Document 37QBm4EnnNZxQ43mo3w7RzJeE

Relation of Particle Dimension to Carcinogenicity in Amphibole Asbestoses and Other Fibrous Minerals1 * * 4 lw* flkk Material May ba Protected by Cmpy law mtk 17 a & Coda) Mead F. Stanton, 3,4 Maxwell Layard, 8,6 Andrew Tegeris,7 Eliza Miller, 3,8 Margaret May, 3-4 Elizabeth Morgan, 7-9 and Alroy Smith 5 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 im 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.--JIMCI 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- to 20-week-old, outbred female Osbome-Mendel rats. In each experiment, 30-50 rats were treated and followed for 2 years, at which time the survivors were killed. All rats were necropsied and all lesions examined histo logically. A positive response was the occurrence of 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 that could cause confusion: the fibrosarcomas of left mam mary glands and the subcutaneous fibrosarcomas in duced by suture material. Vigilance and early surgical removal accounted for most mammary tumors; the use Abbreviations used: alumin aluminum oxide; attapul ~atupulgiie(s); crocid = CTOcidolite(s); dawsondawsonite(s); hallovhalloysite(s); L'1CC = International Union Against Cancer; wollaston* wolJasionitefs). 1 Received November 13, 1980; revised May 6. 1961; accepted June 8. 1981. 1 The guidelines (or the care and use of laboratory animals were followed as set forth by the Committee on Revision of the Guide for Laboratory Animal Facilities; by the Guide for the Care and Use of Laboratory Animal Resources, the National Research Council; and by the National Institutes of Health. ' Laboratory of Pathology, Division of Cancer Biology and Diag nosis, National Cancer Institute (NCI). National Institutes of Health, Public Health Service, U.S. Department of Health and Human Services, Beihesda, Md. 20205. 4 Deceased. * Biometry Branch, Division of Cancer Cause and Pretention, NCI. ` Address reprint requests to Dr. Layard at his present address: Veterans Administration Medical Center, 3801 Miranda Ate.. Palo Alto. Calif. 94304. ' Pharmacopathics Research Laboratories, Inc., 9705 North Wash ington Blvd.. Laurel, Md. 20810. ` Present address: 5524 Trent St., Chevy Chase, Md. 20015. * Present address: Triangle Resource Industries, P.O. Box 599, Laurel, Md. 20707. INFORMATION EXCHANGE CENTER PRICE GILBERT MEMORiAL LIBRARY 905 GEORGIA INSTITUTE OF TECHNOLOGY ATLANTA, GEORGIA 30332-0900 jnci. vol. 67. no 5. November 10002962 WSWIIIilUIi IJUJ 966 Stanton, Layard, Tegerls, el al. of 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 subtractions of the fibrous materials were obtained by ball milling in a steel ball mill and consequently 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 multiple 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 (wollaston 1-4). Seven additional types of particles 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 1, 2), crystals of silicon carbide and potassium titanate (titanate 1, 2), and nickel titanate (titanate 3). All of these materials have been described elsewhere (4, 6, 10, 11, 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 1, 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 non to obtain mutually exclusive size ranges from th same sample (24). The remaining sample, crocid 2, wa obtained from Dr. J. C. Wagner (Medical Researc! Council Pneumoconiosis Unit, Penarth, Wales) a representative of the material used by him in hi original experiments (25). It was our impression tha any mechanical manipulation of these samples coulc both reduce the size of the particles by fragmentatior and effectively increase the size of the particles bi clumping. For this reason, probably the dimensiona measurements on crocidolite are the least representaiivt of all the fibers measured. Glass (glass 1-22).--The first 18 of the 22 glasse* were borosilicate glasses that have been previoush reported and can be recognized from those publication^ by their letter designations (4, 10). Glasses 12, 14, 15 and 18 were preparations of typical large-diameterec insulation glass fibers that were coated with a phenolformaldehyde binder. In the early experiments, glass 1 was used as a control and also served as a vehicle fo the implants. Glasses 19 and 20 were preparations oi large-diametered fibrous glass that was leached tc remove all elements except SiOj. These two glasse were exceptionally fragile and contained many irreg ular fragments. Glasses 21 and 22 were large-diameteret extruded fibers with a microcrystalline aluminum ox ide content greater than 80% (glass 21) and with a microcrystalline zirconium oxide content greater thar 90% (glass 22). Aluminum oxide (alumin 1-8).--The 8 samples of aluminum oxide were all crystalline sapphire whisker 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 exceptionally noteworthy. Sample alumin 8 was nonfibrous, sample alumin 3 was exceptionally fine but tended to cluster in nonfibrous balls, and sample alumin 4 contained whiskers of aluminum nitride a* well as aluminum oxide. Talcs (laic 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 sam ples (crystalline dehydroxy sodium aluminum car bonate [XaAlfOHhCOj] were from several sources. The characteristics and synthesis of dawsonite can be found in (27, 28). Samples dawson 2 and 3 were synthetic crystals prepared by a commercial company (for dawson 2) and by the Bureau of Mines, U.S 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 especialh crystallized and sorted to achieve narrow ranges of size. JNCI. VOL 67. NO 5. NOVEMBER 198 10002963 Carcinogenicity of Fibrous Minerals 96 J (lITte'wtvl >8.0 96% > 4.06.0 C0Thant>2 100% CSSiCwtMte 1WH 152 1.35 ) >254 0 ->152.5 152 1.65 1.31 225 2.04 1.95 1 >50-15 3.75 4.11 3.70 3.40 252 3.08 3.03 324 155 >2550 3.57 4.41 4.79 457 170 4.18 452 4.18 329 389 420 440 >.10.25 3JP 60S 654 6.17 388 4.99 6.07 448 0.81 4.01 4.19 4.56 >55.10 327 5.02 502 4.44 425 451 4.78 451 421 441 4.74 >.01-56 3.40 3.76 4.48 4.54 4.70 44) Dvwwn5 >4 08.0 >254.0 >1525 > J5015 >2550 324 >.102S >55.10 >01.06 218 423 283 >8.0 157 157 42S 173 202 250 4.53 4.93 SI TnmoTKS 1 3.84 4.10 420 3.69 3.14 4.47 425 4.14 3.54 114 392 214 244 3.14 3.14 352 214 3.44 114 16) Trwrofrmz 264 3 01 3.14 2B4 254 405 368 3.77 3.44 386 338 331 4,09 4.04 328 264 114 2B4 17) Onvaon 1 95% >4 08.0 >254.0 >1.525 1.44 >50-1.5 1.92 152 >2550 1.75 3.16 > .1025 262 270 >05.10 340 362 >01.05 257 3.57 >80 1.44 222 132 3.90 362 214 300 1.75 368 42S 466 262 BICrodd.1 SN% 4.07 4.07 5.75 569 467 6.52 562 467 587 4.67 6) Crood. 2 SM% 3.53 4.53 427 4.30 4.56 422 4.83 4.70 422 5.15 S.U 427 5.71 4.96 4 00 4.57 3.83 oa HO) Crocid. 3 >80 >4080 >254.0 >1.52.5 >50-15 4.10 4.10 >25.50 4.10 . .4 40 >.102S 6.12 5.71 455 SOI >05.10 652 5.B4 4.10 >.1006 5.62 4.58 4.10 (ll) A/nottht 94% (12) Ocod. 4 213 243 213 213 3.17 3.36 3.54 261 313 364 3.56 272 353 3.76 321 3.47 258 298 303 261 276 324 384 306 424 3.61 4.36 324 423 4.66 4.92 2.76 4.31 4.53 4.70 324 236 354 306 113) Gtas* 11MOU >4.08.0 >254.0 >15-25 >50-15 223 >25.50 306 >.1025 293 >05.10 >01.05 >8.0 2.53 3.35 293 346 323 308 495 453 4.79 466 (14) Oood. 6 359 4.35 484 4.74 442 281 4.19 425 465 463 426 781 2.81 329 342 366 377 342 281 261 377 2 61 359 329 (IS) Qm 2 (M6D) //% 291 321 3.38 2.91 369 361 408 369 351 3.51 321 399 3 61 4.11 4 02 381 3.38 (161 Glut 3 (KU >4 08 0 >2540 >15-25 >5015 1.45 295 >2550 256 3.16 >.1025 303 316 >.05.10 285 409 >01-05 303 373 pm >01-1 >14 >80 067 067 240 333 3.76 303 >48 0 67 1.52 057 203 219 3 42 274 363 303 325 303 >064 >84 (17) Gian 4 (M6L) 281 244 244 244 >J0M 2.05 244 3.52 344 344 >M 154 259 316 344 356 335 >44 1.53 123 276 213 3.70 290 3.39 262 370 3.14 >M4 >64 (IS) AJumin 1 120 312 226 >01-1 >14 1.10 152 1.61 1.93 080 257 1.73 >4-8 120 1 80 0 SO 1.90 1 43 217 1 68 2.78 221 242 3 51 273 263 >664 >64 Length TiXT-ncrHt 1.--Fiber distribution In common log of the number of panicles per miaogram in each of 34 dimensional caie^ories. JNQ. VOL. 6? NO i. NOVLMBLR IVMi 10002964 968 Stanton, Layard, Tegeris, at al, (19) Claaa 6 OCW) 69* > 4M0 > IM.0 15-25 0.66 >50-15 2.48 1.80 >*JO 1.06 067 >.10-25 >96-.10 >91-95 >8.0 0J7 129 1.70 1.67 1.67 0.06 0.06 0.36 0.76 0.06 1.08 1.41 198 2.42 251 2.75 271 2.51 221 GOI D mav 7 68% 1.77 3.11 272 207 3.01 4.18 4.40 3.70 272 326 4.10 4.64 3.42 1.77 3.33 3.81 425 4.70 272 (21) Dawson. 4 06* 211 3.38 620 5.03 127 222 211 342 4.05 6.00 3.47 1.56 157 222 324 442 4.12 3.99 1.67 206 ' 151 224 248 280 4.01 -- (22) (Wien. 3 go19 >4.58.0 >254.0 > 15-25 >50-15 263 >2550 4.18 4.50 >.10-25 3.91 4.74 >05.10 58S 5.18 >.01-.06 5.90 6.17 >89 307 428 4.41 4.88 428 199 229 323 497 521 529 5.14 B31 Gw 8 (M6WI 0*1* CM) Craod. 1.10 1.68 1.10 1.09 155 1.10 1.10 218 274 201 228 3.17 321 263 284 3.47 456 422 402 1.40 284 3.11 254 4.17 5.07 4.71 4.45 1.70 322 325 3.69 5.10 5.45 4.86 457 210 356 3.47 276 4.74 490 421 247 271 295 271 280 3.66 4.07 (25) Otxid. 7 M* >4.08.0 >254.0 >1525 >50-15 >25.50 1.69 >.1025 205 k >.05.10 I >01-06 Om (28) Alumin. 3 41% 1.86 > 4.08.0 >254.0 1.09 > 152.5 1.59 >5015 236 242 >25.50 0.97 >.1025 0.97 232 >.05.10 127 >.01-.06 >90 0.91 1.69 191 1.91 9.91 091 121 1.51 196 225 202 229 234 251 205 091 >20 1.72 1.19 1.75 0.19 157 192 1.70 1.70 - 233 261 0.67 1.15 1.49 219 241 0.97 1.59 (26) Crteid. 8 D4* 4.45 508 5.61 4.56 523 3.75 4.75 486 5.09 4.59 4.15 (29) Creod. 9 4J* 345 3.45 3.75 275 3.13 3.13 3.13 313 343 3.13 451 425 3.43 524 4.92 425 409 548 5.14 424 174 07) Aturnin. 2 237 1.99 197 1.48 204 1.11 120 0.91 1.15 1.18 190 1.70 150 120 120 (30) WoRanen 1 31 3.15 3.15 328 3.41 211 211 276 325 322 328 2.81 306 215 211 241 122 0.37 1.32 037 1.53 i.m 1.61 1.18 227 1.83 302 289 156 120 211 3.19 211 259 OD Ahmn. 4 > 4.0-8.0 >264.0 215 >'1.62.5 2.80 >.50-15 292 2.83 >26.50 3.57 3.01 >.10-25 229 >96.10 >91-95 1.99 >80 198 248 230 245 3.19 159 229 2.69 1.75 158 215 237 128 275 0.98 2.83 229 229 02) Ooci<L 10 37% 3.79 4.69 603 5.13 4.96 3.10 210 3.57 4.57 501 4.59 446 340 357 179 110 157 3.10 110 03} Akimin S 441* 1.48 266 3.59 277 3.55 424 4.13 402 300 3.70 1.78 218 271 341 422 424 192 252 206 208 2.59 326 1 43 3.92 3 57 123 Text-figure I (continued). (35) Gbas 7 IKCP1 41 * 300 324 324 250 720 >91-1 1.44 2.05 359 388 328 290 298 >14 151 205 217 317 210 210 267 >44 0.97 291 151 1.81 231 0.97 286 201 256 250 >864 >64 061 WbHasien. 3 19% 3.12 332 318 342 258 358 4.09 3.32 3.32 348 258 258 3.12 346 228 228 258 276 >91-1 >14 >46 >864 >64. lanptft Fiber distribution by common log of the number of particles per microgram in each of 34 dimensional categories. JNCI. VOL 67. NO 5, NOVEMBER 1981 10002965 Carcinogenicity of Fibrous Minerals 9( y-l 20% >4.08.0 > 25-4.0 > 15-2.5 > 8.0 6.47 607 626 655 656 621 455 6.16 (401 Oocid. 11 19* >4.08.0 >254.0 238 >15-25 203 >5015 3.15 296 >2550 3.06 150 >.1025 297 >55.10 202 > JD1-.06 >8.0 1.72 242 1.72 150 sin. 6 >4.08.0 >254.0 0.48 > 1.525 082 0.75 122 210 0.82 >80 0.12 0.67 0.12 022 > 95.10 O3E > 91-95 (46) Wollaston. 2 12% >408.0 >25-4.0 278 > 1.525 278 >5015 3.32 380 >2550 343 3.32 >.1025 3.96 326 >55.10 4.37 4.00 > 51-.05 >80 27B 3.16 3.1B 296 248 (38) Hsboy. 2 215 250 1.42 656 6.16 653 4.61 4.51 6 47 6.14 689 (41)GUsa19(RD) 13% 256 267 156 216 292 237 4.51 4.51 4.81 226 1.73 126 243 126 057 022 122 1.44 0.67 0.70 1.54 1.49 057 0.63 1.64 0.82 (44) Demon. 6 13% 6.17 656 7.86 6.47 6.86 6.78 240 (47) Crod. 12 10% 287 3.31 4.00 400 3.67 227 3.31 3.73 421 4 00 3.50 227 3.56 3.39 353 3.45 3.17 273 257 - 2-27 3.39 3.05 3.69 3.57 3.17 229 091 Ctaaa 8 0CUP1 191b 288 3.36 3.36 3.17 211 273 353 3.54 3.50 352 231 126 252 271 3.54 284 261 263 M2) Glass 9 (M8U 14% 1.49 260 15S 0.77 1.12 1.45 (45) Dawson. 2 12% 3.80 4.57 492 422 4.39 294 3.43 3.46 3.31 3.10 (48) Atapti. 2 11% 244 282 274 244 l.>4 6.31 6.16 7.05 6.14 6.54 622 (39 6.51 2.04 1.74 251 126 2.44 3.19 1.96 257 231 280 251 0.17 1.73 202 211 227 235 242 238 155 269 244 1.74 (491 Glass 10 IMOS % >4.08.0 >254.0 297 > 15-2.5 3.43 >5015 388 3.91 >2550 4.34 4.02 >.102S 443 388 >55.10 550 4.19 >51-56 677 4.63 >8.0 246 299 237 276 369 272 1.76 1.17 246 3.37 (SO) Glass 11 0C2P) % 4.12 428 4.11 351 289 3.64 4.53 302 289 3.09 3.56 3.60 1.16 2.41 266 1.81 (511 Tfcsnsla 3 8% 308 1.78 "33T 2 76 3.10 3.54 295 3.35 (52) Anapul. 1 0% >4 08.0 >254.0 >1.52.5 >5015 5.12 >2550 548 >.1025 5.7S 650 > 55.10 662 665 > .01-.05 696 665 sun >91-1 >1-4 >8.0 4.64 454 >48 >564 >64 B3)Tslc 1 7% 4.00 4.15 4 19 3.70 364 4.02 456 3.97 3 91 327 3.09 3.70 327 279 >3)1-1 >1-4 >4-8 279 309 >64 (54) Clan 12 I02P) 7% 232 1.78 2.17 043 0.4S 2.06 248 2.09 043 249 1.95 1.94 043 283 2.83 1.91 1.84 043 2.44 0.48 208 1.65 241 > JD1-1 >1-4 >4-8 > 6-64 >64 length TtXT nci RL 1 (continued).--Fiber distribution by rommon log of ihr number of particles per miaogram in each of 3( dimensional categories JNC1. VOL 67. MO. 5. NOVFMBF.R 1981 970 Stanton, Layard, Tegeris, et al, Dlamatar 66IGW13 0CFP) > 4.04.0 >254.0 289 >15-25 220 > 50-15 219 248 >250 296 122 >.10-2 230 >2.10 227 > .01-.06 >80 267 273 152 152 203 203 052 - 052 (561 GW 14 |P2P) 256 121 236 156 230 228 251 129 081 155 208 051 121 200 221 1.70 069 081 67) GW 15(V2n 1.08 6% 1.08 154 -212 258 1.74 3.42 209 1.60 212 1.60 156 224 . 2)6 1.64 1.60 1JC 681 AJumn. 7 6% > 4.04.0 > 254.0 252 >15-25 0.02 >50-15 > 25-50 0.64 TTT 1.17 ~nr >.1025 > JS-.10 > .01-.05 >ao 1.19 TsT T19 T.70 Toe" TzT Tm 0.31 Ta TaT TTs 0-31 TST 0.51 69) GW 16 IM8S) 241 363 381 28S 221 217 292 248 2.45 275 243 3.46 214 290 259 281 160) Tiic 9 137 281 Q4 4.13 4.85 3.76 3 4fi 3.81 3.81 328 241 281 1611 Talc 4% >50-1.5 > 25-50 >.10-2 > .06-.10 >01.05 > 4.04.0 254.0 258 1.525 3.06 228 326 >8.0 276 321 2.45 245 268 256 228 1.96 162) Talc 4 406 340 438 262 4.19 4.81 4.58. 422 332 332 3.32 3.79 3 02 163) Alumin. 8 272 286 271 222 242 287 2.46 212 159 225 234 1.70 1.72 1.86 0.74 164) Gte 21 (SI) >8 0 e* > 4 04.0 1.15 1.15 1.46 165) GW 22 IS2) 1.47 1.17 1.17 166) GW 17 (M6S > 254.0 1.15 1.85 2.36 200 1.47 1.77 2.07 -1.87 > 1.5-2.5 1.15 1.63 1.46 1.47 1.17 225 1.77 245 2.85 > .50-1.5 1.46 2.06 1.87 1.47 4.43 4.83 438 379 >2.50 1.15 1.46 1.15 6.17 4.88 1.87 2.15 >.10-2 6.60 4.73 358 > .05-.10 545 438 > -01.06 6.16 167) GW 18 (VW> 0* > 4.04 0 > 2.54.0 >15-25 > JO-15 >.2.50 >.102 > .05.10 > .01.06 >8.0 083 092 080 1.00 1.10 0.34 0.40 0.30 Oil 0.41 168) OockL 13 4.30 468 450 468 431 258 4.00 446 4.00 263 2.68 2.98 298 316 268 (69) Wollaston. 4 V7 1.43 0.95 1.43 125 2.07 156 055 1.86 0.95 1.99 0.95 095 1.43 1.56 1.80 151 09 12 rroj T*'c 5 >80 (71) Talc 6 (72) Talc 7 >4.08.0 >254.0 343 > 1.625 4.33 >.501.5 4.62 498 >25.50 4.56 430 >.10-25 427 390 > .05.10 397 360 X01-.05 390 3.13 pm >51-1 > 14 >48 >064 >64 463 483 468 465 4.10 >J01-1 325 3.95 4 83 443 3.95 385 325 >14 325 325 356 355 325 325 355 356 325 > 8-64 >64 4.95 533 518 5.14 482 >.01-1 4.37 509 4.93 4.37 4.63 >1-4 367 367 367 367 >44 >8-64 >6* length Tiat hc.i ri 1 (continued) -- Fiber distribution by common log of the number of particles per microgram in each of SH dimensional categorit JNCI. VOL 67. NO 5. NOVEMBER 1981 10002967 Carcinogenicity of Fibrous Minerals Expt No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 TABLE 1.--Summery of 72 experiments with different fibrous materials Compound Actual ... tumor incidence Percent tumor probability SD Common log fibers/jig, <0.25 fim x >8 jim Expt No. Compound Actual _tumor incidence Percent tumor probability SD Titanate 1 Titanate 2 Si carbide Dawson 5 Tremolite 1 Tremolite 2 Dawson 1 Crocid 1 Crocid 2 Crocid 3 Amosite Crocid 4 Glass 1 Crocid 5 Glass 2 Glass 3 Glass 4 Alumin 1 Glass 5 Dawson 7 Dawson 4 Dawson 3 Glass 6 Crocid 6 Crocid 7 Crocid 8 Alumin 2 Alumin 3 Crocid 9 Wollaston 1 Alumin 4 Crocid 10 Alumin 5 Glass 20 Glass 7 Wollaston 3 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 4/25 5/28 3/21 954.7 100 100 100 100 100 9514.8 9416.0 9316.5 9316.9 9317.1 8619.0 85113.2 78110.8 77116.6 7418.5 7119.1 70110.2 6919.6 6819.8 66112.2 66H3.4 64117.7 63113.9 56111.7 53112.9 44111.7 41110.5 3319.8 31112.5 28112.0 37113.5 2219.8 22110.0 2118.7 19110.5 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 2.47 4.25 0 2.60 3.09 3.73 0 2.50 0 37 Halloy 1 ' 4/25 38 Halloy 2 5/28 39 Glass 8 3/26 40 Crocid 11 4/29 41 Glass 19 2/28 42 Glass 9 2/28 43 Alumin 6 2/28 44 Dawson 6 3/30 45 Dawson 2 2/27 46 Wollaston 2 2/25 47 Crocid 12 2/27 48 Attapul 2 2/29 49 Glass 10 2/27 50 Glass 11 . 1/27 51 Titanate 3 1/28 52 Attapul 1 2/29 53 Talc 1 1/26 54 Glass 12 1/25 55 Glass 13 1/27 56 Glass 14 1/25 57 Glass 15 1/24 58 Alumin 7 1/25 59 Glass 16 1/29 60 TalcS 1/29 61 Talc 2 1/30 62 Talc 4 1/29 63 Alumin 8 1/28 64 Glass 21 2/47 65 Glass 22 1/45 66 Glass 17 0/28 67 Glass 18 0/115 68 Crocid 13" 0/29 7 69 Wollaston 4 0/24 70 Talc 5 0/30 71 Talc 6 0/30 72 Talc 7 0/29 20+9.0 2319.3 19110.3 198.6 15+9.0 14+9.4 13+8.8 13+6.9 12+7.9 12+8.0 107.0 11+7.5 815.6 8+5.5 818.0 815.3 716.9 715.4 615.7 615.5 615.9 515.1 514.4 414.3 413.8 514.9 313.4 614.4 212.3 0 0 0' 0 0 0 0 Common log fibers/jig, SO.25 jim x >8 jim 0 0 3.01 0 0 1.84 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. Wollastonite (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 sire by the designation A, B, D, and F. it was apparent at low-power magnification that only grade F was completely fibrous and that these fibers were relatively large. Tremolite (tremolite 1, 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 sire for carcinogenesis. Compari son of these fibers indicated that they were distinctly smaller in diameter 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 U1CC standard reference samples. No efforts were made to alter this as received, and descriptions of this sample as published should apply (19, 21, 22). Attapulgite (attapul 1-2).--Of the natural fibers, the clay attapulgite was of particular interest because of its use in many household items that generate respirable dust. Two different samples of this complex hydrated magnesium silicate were obtained from sources in Attapulgus, Decatur County, Georgia. Both samples were considerably refined, and by electron microscops they were seen to be composed entirely of short fibers of consistently small diameter (30). These refined clays were considered by the U.S. Bureau of Mines to be 90% or greater in purity, with the remaining 10% being quartz. Halloysite (halloy 1-2).--Halloysite is a natural fi brous hydrated aluminum silicate, which is respirable and of minute size. The 2 samples were obtained from Dr. Waller Parham, who recovered them from the raw water supply of Hong Kong. On examination these samples were seen to have a tendency for clumping in water. In an effort to disperse the minute fibers, the second sample was sonicated and treated with sodium JNCI. VOL. 67. NO S. NOVEMBER l<Mtl i0002968 972 Stanton, Layard, Tegeris, et al. hexametaphosphate. Clumping persisted in this second sample, 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 oclalilanate (titanate 1-3).--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 these 2 samples was nonfibrous, finely ground nickel titanate (titanate 3). The 72 experiments represent all 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-covered, 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 1-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 fim, 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 34 dimensional ranges as indicated in text-figure 1, and the number of particles per microgram in each category was calculated. Dupli 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. 1). 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- Table 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 Untreated* U113 0/15 0/26 0/68 0/26 0/98 1/66 0/27 0/27 1/22 a'488 06 Noncar cinogenic pulmonary implants* 0/49 2/26 4/50 1/70 1/72 1/162 0/3 9/432 2.1 Noncar cinogenic pleural implants* 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. JN'CI. VOL 67. NO 5. NOVEMBER 1941 10002949 Carcinogenicity of Fibrous Minerals 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 1, 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/(l-p)], where Jn 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 panicles equal to or less than 0.25 /im in diameter and greater than 8 pm in length. There was no correlation with panicles equal to or less than 4 pm 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 nitb common logarithm of number of particles per microgram tn different dimensional ranges Fiber diameter >4 >1.5-4 >0.25-1.5 0.23 4 _ -0.45 0.01 0.20 Fiber length, um >4-8 -0.28 -0.24 0.45 0.63 >8 -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 +bi x> +......... + 5* Xk, where X|................x* represent the common logs of numbers of the particles per microgram in the size categories of table 3, and a, hi,......... 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: ln[p'(l-p)]-2.62 + 0.9S05x. (0.24) (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 which particles are carcinogenic and outside of which they are not is that the probability of tumor falls as particle diameter increases and length decreases. Of the 72 experiments, 7 had tumor incidences that deviated markedly from those predicted by the esti mated regression line. These were: experiments 5 (tremolite 1), 6 (tremolite 2), 26 (crocid 8), 29 (crocid 9), 33 (alumin 5), 47 (crocid 12), and 71 (talc 6) (see table 1 and text-fig. 2). For the first 3 of these experiments the observed responses were higher than the predicted responses, but the high responses can in part be explained by the fact that there were substantial numbers of fibers in size categories adjacent to the category used in the regression equation. For the remaining 4 experiments, the observed response was substantially lower than the expected response; al though no apparent explanation existed for these deviations, they were possibly due to inaccuracies in the assessment of functional particle size. In prepara tions of amphibole asbestoses (which included the crocidolites and tremolites), we observed that both JN'CI, VOL 67. NO 5. NOVF.MBLR IMHI 10002970 974 Stanton, Layard, Tegerls, et al. C * crocidolito 1.0 r G * gisss i 0 * dawsoniie 0.9 r L " ilurninum oxide S silicon carbide 0.8 f- A " attapulgite GC O 5 Dh- ! 07 r P tiianate T talc M " tremolite W wollastonite LL 0.6 !- H hallovsite o 0 amojite 0.5 MM 0 p D* o C y- o c a o< 0.4I oa a. 0.3 ifw 0.2 ' HG `hCW c , A09W 0.1 *AGG*T t !LTGGGG I ITTGG G | CWTTCC t 0.0 0.0 0.5 t- > 1.0 1.5 2.0 J- - 1 _T J---------- 1 2.5 3.0 3.5 4.0 _i_____ i 4.5 5.0 5.5 6.0 LOG NUMBER PARTICLES MEASURING < 0.25 x > 8 nm PER MfCROGRAM TtxTHOlRE 2.--Regression curve relating probability of tumor to logarithm of number of panicles per pg with diameter <0.25 pm am 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 /zrnX>8 pm, 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 clear 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 preneoplastii pleural lesions show avid phagocytosis of both shot fibers and large-diameter fibers but negligible phago cytosis of longj, thin fibers. Consequently, in thesi experiments we may simply be measuring the ef ficiency of phagocytosis. Doubtless, we have little rea. knowledge of the way that long, thin fibers can caus< cancer, but as Rous (31) once said, "Since what wt think largely determines what we do, it is well that w< think something." In the spirit of this quote, it might be profitable to consider potential mechanisms ol cancer production by long, thin fibers. Of first impor tance are those hypotheses in which the progenitor o the cancer cell is not directly affected by the fiber. Tht long latent period would suggest that a generalize! alteration either in local milieu or systemic environ ment might be at fault. In this regard, the abundan: collagen in the preneoplastic pleural scars should lx noted. Consideration of a relationship between thi phenomenon and "solid-state" carcinogenesis is attrac live, 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 intc account the fact that both short fibers and thick fiber: are less carcinogenic than fine, long fibers. Since dost was fixed in weight, but was different in dimension for all experiments, one might consider the surface area a: a possible factor. If this were the case then fibers frotr the same pool that were modified only by shortening should be equal in tumor-producing capacity. Clearh this is not true in the following experiments: 13 (glass 1, MOL) vs. experiment 49 [glass 10, MOS see (4, 10)}. J.NCI. VOL 67. NO 5, NOVEMBER I9(T 10002971 Carcinogenicity of Fibrous Minerals and in experimeni 23 (crocid 6) and experiment 25 (crocid 7) vs. experiment 40 (crocid 11), experiment 47 (crocid 12), and experiment .68 (crocid 15). 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 mycelia 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 mycelia that might be critical. REFERENCES (/) Pott F, Hith F, Friidrichs KH. Tumorigenic effect of fibrous dusts in experimental animals. Environ Health Perspect 1974; 9:313-315. (2) Pott F, Friidrichs 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 Bakieriol [B] 1976; 162:467-505. (J) Pott F, Friidrichs KH. Tumoren der Ratten nach i. p. Injec tion faser forminger Suube. Naturwissenschaftrn 1972: 59:318. <#) Stanton MF. Layard M, Ticiris A. Millir E, May M. Kent E. Carcinogenicity of fibrous glass: Pleural response in the rat in relation to fiber dimension. J Nail Cancer Inst 1977; 58: 587-603. (5) SMITH WE. Miller L, Elasser RE. Tests for carcinogenicity of asbestos. Ann NY Acad Sci 1965;' 132:456-488. (tS) Stanton MF. Wrinch C. Mechanisms of mesothrlioma induc tion with 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. (8) Wagner JC, Berry C, Timbrell V. Mesotheliomata in rats after inoculation with asbestos and other materials. Br J Cancer 1973; 28:173-185. (9) Stanton MF, Layard MW. Carcinogenicity of natural and man-made fibers. In: Margison CP, ed. Carcinogenesis. Ad vances in medical oncology, research and education, Vol. I. Oxford and New York: Pergamon Press, 1979:181-187. (10) ----. 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). ill) Stanton MF. Some etiological considerations of fibre carcino genesis. In: Bogovski P. Timbrel! Y. Gilson J, et a)., eds. Biological effects of asbestos. Lyon. France: WHO. 1973: 289-294 (1ARC publication No. 8). 02) Chvrg J, Rosen SH. Mooi.ten S. Histological characteristics of mesothelioma associated with asbestos. Ann NY Acad Sci 1965; 132:614-622. <fJ) Armitace P. Statistical methods in medical rrsearch. New York: Wiley, 1971:576-377, 410-414. Ol) Pilgrim HI, Dowd JE. Correcting for extraneous death in the evolution of morbidity or mortal its from tumors. Cancer Res 1963; 23:45-48. Ilf) Olson RH. Introduction. In: Lrfond SS. t-d. Industrial minerals and rocks. 4th ed New York: American Institute of Mining, Metallurgical and Petroleum Engineers, Inc.. 1975. <J8) Brick DW. Synthetic reolites: Properties and application. In: Lefond SS. ed. Industrial minerals and rocks. 4th ed. New. York: American Institute of Mining, Metallurgical and Pe troleum Engineers, inc., 1975. (77) Sheppard RA. Sedimentary rocks. In: Lrfond SS, ed. Industrial minerals and rocks. 4th ed. New York: American Institute of Mining, Metallurgical and Petroleum Engineers, Inc., 1975. (IS) Mlmpton FA. Commercial utilisation of natural zeolites. In: Lrfond SS, ed. Industrial minerals and rocks, 4th ed. New York: American Institute of Mining, Metallurgical and Pe troleum Engineers, Inc.. 1975. (19) Rendall RE. The data sheets on the chemical and physical properties of the UICC standard reference samples, in: Sha- pito 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.. July 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 Against Cancer standard reference samples of asbestos. In: Shapiro HA, ed. Pneumoconiosis: Proceedings of the international 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. Ini J Cancer 1968; 3:406-408. (23) Timbrell V. Physical factors as etiological mechanisms. In: Bogovoski P, Timbrell V. Gilson JC, et al., eds. Biological effects of asbestos. Lyon, France: WHO, 1973:295-303 (IARC publication No. 8). (24) Speil S. Leineweber J. Personal experiences with making sam ples of fibers for biological experiments. In: Pelnar PY. ed. Fibres for biological experiments. Montreal: Institute of Oc cupational and Environmental Health, 1974:45-50. (23) Wagner JC.-The pathogenesis of tumors following the intra pleural injection of asbestos and silica. In: Neuesheim P, Hanna MG. Deathergate JW, eds. Morphology of experimental respiratory carcinogenesis. Oak Ridge, Tennessee: Oak Ridge National Laboratory, 1970:347-558 (Atomic Energy Commis sion symposium series No. 21). (26) Harington JS. Allison AC, Badami DV. Mineral fibers: Chem ical, physiochemical and biological properties. Adv Pharmacol Chemother 1975; 12:291-402. (27) Jackson J, Huggins CW, Amman SG. Synthesis and characteri zation of dawsonite. Washington, D.C.: U.S. Dept of Interior. 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 trrmolite talc on hamsters. In: Proceedings of the symposium on ulc, Washington, D.C., May 8. 1973. Washington, D C.: U.S. Bureau of Mines 1974:43-48 (Bureau of Mines information circular 8639). (30) Higcins CW, Denny MV, Shell HR. Properties ol pohgorskite, an asbestiform mineral. Washington, D.C.: U.S. Dept of Interior, 1962 (Bureau of Mines information circular R1607I). (31) Rous P. The virus tumors and the tumor problem. In: The Hawey lectures series. No. 31. Baltimore: Williams Sc Wilkins. 1935-1936:74-115. (32) Brand KG. "Solid-state" or "foreign-body" carcinogenesis. In Symington T. Carter RL, eds. Scientific foundations ol oncology. London: William Heinemann Medical Books. 1976 490-495. (33) Wade MJ, Lipkin LE, Stanton MF, Franks AL. In vitro ryiotovicily assay as applied to asbestos and other minerals Its possible relevance to carcinogenicity. In: International work shop on the in viiro effects of mineral dusts. Medical Research Council Pneumoconiosis Unit. Penarth, Wales, Sept 4-7. 1979. Penarth, Wales: Medical Rrsearch Council, 1975 (PS88DI). JNCI, VOL 67. NO 5. NOVKMHFR IVXI i0002972