Document oeOQz9x2wpO2ROqQdram52kQD

GRACE Cv^drvcfitA Oiv'iitn TO: FROM: CC: C O K f J 3 ! I! * : 06151311 R. J. Battaeehl H. C. Duacksr H. A. Esehsr.bach 0. M. Favorito R. M. Vlning R. C. Walsh J. C. Tang Fils 1184 DATE: Saptaaber 7. 953 SUBJECT: Arthur D. Doeunoncs Attachad It eonpltts flla of cha Arthur D. llttla docuaanta lssuad froa Juno 27 to July 13. 1903 toncarniag eha analysis of M0N0K0TE frots Easthaapton (Lot 1F1) and Trantoa (Lot 2FI). Tha tat Includaa cha following lattara aad rtports: 1. 6/27/83 Easthaapton MX analysts 2. 7/U/83 Rasults for wipa out cachnlqua and alcroscople txaa of Trantoa MX 3. 7/12/83 Short ona-paga lattar using Craea/ADl aathod to analysa Traatoa MX. 4. 7/13/83 Sntallad aathodology to analysa Trantoa MX. JCT/ph Actaehwata *47 M4.)77Tt4n2UM Al Arthur D. Little, Inc 06151312 June 27, 1983 Dr. Julie Yang Manager - Research Technologies w.R. Grace l Co. 62 Whittoore Avenue Cambridge, MA 02140 Dear Julie; C76*94 Ir accordance with your request of June 23. 1993. we heve anal ysed a commercial sample of Monokote WX-S (V.R, Grace 6 Co.) for the presence and quantity of smphlbole fibers. For this purpose, we have defined fiber as any particle with approximately parallel sides and with an aspect ratio of 3 or greater. PROCEDURE An unopened bag of Monokote (Lot 1F1. Easthaaptoo, HA) delivered by you was poured on a polyethylene sheet, coned and quartered until reduced to about four pouade. From this sample, one quarter vae reserved (about l pound) end the opposite quarter was coned and quartered to give a sample of about lOOg. This sample was digested with II BC1 at 60*C to dissolve the plaster of parls (che major component). Each digestion utilised about 900 el of acid with 13 mla. of stirring. The suspension vae allowed to settle for 10 minutes efter which the supernote was decanted with a squeese bulb and reserved. This was repeaeed for a total of seven digestions, after vhleh x*ray diffraction felled to show a plaacer of parls (or gypaun) peak. As you recall, a previous analysis hmd shown chat a Ig sample of Montana Trenolite (esbestlform variety) wet ground with a mortar and peetla wee quantita tively recovered after e similar digestion treatment. after digestion, the residue wee wet screened and waahed through e 20# scream. The filtrate wee added to the previous decants. /ti Arthur 0- Uttl. Inc. June 27. 1983 -2- Or. Julie Vang Manager - Research Tachnclogi** V.R. Crace 6 Co. C61S1313 neutralized with SeOH and aade up to iOOO jil. After agitating, templet were withdrawn from the cntr of the suspension to provide separate saaples o( l, - and 16 al. these were each diluted to 100 al and filtered through 0.1 ye pore sire, -7 as dia. Nuelepore filters which were ehaa carbon coated. Saaples (l or 3) were Incised froa each aad placed oa 200 aesh copper tleetroa microscope grids. The filter aedla was cleared with chlorofora using the aodlfled Jaffee wick procedure. when cleared, optical microscopy at 430X Indicated the 4 end 16 al saaples vers boeh heavily loaded. The 1 al saaple was used for electron aleroseopy. For analysis. 20 randoa pore openings were traversed at 10.000X and any partlela aeeelng the fiber definition wee measured (length and width) and analyzed by selected area electron diffrac tion (SAD). Photographs of several typical fibers and their SAEO's ware recorded. USIT.TS For che 20 pore opening, a total of 39 fibers ware observed, with the following Identifications. Catstorr Suaber of fibers Vecalculite (laage aad SAB>) Probable vetaleuliee (laage only) Treaoliee (laage and SAID) Probable treaollte (Image only) Posaibla treaollte (nablguous laage) 29 7 2 12 9 39 laead ea the aaaaurad flbar dlaeaeioaa aad the conservative aseueptioa ef fiber chlckseee being the eaae aa che fiber width, fiber weights were calculated on the beala of both aaaa and aadlna fiber aaaa. The laccar le a sort useful indication of central tendency in cases where a particularly large fiber found la the saaple area analysed nay not be repraeeatatlva of the whole. The calculated values are Arthur 0- Little. Inc. June 27. 1983 t Or. Julie tni Manager - Research Technologies V.R. Coca 6 Co. 06151314 Category Fiber Mao (t) x 10~tZ Wean WedIan Treaolite Treaollce * probable treaolite Treaolite probable ereaolite - possible crraoUt* 2.84 8.44 13.17 2.84 2.40 2.81 Fiber naas concentrations arc calculated by Fiber* a Fiber aa (t) % 4000 ml (saaol*) r 960 m} (filter) 100.2 g (saaple) 1 al (aliquot) 81 a 10-4 m (pore area) a ....... -- a 10* (ppn) 0.24 a 10*^ a Fibers a Fiber ease (|) 20 pores This gives the following results for observes concentration* of treaolite fibers la Hoookoce JflC-5. basis; Avaraae Hass Median Has* Treaolite Treasure * probable treaolite Treaolite probable treaolite possible treaolite 1.4 ppa 28.4 76.8 1.4 ppa 8.1 14.4 /k Arthur D. Little. Inc. June 27, 1983 -3.0.1- 0c. Julia Yang Manager - Research Technologist .R. Craca 1 Co. C6151315 In ay opinion, cha nedian valuat should ha usad aa the bait Indication o sal* concentration. Rasas upon owe laboratory exp*tlane*, replicate analysis agrea within a {actor of two tlaas, indicating a eontimad craaolit* concentration of l to 3 ppa and a possible tranolit* concentration of 7 to 30 ppa. ?lease call >* if you have any questions. Vary truly youra. Eduard I. Paters Tri* :wj* ____ 06151316 /L Arthur D. Little, Inc. -uiv ;, :?S3 Dr. Julie Vang Manager - Research Technologies w.R. Grace 4 Co. 62 -hiccmore Avenue Cambridge. MA 021-0 Dear Julis: C In accordance with your latter of July l, 1983, we have analyzed a commercial sample of Konekote (lot Ho. 2FI) taken from the job site of the ABC building at U. 66ch street. Hew York City and hand delivered by you for the presence and quantity of amphlbole fibers, for this purpose, we have defined fiber as any particle with approximately parallel sides and with an aspect ratio of 3 or greater. PROC-DISE A "neat" sample was eaken for optical eicroscopy. This sample was suspended in acetone, and four subsamplas ware ceken for preparing alldcs - floaters, ssttled solids, suspended solids and a fiberous "lint". Thesa were lasersed in n_ - 1.640 oil and scannad at lOOX under central stop dispersion staining conditions which would reveal a characteristic dark blue appearance for tremollte. Any particles having an approximate color match were individually examined at 500X for shape and ladax of refraction. Both sampling tschnlque and area of observation wera represent ative and ehsractaristie of typibal practice. Secondly, we prepared a sample for electron microscopy by the wlpo^oue method, as you know, this is e very useful method for quickly preparing an unknown sample for component particle Identification by cranamlssloo electron alereecopy. A major drawback la chat it la not quantitative and further, particles cam bo altered by selective fracturing during the wipeout. The as-received sample wee dry sieved through e 35 nesh screen with e portion of the *33 fraction tskeo for analysis. The two frac tions were then recombined end reserved. A pinch of the sample (670 ug) was worked with nitrocellulose in amyl acetate with s spatula on e glees slide and when shiny, wee wiped with a clean Sriais Tier* III--n UsesJtMW* T>r,H Lmtm Sss FrswMe wnMapas Mains Ms Pm* Arthur U Little, Inc. July 11, 1983 G151317 Dr. Yang Manager - Research Technologies W.R. Orec# & Co. tilde. After revetting with a drop of solution, the second slide vas vlped with a third to provide an appropriate "by eye" distri bution. The ilia v*s floated off on pretiltered, distilled water, picked up on a screen holding electron microscope grids and the nitrocellulose was cleaned by the Jeffee vlek method using acetone. For analysis, three pore openings each on two grids vere examined at 10.000X and any particle meeting the fiber definition ves measured (length and width) end analyzed by selected area electron diffraction (SaD). Photographs of several typical fibers and their SAED's vere recorded. RESULTS For the optical microscopy analysis of the starting material, no fibers charaetsristic of tremoiite were observed. The procedure employed Is expected to have a detectability of l part per million (ppm). For tha electron microscopy analysis, a total of thirty fibers were observed. They are identified as follows: Cypsum (morphology and SAD) Probabla gypsum (morphology alone) Veralcullte (morphology and SAED) Cubic s^.nmral (morphology and SAD) Probabla cubic mineral (morphology alonm) 16 6 2 4 2 30 Therm vas no evidence for enphibole fibers or la feet, any fibers that could even possibly be identified as ampblbolem. The eaciaammd detection lisle for ehe sample concentration prepared end the area of sample aaelyted is about 10 ppm. Xn lUMary, examination of tho supplied Honehote sample by optical mad electron microscopy (using the vipe-out method) ebovmd no evidence for the presence of amphlbole fiber. Vary truly yours. Edward T. Petare m* <*4.$~o Twi i:hj* 0615131$ /tl Arthur D. Little. Inc July i;, Dr. Julie Vang Manager - Research Technologies W.R. Craca a Co. 62 Whictecore Avenue Cambridge, MA 021-0 Dear Julie C 76494 In accordance with your lecter of July 1. 1983, w have analysed a eoBercial sample of Monokote (Lot Mo. 2F1) caken from the job site of che ABC building at U. 66ch Screec, Mew York City and hand delivered by you for che presence and quantity of aaphibole fibers. For this purpose, we have defined fiber as any particle with approximately parallel sides and wlch an aspect ratio of 3 or greater. The sample was .coned and quartered wlch opposite quarters combined, weighed, acid digastad to dlssolva plaetar of parls and wet sereenad. Prior studies have shown quantitative recovery of tresollte eabestoa fiber from the Libby mine for the digestion procedure used. The digestion liquors end residue washings were combined, made up to <>000 ml and a l ml sample was taken, filtered through a Muclepore filter from which several electron nlcroseope grids wore prepared by che Jeffee wick meehod. For analysis. 20 random porn openings wars cravereed at 10,0001 and any paxtlelt meating che fiber definition was measured (length end width) end analysed by saleetad area electron diffraction (SAB). Photographs of several cyplesl fibers and Chair SAD's war* recorded. A total of 32 fibers was observed, eight of which were identified es confirmed craaollce or possible tremolice. Based on the dimensions of these fibers, the sample eaelyeed yielded a aasa concentration of 9 ppm. Baaed upon our laboratory experience, replicate analyses agree withla e fester of two times, indicating conservatively e likely range of 3 to 20 ppm tresollte fiber in the Honokote temple examined. In my opinion, this concentration Is totally insignificant. very truly yours. Edward T. Psters SniliiU Hiw L**a Ma*rM Me amSi Juan SMlrwOn Uthu, TSy hum WmMeSMi liemil /ti Arthur D. Little, Inc. nr **4.5*11 rf(,, :u* 06151319 Ju 13. 1983 Or. Julie Yang Manager - Research Technologies W.R. Craee Co. 62 Vhictaor* Avenue Cambridge, Massachusetts 02140 Dear Julia: C76494 In accordance wish your letter of July 1, 1983. ve hava enslyted a cermaarcial ssmple of Konokoce (Lot Mo. 211) takes fro* the job *ica of the AiC building at West 6bch Street, Mew Tork City and hand delivered by you (or the presence and quantity of aaphibole fibers- For this purpose, we have defined fiber as any particle vicn approxiaacely parallel aides and with an aspect ratio of 3 or greater. PROCEDURE The sample. contained la two 300 si polyethylene bottles, was poured on a glaselse sheet, cosed and quartered. Opposite quar ters were reserved and the two renaialng quarters were combined for analysis. This eample was digested In ID HC1 at 60C to dissolve the plaster of psrls (the aajor component). Sequential digestions utilised about 800 ml of add and 15 aisuces of stirring. The suspension was than allowed to settle for 10 minutes after which the supernote wee daeaated with a aqueata bulb and reserved. Alter four digestlone, x-ray diffraction of che residue felled to show s plaster of psrls (or gypsum) psak. it you recall, a previous eaelysle had shawm that e Ig sample of Hoataaa Tremollte (asbeetlfora variety) wet ground with e mortar and peatla vna quantitatively recovered after a elnllar digestion treatment. After digestion, che residue wee pieced la e 20 mesh screen end washed with Isopropyl slcohol. The filtrate wee added to the previous decants, neutralised with KH.OH and asds up to 6000 al. Aftsr agitating, and la accordance* with the standard Arthur 0. Little, lac., protocol for suspensions, samples vers withdrawn !** Arthur U- uuic. me. July 13, 1983 7- Dr. Julie Yang "anagar - Research 7*chr.olagis .X. 0rc 4 Co. 06151320 from eh* center of the sustention to provide separate samples of 0.5 and l si. Tneae were each diluted to 100 al and filtereo through 0.1 - pore size. -7 as dia. Nuclepore filters vr.ich were then carbon coated. Samples (4 or 5) were incited :r:a taco and placed on 200 vein copper electron Microscope grids- The filter sedia was cleared with chloroform using the modified Jaffee wick procedure. The 1 ml sample was selected for electron microscopy. For analysis. 20 randoa pore openings vers traversed at 10.0Q0X and any particle resting the fiber definition was Measured (length and width) and analysed by selected area electron dif fraction (SAD). Photographs of several typical fibers and their SA's were recorded. In addition, as you requested, a portion of the digestion residue caught in the 20 mesh screen vat taken for examination by optical microscopy to ascertain that any respirable fibers had been removed in the vashout process. This sample was suspended m acetone, and threa subsamples were taken for preparing slides settled eollds, suspended solids and a flberous "lint". These were inserted In n_ 1,640 oil on s cover slip and scanned it 100X under centre! stop dispersion staining conditions which would toveal a characteristic dark blue appearance for trcmollte. Any pertides having an approximate eolor match were individually examined at 5OCX fox shape and index of refraction, both sam pling technique and area of observation wars representative end characteristic of typical practice. The optical and electron Microscopy samples described above art referenced as the Arthur 0. Little, lac. Method. Also, a portion of the original sample was taken "neat" for optical mieroseopy. This sample ves simil arly prepared sad analysed, except chat there was ee additional subsample of floaeera. Ad you roduosted, we prepared an additional sample by the wipe out method. As you kaow, this Is s very uaeful method for qulehly preparing aa unknown sample for component particle identification by transmission elseeron microscopy. A major drawback is that it la not quantitative and further, particles eaa be altered by selective fracturing during the wipeout. For our preparation, the reserved sample ves dry sieved through s 35 mesh acreea with a portion of the -35 fraction taken for analy sis. The two fractions vers then recombined and reserved. July 13, 1983 -3- Dr. Julie Yang Manager - Research Technologies V.R. Crsce 6 Co. 06151321 A pinch of the sample (6*0 ug) was worked with nitrocellulose in amyl ecetsce with a spatula on a glAst slide end when shiny, -* wiped with s clean elide. Attar revetting with a drop of solu tion, the second slide was wiped with a third to provide en Appropriate "by eve" distribution tnd loedlng. After drying, the fils was floAted off on prafiltArad, distilled wster. picked up with a ecreen holding electron microscope grid* And the nitro cellulose was cleared by the Jaffee wick aethod using acetone. For analysia, three pore openings each on two grlda were examined at 10.000X as described previously. This ssapls is referred to as the Wlpa-Out Mechoa. RESULTS Arthur P. Little, Inc, Msehod for electron microscopy, a total of 32 fibers were observed and identified as fellows: Category Number of Fibers Varaleullte Usage and SAID) Probable veralculite Usage only) Treaollte Usage and SAID) Probable treaollte (Image only) Possible treaollte (ambiguous Image) Other minerals (Image and SAID) 18 2 2 4 2 4 32 The observed dimensions of the cremeillte, probable tremollte possible tramollte fibers sre as follows: met t(om) Hass(a) Trsmolite 1 Tremollce 2 Probable tremollte i Probable tremollte 2 Probable tremollte 3 Probable tremollte 6 Possible tremollte 1 Possible tremollce 2 44.2 23.2 2.9 1.3 4.1 6.0 1.2 1.1 1.66 0.33 0.28 0.08 0.47 0.94 0.09 0.13 381 8U10'U 20.69 0.72 0.38 3.02 16.60 0.03 o.os *23.30  U At iiiui ^ July 13. 1983 Or. Julie Yang Manager - Research Technologies u.a. Grace k Co. 06151322 where It is assumed thee.thickness equals width end the density of treaolite is 3.13 g/ea . Typical fiber weights for the three fiber categories of Interest were calculated on the basis of both seen and median fiber sets. The latter is a sore useful indication of central tendency in cates where one is looking at an extremely small portion of the eotal sample and one large fiber found in the sample area analyzed say not be representative of the whole. For example. the treaolite l fiber reprecente 90 percent of the total fiber mass encountered in the 20 pores. The finding of this pertleuler fiber in the eroa examined increesee computed fiber concentration by an order of magnitude. Typical values for fiber mesa are: Category Treaolite Treaolite * probable treattilts Treaolite probable treaolite - poseible treaolite Fiber Mass _(l) x 10*U Msen Median 201.25 201.25 70.54 9.61 52.92 1.87 The very large median values for trsmollts and tromellte probable treaolite categories are associated with e few large fibers la tho vary snail population observed for chase cate gories, 2 and 6, respectively. fiber mass concentrations srs calculated by Fibars a Fiber maee(a) 1000 ml (sample) 31.1 (seaple) * 1 ml (aliquot) * *raaI * ---- * i6 pp 50 a 10~* am* (pore arse) "20 pores - 0.63 x 1012 x Fibers x Fiber sees (g) July 13. 1913, -5 Or. lulxt Yar.g Manager - Research Technologies '.s, Grace Co. 0Gf5T3Z3 This gives Che following results for observed concentration* of treaoiice fibers in Monokote MK-5. Seals: oversea Mass Median Mats Trvaclice Treaolice * probable creaolice Tremolite * probable treaolite 253 ppa 267 267 233 ppa 37.1 9.4 la ay opinion, the eedian values should be used as the best indication of aass concentration. Baaed upon our laboratory experience, replicate analysis agree within a factor of two tlaes. indicating a possible creaelice concentration of 5 to 20 ppa. for the optical aicroscopy analysis of the washed residue and tha as-received ''neat" saaples, no fibers eharaeccriseic of creaelice were observed. The procedure eaployed is expected to have a detectability of 1 ppa. Mine-Out Method The electron aicroscopy analysis revealed a cotel of chircy fibers. They are identified as follows: Gypsua (aorpholegy sad SAID) Probable gypsua (aorpholegy alone) Veraiculite (aorphology and SAD) Cuble ainaral (aorpholegy and SAD) Probable cubic alneral (aorphology alone) 16 6 2 <* 2 There was no evidence for aaphlbole fibers or la fact, aay fibers chat coaid even possibly be ldeatlfled as aapblboles. The eeciaaced detection limit for the ssapie concentration prepared sad che area of saaple analyzed la about 10 ppa. July 13. 1983 -6- Or. Julie fang Manager - Research Technologies W.R. Grace 6 Co. 06l5i32rJ DISCUSSION There are several caveats to the methodologies utilised lor indicating the vteser.ee and amount o trace level components in a bulk, heterogeneous sample, including the following: e Representativeness of the analytical sample compared to the whole o Sample alteration during spmelman preparation o Small amount of total sampla actually examined o Computation of concantratlon on basis of very low number of fibers observed and categorised. A few comments will be made regarding each of these. Rentesentativentss - Standard methods exist for reducing a large, bulb sample into a smaller represents-Iva, analytical sampla. Principal methods Include coning anc quartering and riffling. One should start with the whole sample (bag or container). A scooped subaampla can be very mlarepresentative of the whole. Sample Alteration - The purpose of sample analysis la to provide e characterisation of ehm product am it la used la cos--tcial Bracelet. Therefore, precaution moot be taken to avoid alteration of the alas, shape or mature of the product components. If a single fiber bundle is broken down into many fibers during a rub-out, the "apparent" concentration la greatly altered. Further, during acid dlgtatlon and stirring, it la possible ehat contaminant asbeetoe fibers intercalated within the vermicullte plate* may be exfoliated, thereby resultlag in free etbestoe fibers la the analytical sample that vara not present in the bulk sample. This could lead to an overastlaate of asbestos fiber content of the sample, as an Intercalated asbestos particle would not be seen as a fiber but rather am a vervlcullte plats. Thm ASTM Asbestos Taak Force Comlttee demonstrated that the Mucldpere filtration - modified Jaffoe wick preparations procedure does noe result in fiber loss. Therefore, the results of the Arthur 0, Little, Inc. method, as /ll Artnur u- u>uc, iuv. PAGE 18 July 13, 1983 Dr. Julie Tong Manager - iesearth Techno logic* w.l. Crece 4 Co- 06151325 described above, are conservative, being an overestiaace on the sale side. Snail Sample Site - As described earlier, the analytical sample is a very suii portion of the original sanple (the bag or arum of solids), furthermore, an extremely small portion of the analytical sanple is actually examined. For example, by the Arthur D. Little, Inc. Method, only about two parte in a billion of the initial sanple are analysed, by the wipe-Out Method, about a hundred tines smaller sespls is examined. The only method chat can possibly be representative of the subsenple is an aliquot taken from a suspension and the analysis of a reasonable field of view. This is the basis of the Arthur D. little, Inc. Method, which is quantitative. Computation of fiber Concentration - Two factors must be considered in computing fiber concentration - the means by which a typical fiber dimension is specified and the representativeness of the smell number of flbore chat ere actu ally analysed sad ideatlfitd. la s previous study, we have shown chat mess balances can be obtained when median fiber dimensions are used but cam be overestimated by 10 or mors class when aeon dimensions art used (Peters, C.T. sad T.E. Doerflsr. "Aaphibole Hiattal Tiber Analysis by electron Microscopy: Comparison of Sample Preparation Procedures", Chapter 17 m Electron Microecopy end X-rsy Applications to Emvlronaeatsl and Occupational Health Analysis, P.A. Xussall and A.E. Hutchings, Eds, Ana Arbor Science, 1970]. This con he demonstrated by combining ell the possible crenelles fiber date from this analysis end chat re ported an Jane 27, 1983, conprislng 31 fibers (see Figure l). A name balance should be achieved whether one calculates masses for mack fiber sad takes the average or if one calculates average ill menslees sad then computes mess. A comparison between mean (overage) end median calculations gives Mein Median Average fiber ness Maes from average fiber dimensions 13.94 7.07 2.42 2.44 /tl Arthur u i-iiiic. mu July 13, 1983 -8- Or. Julie Yang Manager - Research Technologies R. Grace 4 Co. CS152.326 it is clear char aedian value Is preferred. Relating to the representativeness of a particular analysis, statistical approaches are available for estimating the range of results that could be expected for a series of replacate analy ses. The Poisson distribution function is generally used for comparing the results of repllcaee analysis or for computing a confidence Interval. The usual confidence Interval is the range in which 93 of 100 replicate analyses would fall assusing the fibers analysed satisfy the Poisson distribution. This esn bo exaslnod for tho tvo recently completed analyses. June 27 July 12 Catssory Confidence fibers Interval Confidenea Fibers Intsrvsl Treaolire 2 0.24 - 7.22 2 0.24 - 7.22 Treaellts * Probable tresolits 14 7.65 - 23.49 6 2.2 -13.06 Treaollte * Probable tresellta * possible tresollce 23 14.58 34.31 9 3.43 -13.76 For the conversion forsula end the sedien fiber nesses given earlier, the results rounded off to the nearest pps art Tiber Concentration (p^> June"27 J^Tt 12 Tresollce 0-5 30 - 913 Treselica * probable tresellta 4-1* 14 - 81 Tresellta probable tresellta * possible tresellta 9-22 4-19 July 13. 1983 -9- Or. Julie Vang Manager - Research Technologies V..R. Jracc 1 Co. 06151327 '.z is observed char far reference to the treaolice data above, the ranges for cne eve samples ho not overlap, l.e., the Poisson condition at unifora fiber distribution i* not aet. Very good agreement is seen for cne category of all possible trendies fibers based upon the larger fiber population which does satisfy the Poisson distriouticn. It should also be mentioned that In reference to Figure l. 3 trerclite fibers. 3 probable treewlltc and 3 possible tremolite fibers fall In a range that could be teen by optical microscopy. This points out i) a non-representative sample with reepect to the two very large crenolice fibers end 2) a probable over estimate of probable end possible treaolite fibers, ee none were seen by optical microscopy. CJNC'-'-S ION 3ated upon eur examination of supplied Stonakote samples by optical and electron microscopy according to established proto cols. and in light or the discussion above, confined and possi ble trenollce fibers (by electron microscopy) in Honokote corre spond to e mess concentration range of S ra 20 ppm. In my opinion, this concencretlon la totally insignificant. Vary truly yours. ^. ____t _ __f '_,__i _- __ Edward 7. Peters V T IS i | 3 1 ^4 ! = 1T T*0*CVtTt "t ! *:s}tji ^9w.rr: | -n:*' n* u. rritw j 06151328 rtirtt 1. Nmltnw ( (lkr tUMilm M iliHtllciiut litNtrr Iim nalf: Jm* 23. 1M) aw*is flUrt Mr 1. 1W Mil I