Document 1Qb1B8078G21ZaJ4vLmEeLRwd

Relation of Particle Dimension to Carcinogenicity in Amphibole Asbestoses and Other Fibrous Minerals 12 * * Meari F. Stanton,3,4 Maxwell Layard,5,6 Andrew Tegeris, 1 Eliza Miller,3,8 Margaret May, 3,4 Elizabeth Morgan, 7**9 and Alroy Smith65 ABSTRACT--In 72 experiments, durable minerals In the form of particles on respirable size and of wide chemical and structural varieties, were implanted in the pleurae of outbred female OsborneMendel rats for periods of more than 1 year. The incidence of induced malignant mesenchymal neoplasms correlated well with the dimensional distribution of the particles. The probability of pleural sarcoma correlated best with the number of fibers that measured 0.25 nm or less in diameter and more than 8 pm in length, but relatively high correlations were also noted with fibers in other size categories having diameters up to 1.5 pm and lengths greater than 4 pm. Morphologic observations Indicated that short fibers and large-diameter fibers were inactivated by phagocytosis and that negligible phagocytosis of long, thin fibers occurred. The wide variety of compounds used in these experi ments suggested that the carcinogenicity of fibers depended on dimension and durability rather than on physicochemical prop erties.--JNCI 1981; 67:965-975. Work in several laboratories has indicated that di verse varieties of minerals are carcinogenic when ap plied directly to the pleura of the rat or hamster in the form of microscopic fibers, i.e., particles with dimen sional aspect ratios of 3:1 or greater (1-9). The same minerals are much less carcinogenic when applied at equal weight and size in nonfibrous form. Further, preliminary experiments indicate that carcinogenicity correlates best with increasing numbers of fibers having both diameters of 0.25 pm or less and lengths of more than 8 pm and that the correlation diminishes with fibers of greater diameter or lesser length. Conse quently, a reasonable conclusion is that the long, thin, fibrous structure is critical to the carcinogenicity of these minerals. Studies on fibrous samples within very narrow dimensional ranges would be valuable in the establishment of this hypothesis, but these ideal sam ples are not available. Consequently, we are faced with the correlation of carcinogenicity with fiber samples of widely mixed dimension. The purpose of this report is to correlate our best estimate of fibrous dimension with carcinogenicity for all those minerals that we have studied that are both durable and within the size range of respirable particles. This involves 72 experiments with minerals of wide chemical and structural variety. Of special interest are the data on the amphibole asbestoses: amosite. tremolite, and crocidolite, though estimates of the dimensions of the asbestoses are especially liable to error. Chrysotile. although as car cinogenic as the amphiboles at comparable dimen sions, could not be included since it has proved difficult to be measured with any degree of precision. MATERIALS AND METHODS None of the methods were appreciably different from those described in earlier papers (4, 6, 9-11). Con sequently. only modifications of methods are detailed here. A standard 40-mg dose of particles uniformly dispersed in hardened gelatin was applied by open thoracotomy directly to the left pleural surface of 12- tc 20-week-old. outbred female Osborne-Mendel rats. In each experiment, 30-50 rats were treated and followed for 2 years, at which time the survivors were killed. AH rats were necropsied and all lesions examined histo logically. A positive response was the occurrence ol pleural sarcomas that resembled the mesenchymal meso theliomas of man, developing after the 1st year (12) Three types of controls were considered: untreated rats rats that received thoracotomies but no pleural implant and rats with pleural implants of nonfibrous material There were two types of spontaneous tumors thai could cause confusion: the fibrosarcomas of left mam mary glands and the subcutaneous fibrosarcomas in duced by suture material. Vigilance and early surgica. removal accounted for most mammary tumors; the use Abbreviations used: alumin = aluminum oxide; attapul "attapul gite(s); crocid=crocidoiite(s); dawson =dawsonite(s): halloy = hallossite(sl; UICC= International Union Against Cancer: woilaston-' wollastonitets). 1 Received November 13. 1980: revised May 6. 198L: accepted Jun 8. 1981. ` The guidelines for the care and use of laboratory animals wer followed as set forth by the Committee on Revision of the Guide fc Laboratory Animal Facilities; by the Guide for the Care and Use c Laboratory Animal Resources, the National Research Council: an by the National Institutes of Health. ! Laboratory of Pathology. Division of Cancer Biology and Dias nosis. National Cancer Institute (NCI), National Institutes of Healtl Public Health Service. U.S. Department of Health and Huma Services, Bethesda, Md. 20205. * Deceased. s Biometry Branch. Division of Cancer Cause and Preventioi NCI. 6 Address reprint requests to Dr. Layard at his present addres Veterans Administration Medical Center. 3801 Miranda Ave., Pal Alto. Calif. 94304. ' Pharmacopathics Research Laborato. nc.. 9705 North Was! ington Blvd.. Laurel, Md. 20810. ` Present address: 5524 Trent St.. Chevy Chase, Md. 20015. ' Present address: Triangle Resource Industries. P.O. Box 59" Laurel, Md. 20707. 965 JNCI. VOL. 67. NO. 5. NOVEMBER 198 966 Stanton, Layard, Tegerls, et al. ipf synthetic, biodegradable, polyglycolic acid sutures largely eliminated suture sarcomas. An equivocal diag nosis for the origin of a tumor was necessary in less than 1% of the tumors. The probability of pleural sarcoma in each experiment was calculated by an actuarial life table method that accounts for early deaths without pleural sarcoma and provides a good means of making quantitative comparisons of one experiment with another. Details of this method are given in (13, 14). The fibrous materials used in these experiments were mostly commercial products that were submitted by the manufacturers from an interest in their potential car cinogenicity. Consequently, they were used as received and were not especially refined except in our efforts to separate particles by size. None of the preparations appeared overtly contaminated by other materials when examined in the electron microscope. A few of the small-fibered subfractions of the fibrous materials were obtained by ball milling in a steel ball mill andconsequently were contaminated with fragments of steel. In general, subfractions were obtained by simple gravimetric methods in aqueous media to separate fibers of different dimensions. These maneuvers in cluded sedimentation, centrifugation, and filtration, which in some instances were also responsible for the reduction of the size of the particles but did not otherwise alter the particles physically or chemically. Eleven chemically and structurally different groups of fibers were available for study, and samples studied are listed in text-figure 1 and table 1. Six major groups of particles had muidple dimensional ranges; these in clude: crocidolites: (samples crocid 1-13), glasses (glass 1-22), aluminum oxide whiskers (alumin 1-8), talcs (talc 1-7), dawsonites (dawson 1-7), and wollastonites t wollaston 1-4). Seven addidonal types of parucles had only one or two dimensional ranges. These were the amphibole asbestoses tremolite (tremolite 1, 2) and amosite, the clays attapulgite (attapul 1, 2) and halloysite (halloy l, 2). crystals of silicon carbide and potassium dtanate (titanate 1, 2), and nickel titanate (titanate 3)'. All of these materials have been described elsewhere (4, 6, 10, ll, 15-18), but the following information is pertinent. Crocidolite (crocid 1-13).--These 13 samples of South African crocidolite (an amphibole asbestos) were from four different sources. Samples crocid l, 3, and 9 were prepared in our laboratory from a single sample of hand-cobbed, unmilled ore. The ore sample was hand milled without exposure to any metallic ma terials and reduced to the approximate size of com mercial crocidolite. Samples crocid 6, 7. 8. 11, 12. and 13 were all prepared in our laboratory by various milling, sedimentation, and flotation methods from a single lot of standard UICC crocidolite designated crocid 5. Differences in dimension were the result of different milling times. Crocid 5, the original UICC sample, has been characterized in (19. 20-23). Samples crocid 4 and 10 were specimens prepared in a com mercial laboratory from a single separate sample of South African crocidolite and separated by centrifuga tion to obtain mutually exclusive size ranges from the same sample (24). The remaining sample, crocid 2, was obtained from Dr. J. C. Wagner (Medical Research Council Pneumoconiosis Unit, Penarth, Wales) as representative of the material used by him in his original experiments (25). It was our impression that any mechanical manipuladon of these samples could both reduce the size of the particles by fragmentation and effectively increase the size of the particles by clumping. For this reason, probably the dimensional measurements on crocidolite are the least representative of all the fibers measured. Glass (glass 1-22).--The first 18 of the 22 glasses were borosilicate glasses that have been previously reported and can be recognized from those publications by their letter designations (4, 10). Glasses 12. 14. 15, and 18 were preparations of typical large-diametered insulation glass fibers that were coated with a phenolformaldehyde binder. In the early experiments, glass 18 ' was used as a control and also served as a vehicle for the implants. Glasses 19 and 20 were preparations of large-diametered fibrous glass that was leached to remove all elements except SiO:. These two glasses were exceptionally fragile and contained many irreg ular fragments. Glasses 21 and 22 were large-diametered .extruded fibers with a - microcrystalline aluminum ox ide content greater than 80% (glass 21) and with a microcrystalline zirconium oxide content greater than 90% (glass 22). Aluminum oxide (alumin 1-8).--The 8 samples of aluminum oxide were all crystalline sapphire whiskers prepared by General Technologies Corporation. Reston, Va., or by Thermokinetics Fiber Incorporated, Nutley, N.J. (15-18, 26). All of the samples were processed and selected for dimensional ranges. Of the samples. 3 were excepdonally noteworthy. Sample alumin 8 was nonfibrous, sample alumin 3 was excepdonally fine but tended to cluster in nonfibrous balls, and sample alumin 4 contained whiskers of aluminum nitride as well as aluminum oxide. Talcs (talc 1-7).--All seven talcs were refined raw materials for commercial products. Each was from a separate and diverse source and selected to include all extreme ranges of dimension. Platelike structure was consistent and was considered in the calculation of the volume (15-18). Dawsonite (dawson 1-7).--The 7 dawsonite sumpies (crystalline dehydroxy sodium aluminum carbonate [NaAKOHhCOs] were from several sources. The characteristics and synthesis of dawsonite can be found in (27, 28). Samples dawson 2 and 3 were svnthedc crystals prepared by a commercial company (for dawson 2) and by the Bureau of Mines, t Department of Interior (for dawson 3). Sample dawson 4 was a natural crystalline dawsonite from the Olduvai Gorge. Tanzania. The remaining 4 samples (dawson 1. 5. 6. and 7) were synthetic crystals from a second commercial company. These 4 samples were especially crystallized and sorted to achieve narrow ranges of size. . 1 ' . . ' . j i 1 .] j 1 ' j j , j JN'CI. \Ot.. o7. NO. S. NOVEMBER 1981 uea- trch as bis that Hild tion bv >nal i rive ISSFS uslv tons . 15, vied nol-s 18 for s of I to t.vses regered oxha ban . ! I kers ton, tley, and were lonbut npie c as raw ma all was the i 1 am- res. t be vere mnv U.S. .son uvai n 1. i I Carcinogenicity of Fibrous Minerals 01 Titantfr 1 95,1 > 4.CF8.0 >25-4.0 > t.5-2.5 >50-t.5 > JS-.50 3.57 4.41 >.I0-2S 187 6.08 > .06-.10 377 5.02 > ,0J -.05 >8.0 3.7S 4.79 634 5.02 4.11 AST 6.17 AM (2) Tirarurtt 2 1W% 170 188 475 3.40 4.18 4.93 431 3.70 4.52 5.07 4.78 140 4.18 4.48 431 (3) Si Carbide 100% 275 232 3.08 373 3.83 0.81 4.01 471 I 3.75 4.48 132 132 2.04 3.03 470 4.19 4.41 4.54 135 1.65 1.2 1.95 324 12 4.40 4.66 4.74 4.70 (4) Dvwscn 5 100* >4.08.0 >25-4.0 >15-25 >50-15 >25-50 124 218 473 > .10-75 > .05-.1Q 283 > .01-.05 >8.0 137 137 47S 173 102 ISO 4.53 4.93 6) Trwnofh* 1 IW% 184 4.10 420 3.63 114 4 47 475 4.14 3.54 114 3.92 3.14 144 3.14 114 332 114 3.44 14 (61 Tromofta 2 44 4.09 114 3.14 4.05 386 4,04 2.84 254 234 3.68 338 373 ... t 3 01 254 3.77 3.31 284 (71 Dawson 1 95% >4.08.0 >254.0 >1.525 1.44 >.50-1.5 1.92 132 >75.60 1.75 116 > .10-7S 262 3.70 >.05.10 3.40 3.62 >.01-.05 3.57 3.57 >8.0 1.44 272 132 3.90 3.62 214 3.00 1.75 3.63 425 4.66 262 (8) Crood. 1 94% 4.07 4.07 5.75 5.63 4 67 6.52 5.62 4 67 5.87 4.67 (9) Croext2 3.53 4.53 477 4.30 4.56 422 4.83 4.70 4J22 515 5.55 5.57 4.73 TOT- 477 5.71 4.96 4.00 4.57 3.83 00) Crock! 3 >4.08.0 >254.0 >1.52.5 >.50-1.5 >25.50 >.10-75 6.12 - 5.71 >.05.10 6.52 5.84 >.10-.05 5.62 4.58 >ao 4.10 4.10 ASS 4.10 4.10 4.10 4.40 - S.01 Si) A^cah* 33% 02) Oocid. 4 Bt>% 213 243 213 213 117 3.36 3.54 261 3.13 3.64 3.56 353 76 31 172 3.47 278 298 3.03 261 276 324 3.84 3 ( 424 3.61 4.36 3-24 4.33 4 60 4.92 2.76 4.31 4.53 4.70 32* 136 3.54 3.06 (131 GUs* 1 (MOU >4.08.0 >25-4.0 >1.52.5 >30-13 273 >75.50 308 >.10-75 293 >.05*.10 >.01-.05 >8.0 253 3.35 3.93 3.46 3.23 3.08 4.95 4.53 4.79 4 (141 Crock!. 5 /o% 153 4.3S 484 4.74 4.42 2.81 4.19 4.3S 4.65 4.63 426 2811 2811 3291 3.42 366 3.771 3 421 2811 281 3.77 2.81 3.53 323 (15) Glass 2 (M6DI 77% 2.91 321 3.38 2.91 3.69 3 61 4.08 3.69 3.51 151 321 399 3, 4.11 4.02 3.81 138 (16) Glass 3 (K!) >4.08.0 >25-40 > T3-25 1.45 >30-15 2.95 >75.50 255 3.16 >.10-7S 3.03 116 >.05.10 2 as 409 >.01-.05 303 3.73 pm >.01-1 >1-4 >8.0 0-67 0.67 2.40 3.33 3.76 303 >48 0.67 1.52 0.37 203 219 3.42 274 363 3.03 37S 3.03 >884 >64 (17) Glass 4 (MED 2.81 244 244 244 >71-1 2.05 244 3.52 144 344 >1-4 174 269 316 3.44 3.56 33S >4-8 1.53 123 276 213 3.70 290 3.39 162 3.70 3.14 >864 >64 08) Alum in. I 70% 120 3.12 226 >.01-1 >1-4 1.10 1.82 1.61 1.93 0.80 257 1.73 >4-3 120 1.80 0 1.90 l 217 l 2.78 *1 242 3.51 273 2.63 >8-64 > Length Text-figure 1.--Fiber distribution by lommon log of the number of particles per microgram in each of 34 dimensional t.uegor sue. JNCI. VOL. W, NO. 5. NOVEMBER 968 Stanton, Layard, Tegeris, et al. Dlsmstaf (191 Glen S OCW) 63% > 4.00.0 > 254.0 15-2.5 0.66 >-60-1.5 > 55-.50 2.48 1.06 150 0.67 >.10-55 >.05-.10 > 91 -.05 >8.0 057 1.70 1.67 1.67 0.06 0.06 0.36 0.76 0.06 1.06 1.41 156 2.42 251 2.75 2.71 251 251 GO) Dawson. 7 68% 1.77 3.11 272 207 3.01 4.18 4.40 3.70 272 356 4.10 4.64 3.42 1.77 3.33 3.81 455 4.70 272 (21) Dawson. 4 o% 211 3.38 650 5.03 157 252 211 3.42 4.05 5.00 3.47 1.66 157 7 77 354 4.42 4.12 3.33 1.87 206 1.81 254 248 2.80 4.01 (22) Dawson. 3 dO% >4.08.0 >254.0 >15-25 >50-15 269 >55-50 4.18 4.50 >.10-55 3.91 4.74 > .05-.10 5.8S 6.18 >.01-.0S 5.90 5.17 >80 3.07 458 4.41 4.88 458 159 259 353 457 551 559 6.14 23) Glass 6 (MOW) 04 ^ 258 1.40 1.70 210 271 1.10 1.10 317 284 352 3.55 295 1.09 218 351 3.11 35S 3.47 271 1.10 1.58 155 1.10 274 201 269 284 254 3.69 276 280 (24) Crocid. 6 63% 3.47 4.17 5.10 4.74 265 456 5.07 5.45 4.90 4.07 1-------- 452 4.02 4.71 4.45 4.86 4.57 451 2. 47 1 25) Good. 7 &6% > 4.08.0 >25-4.0 > 15-2.5 >.50-15 >55.50 1.69 >.10-55 205 >.05-.10 1.86 >.01-55 >80 051 1.0 151 191 9.91 091 151 151 196 255 2Q2 259 234 251 2.05 091 US) AJumtn. 3 41% > 4.08.0 >25-4.0 199 >15-2.5 1.59 >.50-15 236 242 >55-50 0.97 >.1055 097 232 > .05-.10 157 > .01-.05 >80 1.19 157 1.70 233 1.15 219 0.97 1-72 1.75 0.19 1.82 1.70 261 0.67 1.49 241 159 PI) AlumIn. 4 2B% >4.08.0 >254.0 2.15 >'1.5-2.5 2.80 >.50-1.5 292 2.83 >55.50 3.57 3.01 >.1555 259 >.05.10 > .01-95 199 >8.0 198 248 230 245 3.19 199 2J29 269 1.75 198 215 237 158 275 0.96 283 2-23 259 (25) Good. 8 W% 4.45 5.06 5.61 4.55 553 3.75 4.75 4.86 5.09 4.59 4.15 (29> CrocSd. 9 43% 3.45 3.4S 3.75 3.75 3.13 3.13 213 3.13 3.43 3.13 4.51 455 3.43 554 4.32 45S 4.09 5.43 5.14 4.34 3.74 (32) Good. 10 37% 3.10 3.10 3.57 3.79 3.57 3.40 3.10 4.69 4.57 3.57 6.03 5.01 3.79 3.10 5.13 4.59 3.10 4.96 4.46 (27) AJumin. 2 +4% 1.07 1.43 237 204 199 1.11 150 091 1.15 1.18 1.60 1.70 1.50 150 150 GO) Wollaston. 1 31% 3.15 3.15 358 3.41 211 3.11 3.76 355 352 358 2.81 3.06 115 211 2.41 (33) Alun-iin. 5 22% 1.48 266 3.59 277 3.55 454 4.13 4.02 300 3.70 1.78 218 271 3.41 452 454 3 92 252 152 0.37 1.32 037 1.53 1.02 1.61 1.18 237 1.83 302 283 1.98 150 211 3.19 2.11 259 206 206 2.59 356 *, 43 3.92 357 353 ) :\ V.f i (34) Glass 20 (RT) 22% > 4.58.0 > 2.54.0 >1.525 >-50-15 >55.50 >.1555 >.05.10 > .01-.05 fim >91-1 >14 >80 >48 >884 0.88 0.98 >64 C3S1 Glass 7 (KCP) 21% 3.00 354 354 250 250 >91-1 1.44 2.05 359 3.88 358 290 298 >14 181 205 217 3.17 210 3.10 267 >4-8 097 201 1.81 1.81 2.31 0.97 285 2.01 2S5 250 >564 >64 (36) Wollaston. 3 1% 258 3.58 3.12 4.09 3.32 332 3.18 3.32 3.42 34a 258 26a 3.12 3.46 258 253 258 276 >.01-1 >14 >48 >564 > 54- LngU> Texr-HCt'RE ! fcontinued).--Fiber distribution bv common log of the number of particles per microgram in each of 34 dimensional categories- JNOl. VOL. 07. NO. 5. NOVEMBER 1981 Carcinogenicity of Fibrous Minerals COT Halley. 1 2* >438.0 > 25-4.0 > 1.525 >504.5 >.25-50 628 >.10-25 655 > 35.10 6.47 658 >51-.05 6.07 621 > 8.0 455 8.16 C8) Hailey. 2 636 6.16 633 4.51 451 6.47 6.14 6.89 451 4.51 4.81 39) Glass 8 OCUP) 2.68 136 3.36 3.17 | 2.11 273 333 3.54 3.50 332 231 126 252 271 3.54 284 261 263 2.04 1.74 231 126 2.44 3.19 1.96 257 2.31 280 231 (40) Crcod. 11 19% >4.08.0 >25-4.0 238 > 15-25 203 >50-15 3.15 298 >25-50 3.05 150 > .10-25 297 >.0S-.10 202 > 51-.05 >8.0 1.72 242 1.72 150 215 250 1.42 (41) Gta 19 tHD) 15% 256 267 156 216 292 237 226 1.73 126 243 126 273 (42) Glass 9 (M8U 1.49 260 1.8S 0.77 1.12 1.45 0.17 1.73 202 211 227 23S 242 233 155 {43} AJumin. 6 13% >4.083 >25-4.0 0.48 > 1.525 0.82 >50-15 0.75 122 >2550 2)0 0.82 > .10-2S >55.10 > .01-.05 > 8.0 0.12 0.67 0.12 022 037 022 122 1.44 0.67 0.70 1.64 1.49 097 0.63 194 0.82 (44) Dawson. 6 13% 6.17 6.47 696 6.86 7,86 6.78 (45) Dawson. 2 12% 280 4.57 492 422 439 294 3.43 3.46 3.31 3.10 244 282 274 244 1.74 269 244 1.74 (46) Wollaston. 2 >4.08.0 >254.0 278 > 1.525 278 >50-15 - 3.32- 330 > 2550 3.43 332 >.1025 3.96 326 > .05.10 437 4.00 > 31 -.05 >8.0 278 3.18 3.18 . 296 248 240 (471 Crodic. 12 U% 287 331 4.00 4.00 3.67 227 331 3.73 421 4.00 3.50 227 3.55 3.39 3.53 3.45 3.17 257 227 339 3.05 3.69 3.67 3.17 229 (49) Glass 10 (M0S) >4.o-ao >254.0 297 > 15-2.5 3.43 >50-15 3.88 3.91 > 25.50 4.34 4.02 > .10-2S 4.43 3.88 > .05.10 590 4.19 > 31 -.05 6.77 4.63 >8.0 246 299 237 276 3.69 272 1.76 1.17 2.48 137 (50) Glass 11 DC2P) 5% 4.12 428 4.11 39) 2kT 3.84 4.53 3.02 2.89 3,09 3.56 3.60 1.18 241 266 1.81 152) AnopuL 1 8% >4.08.0 >254.0 >1.525 > 50-15 5.12 >25.5O 5.48 > .1025 5.75 6.50 > .05.10 662 6.6S > .01-.05 6.96 6.65 jjm >.01-1 >14 > 8.0 4.64 494 >4-8 >854 >64 (53) Talc 1 7% 4.00 4.15 4.19 3.70 3.64 4.02 4.56 397 3 91 327 3.09 3.70 327 279 279 3.09 >31-1 > 1-4 >4-8 >8-64 LwiCth >64 (Ml Glas112 (02P) 7% 283 244 2,08 241 2.06 2.49 2.83 0.48 1.65 232 2.48 195 191 1.68 1.78 217 2.09 1.34 1.64 >31-1 > 1-4 >4-8 >564 0. 0. 0 0 0 > TtXT.nc.URE I (continued).--Fiber distribution by common log of the number of particles per microgram in each of 34 dimensional categot JNCI. VOL. 67. NO. 3. NOVEMBER 370 Stanton, Layard, Tegeri*, et al. (551 Glow 13 OCFP) O'* > 4.08.0 >254.0 283 > 1.5-25 320 > .5015 3.19 3.48 > 25-50 236 122 >.1025 230 > .06-.10 327 >.01-.05 >8.0 267 273 152 152 203 203 052 052 (56) Gioa* 14 (P2P) O'* 256 121 236 155 230 228 251 129 0.81 155 2.06 0.91 121 2.00 221 1.70 0.69 0.81 (57) Gian 15 (Y2P) 1.06 0% 1.08 154 258 212 1.74 3.42 2.09 1.60 212 1.60 T 156 224 216 1.64 1.60 060 1.30 (58J Altunin. 7 On >4.08.0 >254.0 2.52 > t5*25 0.02 >50-15 0.64 T.15 > 2S-50 1.17 T24 >.10-25 > .05-.10 > ,01-.05 >ao T.19 252 T.19 T.70 T.oe T27 T.64 0.31 T28 1.43 2.15 031 T43 051 (59) Gian 16 (MSS) 9* 3.4) 3.83 381 28S 221 3.17 292 248 2.45 275 343 346 214 290 3.59 2.81 (60) Talc 3 137 2.81 424 4.13 4.85 3T7S 358 3.81 3.81 328 3.41 281 4% > .50-1.5 > .25-.50 >-.10-25 > .C6-.10 >.01-.05 > 4.08.0 254.0 258 1.5-25 3.06 7 7ft 326 >8.0 276 321 245 245 268 256 228 1.98 (62) Tate 4 4.06 3.40 438 362 4.19 4.81 4.58 422 3.32 3.32 332 3.79 302 (63) AJumm. 8 272 286 271 222 242 287 246 212 159 225 234 1.70 1.72 1.86 0.74 Dlarrwtw (641 Gtan 21 (SI) es > 4.08.0 > 25-4.0 1.15 > 1.5-2.5 1.15 > .50-1.5 1.46 2.06 >26.50 1.15 1.46 > .10-25 > .05-.10 > .01-.05 >8.0 1.15 1.85 1.63 1.15 1.46 2.26 2.00 1.46 t.15 (65) Gian 22 <S2) 1.17 1.47 1.17 1.47 1.77 2.07 1.47 1.17 225 1.87 1.47 (66) Gian 17 (M6S1 1.87 1.77 245 4.43 483 4.38 6.17 4.88 1S7 5.60 4.73 3.58 5.45 428 6.16 2 85 3.79 2.15 ' 4 (67) Gtan 18 (YW) >ao 0.89 0.34 0% > 4.08.0 0.92 0.40 > 2.54.0 080 0.30 >15-25 1.00 0.11 >50-15 1.10 0.41 >25-50 >.1025 > .05-.10 >.01-.05 (S3) Crodd. 13 258 4.00 4.30 4.46 4.68 4.00 450 363 4.6S 2.68 431 298 298 316 268 (69) WMsston. 4 0% 1.43 0.95 1.43 125 2.07 156 0.95 1.86 0.95 1.99 0.95 0.95 1.43 1.56 1.80 151 0.95 125 (70) Tate 5 0% >4.08.0 25-4.0 3.43 1.625 4.33 >.501.5 4.62 498 >25.50 4.56 430 >.1025 427 3.90 >.05.10 3 97 360 >..01-05 390 3.13 >51-1 > 1-4 >8.0 >4-8 > 8-64 >64 (71) Talc 6 4.63 483 4.68 465 4.10 >51-1 32S 3.95 489 4.43 3.95 3SS 325 >1-4 32S 325 3.55 355 325 325 3.55 3.55 32S >48 > 8-64 >64 (72) Talc 7 4.95 5.33 5.18 5.14 4.82 > .01-1 4.37 5.09 4.93 437 4.63 > 1-4 367 3.67 367 367 >4-8 >8-64 > 4 Longth Ti-XT.Fir.rRfc 1 (continuedi.--Fiber distribution bv common log of the nnmbei of particles per microgram in each ol 34 dimensional categories. JNCI. VCll.. '>1. NO. S NOVEMBER l<MI Carcinogenicity of Fibrous Minerals Expt No. Compound I Titanate 1 2 Titanate 2 3 Si carbide 4 Dawson 5 5 Tremolite 1 6 Tremolite 2 1 Dawson 1 8 Crocid 1 9 Crocid 2 to Crocid 3 11 Amosite 12 Crocid 4 13 Glass 1 14 Crocid 5 15 Glass 2 16 Glass 3 17 Glass 4 18 Alumin 1 19 Glass 5 20 Dawson 7 21 Dawson 4 22 Dawson 3 23 Glass 6 24 Crocid 6 25 Crocid 7 26 Crocid 8 27 Alumin 2 28 Alumin 3 29 Crocid 9 30 Wollaston 1 31 Alumin 4 32 Crocid 10 33 . Alumin 5. 34 Glass 20 35 Glass 7 36 Wollaston 3 Table 1.--Summary of 75 experiments with different fibrous materials Actual tumor incidence Percent tumor probability SD Common log fibers/pg, <0.25 pm X >8 pm Expt No. Compound Actual tumor incidence Percent tumor probability + SD 21/29 20/29 17/26 26/29 22/28 21/28 20/25 18/27 17/24 15/23 14/25 15/24 9/17 14/29 12/31 20/29 18/29 15/24 16/25 16/30 11/26 9/24 7/22 9/27 11/26 8/25 8/27 9/27 8/27 5/20 4/25 6/29 4/22 4/26 5/28 3/21 95+4.7 100 100 100 100 100 95+4.8 94+6.0 9316.5 93+6.9 9317.1 869.0 85+13.2 78+10.8 77+16.6 74+8.5 719.1 7010.2 69+9.6 68+9.8 66+12.2 6613.4 64+17.7 63+13.9 56+11.7 5312.9 4411.7 4110.5 339.8 31+12.5 2812.0 3713.6 . 22+9.8 22+10.0 218.7 19+10.5 4.94 4.70 5.15 4.94 3.14 2.84 4.66 5.21 4.30 5.01 3.53 5.13 5.16 3.29 4.29 3.59 4.02 3.63 3.00 4.71 4.01 5.73 4.01 4.60 2.65 0 2.95 2.474.25 0 2.60 3.09 3.73 0. 2.50 0 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 . -64 65 66 67 68 69 70 71 72 Halloy 1 Halloy 2 Glass 8 Crocid 11 Glass 19 Glass 9 Alumin 6 Dawson 6 Dawson 2 Wollaston 2 Crocid 12 Attapul 2 Glass 10 Glass 11 Titanate 3 Attapul 1 Talc 1 Glass 12 Glass 13 Glass 14 Glass 15 Alumin 7 Glass 16 Talc 3 Talc 2 Talc 4 Alumin 8 Glass 21 Glass 22 Glass 17 Glass 18 Crocid 13 Wollaston 4 Talc 5 Talc 6 Talc 7 4/25 5/28 3/26 4/29 2/28 2/28 . 2/28 3/30 2/27 2/25 2/27 2/29 2/27 1/27 1/28 2/29 1/26 1/25 1/27 1/25 1/24 1/25 1/29 1/29 1/30 1/29 1/28 2/47 1/45 0/28 0/115 0/29 0/24 0/30 0/30 0/29 20+9.0 239.3 19+10.3 198.5 159.0 149.4 13+8.8 136.9 12+7.9 128.0 10+7.0 11+7.5 8+5.6 85.6 8+8.0 8+5.3 7+6.9 7+5.4 65.7 65.5 6+5.9 55.1 5+4.4 44.3 4+3.8 5+4.9 3+3.4 . 6+4.4 2+2.3 0 0 0 0 0' 0 0 Common log fibers/pg, 0.25 pm x >8 pm 0 0 3.01 0 0 L.84 0.82 0 0 0 3.73 0 0 0 0 0 0 0 0 0 1.30 0 0 0 0 0 0 0 0 0 0 0 0 0 3.30 0 They represent an excellent size distribution for com parison. IVollastonite (wollaston 1-4).--Wollastonite is a na turally occurring crystalline fiber of monocalcium silicate (15-18). Four separate samples of this substitute for asbestos were received from the same Canadian mine. These were graded commercially according to size by the designation A, B, D. and F. ft was apparent at low-power magnification that only grade F was completely fibrous and that these fibers were relatively large. Tremolite (tremolite l, 2).--The second type of amphibole asbestos studied was tremolite, a material that has a close affinity to the talcs. Both of these samples were from the same lot of asbestos and were in the optimal range of size for carcinogenesis. Compari son of the. 'bers indicated that they were distinctly smaller in dia-.eter than the tremolite fibers used by Smith et al. (29). Amosite.--The third amphibole asbestos studied was a single sample of South African amosite from the UICC standard reference samples. No efforts were made to alter this as received, and descriptions of this samp as published should apply (19, 21, 22). Attapulgite (attapul 1-2).--Of the natural fibers, it clay attapulgite was of particular interest because of i use in many household items that generate respirab dust. Two different samples of this complex hvdratt magnesium silicate were obtained from sources Attapulgus. Decatur County, Georgia. Both samp! were considerably refined, and by electron microscoi they were seen to be composed entirely of short fibe of consistently small diameter (30). These refined da were considered by the U.S. Bureau of Mines to be 9( or greater in purity, with the remaining 10% bcii quartz. Halloysite (halloy 1-2).--Halloysite is a natural brous hydrated aluminum silicate, which is respirat and of minute size. The 2 samples were obtained frc Dr. Walter Parham, who recovered them from the r: water supply of Hong Kong. On examination th< samples were seen to have a tendency for dumping water. In an effort to disperse the minute fibers, ' second sample was sonicated and treated with sodit JNCI. VOL. o7. NO. 5. NOVKMBF.R 972 Stanton, Layard, Tegeri*, et al. #hexametaphosphate. Clumping persisted in this second ample, and little different was seen between the 2 samples. Silicon carbide (si carbide).--One metallic crystal line whisker other than alumin was prepared by the General Technologies Corporation. Silicon carbide was a single sample, which was of exceptionally fine, uniform dimension. Potassium octatitanate (titanate !-}).--In addition to the synthetic crystals of dawsonite, aluminum oxide, and silicon carbide, 2 samples of fibrous crystalline potassium octatitanate (titanate 1 and 2) were tested. These were obtained from two different suppliers but they represent a single source. Because of the potential carcinogenicity of metallic nickel, the control for the.se 2 samples was nonfibrous. finely ground nickel titanate (titanate 3). The 72 experiments represent ail of the experiments done in a single dose range and with durable minerals and particles in the respirable range. Additional con trols outside of these limits are mentioned in "Results." Fiber measurements.--An aliquot of each of the 72 experimental mineral samples was placed on a Form- var-covcred, slotted grid with an opening measuring 1X2 mm. This grid was air dried and first examined under the light microscope. If the fibers appeared satisfactorily distributed,- a photomontage of the entire grid was made at a final magnification of X3.000. The slotted grid was then placed in a Siemens electron microscope, Elmiskop I-A, and the entire grid was scanned at low magnification. From this scan, an area that seemed to represent a typical distribution of particles in the specimen was selected for counting. At a final magnification of about X5.000-100,000, a second photomontage was made of that section of the grid selected to include particles typical of the sample. This selected area, which generally measured about 350X150 pm, was then located on the lower magnification montage of the grid and examined to determine whether the area chosen was truly representative of the entire grid. Finally, all fibers in the area were counted and measured individually. For the diameters, a compara tive scale at the final magnification was used to measure magnified diameters that measured less than 1 mm. In most cases, the selected area counted included at least 1,000 fibers, but the actual number varied with the overall size of the particles. Subsequently, with the aid of the IBM system 370 computer, assuming the fibers to be of cylindrical shape and using the density of the material, we were able to estimate the weight of the counted samples and the number of particles of a given dimension in the 40 mg dose administered. For the purpose of calculation, particles were grouped into 3t icnsional ranges as indicated in text-figure 1, and the number of particles per microgram in each category was calculated. Dupli A cate counts on the montages were done on most samples and were surprisingly similar, as were counts on different areas of the same montage. However, when studies of repeat samples from the original fibers were made, considerable variation in counts occurred. Clearly, the method is subject to several errors; calibration of the electron microscope, deviation of particles from the assumed cylindrical shape, and sampling errors, es pecially where large particles are concerned, represent the major problems. Nevertheless, the estimates are probably valid to within one order of magnitude. Consequently, the counts are reported as the common log with the characteristic of the log representing the probable limit of accuracy (text-fig. I). ' RESULTS Controls have been discussed in previous publica tions (4, 6, 9-11), but they were approached here in a slightly different way. In addition to untreated controls we studied rats in which open thoracotomy was per formed and a noncarcinogenic material was either applied to the pleura or implanted in the lung. These 3 groups (table 2) were rats from numerous experi ments that were of the same species, sex, and age and that were housed in the same quarters. The incidence of clearly apparent pleural neoplasms in untreated, aged outbred Osborne-Mendel female rats was essentially nonexistent. However, a few pleomorphic sarcomas that might be confused with pleural tumors occurred in the left thorax of both treated and, - to a lesser degree, untreated controls. Although these tumors involved the thickness of the chest wall, in most cases the tumors appeared to be derived either from mammary gland fibroadenoma or from suture granuloma in the subcu taneous tissues. But there remained a few tumors for which no definite origin could be determined and which were histologically comparable with pleural sarcomas. In both the experimental groups and the control groups these questionable tumors were counted as pleural sarcomas. These essentially confusing tu mors observed in the controls need to be taken into account in the assessment of the carcinogenicity of the experimental materials. The incidence of pleural sar- i Txble 2.--Incidence of pleural sarcomas in outbred female Osborne-Mendel control rats Time, wk 12-52 53-65 66-78 79-91 92-104 105-120 121-130 131-143 144-156 156 Total Percent Untreated1' Noncar cinogenic pulmonary implants11 1/113 0/15 0/26 0/68 0/26 0/98 1/66 0/27 0/27 1/22 3/488 0.6 0/49 2/26 4/50 L/70 1/72 1/162 0/3 9/432 ' 2.1 Noncar cinogenic pleural implants'1 0/47 1/72 3/64 2/85 10/294 1/36 17/598 2.8 Combined controls' 1/209 3/113 7/140 3/223 11/392 2/296 1/69 0/27 0/27 1/22 29/1,518 1.9 * No. dead with pleural sarcomas/No. dead without pleural sarcomas. JNU. VOI,, (,7. NO. 5. NOVEMBER IMHt Carcinogenicity of Fibrous Minerals riv. * esem arc * ide. ion the ica- | na rols '.crher lese , eriand nee ted, Uy hat the > ree, | the | tors and " { trai ) the tted tunto the ale. ted isJ i t i t ) \8 comas in all 3 control groups combined, calculated by the life table method (13), was 7.74.2%. Comparison of this incidence with the pleural sarcoma incidence in the 72 individual experiments showed that the inci dence of pleural sarcomas in a particular experimental group was significantly greater than that in the com bined control group only if it exceeded 30% {see expts 1-29 in. table 1). In regard to the controls, some negative experiments with intrapleural implants not used as controls should be mentioned. These experiments included intrapleural implants that did not conform to the type of materials under consideration because the particles were either nondurable (cotton lint, gypsum, and carrageenan), were of greater than respirable size (steel shavings, steel wool, vermiculite, polyurethane, tungsten carbide, and infusorial earth), or were exclusively nonfibrous (poly acrylic nitrile, antigorite, silicon dusts, and several glasses). None of these experiments had an incidence of pleural sarcoma that was significantly greater than the 7.7% incidence of the combined control group. From the summarization of the 72 experiments in table 1 and text-figure l, even cursory examination of the fiber distribution suggested that particles in the relatively thin- and long-dimensional categories were associated with higher tumor probabilities. This obser vation was confirmed by the statistical correlation and regression techniques that were used in previous papers (4, 9, 10). The logit transformation (13) was applied to the estimated tumor probabilities (p) according to the formula: Logit=ln [p/(\~p)], where In denotes the natural logarithm. The 34 dimensional categories indi cated in text-figure 1 were arbitrarily, grouped into 11 larger categories, and the simple correlation coefficients of the logit of tumor probability with the common logarithms of numbers of particles per microgram in each of these categories was calculated (see table 3). The maximum correlation coefficient, 0.80, was with particles equal to or less than 0.25 /am in diameter and greater than 8 /am in length. There was no correlation with particles equal to or less than 4 /am in length or with particles greater than 1.5 pm in diameter, but relatively good correlations were noted with log num bers of fibers in categories greater than 4 pm in length and up to 1.5 pm in diameter, with correlation coefficients of 0.45-0.80. The possibility of the existence of relationships between the particle size distributions and tumor prob- Table 3.--Correlation coefficients of logit of tumor probability with common logarithm of number of particles per microgram in different dimensional ranges Fiber Fiber length, pm diameter turn <4 >4-8 >8 >4 >1.5-4 >0.25-1.5 <0.25 -- -0.45 0.01 0.20 -0.28 -0.24 0.45 0.63 -0.30 0.13 0.68 0.80 abilities, which are not disclosed by the simple correla tion coefficients in table 3, was explored by multiple regression methods. These methods were used to find the best-fitting function of the form: logit = a -rb i xt +......... r bk Xk, where ................... .... represent the common logs of numbers of the particles per micro gram in the size categories of table 3, and a, bi............ bk are the regression coefficients to be estimated. The analysis indicated that the addition of further dimen sional categories to the category with diameter equal to or less than 0.25 pm and with length greater than 8 pm did not significantly improve the explanation of the variation in tumor probability. The regression equation for the single variable (x) representing the common log of number of particles per microgram with diameters equal to or less than 0.25 pm and lengths greater than 8 pm was: Int/vO--p)] = -2.62 + 0.9305*. f0.24) (0.0834) The numbers in parentheses beneath the regression coefficients are their estimated standard deviations. The relationship expressed by the above equation is highly significant (P<0.0001). The estimated regression curve is illustrated in text-figure 2. The fact that the use of additional dimensional categories did not significantly improve the fit of the regression equation does not indicate lack of carcino genicity in other categories. The regression of logit of tumor probability on common log of numbers of particles in other categories with a diameter up to 1.5 pm and a length greater than 4 pm would also indicate a highly significant relationship. The difficulty here is that the numbers of particles in adjacent size categories were highly correlated. Better definition of the critical range of carcinogenicity would require more narrowly defined samples (i.e., particles in a narrower dimen sional range). What is perhaps more likely than the existence of a narrow range of sizes within whicr pardcles are carcinogenic and outside of which they an not is that the probability of tumor falls as partich diameter increases and length decreases. Of the 72 experiments, 7 had tumor incidences tha deviated markedly from those predicted by the est; mated regression line. These were: experiments 5 (treir olite 1), 6 (tremolite 2), 26 (crocid 8), 29 (crocid 9), 3 (alumin 5), 47 (crocid 12), and 71 (talc 6) {see table and text-fig. 2). For the first 3 of these experiments th observed responses were higher than the predicte responses, but the high responses can in pari t explained by the fact that there were substand: numbers of fibers in size categories adjacent to tf category used in the regression equadon. For tf remaining 4 experiments, the observed response w; substantially lower than the expected response: a though no apparent explanadon existed for the deviadons, they were possibly due to inaccuracies i the assessment of functional particle size. In prepar; tions of amphibole asbestoses (which included tf crocidolites and tremolites), we observed that boi JNCI. VOL. 67. NO.. 5. NOVEMBER 974 Stanton, Layard, Tegerls, et al. C cfocidolite i.Or G * glass 0 dawsonite 0.9 L * aluminum oxide S - sificon carbide 0.8 f- *\ - attapulglre cc o 5 0.7 D P - tftanate t = talc M * tremolite VV v/ollastonite 0.6 f" H = haNovsite 0 amosite >; 0.5 <5 0.4 Q<3 aocc. 0.3 bw I /) 9 ILHnQew !.- AaDOW o.l rAGGPP -- ! LrGGGG I LTTGG CWTTGG | 0.0 0.0 0.5 1.0 1.5 s Si. P os Op c y ,, XG SLX s G oc / _L 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 LOG NUMBER PARTICLES MEASURING < 0.25 x > 8 urn PER M/CROGRAM Text-figure 2.-- Regression curve relating probability of tumor to logarithm of number of particles per fig with diameter <0.25 pm and length >8 pm. clumping and fragmentation of the particles were greater than those in the other'minerals, and'estimates of particle size distribution in duplicate samples varied most for amphibole asbestoses. DISCUSSION The results show that a wide variety of compounds that seem to have only dimension and durability in common are carcinogenic for the pleura of the rat. Our conclusions regarding those dimensional categories that correlate strongly with probability of pleural tumor remain essentially the same as in previous studies, namely, that probability of pleural sarcoma correlates best with fibers that measure <0.25 /amX>8 Hm, but that relatively high correlations were also observed with fibers in other categories having a diameter up to 1.5 pm and a length greater than 4 pm. A more refined estimate of critical carcinogenic dimen sion may be possible if the parameters of the experi ments were changed. A different animals species, lower dose, more precise means of fiber measurement, more accurate volumetric calculations, and samples with narrower dimensional ranges all might be determining factors in better assessment of the particle dimensions critical to carcinogenicity. However, we should keep in mind two points: a) the dimensional limits are prob ably far from absolute, and b) we are dealing with cancer in the rat and thus extrapolation to man may not be precise. It is t.lear from the histologic studies of these experiments and of previous studies that our data offer an explanation more for the lack of carcinogenicity of short fibers and thick fibers than for the carcino genicity of long, thin fibers. Sections of preneopiastic pleural lesions show avid phagocytosis of both short fibers and large-diameter fibers but negligible phago cytosis of long, thin fibers. Consequently, in these experiments we may simply be measuring the ef ficiency of phagocytosis. Doubtless, we have little real knowledge of the way that long, thin fibers can cause cancer, but as Rous (31) once said, "Since what we think largely determines what we do, it is well that we think something." In the spirit of this quote, it might be profitable to consider potential mechanisms of cancer production by long, thin fibers. Of first impor tance are those hypotheses in which the progenitor of the cancer cell is not directly affected by the fiber. The long latent period would suggest that a generalized alteration either in local milieu or systemic environ ment might be at fault. In this regard, the abundant collagen in the preneopiastic pleural scars should be noted. Consideration of a relationship between this phenomenon and "solid-state" carcinogenesis is attrac tive. though the reduction of plastic sheets to small particles tends to reduce carcinogenesis. Mechanisms of solid-state carcinogenesis have been thoroughly re viewed by Brand (32), and little more need be added. Any hypothesis concerning fibers must take into account the fact that both short fibers and thick fibers are less carcinogenic than fine, long fibers. .Since dose s fixed in weight, but was different in dimension for an experiments, one might consider the surface area as a possible factor. If this were the case then fibers from the same pool that were modified only by shortening should be equal in tumor-producing capacity. Clearly, this is not true in the following experiments: 13 (glass 1. MOL) vs. experiment 49 [glass 10, MOS see (4. 10)]> JNOI. VOL. t,7. NO S. NOVEMBER I OKI Carcinogenicity of Fibrous Minerals and ne. we'we (ht of <>rof he ed m- mt be t is ill I of to IS >c > ol . I .IS in )Z I- and in experiment 24 (crocid 6) and experiment 25 (crocid 7) vs. experiment 40 (crocid 11), experiment 47 (crocid 12), and experiment .68 (crocid 13). However, in these examples the phagocytosis variable cannot be ruled out. A provocative explanation relates to the ability of fine, long fibers to penetrate cells without killing them. That this can occur is evident from in vitro studies (33). However, simple penetration of cells by myceiia of fine dimension (a notable aspect of con tamination of cell cultures by fungi) rarely produces transformation of cell cultures and thus is unlikely to produce cancer. However, mineral fibers differ from fungi in their rigidity as well as chemical content, and one easily could conceive of physical differences be tween the mineral fibers and myceiia that might be critical. REFERENCES (/t Pott F, Huth F. Friedrichs KH. Tumorigenic effect of fibrous dusts in experimental animals. Environ Health Perspect 1974; 9:313-315. (2) Pott F, Friedrichs KH, Huth F. Results of animal experi ments concerning the carcinogenic effects of fibrous dusts and their implication with regard to carcinogenesis in humans. Zentralbl Bakteriol (B| 1976; 162:467-505. (3) Pott F, Friedrichs KH. Tumorcn dcr Ratten nach i. p. Injec tion faser forminger Staube. Naturwissenschaften 1972: 59:318. (7) Stanton MF, Layard M. Teceris A, Miller E. May M. Kent E. Carcinogenicity of fibrous glass: Pleural response in the rat in relation to liber dimension. J Natl Cancer Inst 1977; 58: 587-603. (5) Smith WE. Miller L. Elasser RE. Tests for carcinogenicity of asbestos. Ann NY Arad Sci 1965: 132:456-488. .-(<>) Stanton..ME, Wrench.C. Mechanisms of mesothelioma inductionwith- asbestos and fibrous glass. J Natl Cancer Inst 1972: 48:797-821. (7) Wagner JC. Asbestos carcinogenesis. Am Chem Soc Monogr 1976; 173:729-736. (S) Wagner JC. Berry G. Timbrell V. Mrsotheliomata in rats after inoculation with asbestos and other materials. Br J Canter 1973; 28:173-185. (9) Stanton MF. Layard MW. Carcinogenicity of natural and man-made fibers. In: Margison GP. ed. Carcinogenesis. Ad " vances in medical oncology, research and education. Vol. 1. Oxford and New York: Pergamon Press, 1979:181-187. I/O) --------- . The carcinogenicity of fibrous minerals. In: Proceedings of the workshop on asbestos: Definitions and measurement method: held at the National Bureau of Standards. Gaithers burg. Md,, July 18-20, 1977. Washington. D.C.: National Bureau of Standards, Nov 1978:143-151 (NBS special publica tion No. 506). (.It) Stanton MF. Some etiological considerations of fibre carcinosencsis. In: Bogovski P. Timbrell V. Gilson J. et al., eds. Biological effects of asbestos. Lyon, France: WHO. 1973: 289-294 (IARC publication No. 8). (/2) Churg J. Rosen SH. Moolten S. Histological characteristics of mesothelioma associated with asbestos. Ann NY Acad Sci 1965: 132:614-622. (13) ArmitagE P. Statistical methods in medical research. New York: Wiley. 1971:376-377. 410-414. (14) Pilgrim HI. Dowd JE. Correcting for extraneous death in the evolution of morbidity or mortality from tumors. Cancer Res 1963: 23:45-48. (15) Olson RH. Introduction. In: Lefond SS, ed. Industrial minerals and rocks. 4th ed. New York: American institute of Mining. Metallurgical and Petroleum Engineers. Inc.. 1975. (16) Breck DW. Synthetic zeolites: Properties and application. In: Lefond SS, ed. Industrial minerals and rocks. 1th ed. New York: American Institute of Mining, Metallurgical and Pe troleum Engineers. Tnc., 1975. (17) Sheppard RA. Sedimentary rocks. In: Lefond SS. ed. Industrial minerals and rocks. 4th ed. New York: American Institute of Mining, Metallurgical and Petroleum Engineers, Inc., 1975. (18) Mumpton FA. Commercial utilization of natural zeolites. In: Lefond SS. ed. Industrial minerals and rocks, 4th ed. New York: American Institute of Mining, Metallurgical and Pe troleum Engineers, Inc., 1975. 119) Rendall RE. The data sheets on the chemical and physical properties of the UICC standard reference samples, in: Sha piro HA. ed. Pneumoconiosis: Proceedings of the interna tional conference. Johannesburg. Capetown, Union of South Africa: Oxford Univ Press, 1970:23-27. (20) Ross M. The asbestos minerals: Definitions, description, modes of formation, physical and chemical properties, and health risks to the mining community. In: Proceedings of the work shop on -asbestos: Definitions and measurement methods: held at the National Bureau of Standards. Gaithersburg. Md.. Juh 18-20. 1977. Washington. D.C.: National Bureau of Standards. Nov 1978:49-63 (NBS special publication No. 506). (21) Timbrell V. Characteristics of the International Union A gains: Cancer standard reference samples of asbestos. In: Shapirc HA, ed. Pneumoconiosis: Proceedings of the intemationa. conference. Johannesburg. Capetown. Union of South Africa. Oxford Univ Press, 1970:28-36. (.22) Timbrell V. Gilson JC Webster I. UICC standard reference samples of asbestos. Int J Cancer 1968: 3:406-408. (23) Timbrell V. Physical factors as etiological mechanisms. In Bogovoski P, Timbrell V. Gilson JC et al,, eds. Biologtca. effects of asbestos. Lyon, France: WHO. 1973:295-303 IIARC publication No. 8). (24) Speil S. Leineweber J. Personal experiences with making sum " pies of fibers for biological experiments. In: Pelnar PV. ed Fibres (or biological experiments. Montreal: Institute of Oc cupational and Environmental Health, 1974:45-50. (25) Wagner JC The pathogenesis of tumors following the intra pleural injection of asbestos and silica. In: Nettesheim P Hanna MG. Dcathcrgate JW-, eds.- Morphology of experimenta respiratory carcinogenesis. Oak Ridge. Tennessee: Oak RidgNational Laboratory, 1970:347-358 (Atomic Energy Com mi t sion symposium series No. 21). 126) Harington JS. Allison AC, Badami DV. Mineral fibers: Chem icai. phvsiochemical and biological properties. Adv Pharmaco Chcmother 1975; 12:291-402. (27) Jackson J, Huggins CW, Ampian SG. Synthesis and characteri ration of dawsonite. Washington. D.C: U.S. Dept of lntertoi 1972 (Bureau of Mines report of investigation, 7664). (28) Huggins CW, Green TE. Thermal decomposition of dawsonite Am Mineralog 1973; 58:548-550. (29) Smith WE. Experimental studies on biological effects of tremr lite talc on hamsters. In: Proceedings of the symposium o talc, Washington. D.C, May 8, 1973. Washington. D.C.: (I..Bureau of Mines 1974:43-48 (Bureau of Mines informano circular 8639). (30) Huccins CW. Denny MV. Shell HR. Properties of polygo skite. an asbestiform mineral. Washington. D.C.: U.S. Dept Interior. 1962 (Bureau of Mines information circular R16071 (31) Rous P. The virus tumors and the tumor problem. In: Th Hawey lectures series. No. 31. Baltimore: Williams & Wilkin 1935-1936:74-115. i.>2) Brand KG. "Solid-state" or "foreign-body" carcinogenesis. 1; Symington T, Carter RL, eds. Scientific foundations oncology. London: William Heinemann Medical Books. 197 490-495. (33) W' MJ. Upkin LE. Stanton MF. Franks AL. In vitro ryt. to ' assay as applied to asbestos and other minerals: 1 possible relevance to carcinogenicity. In: International wori shop on the in vitro effects of mineral dusts. Medical Researc Council Pneumoconiosis Unit. Penarth. Wales. Sept !-' 1979. Penarth. Wales: Medical Research Council, 197 t'P388Dl). JL JNC1, VOL. b7. NO. ?. NOVEMBER ENVIRONMEMM. RESEARCH 31, 32-53 (1983) Characterization of Three Types of Chrysotile Asbestos after Aerosolization Kent E. Pinkerton,*'1 Arnold R. Brody,t Daniel A. McLaurin.? Bernard Adkins, Jr,S Robert W. O'Connor.S Philip C. Pratt." and James D. Crapo* 'Departments of Pathology and Medicine. I)nke University and Durham \'eleran\ Administration Medical Center. Durham. Xorrh Carolina 27705; *Laboratory of Pulmonary Function and Toxicology, ,Vational Institute of Environmental Health Sciences. xBeclon-Dickinson Research Center, and ^Northrop Services Inc. Research Triangle Park, \tirrh Carolina 27711 Received December 22, 1981 Jeffrey Mine and Coalinga Mine chrysotile. two asbestos samples prepared for experi mental research by the National Institute of Environmental Health Sciences, and the LTCC B chrysotile reference sample have been characterized in the aerosolized state using gravimetric measurements, light microscopy, scanning electron microscopy, and x-ray energy spectrometry. These methods revealed (11 a greater "respirable" mass fraction in the Jeffrey and UICC B preparations compared to the Coalinga sample, (2) for fibers greater than 5 pm in length and less than 3 pm in diameter. Jeffrey Mine chrysotile contained a significantly greater fraction of fibers longer than 40 pm in length compared to the UICC B or Coalinga Mine chrysolites, and (3) Jeffrey and UICC B chrysotile contained no fibers or fiber clusters which exceeded 2 pm in diameter while Coalinga chrysotile contained numer ous fibers and fiber clusters which were greater than 2 pm in diameter. The characterization of these chrysotile preparations in the aerosolized state, in particular the Coalinga Mine chrysotile, demonstrated different fiber length and fiber width distributions when compared with previous characterizations of samples that had been dispersed in a liquid medium ">y ultrasonification. These observations emphasize the importance of determining the size distribution of fibers in the aerosolized state for inhalation studies and the size distribution of fibers in a liquid suspension for oral ingestion, instillation, or injection studies. Because of differences in length-width distributions, each of the studied chrysotile preparations would be expected to have different patterns of deposition in the alveolar regions of the lung after an inhalation exposure. INTRODUCTION There is increasing evidence that asbestos-induced pulmonary fibrosis neoplasia are due to the physical and chemical characteristics of the inhaled ftb (Wagner, 1965: Seaton, 1975: Stanton and Layard. 1978: Pott. 1978). This mak important to use well-characterized fiber preparations in experimental research more clearly identify those factors leading to lung injury. The purpose of paper is to describe two new samples of chrysotile which have been prepared experimental research by the National Institute of Environmental Health Sc en (N1EHS). A third chrysotile. UICC B. has also been characterized in this s ,tp To whom correspondence and reprint request, should be addressed: Box 3P". Duke l.'nr.e Medical Center. Durham. N.C. 2"'"710. 0013-9351 8? S3.00 i >l\v.rtcht I'lK' n Actdcnn- Pvjsv Ins. \i! tights i} JdPMduwtion .tm vnr. <> CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 33 ile Asbestos \. McLaurin ip C. Pratt.'- nrhatn Veterans laboratory of 'mental Health \ervire\ Inc.. n epared for experiloes, and the L'ICC .olized state using roscopy. and x-ray mass fraction in the -) for fibers greater rysotile contained a >ared to the UICC B mtained no fibers or le contained nuraer!`he characterization the Coalinga Mine ons when compared a liquid medium bv etermining the size e size distribution of studies. Because of preparations would ons of the lung after nonary fibrosis of the inhaled fibei 1978). This makesil irimental research ti The purpose of this] . -e been prepared fot ntal Health Sciences] erized in this paper. \ M"1", Duke l`ni\,-rsilV1 ^sotile represents over 90% of the world's asbestos production and is ubiq- illin our environment. The properties of fire retardance, chemical inertness, S Strength, and flexibility make chrysotile important in numerous applicaas insulation, ceiling tiles, brake linings, and cement products. These i^ons of asbestos enhance the frequency of inhalation of chrysotile by Sii'ljnot directly involved with the mining or processing of this material. 5tile is a member of the serpentine family of minerals, whereas all other i>s minerals belong to the amphibole family (Sinclair, 1959). Mechanical Silin of chrysotile can lead to disruption of the fiber at weak points along the ifiThis disruption can cause the fiber to "open up" into its fibrillar subunits, new fibers of smaller diameter and/or shorter length, fibers with splayed 5'fibers with an uneven diameter along the length, or combinations of the e'(Assuncao and Corn, 1975). These features make chrysotile more complex /ze in terms of fiber size and number than the amphibole types of asbestos iich fibers are basically straight and uniform in diameter. characterization of a chrysotile preparation in terms of particle and fiber ilstribution is influenced by a number of factors. These include (1) the state in f|the chrysotile is found, i.e., in bulk, in suspension, or in an aerosol, (2) the of sample collection, i.e., on a slide or filter, (3) the manner in which the jjiffed sample is prepared for examination, i.e., by ultrasonification, transfer or preparation of the filter or slide, (4) the instrument used to measure the *|b]s and fibers in the sample, i.e., the optical microscope, transmission elec troscope, or scanning electron microscope, and (5) the criteria used for ^fibers, i.e., a minimum fiber length, a 3 to 1 aspect ratio, or characteristics ^morphology. I^cterization of chrysotile in the aerosolized stat.e has been done for the and B reference samples (Timbrell, 1970a: Beckett, 1973). Two new &h samples of chrysotile have been prepared in bulk (approximately 1000 is of each) and are available for experimental research through the National $te of Environmental Health Sciences, Research Triangle Park, North jjjfaa. These new preparations (Jeffrey and Coalinga chrysotile) have been Sterized for elemental composition, mineral composition, particle surface |i|nd density, and thermal properties using optical emission spectrography, //diffraction, thermogravimetry, pycnometry, and a number of petrographic SScopic techniques (Campbell et at.. 1980). Particle size analysis has been samples dispersed in a liquid medium in which measurements of fiber and diameter were made (Wylie, 1979; Campbell el al., 1980; Siegrist and IS, 1980). Particle size analysis of these preparations in the aerosolized state |ihbt done. The purpose of this study is to characterize in the aerosolized state ^Jeffrey and Coalinga chrysotile preparations along with the UICC B reference ^ample. To characterize each preparation, gravimetric dust measurements, optical -.microscopy, scanning electron microscopy (SEM), and x-ray energy spectrometry ' Wete used. Information obtained using these techniques has helped identify char acteristics of these asbestos preparations which may influence their deposition Jwtem when inhaled as an aerosol and which could potentially contribute to their ability to cause pulmonary injury. 34 PINKERTON ET At.. MATERIALS AND METHODS Fiher Prepara lions Chrysotile samples were obtained from the following locations: the Coal] Mine in California (Union Carbide), the Jeffrey Mine in Quebec. Ca (Johns-Manville). and the Canadian reference sample prepared by the tional Union Against Cancer (UICC B). A brief history for each preparati given. Coalinga Mine fiber. Identified as COF-25, this unique form of chrysoti obtained from the New Idria serpentinite mass located in the Diablo California. The deposit is unusual in that the fibers are randomly oriented mat, rather than as parallel fibers running in veins, and the mining process is with bulldozers. The deposit is almost pure chrysotile. The fibers are sho length and are not of spinning grade. Because of the absence of long fibers in chrysotile, a great deal of interest has been generated in using this chrysotile "short-range fiber" preparation. Preparation of this short fiber material for experimental research was doi the following manner. The ore from the mine was first screened to remove laminating rocks. From this point the material was processed in water as as of 1 % solids and 99*^ water. The slurry was passed through a grinder a magnetic separator three times to open the fiber bundles and to remove iron-containing minerals. Between each grinding the slurry was passed throu' particle size separator (hydroclone) under pressure. The aqueous slurry w tated inside the hydroclone forming a vortex. Heavy particles escape fror hydroclone through a side port located near the bottom of the hydroclone. "| coarse particles were reground and put through the hydroclone again. Thei particles leave the hydroclone through a side port located near the top o vortex. These particles were fed into a series of hydroclones in which the top ports (overflow port) are progressively smaller to allow for the collection of, and finer chrysotile fibers. The final hydroclone port had an external diame 25 mm from which the chrysotile preparation was collected. (For industrial poses the final hydroclone port is usually six inches (personal communic Asbestos Group. Union Carbide. Niagara Falls. N.Y.l.) Langer el al. (1978) described in detail a chrysotile sample also obtained the Coalinga Mine deposit, referred to as Calidria RG-144. The difference tween COF-25 and RG-144 are a result of the differences in the processing raw material. COF-25 has a finer particle size than RG-144. This finer particl is a result of differences in the pressure, vortex characteristics, and overflo' diameters used in the hydroclone. COF-25 is not derived from RG-144 by fi` grinding or pellet milling. j Jeffrey Mine fiber. Identified as Plastibest-20. this form of chrysotile is agra asbestos used by the plastics industry. The fibers run in parallel bundles, orie crosswise in veins in serpentine rock and for this project were purified by r' milling and air separation using standard industrial techniques. The (:ial preparation by volume is greater than 960 chrysotile tCampbell ei 1 Preparation of the material for experimental research was accomplished by sing the material through a hurricane pulverizer three times to open fiber bun CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 35 B chrysotile reference sample. This preparation is a grade 4 chrysotile I et al., 1968) obtained from eight different Canadian chrysotile mines and according to the proportional production of each mine during the year of : the Coalingaj The approval of this preparation was made in 1966 by the Union Inter- sbec, Canadal i<...w|n^nale Contre Cancer (UICC) to standardize asbestos samples used in experi- >y the Interna- research. The literature is replete with information regarding the physical preparation is' ||'i|?S|i;chemical makeup of this chrysotile preparation (Timbrell et al.. 1968; Tim- 1970a, b; Rendall, 1970, 1980; Morgan and Cralley, 1973; Beckett, 1973). We jf chrysotile is j reexamined UICC B chrysotile in the aerosolized state with a twofold pur- uablo range in PSjlpk t0 compare our results for the UICC B fiber size distribution with those /.oriented as a '' "jjd in prior studies in the literature, and (2) to correlate the size distribution of process is do ne;|.Mfe'Coalinga and Jeffrey Mine fibers to the UICC chrysotile fiber preparations ||Lrs are short in ag fibers in this srass : for iterations experimental inhalation research. No manipulation was done prior to aerosolizing the fibers. of these chrysotile chrysotile as a Aerosolization ch was done in m. ' ^modified Timbrell generator (Timbrell, 1968) was used to create a dust cloud to.remove con- n a 5-m3-stainless steel-exposure chamber. Each asbestos preparation was ater as a slurry m gjinder and a mr /compressed with a plunger in a 2.8 x 9.0-cm cylinder to form a plug. This |jmechanicaily advanced into the pathway of a blade rotating 1500 rpm to to;, remove any an aerosol of fibers within the dispersing bowl of the generator. A copper assed through a jnnecting the dispersing bowl to the exposure chamber facilitated the pas- ,'Jjurry was ro- aerosolized fibers into the exposure chamber. The concentration of the se|pe from thej |ps dust cloud was regulated by adjustment of the airflow through the expo- .iraplone. Thesi 3|$n. The light Jjhamber. This resulted in the flow of air being maintained between 200 to 400 Ifter minute. ' '' : top of the S the top exit lotion of finer) $Dust Mass Concentration A$ce the exposure chamber had stabilized (normally after 1 hr of dust genera- diameter of| jjj| a gravimetric measurement of the dust concentration within the exposure industrial pur- taber was made. This was accomplished by sampling 100 liters of chamber air i$|hmunication, In through a 0.8-ju.m Nucleopore filter housed within a Gelman filter holder. Unsample time was 10 min at a flowrate of 10.0 liters per minute. Tojbtained from] mi. , " differences be- Mass Concentration t^pssing of the | lpiive resp.irable mass concentration in the exposure chamber for each chrysotile particle size! Iteration was measured using two different instruments: a Casella sampler and ^/overflow port | !|p|||scade impactor. The term "respirable mass concentration" is arbitrary in ^44 by further1 rf. was decided by totally different parameters for each apparatus used. Using .||S;Casella sampler, a gravimetric estimate of the respiratory mass concentration ;ofile is a grade 4 wte made by collecting fibers and particles with an equivalent aerodynamic diam- andles, oriented "Hir of 7.1 fj.m or less on a glass fiber filter. Fibers and particles larger than this nnfied by roller Mflivalent diameter were prevented from depositing on the filter by a multichan -/Ifhe final fiber nel horizontal elutriator. Each sample was collected over a 6-hr period al a flow Met al.. 19'-'01. rate of 2.5 liters per minute with the Casella sampler placed inside the exposure iplished by pas chamber. ?! fiber bundles. Using the Cascade impactor. the respirable mass concentration was defined to 36 PINKERTON ET AL. include filters on which the majority (>50%) of fibers collected were less than! in length. This was determined by examination of the filter by optical mic copy. This respirable mass was obtained by using precutter stages in the Cas impactor to eliminate the longer and thicker fibers and by adjusting the flowrat obtain a fiber size distribution (cut point) of 10 pm or less in length. The volume of air sampled through the Cascade impactor was 200 liters. A sample mass was also collected on the same day by sampling 0.2 m3 of chambe at a flowrate of 10.5 liters per minute for 19 min 3 sec using the same setup the Cascade impactor. All gravimetric measurements were expressed in mg/ Gravimetric measurements of the respirable mass collected in the Cascade im tor were done for the Jeffrey and Coalinga Mine preparations only, because of known similarity between Jeffrey and U1CC B chrvsotile. Collection of Fiber Samples The same apparatus used to collect samples for dust concentration surements was also used to collect fiber samples. Samples to be analyzed by microscopy were collected on Millipore-type AAVVP membrane filters at a fll rate of 0.1 liters per minute for 3 to 5 min depending upon the chamber concenj tion of the chrysotile preparation. Before use. the filters were treated in a Oj solution of Hyamin 2389 and dried at room temperature overnight to prevent st| charging of the filter. Samples analyzed by scanning electron microscopy collected on a 0.2-p.m Nucleopore filter for 2 sec and 5 sec at a flowrate of 10 lij per minute. Preparation of Filters for Examination Light microscopy. After sampling, the filter was placed on a 25 x 75-mm microscope slide, sample side down. The filter was cleared by holding the s over an evaporating flask of boiling acetone. After clearing, a drop of glyc triacetate (Permount) was placed on the dissolved filter and a cover slip app,! Electron microscopy. Filters were secured to the polished side of a G planchet with carbon paint and were gold coated. The thickness of the gold was 100 to 150 nm. Fih er Cha factorization Light microscopy. A Beckett G22 graticule was used. At 500x, the magh tion used for sizing and counting, the graticule was a square, 100 pm on each The lattice on the graticule had a spacing of 5 pm in one direction and 3 /ion perpendicular direction. The following rules were used for fiber characteriza All fibers counted had at least a 3 to 1 length-to-diameter ratio (aspect ratio), fibers 5 pm in length or longer were counted. A fiber bundle which met the 3 aspect ratio requirement was counted as a fiber. Bundles with a diameter gr< than 3 pm were not counted. Only fibers whose midpoint was located within' graticule were counted. At least 200 fibers were counted from 20 to 10; ran graticule fields for each filter sample. The characterization of each chrvs preparation by light microscopy was based upon measurements from filters lected on a daily basis (5 days week for 12 months). Electron microscopy. The magnification used for particle sizing and coun CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 37 llected were less >. .an ; filter by optical microi ter stages in the CascadJ adjusting the flowrate ;ss in length, tor was 200 liters. A toi ling 0.2 m3 of chamber ing the same setup mini ere expressed in mg/n^ ed in the Cascade imp; tions only, because of tl le. 1^000. To facilitate measurement of long fibers and fiber width, magnificaging from 1200 to 18,000x were used. Counting and sizing of particles itne directly on the electron microscope screen. Only particles whose midj&j] within the area of the viewing screen were analyzed. This procedure ^ly reduced all particles to points, thus causing every particle to have an jjppbability of being counted. To assure accuracy in measurement, a stan- made consisting of latex beads 1.099 fim in diameter on a 0.2-/u.m Nui filter. The filter and beads were gold coated and used to verify the magjin prior to each counting session. The area of the filter analyzed consisted jfs randomly selected across portions of the filter. A field was randomly at 10,000 magnification and then moved to the right. Several bands jiounted per filter and a range of 4 to 8 filters was analyzed for each ftile preparation. lust concentration mei es to be analyzed b\ li| embrane filters at a flo' n the chamber concent! -s were treated in a 0.1 ivemight to prevent statij lectron microscopy wei ;C.at a flowrate of 10 Iitei I'following three categories were used to characterize the material on the Sijfor each of the chrysotile preparations: HjPiber: (a) At least a 3 to 1 aspect ratio was required, (b) Generally, a fiber lylinder-shaped with a uniform diameter. However, a slight separation of the gji.along any section of the fiber or at the ends was permissible, (c) The jjr of the fiber was measured at the widest portion of the fiber whether it jjjjcl at the end or at some point along the fiber, (d) Smaller fibrils solidly hto the major body of the fiber constituted a portion of the fiber. It was not !{as a separate fiber. The diameter of the fiber would be measured at this ; constituted the widest portion of the fiber, ijfoer cluster: (a) A fiber mass with at least a 3 to 1 aspect ratio was re- ed!bn a 25 x 75-mm gl: $red by holding the slii ^ing, a drop of glycert Rnd a cover slip applied. Dblished side of a Gou| thickness of the gold coi F' |b) It was composed of small fibrils usually oriented in the same direction, |ally separated. Separation of fibrils may occur at any point along the r.the bundle or cluster, (c) The diameter 'of the cluster was measured at ||est combined width along the fiber mass. A fiber cluster usually has a jpriable width than a "fiber." Sfonfibrous particle: (a) Any particle with less than a 3 to 1 aspect ratio, (b) ijgest dimension was measured as well as the greatest perpendicular meaSfnt. Sr- "floes" or clumps were also occasionally present on the filter consisting Jed fibers and fiber clusters. These were not analyzed because their com- ipht 500 x, the magnific fc-prevented accurate separation into individual fibers and fiber clusters, ijlfe, 100 /am on each sidej jiresentative fibers, clusters, and nonfibrous particles were analyzed for ele- dection and 3 /am in f iJjComposition using x-ray energy spectrometry for each chrysotile prepara- " '(fiber characterization! jlach particle was analyzed for the presence of an element by a scoring |tio (aspect ratio). Onln ranging from 0 to 4+. At 10,000x, 50 nonfibrous particles whose mid- lie which met the 3 to 1 psifell within the viewing area of a randomly selected band across the filter lesiwith a diameter greatei ^ analyzed. In addition, the magnesium-to-silicon ratio was determined for all )H|t was located within th^ icies containing these two elements. t^from 20 to 100 randott prttion of each chrysotiltS '0 RESULTS bments from filter1 coh Concentration Measurements |jlcle sizing and countin A'tir. iI'*A/*''S,-.- r'^The total mass concentration and corresponding respirable mass concentration ' wheach chrysotile preparation are given in Table 1 using both the Casella sampler 38 PINKERTON ET AL. TABLE I Gras metric Measurements for Each CimssotiLE Preparation in the Exposure Chamber ! Preparation (Al Chamber dust mass concentration (mg/nv1)" (B) Respirable concentration tmg.'m3V' Ratio1 B/A] Jeffrey UICC B Coalinga Jeffrey Coalinga 11.36 = 2.18 10.99 2.11 1.76 - 1.46 12.29 r 3.3.V' 15.63 c; 2.41< Casella sampler 9.90 1.63 8.32 s 1.75 3.28 t 0.83 Cascade impactor 2.62 r 0.8 P 0.801' 0.871' 0.75?: 0.423 > 0.213 o.051 ' All daia are means i SD. n = 240 for each ehrysotile preparation. '' For Jeffrey all data are means - SD. n = 12. ' For Coalinga the chamber mass concentration is based on three samples and the respirable col (ration is a single sample. and the Cascade impactor. When comparing the ratio of the respirable concentration to the total dust concentration found in the exposure chamber Jeffrey Mine ehrysotile and UICC B ehrysotile have a relatively high percen] of respirable material based upon the Casella sampler measurements (87% fo Jeffrey ehrysotile and 76% for the UICC B ehrysotile). The ratio of the respii mass concentration to the total mass concentration is significantly low Coalinga Mine ehrysotile (42%). Using the Cascade impactor a substanl smaller fraction of the total dust concentration was found to be respirable. Fol Jeffrey fiber 21% of the total dust concentration was respirable and 5% of the dust concentration was respirable for the Coalinga fiber. Fiber Characterization--Light Microscopy Table 2 lists the percentage of fibers found for each given length interval by microscopy for each of the three ehrysotile preparations. Fibers greater than] /am in length and less than 3 /am in diameter were found in each preparation, percentage of fibers present from 5 to 30 ^m in length was similar for all aerosolized preparations with greater than 70% of all counted fibers falling this length interval. For the fiber-length intervals of 40-50. 50-100. and gri than 100 /u.m, the percentage of fibers contained within each of these length ii vals was significantly greater (P < 0.05) for the Jeffrey Mine preparation th either the UICC B preparation or the Coalinga Mine preparation. Fiber Characterization--SFM The length, width, and aspect ratios for the combined fibers and fiber elusj are illustrated in Figs. 1-3. Figures IA-D illustrate fiber length charac'erisl Those fibers or fiber clusters having a diameter greater than 0.6 /am are show) the crosshatched portions in these figures. This cutoff was chosen becau-e fil \\ 'i i a ,4 1 f J 1 s i CHARACTERISTICS OF AEROSOLIZED CHRVSOTILE TABLE 2 Optical Microscops Fiber Ch\raci i-rization: Percp.niage (*71 Op Ai i. Fibers '5 pi.m in Each Size Class"'' 39 Ratio B/A 0.871 0.757 0.423 Fiber size 5-10 10-20 20-30 30-40 40-50 50-100 Coalinga 32.9 7.4 28.8 2.7 17.1 - 3.4 10.2 2.3 6.0 1.8 3.3 1.3 1.5 1.1 UICC B 31.0 6.8 27.9 2.3 17.6 3.5 11.0 2.5 6.8 1.7* 3.9 1.5 1.9 0.8 Jeffrei 28.5 6.5 ` 25.3 r 2.7-* 17.2 3.0 11.6 2.5` 8.2 = 1.8-* 5.7 1.9-*' 3.6 1.7'* Hi dala are means SD, n = 52 for each chrysolile preparation. Number.of fibers counted per |e<= 1000-2000. libers wtlh a diameter greater than 3 /am were not included in this study. he respirable eonceni npftKirespirable mass! f-t 0.05 when comparing Jeffrey to UICC B using Duncan's multiple comparison test. |r than 0.6 /am have a substantially smaller probability of being respired than are chamber, the! high percentage! ents (87% for the! S3I.'.fibers with smaller diameters (Pooley and Clark, 1979). Figure 1A shows |gth distribution for combined fibers and fiber clusters of Jeffrey Mine JtUe using a normal numerical distribution plot. Figure IB illustrates the of the respirable! '^distribution for Jeffrey chrysotile using a log scale. The length distribution cantly lower foil \ and fiber clusters for UICC B chrysotile, illustrated in Fig. IC, is very r a substantial!TM $to that of Jeffrey Mine chrysotile. Combined fiber and fiber cluster length cspirable. For the! jnga Mine chrysotile, seen in Fig. ID, shows a'distribution with many nd 5% of the total xceeding 30 ftm in length. h the Jeffrey and the UICC B chrysotile, approximately 75% of the ijjijed fibers and fiber clusters, were less than 5 ftm in length, while less than %ithe combined fibers and fiber clusters from the Coalinga chrysotile were h interval by ligh(| n 5 ftm in length. At least 92% of the combined fibers and fiber clusters in > greater than Ijttrey and UICC B aerosols were less than 10 ftm in length, but only 66% of preparation. Thi fmbined fibers and fiber clusters in the Coalinga chrysotile aerosol were less milar for all three) fjjO ftm. hers falling withii jSj-log distribution for combined fiber and fiber cluster width is similar for 100. and great Mine chrysotile and UICC B chrysotile as seen in Figs. 2A and 2B. Few these length inter-^ H>br fiber clusters exceed 0.6 ftm in diameter. In contrast, (Fig. 20 numerous eparation than fol; and fiber clusters exceed 0.6 gm in width in the Coalinga Mine chrysotile 'n. p*sJHn- . I^.^fjljhe log aspect ratio for the Jeffrey Mine chrysotile and UICC B chrysotile show I'^^majority of the fibers and fiber clusters having an aspect ratio less than 100:1 and fiber clusters 3A and 3B). However, both the Jeffrey and UICC B chrysotile have some 'h characterF'.ics. {voters exceeding this aspect ratio. The Coalinga Mine chrysotile had no fibers or am are show a by Bb|r clusters exceeding an aspect ratio of 100:1 (Fig. 30. -on because fibers .^Tables 3-5 contain a detailed description of the length distribution of fibers and 40 PINKERTON ET AL. JEFFREY MINE CHRYSOTILE 100. UICC 8 CHRYSOTILE 500* S 2CC- CO- 05 1 LEN5GT1H0 am 50 CC D COALINGA MINE CHRYSOTILE 300- 05 I 5 10 50 tQQ LENGTH (jiml 1000 05 \ 5 10 SO 'CO LENGTH ( um } Fig. 1. Frequency distribution of length for combined fibers and fiber clusters in each of aerosolized chrysotile preparations. (A) Jeffrey Mine chrysolite plotted on a linear scale for le (B-D) Chrysotile preparations plotted on,a.log -.calc for length. fiber clusters, each as a separate category. The fibers arid fiber clusters are: pressed as numbers counted per length interval, the percentage of the total fo in each length interval, the cumulative percentage, and the mean diamete fibers or fiber clusters in each length interval. The fibers and fiber clusters fore preparation are also expressed in terms of their respective aspect ratio Tables 3 --5. A comparison of Tables 3-5 demonstrates that the mean diameter of fiber fiber clusters for any given length interval is similar for the Jeffrey and L1C preparations. There was no fiber or fiber cluster measured in either the Jeffrey UICC B preparation that exceeded 2 jtm in diameter. The aspect ratio incre' for both fibers and fiber clusters with increasing length in the Jeffrey and LTC preparations. The Coalinga preparation has a similar mean fiber and fiber clu diameter below the 5-/am length interval. However, for fibers and fiber clusf longer than 5 ju.m in length, the mean diameter in the Coalinga prepar.itio significantly greater for both fibers and fiber clusters compared to the Jefirey UICC B preparations. In general, as the fiber or fiber cluster length increase; the Coalinga preparation, the width also increased. The presence of "thick fi", I i characteristics of aerosolized chrysotile 41 m&\ JEFFREY MINE CHRYSOTILE n = 1054 life 1 u UICC 8 CHRYSOTILE n=1013 COALINGA MINE CHRYSOTILE n = 1050 T!LE mm, 02 061 2 005 02 06 1 2 W10TH (jim) WIDTH {um | ... . 2. Frequency distribution of log width for combined Fibers and fiber clusters in each ^^Sraaolized chrvsotilc preparation. mHip|rthick" fiber clusters in the Coalinga preparation is reflected in the smaller list ratios which in any length interval seldom exceeded 60:1. 1000 fe results of fiber counting by optical microscopy cannot be expected to i each ol' the;! ilc For length. ^l.ate closely with those obtained by scanning electron microscopy. The |jHa for fiber counting by optical microscopy includes all fibers less than 3 m ers are exlotai found 'iameter of .TS for each i ratios in| JEFFREY MINE CHRYSOTILE UICC 8 CHRYSOTILE COALINGA MINE CHRYSOTILE >f libers or i! UICC B Jeffrey or increased id UICC B her cluster -T clusters . \iration is effrey .ind - . creased in j|?)' ck" libers lie ion toon ASPECT RATIO ton toon ASPECT RATIO ion ioo ASPECT RATC ' Fig ' f-roquenc'. distribution of aspect ratio for combined tlbeis and fiber clusters in each j-ierosolv/ed chnsoldo preparation, if? - 1054. 1013. and 1050. respeethcK i m oooooooo C\ o o o o Ml ! 79.9:1 SO: 1 99,9: 1 lOO-.J- 199:1 CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 45 jt and greater than 5 /xm in length. The criteria for fiber-fiber cluster !'|by scanning electron microscopy includes fibers of all lengths and diamjjng as they possess at least a 3 to 1 aspect ratio. By these standards, 75% r--fiber clusters in the aerosolized Jeffrey and UICC B preparations ''y scanning electron microscopy would not have been counted by optical my. For the aerosolized Coalinga preparation, 60% of the fiber-fiber --puld not have been counted by optical microscopy (10% of the total r cluster number exceeded the 3-/xm-diameter limit and 50% of the total ' r cluster number was less than 5 ^.m in length). If. Particle Anal,ysis v foments detected in the 50 randomly selected nonfibrous particles for each e preparation are shown in Table 6. One half of the nonfibrous particles in the Jeffrey and UICC B preparation samples demonstrated the pres{jhagnesium and silicon only. The magnesium-to-silicon ratio for fibers and 8ters analyzed in all three preparations ranged from 0.62 to 1.34. Although *ans of identification, particles with a Mg-to-Si ratio in this range are likely (sent fragments of chrysotile. The percentage of particles possessing the 1-to-Si ratio as chrysotile is 40% in the Jeffrey Mine preparation, 52% in C B reference sample, and 20% in the Coalinga Mine preparation. The 8,iWith low magnesium-to-silicon ratios of 0.30 to 0.40, found only in the Inline aerosolized sample, may represent talc particles. y: DISCUSSION i?" the chrysotile preparations designed for experimental research and wzed in this study have not been previously studied in the aerosolized EC B chrysotile has been studied extensively by both Timbrell (1970a) fett (1973) in the aerosolized state. To fit the Coalinga Mine and Jeffrey iples into the spectrum of chrysotile preparations available for experiliahalation research, UICC B chrysotile was used as a reference to link the ,,study to the work carried out by other investigators, in particular, the ation studies of Timbrell and Beckett. Although differences exist in the , of fiber collection from the dust chambers and in the preparation and is of samples, the studies of Timbrell and Beckett on UICC B arecompara|fee present study (Table 7). By light microscopy we found a greater propor- fibers falling into longer fiber length intervals than did the studies of fjtand Timbrell. This difference may reflect differences in fiber preparation ^.'counting techniques. By electron microscopy, the results of this study for |B chrysotile are nearly identical to those of Timbrell for the distribution of ^Ibngths. Beckett (1973) used SEM to determine fiber length and evaluated J'fibers longer than 5 /am. When our present results were recalculated and ,|essed in a similar fashion (Table 7). we found a similar pattern, but proporJUy fewer long Fibers than did Beckett. Some of these differences may be due uferences in the manner in which the aerosols were generated and the samples infected. ^Ehe Jeffrey Mine and Coalinga Mine chrysotiles have been elegantly analyzed 46 PINKERTON ET AL.. TABLE 6 Eli-.mkm u CoMPosmos op fHE Nonfibrous Puuici.p.s in EACH Chiu sorn h Priu'sra iion' Percentage occurrence Elements detected NaMgAlSiCa NaAl NaAISiKCaFe Mg MgAl MgAISi MgAISiKFe MgAISiCaFe MgAISiCrFe MgAISiFe MgSi ratio Mg to Si =. 0.3-0.4 Mg to Si = O.h - 1.1 Mg to Si = 3.0-10 MgSiCa MgSiFe MgK A1 AlSi AlSiCa AlSiK AlSiCr AISiFe AIK AIKCa AlCr SiCr CaCr Cr Fe None Jeffrey *> -- > T 4 14 -- -- ft 50 {6> (401 (41 -- --- 1 (5 -- -- --- . -- -- ___ -- > L'lCC B -- -- -- 1 -- 30 -- -- _ 54 "> -- T 4 -- ___ ___ ___ . __________ ___ ___ ___ ___ ___ ___ ___ 4 <--\ S' 1 Oi Coalinga ___ 6 ___ 4 -- 10 -- -- > 20 <-- (20 (- 6 --j 14 ! 4 fi 2J 4 2 ___ 2 14'' . i 4'' 3 ___ ! -- ' 50 random nonfibrous particles were analyzed for each chrysolite preparation. May represent a contaminant from the rotary blade used to aerosolize Ihe preparation. by others using the sample preparation technique of fiber dispersion in a liijj medium (Wylie. 1979; Campbell ei a!.. 1980; Siegrist and Wylie. 1980). This of analysis is satisfactory for studies in which the preparations are to be ingi or injected in suspension. However, these studies do not provide an a characterization of the asbestos preparations for inhalation studies since the, cess of aerosolization is generally less efficient in fiber dispersion. A case in is the Coalinga Mine chrysotile preparation. This asbestos preparation j! characterized b> Campbell cl at. < 19801 using fiber samples which were Jispi in a liquid medium by ultrasonification for 10 min. The samples were analyzi transmission electron microscopy. They found that 2.IT of the total chrys'j |jj;|||fen CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 47 TABLE 7 Fiber Length Distribution--Optical Microscopy Dispersion Percentage longer than stated length (/xm) Method Study 4 5 10 20 40 Aerosol Present 100 68.2 Aerosol Beckett*1 100 32 Aerosol TimbrelP 100 22.3 Alcohol TimbrelP 100 46.9 Celloidin Timbrell'1 100 53.6 Fiber Length Distribution--Electron Microscopy 40.5 11 11 21.5 16.9 12.6 5 1.7 7.0 3.8 Instrument TEM Study TimbrelP Percentage longer than stated length (/xm) 0.2 1 T 5 10 20 100 73.6 54.4 27.0 9.6 3.2 SEM Present 100 88.2 64.2 25.5 6.7 1.9 SEM Present 100 85.0 50.7 24.9 8.0 2.7 SEM Beckett" 5 10 20 30 50 60 100 45 12 6 1 SEM Present 100 26.4 7.4 2.3 0.8 SEM Present 100 32.1 10.8 5.5 2.7 jgure 2a in Beckett, 1973. [fed from Tables 6, 7, or 8 in Timbrel), 1970a. 1 . les analyzed were greater than 10 /xm in length and had a mean diameter of m. In the present study using aerosolized samples drawn directly upon [ppore filters and subsequently gold coated, it was found that 34.0% of the fibers and fiber clusters.or 28.1% of all particles (fibers, fiber clusters, ionfibrous particles) in the Coalinga preparation were greater than 10 /xm in i and had a mean diameter of 2.92 /u.m. In a subsequent study by Siegrist and jj? (1980), the Coalinga preparation (referred to as short-range chrysotile) was iCterized for particle size distribution by both transmission and scanning ;on microscopy. The samples were prepared by hand swirling in distilled Ip and dishwashing liquid (for SEM analysis) and by ultrasonification in water $D min (for TEM analysis). The results demonstrated that by both SEM and |M more than 90-95% of the particles were less than 10 /xm in length and more 95% of the particles were less than 1 /xm in diameter. In our study using Hl^osolized samples 71.9% of the total number of particles were less than 10 /xm in and 68% of the total number of particles were less than 1 /xm in diameter. . The differences between these two methods of sample preparation and analysis 48 PINKERTON ET AL. Fro. 4. (A) Aerosolized Jeffrey Mine chrysotile collected on a fl._-p.rn Nucleopore filter an-.! golj coated. Bar = 10 p.m. (B) Aerosolized L'lCC B chrysotile. Bar = 10 pm. iQ Aerosolized Coaling Mine chrysotile. Bar * 10 /an). Examples of a fiber (short arrow, i and a fiber cluster (long arrow l alj! indicated in this micrograph. (Dl Higher magnification of an aerosolized Coalinga Mine chrysotile fibJ cluster demonstrating the fibrillar subunit composition. Bar - 1 um. leopore filler and ^ "J Aerosolized Coal: ea luster dong arrow 1 re Mine ehrvsotile liner 50 PINKERTON ET AL. suggest that either ultrasonification breaks down fiber bundles into smaller! dies and fibrils or that fibers tend to cluster in the aerosolized state. Other in tigators have shown that ultrasonification can result in the breakdow] chrysotile into smaller fibers (Spurny et al.. 1980). In addition, clustering of ` caused by the aerosolization does not seem likely since the Jeffrey and UI chrysotile preparations collected in the same manner as the Coalinga chrys1 have distinctly different particle size distributions from that seen aerosolized Coalinga preparation. The illustration of each fiber preparation o Nucleopore filters (Fig. 4) also show a distinct fiber morphology for the Co~ chrysotile compared to the Jeffrey and UICC B chrysotile. Some of the differences between each of the chrysotile preparations weref reflected by the gravimetric measurements taken of each chrysotile aerosol iJ' exposure chambers. Comparisons made between each chrysotile preparati' the ratio of respirable mass concentration to total mass concentration demons a distinct difference between the Coalinga chrysotile and the other two pre'1 tions of Jeffrey and UICC B chrysotile. It would appear that compared Jeffrey and UICC B fibers the Coalinga preparation contains a greater proper' of particles which are captured in the elutriator system of the Casella sample! in the precutter stages of the Cascade impactor. These fibers and particlef captured because of their greater mass. Using scanning electron microscopy we found that the range of diamete' fibers and fiber clusters was greater for the Coalinga preparation than fo1 Jeffrey and UICC B preparations in the aerosolized state. The trend of incre' diameter with increasing length was found only in the Coalinga preparatio bers and fiber clusters greater than 10 /xm in length with diameters exceeding! were present in the aerosolized Coalinga preparation while in the Jeffre UICC B preparations no fiber or fiber cluster of any length exceeded 2 diameter. The characterization of chrysotile in the aerosolized state in terms of length and fiber diameter is paramount in understanding the nature of any p'1' ration used for inhalation studies. Although the physical characteristics c may not be the only factor in causing lung injury, fiber diameter and, to al degree, fiber length play key roles in the potential for a fiber to reach the alv regions of the lung where injury as a result of asbestos inhalation appears more severe. The potential for each of these chrysotile preparations to cause lung injur inhalation can be best assessed by a review of what is known about the ph; characteristics a fiber must possess in order to reach the alveolar portions < lung. Spherical particles below a certain diameter (approximately 3--4 fin reach the alveolar airspaces of the lung (Lippman and Albert. 1969). Larger! cles are eliminated by deposition in the upper airways primarily thrh sedimentation and impaction as a result of their greater mass. Timbrell. Harris' Fraser have examined the more complex aerodynamic properties of fiber comparing the deposition pattern of fibers to those of artificial spheres l imb 19651. In general they found that fibers have a similar deposition pattern to tfr spheres of three to four times greater diameter. This would suggest that fi CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 51 dies into smaller bu. ed state. Other invei n the breakdown n, clustering of fib& Jeffrey and UICcl e Coalinga chrysoti m that seen in til ier preparation on u logy for the Coaling ^ater than l /tm in diameter would not be likely to reach the alveolar JWthough the curvature of the fiber should also be taken into consideram^rell, 1970b). Additional studies have shown that an equivalent fiber i|)f 0.5 to 2.0 yarn results in deposition within respiratory bronchioles and ll^lveolar ducts, while Fibers of smaller equivalent diameters result in `"'jleposition in the more distal portions of the lung (Harris and Fraser, |pley and Clark (1979) measured the diameter of chrysotile fibers recovSvlung specimens after tissue digestion. They found that only 0.06% of the |d a diameter exceeding 0.5 yum. Although it is possible that these eparations were alsi ysotile aerosol in th sotile preparation c jjtfcfibers have fragmented longitudinally while//; vivo (Suzuki and Churg, Jese results suggest that few fibers with a diameter greater than 0.5 yam *$b alveolar regions of the lung. If this limit in fiber diameter is true for _ lung deposition and the length of the fiber plays the major role in fi- ntration demonstratl, i other two preparal. nat compared to th(jj a greater proportion Casella sampler anL ;rs and particles are| #sis and cell injury, the potential for the three asbestos preparations reported f'study to cause lung injury by inhalation will be different. Only the Jeffrey JfCC B chrysotile preparations have fibers which are less than 0.6 yam in fcr when length is greater than 30 yarn. The Coalinga Mine chrysotile does Isess this property; instead, with increasing fiber length, fiber diameter also |jss. Most fibers longer than 20 yam in length in the aerosolized Coalinga mge of diameters aration than for the, i trend of increasing) nga preparation. Fi lers exceeding 2 yim| in the Jeffrey and) i exceeded 2 y/.m in fions are also greater than 0.6 yam in diameter. Thus, in an inhalation more fibers of greater length distribution would be deposited in the alveolar pf the lungs for the Jeffrey and UICC B chrysotile preparations than for the ; chrysotile preparation. |i point one may ask the question, what constitutes a short-range fiber prepa re have seen that the Coalinga preparation originally thought to be a shortsparation contains numerous long fibers in' the aerosolized state. The J to this question should be based on the potential of a fiber to reach the P'vin terms of fiber regions of lung. In other words, is the fiber respirable? The work of gators who have studied the aerodynamic properties of fibers (Timbrell, are of any prepalleristics of a fiber Hi'and, to a lesser |||#ch the alveolar 'Mi appears to be Saw- plung injury by ||1973) and the physical dimensions of respired fibers (Pooley and Clark, i; would suggest that the Coalinga chrysotile in the aerosolized state repre|fa short-fiber preparation since only fibers less than 30 yarn in length are likely respirable. In contrast, the Jeffrey and UICC B chrysotile preparations pin fibers greater than 30 yam in length possessing a diameter which would |t penetration of the fiber into the alveolar regions of the lung. Based on these Nations, the Coalinga Mine chrysotile preparation can be considered to rep the physical it a "shorter" fiber preparation in the aerosolized state. portions of the ! j3-4 yam) can ((summary, the three chrysotile preparations characterized in the aerosolized in this study demonstrated distinct properties. Gravimetric measurements of Larger parti- rj?h chrysotile preparation revealed that both the Jeffrey and UICC B prepara- jgarily through ptths have a significantly greater portion of the total chamber dust concentration |Hl, Harris and Kji^l^ieh is respirable compared to the Coalinga preparation. By light microscopy it g|pf fibers by found foat for fibers greater than 5 yam in length, the Jeffrey preparation in the gs (Timbrcil. ||^^f!:,osolized state has a significantly greater fraction of fibers exceeding 40 yam in teijii to that of |v.(length than does UICC B chrysotile or Coalinga chrysotile. Finally, by scanning *% that fibers p1'.'^iectron microscopy it was found that the Jeffrey and UICC B preparations pos- icyytro>s .AMt fo< . .&-L-i PINKERTON ET AL. sess fibers and fiber clusters which cross a large length range (the longest sured was 150 /am), but none which exceeded 2 /am in diameter. The Coalij preparation also possessed fibers and fiber clusters across a similar length rai but many exceeded 2 /am in diameter. In terms of respirability, the Jeffrey UICC B aerosolized fibers constitute a mixed short-range and long-range fiber pi ration, while the Coalinga aerosolized fibers represent a somewhat short-fiber pi ration in which very long respirable fibers are not present. In applying the ab| fiber characterization studies to experimental research it is essential to rememl that further manipulation of these chrysotile preparations by grinding (Langei ctl.. 1978) or by fiber separation techniques may alter the fiber size distribui from that presented in this paper. ACKNOWLEDGMENTS This work was supported in part by N1EHS Contracts NOl-ES-0-0004 and NOl-4-2l-64-BCD|| REFERENCES Assuncao, J.. and Corn. M. (1975). The effects of milling on diameters and lengths of fibrous glas^ chrysotile asbestos fibers. Amer. hid. Hyg. Assoc. J. 36. 811 -819, Becked, S. T. (1973). The evaluation of airborne asbestos fibres using a scanning electron microsi Win. Ocatp. Hyg. 16. 405-408. Campbell. W. J.. Huggins, C. W.. and Wylie. A. G. (1980). "Chemical and Physical Characteri; of Amosite, Chrysotile, and Nonfibrous Tremolite for Oral Ingestion Studies by the Nal Institute of Environmental Health Sciences." U.S. Bur. Mines Rep. Invest. 8452. Harris. R. L.. and Fraser. D. A. (1976). A model for deposition of fibers in the human respii system. Amur. hid. Hyg. Assoc. J, 37, 73. Langer. A. M.. Wolff, M. S.. Rohl, A. N., and Selikoff, 1. F. (1978). Variation of properti] chrysotile asbestos subjected to milling. J. Toxicol. Environ. Health 4, 173-188. Lippman. M.. and Albert, R. E. (1969). The effect of particle size on the regional deposition of ii aerosols in the human respiratory tract. Amer. hid. Hyg. Assoc. ./. 30, 257-275. Morgan, A., and Cralley. L. J. (1973). Chemical characteristics of asbestos and associated trai mcnts. hi "Biological Effects of Asbestos" (J. C. Wagner. Ed.I. pp, 113-118. IARC Scie) Publication. Lyon. Pooley. F. D.. and Clark, N. (1979). Fiber dimensions and aspect ratio of crocidolite, chrvsotil| amosite particles detected in lung tissue specimens. Ann. S.Y. Acad. Sci. 330, 711 -716. Pott. F. (1978). Some aspects on the-dosimetry of the carcinogenic potency of asbestos and fibrous dusts. Staub -Iteinhoh l.itfi 38, 486--490. Rendall. R. E. G. (1970). The data sheets on the chemical and physical properties of the UICC si reference samples. In "Pneumoconiosis'. Proceedings of the International Conference. Johi burg. 1969" (H. A. Shapiro. Ed.), pp. 23-27. Oxford Univ. Press. London. Rendall. R. E. G, !|980). Physical and chemical characteristics of L1CC reference sampli "Biological Effects of Mineral Fibres. (J, C. Wagner. Ed.I. Vol, 1. pp. 87-96. IARC Scii Publication No. 30. IARC, Lyon. Seaton. A. (19751. Asbestosis. In "Occupational Lung Diseases" (Morgan. C. Keith, and A. Si Eds.i. Saunders. Philadelphia. Siegrist. H. G,. and Wylie. A. G. ()980l. Characterizing and discriminating the shape of asl particles. Environ. Rc\. 23. 348-361. Sinclair. W. E. 119591. " Asbestos--Its Origin. Production, and Ciili/ntion." Mining nublii London. Snurny. K. R., Opiela. H.. and Weiss. G. 11980). On the milling and ultrasonic treatmer < biological and analytical applications In "Biological Effects of Mineral Fibres" .! t Ed.1, pp 931 93n IaRC -Scientific Publication No. 30, I \RC. Loon. -< 'AJdi't'`Vf ngth range (the longest mej n in diameter. The Coalin, lcross a similar length rang respirability, the Jeffrey _ ige and long-range fiber pre,, a somewhat short-fiber pren esent. In applying the aboy ;frit is essential to rememb, itions by grinding (LangerL er the fiber size distributiol 1^0004 and N01-4-21-64-BCD //. ' |(> ii jg^td lengths of fibrous glass aoj fanning electron microscopjj g||fld Physical Characterizatio| Studies by the Nation jhvest. 8452. ||.in the human respirator! iiation of properties of! |473-188. Hynal deposition of inhale 1257-275. nd associated trace elfr -118. IARC Scientific^ |i... iwidolite, chrysotile an ||i.330, 711-716. f|jpf asbestos and other V,. i Mflfthe UICC standard JSpnference, Johannes- ^rence samples. Inii . IARC Scientific]] ^pith. and A. Seaton, ^ shape of asbestos ^lining publication, Kfii jtment of fibres for fJ. C. Wanner. CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 53 P' Ji and Layard, M. (1978). "The Carcinogenicity of Fibrous Minerals. Workshop on ^Definitions and Measurement Methods," pp. 143-151. NBS Special Publication 506. fl/and Churg, J. (1969). Structure and development of the asbestos body. Amer. J. Pathol. tl965). The inhalation of fibrous dusts. Ann. N.Y. Acad. Sci. 132, 255-273. lp68). A simple dispenser for generating dust clouds from standard reference samples of IfjMhrt. Occup. Hyg. 11, 273-281. Ail970a). Characteristics of the International Union Against Cancer standard reference lof,asbestos. In "Pneumoconiosis: Proceedings of the international Conference, Johanjjfejjlg9" (H. A. Shapiro, Ed.), pp. 28-36. Oxford Univ. Press, London. if(1970b). The inhalation of fibers. In "Pneumoconiosis: Proceedings of the International Uie, Johannesburg, 1969" (H. A. Shapiro, Ed.), pp. 3-9. Oxford Univ. Press, London. pp^j973). Physical factors as etiological mechanisms. In "Biological Effects of Asbestos" fwjlgner, Ed.), pp. 295-303. IARC Scientific Publication. Lyon. Ifu'&ibson, J. C., and Webster. 1. (1968). UICC standard reference samples of asbestos, hit. iHber 3, 406--408. li'C. (1965). The sequelae of exposure to asbestos dust. Atm. ,Y. 1'. Acad. Sci. 132, 691. jfej, (1979). Fiber length and aspect ratio of some selected asbestos samples. Atm. X.}'. Acad. P30, 605 - 610. W P P K m i/si/v ll'HCK- j!