Document zoGyDXMbvz39ppkEqORLEz0jR

FILE NAME: Oil Industry and American Petroleum Institute (API) DATE: 1956 DOC#: API131 DOCUMENT DESCRIPTION: Unpublished Conference Presentation Relationships Between the Composition and Carcinogenic Properties of Various Refinery Streams - Kettering Lab > * RELATIONSHIPS BETWEEN THE COMPOSITION AND CARCINOGENIC PROPERTIES OP VARIOUS REFINERY STREAK bj At Wtoaley Horton, Ruaaell Tye, and Mary Jane (S*af THE KETTERING LABORATORY in the Department of Preventive Medicine and Industrial Health College of Medicine U N IV ER SITY O F C IN C IN N A T I, C IN C IN N A T I, O H IO ! 1 RELATIONSHIPS BETWEEN THE COMPOSITION AND CARCINOGENIC PROPERTIES OF VARIOUS REFINERY STREAMS by A. Wesley Horton, Russell Tye, and Mary Jane Graf Kettering Laboratory University of Cincinnati to be presented at the Symposium on Polycyclic Hydrocarbons National Meeting Of American Chemical Society Atlantic City, New Jersey September, 1956 API 06533 0 RELATIONSHIPS BETWEEN THE COMPOSITION AND CARCINOGENIC PROPERTIES OF VARIOUS REFINERY STREAMS by A, Wesley Horton, Russell Tye and Mary Jane Graf Kettering Laboratory University of Cincinnati For a long tine, cancer of the skin and particularly that of the scrotum, has been known to occur with relatively high frequency in association with certain occupations (1), The reported instances in which this problem has developed in the American petroleum industry have been rare (2), The majority resulted from exposure to un cracked waxy distillates during their filtration in hydraulic presses. Historically, most of our information on occupational cancers of the skin has derived from Great Britain and Germany. Chemical research in these countries on active fractions of coal tar centerec interest on polycyclic aromatic hydrocarbons. In parallel investiga tions, organic synthesis made available a large number of 3 -6 ringed compounds, many of which proved to be carcinogenic for the skin of mice (3). Isolation of one of these, the 5-ringed benzo[a]pyrene from coal tar appeared to round out the picture (4). It was shown that polycyclic aromatic hydrocarbons were indeed present in each of the various industrial materials which had been associated with a significant increase in the incidence of cancer of the skin of the exposed workers(e*g o ,5)o Fractions of these materials free of the ! I - 2 - 3-6 ringed aromatic compounds, when applied upon the skin of mice, produced no tumors (6 ). Because of the demanas of the Air Force for high-octane fuels in World War II, the practice of catalytic cracking grew rapidly in American oil refineries to large proportions. Individual research programs of certain of the oil companies showed that the concentra tion of polycyclic aromatics was markedly higher in heavy catalytic gas oils than in any stocks previously hanalec in the refinery. Preliminary tests on the skin of mice confirmed the suspicion that the relative potencies of catalytically crackea residua might be much higher than that of straight run products or thermal tars. a s a result the industry initiated a PI Research Project MC-1 at The Kettering Laboratory to investigate fully any potential hazards of cancer which might result from contact with these or other inter mediates or products of petroleum refineries. The major objectives of the experimental phases of this project are, 1 . to aetermine the relative carcinogenic potencies of these oils and tars for the skir. of mice, and 2 . to ae\op chemical or physical procedures for estimating the relative carcinogenic activ ity of such intermediates and products. Other'goals, such as the development of methods for removing oils from the skin after contact, will not be referred to again in this paper. Technique of Bioassav The biological testing involved the repeated application of the oil in question upon the skin of mice under selected conaitions API 06535 s - 3 with respect to dosage and frequency of application (7). Under suit able conditions the average period of exposure required to induce visible tumors furnished a satisfactory means of comparing the poten- cies of different oils. Examples of the average duration of tumor- free life in experiments of this type on selected refinery streams are shown in Tables 1 and 2 . For convenience in thinking of the biological result in terms of the concentration of a standard carcino gen in a standard medium required to produce it, the corresponding values of the relative potency, Pj/jq, are shown. The potency, , is aefined as that concentration of 3 -methylcholanthrene in benzene which will produce tumors a.t the same rate as the oil in question under comparable experimental conditions. The value of P-^q for a given oil is related to the average expectancy of tumor-free life, e, by the following equation: where f is equal to the number of applications each week. Composition vs Relative Carcinogenic Potency of Complex Oils The initial approach to this problem, for reasons apparent from tne introduction, was to examine the polycyclic fractions of various carcinogenic oils. It was soon apparent that there were numerous exceptions to any attempted correlations of relative biological activ ity with either total content of polycyclic aromatics or with the physical properties of selected fractions thereof. Parallel experi ments involving solutions of pure carcinogens in solvents of lower volatility than benzene suggested a reason for this lack of correlation. A p I 06536 0 4 -u - By the use of n-dodecane, dodecylbenzene and other solvents of simi lar molecular length, the rate of tumor production was accelerated to a level as high as that displayed by much more concentrated solu tions in benzene (8 ). The attempt was made therefore to determine the amounts and types of both the carcinogens and the accelerators in the oils, and to measure their relative contributions to the car cinogenic potency of various refinery streams. The following equation is being used as an expression of a tenta tive hypothesis to orient the collection of pertinent information: Relative .potency, Pj^q = M a (C]_ + C2 + ), where Ma is a multiplication factor due to the contribution of accel erators. This factor is not only a function of the concentration and activity of the accelerators but also a function of the concen tration of the carcinogens and of the quantity of oil applied upon the skin (dee Table 3). Obviously, the concentration of carcino gen may reach such a high level that maximal activity obtains with out acceleration (average time of appearance of tumors approximately 6 weeks in C3H mice). The concentrations, C]_, C2 , and C3 , of the important types of carcinogens must be expressed in terms of methylcholanthrene equivalents. The validity of the concept of the summation of the potencies of the different classes of carcinogens, so long as the maximum activity for any given class has not been reached, is supported by a certain amount of experimental evidence such as that shown in Table 4. In previous experiments benzoCaJpyrene has been found to API 06537 I - 5 - be about 60 per cent as active as methylcholanthrene. Comparison 0f the results of Experiments 623 and 624 indicates that the poten cies of these two strong carcinogens summate satisfactorily. BenzCaJanthracene (BA) is a very weak carcinogen which shows significant potency only when tested in a powerful accelerator. There have been reports in the literature that this polycyclic hydro carbon inhibits the potency of certain' strong carcinogens,~ par ticularly 7 ,12-diraethylbenz[a]anthracene (DMBA),(e.g.,9) However, in Experiments 625 and 626 it was found that BA does not influence the potency of methylcholanthrene to C3H mice significantly. Similar indications were obtained by extracting a distillate, Sample 3-12, rich in BA, with maleic anhydride and testing the non-adduct. Some carcinogens are relatively susceptible to oxidation, DMBA being particularly reactive from this standpoint. It may be that this compound's extremely high level of potency is related to this chemical property. By inhibiting the oxidation under certain cir cumstances BA might be able to reduce the potency of DMBA to some basic level. Some of the available evidence on straight run products has been interpreted as suggestive of the presence of inhibitors in the high boiling fractions of crude oils. Thus, although most raw lubri cating stocks which we have tested have shown some capacity to produce tumors of the skin of mice, the crude oils and reduced crudes such as those in Table 1, rarely showed any activity. The possi bility that some of the high boiling non-hydrocarbons might act as inhibitors is intriguing, but no direct evidence has yet been API 06538 / > - 6- eveloped on this point. The lack of potency of the reduced crudes, even though they contain carcinogens, might be explained partly on the basis of dilution of the carcinogens, ana partly on the basis of the fact that accelerators, which are so important to the potency of lube distillates, are usually present in reduced crudes in concen trations too low to be effective. Composition of Pertinent Polycyclic Fractions Research directed toward the characterisation of the important classes of polycyclic carcinogens has been carried out primarily on catelytically cracked residua. Although some of these cracked oils are not much more potent on the skin of mice than certain straight run distillates, the more active samples have produced tumors in an average time of less than two months. It was expected that the cracking operation would greatly reduce the complexity of the aromatic fraction of these oils as compared to uncracked materials, ana further that the concentration of carcinogens might be significantly higher. Analyses for one specific carcinogenr :benzopyrene (BP), by a method developed in the course of this investigation (1 0 ), confirmed this expectation. As shown in Table 1, the content of this compound in uncracked streams is usually insignificant, although the potencies, in some cases, are moaerately high. In the case of tars from thermal cracking of virgin feeds, the concentration of BP reaches an average level of 0.05 per cent. In residua iron catalytic cracking (moving bed), the content of this carcinogen is significantly greater, 0.30 per cent being an average value. API 06539 -7 - There is some evidence from biological tests that BP may be more refractory to thermal cracking than are the carcinogens of lower aolecular weight in catalytically cracked oils. A series of such comparisons are shown in Table 5, in which the feed stock and the tar produced from a thermal cracking operation were collected at the same time and then tested on comparable groups of mice. From the ana lytical data it will be noted that in each case the concentration of BP wes higher in the tar than in the feed. Further, the ratio of the concentraion of BP to the relative potency was increased in this process. However, one cannot interpret these data as proof that the con centration of other carcinogens was reduced. The potency of most of the feed streams depended to some extent on long-chain acceler ators. Extensive splitting of these chains by the thermal cracking would be expected. The analysis for BP furnished one of the parameters, C]_, of the general potency equation above. Information about other important classes of carcinogens, symbolized by C2 ana C3 , was required. A catalytically cracked residuum of high potency, Oil No. S, was frac tionated in a 25-plate Stedman type of column at 0.03 mm. pressure at the head. From the data in Table 6 , it will be noted that the fractions boiling below 670F. (corr. to 760 mm) showed no activity, confirming the reports of Fischer and coworkers ill). A high level of activity was found in the fractions in the range, 750-800F. estimated, although no significant amount of the benzopyrene of this oil (0.4 per cent in the original) distilled until higher temperatures API 06540 9 - 8 - were reached. The highest boiling distillate, 8-li|., and the residuum, 8-15 contained sufficient BP, 0.9 and 2.0 percent, respectively, to account for all of their carcinogenic potency. It is not surprising that none of the fractions proved to be more potent than the original oil. Most of the non-carcinogenic but im portant accelerators of this oil distil below the point at which significant levels of concentration of polycyclic carcinogen are obtained. The carcinogenic distillates were subjected to further fraction ation by chromatography on alumina, partition between certain volatile hydrocarbons and concentrated sulfuric acid at 5 to 10C., and reaction with maleic anhydride. Most of these techniques are standard practice. One of them, that involving the solution of certain classes of polycyclics in concentrated sulfuric acid without sulfonation (1 0 ,1 2 ), may be relatively unfamiliar. In general, pericondensed molecules, such as pyrene, BP, and perylene, as well as anthracene and BA dissolve. Phenanthrene, fluoranthene, chrysene, benz.o(c_)phenanthrene, and dibenz(a_,h)anthracene do not, though some of their alkyl drivtes do. The utility of this method of fractionation is shown by the absorption spectra of Figure 1 , representing the separation of dimethylpyrenes from a mixture also containing alkylphenanthrenes and fluoranthenes. Unsubstituted benz(a)anthracene (BA) was found to be concen trated chiefly in the distillate 8-12. Since only one of the known carcinogenic hydrocarbons has a lower molecular weight than BA, i.e., 9,10 -dimethylanthracene, and its relative potency is very low, consid erable interest has centered in the identification of the active API 06541 i components of fractions 8-10 a n d -11. Furthermore, these carcinogens are probably the same as those responsible for the potency of sidestreams from catalytic cracking and delayed coking, and possibly even those in certain raw lubricating stocks. The absorption spectra of these distillates indicated the presence of pyrene and its methyl and dimethyl derivatives. The l|.-methyl and i).,9-dimethyl homologues were isolated. In view of the importance of benzopyrene in higher boiling cuts, the possibility that some dimethyl- pyrene might be responsible for the potency of fractions 8 -10 a n d -1 1 seemed plausible. However, such does not seem to be the case. The combined distillates were partitioned between cold concentrated H8S04 and n-heptane. The raffinate, which was essentially free of pyrenes after 5 extractions, proved to have most of the potency of the original distillates. Its spectrum indicated that alkylphenanthrenes were among the major components. In order to have an adequate quantity of this material for detailed examination, a sample of 3*5 kg. of the aromatics from chroma tography of Oil No. 8 has been subjected to a more efficient fractional distillation (column rated at 80 plates at total reflux). The fractions of lowest boiling range which showed significant carcinogenic activity, 8-26-1,-m, and-n in Table 7, were extracted with aqueous acid and base and then subjected to the reaction with maleic anhydride. It was found that the compounds regenerated from the Diels-Alder adduct from 8-26-n had contributed about l/6th of the potency of the original distillate. The following composition was estimated from the mass API 06542 gpectrum of this mixture: n cnH2n-l8 17 18 19 20 21 Total 1 .8* 7.5 7.U 3.0 1.2 21* cnH2n -2 0 1 .0* 8.7 1 0 .1+ 1+.9 1.5 26* CnH2n -2 2 0.6* 6 .1 7.6 2.6 0.8 18* CnH2n-2l4. mm 29.3* 3.9 1.1 0 .1+ 35* The cyclopentano substitution is considered thd most likely structure for the CnH2n _20 hydrocarbons. A six-membered ring would be expected (on the basis of the spectrum of tetralin) to yield a strong fragment peak at m/e of 2QLp by loss of -CHB-CH*- . The intensity of this ion in the spectrum of 8-26-n-l is negligible. Similar extractions of 8-26-1 and -m with maleic anhydride were carried out. The pooled non-adducts were then subjected to a systematic fractionation on alumina (Alcoa F-20). The components in successive fractions, the presence of which was most likely on the basis of UV API 06543 / - 11 - and IR absorption spectra, were as follows: Eluted 0 - 22* 21 - 21% 25 - 55* 35 - 86* 78 - 87* 82 - 9 1* 87 - 91 91 - 100* __________ Apparent Components_________ Saturated hydrocarbons Alkylnaphthalenes Alkylphenanthrenes Pyrenes, chiefly dimethyl Fluoranthenes, chiefly methyl Benzofluorenes, chiefly methyl Chrysene and triphenylene Non-hydrocarbons, including carbazoles No satisfactory evidence for the presence of any significant amount of the carcinogenic benzo(c_)phenanthrene has been found. Further, its highly strained, non-planar structure would seem to militate against its presence in a severely cracked oil of this type. Biologic tests of certain of the more strongly adsorbed fractions have been initiated. The approximate composition of selected cuts has been estimated from absorption and mass spectra as follows: 85.0 - 88.U* Fraction 88.1+ - 90. 5* Fraction n cnH 2n-22 ^n^2n-2l+ ^n2n-22 cnH2n-2l+ 17 2.0* 18 79.8* 3* 19 11+.8* 20 0.8* 1.1+* 68.6* 8.3* 0.7* 21.1* 0.2* Total 97* 3* 79* 21* Suggested structures of principal components: 0 4 - 12 - m nclualons . Various classes of hydrocarbons, including the saturates, contribute to the carcinogenic potency of certain refinery streams for the skin of mice. A useful hypothesis about the relationships between these contributions may be represented by the general equation: PMC S M* {C* + C- + C* ) The relative potency, PMC, expresses the results of a given bioassay in units of concentration of standard polycyclic carcinogen, 3-me thylcholanthrene. The multiplication factor, Ma , is a complex function describing the catalytic influence of certain non-carcinogenic hydrocarbons, such as n-hexadecane and amylnaphthalene. Factors up to 10 may be introduced into the potency equation by such accelerators. The summation of C's represents the basic potency contributed by certain polycyclic aromatic compounds, the most important of these being |- and 5-ringed hydrocarbons. Benzo(a)pyrenes are probably the chief class of carcinogens of catalyilcally cracked residua. The data presented suggest that the most important carcinogens of sidestreams from catalytic cracking or coking operations may be der ivatives of the benzofluorenes or possibly of cyclopentanophenanthrenes. Certain uncracked lubricating distillates also demonstrate significant carcinogenic activity for the skin of mice and men. Acceleration of the effects of a relatively low concentration of carcinogens by long-chain saturates and alkylnaphthalenes is of primary importance for such oils. API 06545 # r ." "T - T T T T T ^ . Henry, -: . A., ?rl* . Red. S u i . . , , 389-kOl (IQ4.6) . Heller, I . , J. Ind. y*., 12, .109-97 (1?0). Hartvrell, . i . , Public P'ealth Service Publication 'Je. lk, Cr:.i l e . , l o i f *> , ". i., and " i e e e r , I . , . Cherr.. S e c . , 1? ? t , - rT - I . a n d : r n o e n t a l , TM 3rit. J. Exp. P a t h . , T u , 2 ^ 2 (i q L * ). "crtA. '.V., m i > rii. J. C a n c e r . 1, l - ' - r , r.r.f.n, " - > C a n c e r Res., 1 5 , ? C l - t ( I b 5 ; H c r t e n , A . >/., I - e n o j a r , r . T . , a n d C r e s s e t , R , ? , , C a n c e r P e a , (in :r:ce S3; . ~~ > * 9 . r , : . Pizze, A, Riegel, 5., Shurik, ?., C -. .V "a n c e r " e s . , 11, v? 2- ? 7 , ( 1^ 51; . ve, R . , ' r a l , ' . J . , are - ' r b o n , A. V /., A nni . Chen-. , i T , 2 . 5- : * (la:;?). . cntrr , Pr .e 7, ., jlrr 9. " ...> \ .e3s , -. am "ce . , > :U - 2-' (Ibri). lr, P ., Src.T.at i ene 'celer.-vaa sserr'a3ttoefff ee,, Per iin, Spri -.--er-"eria API 06546 ABSORBANCE (at concentration indicated) IN I300CTANZ .'.PALaTION 0 [l I'Or,v (TCLJC AROMATICS BY PARTITION BETWEEN ISOOCTANE AND CONCENTRATED SUIFURIC ACID o OU'1 WAVELENGTH, m^Jl 0 0 Table 1 '-elative Carcinogenic lotencies of Various TJncracked Refinery Streams ~escr1 eticn -racticn lelo'.v 753"?. (') denooQT)vrene ~araffinic 73 distillate lube distillate 50 0.0003 O.OOl2 . ; .sohaltic crude oil Lo . *7 Vest Texas reduced crude li - "henol' extract > o m nanhtnenic '.istillate 10 ? Rhenol extract from naor.thenic distillate 30 ^-0.0005 /v0.002 ic.0032 --^ - 0 .0022 Aver arre "'uration of Tumor-Free l i f e ,1 . (weeks) elative ,~are ir.ccan 1 _ "ctency, U> n> 22 0.11 5o 0.02 - 0.00 - 0.00 79 0.01 11 O.'O tests invclvinr three apolic-ati cns each week, "'/.certainty due to interference from perylenes. Table 2 Relative Carcinogenic Potencies of Various Cracked Refinery Streams 'description -enzcQp r-vrene " ) "eleyeu C::<av 'I '. "eiayed Coker c?,s cil _. JJo 0.003 catalytic ally cracked sidestream, r'.3.?. 750J 0.001 C-atalytically residuum Catalytically residuum Catalytically residuum cracked cracked cracked 0 .1+2 0.10 0.53 'nermal tar (no cata?. vt- 0 .01+ icslly cracked com ponents in feed) Aver age duration of Tumor-rr*e fife,1 e (weeks) 9 est.2 23 IP. 5.6 - 3 3 est. 11 est. 2 22 CT- O PeTaTTvPT reinor r cf.?n c~T. 'T 0.79 0.10 0 .2 0.5 - 1. 0.32 *\J *0 .i:3 "or tests involving three 100 mg. applications each week. 'ased cn comparable tests involving 1 or 2 applications each week "ncertainty due tc poor health of mice. # Table 3 Relative Activity of Various Solvents to Accelerate Carcinogenesis experiment Number 502 273 295 269 b 293 263 262 213 625 Solvent n-0ctane n-Decane n-Dodecane n-Hexadecane Hexadecene- 1 1-Cyclchexyldecane 1-Phenyldecane Diamylnaphthalenes. Technical White Oil, 95 3 .S.U./10 0 "?. 1 . Uncertainty due to poor health of mice. Relative Accelerating Activity 1 3 5 li 5 S 7 2 API 06550 Experiments to De terrine the 'ct.ency of fixtures of t'o l /< n li e hydrocarbons for the Skin of Vice Exp e rim en t Number 62l| 623 hydrocarbon 3-lVethylcho lant hrene 3-Me t by l c h o l an l-h re ne OnnzoQ^J p yrene S o lv e n t T e c h n ic a l white o i l , 95 s.-'.n./ioo'r. Technic. 1 w hite o il ' 0: .'n itr a t i oil 0/ f i r c i noe on ( by w eight) 0 .10 0.20 vela t iv e Car-c in o n d ilo `o t e n c y , ' mc1 0 -jr'+.Ol U* - . 0 2 0 .1? .0 2 0 O API 06551 626 d -i? d -L 2 a O -Y e t h y lc h o la n ln r ne 3 -F*e t hy 1 c h01 nnM .re fte dona rajan th rn co ru; 50# , D i s t i l l a t e from c a t a ly t ic a lly cracked o i l No. J ; 50# hc.nzene 50?/, N o n -a d d u c t from D i e l a - A l d e r on " .-12; SO,. uan/.e ne T e c h n ic a l vdiite o i l T e c h n ic a l white o il 0 .10 0 .10 0 . 10X .01 ? 0 . 03;- 0 .2 1X .0 3 o .fil.o i, 1 . fix . I t a of C 011 f I d e a r 0 ; 0 .o ). c lue. cl a i n t y due to ( hr. .1th rT mice Experi ment 'T^rber Tescriction 0 Feed: Cstalyt icaliy cracked 3 ide ctr-earr. ?L " ar pr c lueed i'rcr: 23 ?" 3/ Fred: .'atalyt icaily cracked ;as oil ?0 Tar rrccvced .>U- Teed: Tat'lyt 1:&IT y cracked 2 as oil 35 "ar prcduced > k 4 - '' id: *A^alyt ieslly c.' .ck d sidestrear. L 7 "ar ^reduced A 7 "ce-2 : Cc-calyt icaily cracked ras oil 61| T rV*r,iuc 3'''? /*\ -a Te 1 : Catalyt ice 1 ly cracked cirestreen 9? "ar r.r'-C'rced lenze(a)ovrer.e ) Relative Carc inr ye:-.' r ?r`er.c", ?vc O.jl* 0.22 C.Tii. ' '4 T.at io of ^ pyre ne ^ 0.5 <J * xO C 2 3 c- ?1 0 .2 ^ -* 7 0.7 :. oil* :.i - - 0.15 0.15 s ` i C *'-- 0.15 0 .~2 J .2 *- . i r* y ~ o,17 0.3 0 .11 C ,17 - T . 'i* 0.07 -- 0.31 lw/` * API 06552 Table 6 -ractional Distillation of Catalyt ically Cracked Residuum, Oil No. 8 Experiment Number S-6 1 2 3-7 8-8 3-9 3-10 3-11 3-12 3-lli1 3-15 Estimated Boiling Range (in F. at 760 mm.) 1+30 - 550 550 - 620 620 - 670 670 J 635 635 - 730 730 - 755 755 - 780 730 - 305 305 - 830 > 830 Percent of Original Oil . 0 - 9.3^ 9 .3 j- 1 9 19.4 - 30.3* 30,3 - 79.9^ 3 9 .8 - 4.9.65? ii.9.6 - 59.2? 59.2 - 69. 69.1 - 73.55 73.5 - 30.3* 80.3 - ioo O O (\i rH elative Careino- genic Potency, __ PVC 0 0 0 rO 0.03 0.1U 0.17 (0.15) 0.11 --1 O 1. All fractions were tested as 5C$ solutions in benzene except 3-lL, which was used |is a 33# solution. 2. Uncertainty due to high mortality among mice. API 06553 Table 7 Tumor-Inducing Activity of Products of Fractional Distillation of the Polycyclic Aromatics of a Catalytlcally Cracked Oil Experi ment Number 8-26-c2 8-26-d 8-26-e 8-26-f 8-26-g 8-26-h 8-26-1 8-26-j 8-26-k2 8-26-1 8-26-m2 8-26-n Fraction Distilled 32 - 3+.1* 3+.1 - 36 36 - 38 38 - 1+0 k - +3.9 +3.9 - +6 i+6 - $0.3 50.3 - 52.5 5 2 5 - 5+.8 51+.8 - 57.3 $7 .3 - 60.1+ 60.1+ - 62.1+ Estimated Boiling Range (In 0 F. at 760 mm.) Incidence of Tumors {% - week) Relative Carcino genic Potency, PMC 651 - 6$6F. 656 - 662 662 - 668 0* 44. wks 0 65 0 66 668 - 671+ 7 80 671+ - 686 686 - 693 693 - 709 709 - 717 717 - 727 15 52 12 kS 28 52 71 72 83 19 v 0.01 o.03.o i = 0. 212 727 - 736 736 - 71+7 7k7 - 759 90 51 87 lb 100 39 0 .9:| 0.292 0 ld+ , 02 O123?-.07 Contribufcions of acclrt ors to potencies 8- 26-2 8-26-4. 3 Solution of Benzopyrene In fraction"^ of 8-26-g, -h, and -1; plus 20% benzene Solution of Benzopyrene In fraction^ of 8-26-1, -m, and -n; plus J>0% benzene Acceler ating Factor 5 2 1. During Indicated period of exposure. 2. All fractions tested as $0% solutions In benzene except 8-26-k and -m, which were accelerated with $0% dodecylbenzene. 3. Chromatographic fractions including saturates through naphth alenes, 27% of 8-26-g, -h, and -i, of 8-26-1, -m, and -n.