers. SUMMARY Considerations in evaluating emissions from con sumer products have been discussed with reference to eight that should be accomplished when conduc ting such a study; (1) defining the potential health hazard; (2) acquiring application technique and usage pattern information; (3) establishing study goals; (4) assessing instrumental and facility capabilities; (3) writing protocol; (6) pretesting; (7) conducting the study; and (8) post-study considerations. These con siderations are using examples provided by an ongoing study of methylene chloride emissions from paint removers and aerosol finishes. Emphasis is given to those considerations that are unique to studies of consumer products. These considerations include the need to use the products realistically to obtain valid data, the difference between personal exposure and exposure based upon the average concentration and the factors that affect this difference, and the need for personal protection of the researchers. Acknowledgments--This work was supported by the Assistant Secretary for Conservation and Renewable Enogy. references Anonymous (1983) Resource Center. Poilut. Engng 17,48-31. De Bortoli M,, Knoeppei H,, Pecchto E- Pl A_ Rogon L. Schauenburg H_ Schlitt H. and Visscn H. 11985) Measurements of indoor air quality and comparison with ambient air. Commission of the European Communities Joint Research Center Report EUR 9636. Office for Official Publications of the European Communities. Luxembourg. Girman J. R. and Hodgson A. T. 11983) Source characteriz ation and personal exposure to methylene chloride from consumer products. Lawrence Berkeley Laboratory Report LBL-20227. Lawrence Berkeley Laboratory, Berkeley. CA. Girman J. R,, Hodgson A. T,, Newton A. S. and Winkes A. W. (1987) Emissions of volatile organic compounds from adhesives with indoor applications. Em-ir. Int. iin press). Johnson D.. Billiek 1,, Moschandreas D. and Reiwoani S. (1984) Emission rates from unvented gas appliances. Proc. 3rd Int. Conf. on Indoor Air Quality and Climate. VoL 4. pp. 367-373, Swedish Council for Building Research, Stockholm. Sweden. Lcadcrcr B. P. (1982) Air pollutant emissions from kerosene space heaters. Science 218, 1113-1113. Lebret E, van de Wiel H. J, Bos H. P_ Noij D. and Boleij J. S. M. (1984) Volatile hydrocarbons in dutch homes. Proc. 3rd Int. Conf. on Indoor Air Quality and Climate. Vol. 4. pp. 169-174. Swedish Council for Building Reseantii, Stockholm. Sweden. Moihave L (1982) Indoor air pollutants due to organic gases and vapors of solvents in building materials. Emir. Inc. 8, 117-127. National Research Council National Academy of Sciences (1981) Indoor Pollutants. National Academy Pros, Washington. DC. Peer Review Panel (1985) National Toxicology Program Board of Scientific Councilors, 29 March 1985. Research Triangle Park, NC. Ramey R. S, Wegman D. H. and Elkins H. B. (1984) in moo conversion of methylene chloride to carbon monoxide. Arch. Emir. Hlth. 28, 223-226. Stewart R. D, Fisher T. N,, Hosko M. J, Peterson J. E, Baretta E D. and Dodd H. C (1972) Experimental human exposure to methylene chloride. Arck Emir. Hick 25, 342--348. Traynor G. w, Anthon D. W. and HolloweU C. D- (1982) Technique for determining pollutant emissions from a gasfired range. Atmospheric Environment 16, 2979-2988. Traynor G. W, Girman J. R. Apte M. G,, Dillworth J. F. and White P. D. (1985) Indoor air pollution due to emissions from unvented gas-fired space heaters. J. Air Poilut. Control Ass. 35, 231-237. SL 037200 Atmommnt Printed m G f L ( I 1 1 1 The stut consuitii envirom air quali can pro' nation < sources that sin 1984, 1! 1982a;] tenantpound; sions fi recent .< of a bi 1982b;' by De have al sources The methoc strongl for the 'Rest tion Agi Marun