Procedures for gross separation into basic, acidic, phenolic, and neutral fractions and for further processing of these fractions vary from laboratory to laboratory. The criteria upon which identification is based also vary. The most reliable identifications are based upon an ultraviolet absorption spec- trum and/or a fluorescence spectrum in good agreement over the entire range \tith that of an authentic sample and include one or more of the following: Rf value observed in a paper chromatogram 111) ; order of elution from alumina; mass spectrometry. COMPOUNDS OF THE PARTICULATE PHASE OTHER THAN HIGHER POLYCYCLICS This brief summary is based largely on the comprehensive review by Johnstone and Plimmer of the Medical Research Council at Exeter Uni- irraity. England ( 24 I. It should be noted that water constitutes 27 percent Ilf the particulate phase. Th e major groups of compounds included are .ho\tn in Table 1. ALIPHATIC AND ALKYCLIC HYDROCARBONS Almost all of the possible hydrocarbons, C, through C,, saturated and urr+aturated, straight-chain and branched-chain, have been reported to be prcaen, in tobacco smoke. Intermediate, normally liquid paraffins are pres- ent. All the C,, through C,, n-a lkanes have been identified, as well as the CZ: and C,!,-c',, isoparaffins. T4BL.E I.--Major classes of compounds in the particulate phase of cigarette smoke Percent in Sumber 01 particu- comwmd late* phase - 7.7-12.8 1 25 5.3-8.3 18 8. 5 21 4. 9 0.44 El 1. G-3. 8 45 Toxic action on lung Some irritant Possible irritation Some irritant Some irritant Some carcinogenic Irritant and possibly cocarcinokxnic TERPENES AND ISOPRENOID HYDROCARBONS isoPrene, the basic unit of the terpenes and of higher terpenoids has been !dcntified in mprrthadiene. cigarette smoke (34) as have its dimers, dipentene and 1,8-p- and shown t The triterpene squalene, consisting of six isoprene units o b hilit! of its be* e present in smoke (47) is of interest because of the possi- mg cyclized to polycyclic compounds and because of its ready 51 CHa CE4 CHa C& HaC CH: CHa CHa Squalene reaction with air to form hydroperoxides (which would be destroyed during attempted isolation ) ; a hydroperoxide derived from cholesterol has been shown to be carcinogenic i cancer-causing) : at least under certain conditions of administration I 12) . Phytadienes. products of the dehydration of the diterpene alcohol phytol, are also present in smoke and subject to air oxida- tion to hydroperoxides. C& Crt CH: CHIOH HsC Phytol ALCOHOLS ANT) ESTERS A wide variety of mono- and dihydric alcohols, both aliphatic and aro- matic, are present in tobacco smoke. Solanesol, a primary alcohol con- taining 9 isoprene units, has been found in both tobacco and tobacco smoke; 20 g. of pure material was isolated from 10 lbs. of flue-cured aged tobacco (0.44 percent). Grossman et al i 13) found that pyrolysis of solanesol at 500" C. gives isoprene, its dimer dipentene, and other terpenoid products and concluded that the alcohol is the source of terpenoid compounds which are important factors in the flavor of tobacco smoke. Ethylene glycol and glycerol have been found present in smoke, but it is not clear from the literature whether they are present in smoke from un- treated tobacco or arise from addition of these humectant substances to tobacco to improve moistness. Many common esters, such as the ethyl esters of the C2, C,, and C, fatty acids, are present in smoke. Higher fatty- acids are found both as free acids and as esters. STEROLS Stigmasterol, p-sitosterol, and r-sitosterol have been isolated from to- bacco smoke. Indeed the sterol fraction is reported (29) to constitute approximately 0.15 percent of whole tar. The sterols are of interest as possible precursors of polyc)-clic aromatic hydrocarbons and because of the evidence, noted above. that sterol hydroperoxides can be carcinogenic. ALDEHYDES AND KETONES Most common aldehydes of low molecular weight (acetaldehyde, pro- pionaldehyde, acetone, methyl ethyl ketone, etc.) have been found present 52 -4 CHlOH HIC I 5oo"= Ad + Isoprene I-&C `-.2Hz Dipentene (major product) Solanesol \ 6 (3 HsC' CHa C& &C /`hII., CH, I H:C J&C HIC CHa in tobacco smoke, as have such dicarbonyl compounds as glyoxal and di- acetyl. Dipalmityl ketone exemplifies ketones of high molecular weight isolated from tobacco smoke. 0 16' C& Dipalmityl ketone ACIDS A large number of volatile and nonvolatile acids of low molecular weight are present in tobacco smoke. Fatty acids of chain length C,, to C,, are reported to constitute 1 percent of the whole tar and the bulk of these acids are present in the free form (46). Unsaturated fatty acids and keto acids `e.g., pyruvic acid) are also present. 53 PHENOLS AND POLYPHENOLS Since the phenols and polyphenols present in tobacco leaf play an im. portant role in the curing and smoking quality of tobacco, a great deal of investigative work has been done on the estimation, separation, and ident& cation of complex tobacco phenols such as rutin and chlorogenic acid. The presence of simple phenols in tobacco smoke was established as early as 1871. The phenol content of smoke became of increasing importance with OH HO H? \ - 0 CH- CH~O~co,H - ir HOi OH Rharnnoae Rutin Chlorogenic acid the demonstration that phenol and substituted phenols can function as cocarcinogens; that is, they promote the appearance of skin tumors in mice following application of a single initiating dose of a known carcinogen (4). Furthermore, the smoke from one cigarette contains as much as 1 mg. of phenols (7). In add t i ion to simple alkylphenols, naphthols, and the poly. phenols, resorcinol and hydroquinone are also present. ALKALOIDS, NITROGEN BASES, AND HETEROCYCLICS Pyridine, nicotine, nornicotine, and other substituted pyridine bases con. stitute some 8-15 percent of whole tar; nicotine and nornicotine constitute about 7-8 percent of the total tar. The companion bases are products of the pyrolysis of the alkaloids present in tobacco leaf. Quinoline and three poly-cyclic heterocyclic compounds have also been identified in smoke (45) and will be discussed later since the three polycyclic compounds are carcino- genic. `4 pentacyclic compound related to xanthene, namely 1,8,9peri- naphthoxanthene. has been identified in smoke (45). 1,8,9-Perinaphthoxanthene AMINO ACIDS Although tobacco leaf contains a number of amino acids, relatively few have been found present in smoke; among these are glutamine and glutamic acid. 54 INORGANIC COMPONENTS It is estimated that the main-stream smoke from one cigarette contains about 150 1j.g. of metallic constituents. which are mainly potassium (90 tIercent I _ sodium (5 percent I, and traces of aluminum, arsenic, calcium. and copper. Arsenic is reported to be present to the extent of 0.3-1.4 pg. in the smoke of one cigarette. Th e inorganic compounds are most likely chlorides. but metals themselves may be present. Apparently bery-ilium is present in tobacco in trace quantities. but is not 1 olatilized in the smoking process ( 4s ) . Nickel is present in cigarettes in trace amounts and may occur in main-stream smoke to a small extent, l:robably as the chloride (31 t . Spectrographic analysis has shown the presence of chromium in smoke at a level of less than 0.06 ;tg. per cigarette. This level appears too low to represent a hazard 148). IVONCARCINOGENIC AROMATIC HYDROCARBONS The aromatic h>-drocarbons present in tobacco smoke have received an enormous amount of attention since some of them are carcinogenic. Toncarcinogenic hydrocarbons of smoke containing one to three rings include benzene. toluene and other alkplbenzenes, acenaphthene, acenaph- thylene. flnorene. anthracene. and phenanthrene. Hydrocarbons of estab- lished carcinopenicity to mice all contain from four to six condensed rings. Ifowever. no less than 27 hydrocarbons containing four or more `condensed rings which have been tested for carcinopenicity with negative results have heen isolated from tobacco smoke tar. As methods of separation and identification improve, it is almost certain that additional hydrocarbons will be found present in smoke, because almost every conceivable ring system has been demonstrated to be present and the number of possible alkylated polycyclics is very large indeed. CARCINOGENIC HYDROCARBONS AND HETEROCYCLICS IN TOBACCO SMOKE In 1925-30 Kennaway et al. in seeking to identify the active substance in high-boiling fractions of coal tar distillates of established carcinogenicity to mice, discovered that dibenzo(a,h)anthracene (for formula, see Table 21 prepared by synthesis evokes skin cancer when applied to the skin of mice (11). The hydrocarbon was recognized as different from the carcino- gen of coal tar because its fluorescent spectrum did not match the character- istic three-banded spectrunr of the tars. In 1933 Cook and co-workers i 11) isolated the coal tar constituent responsible for the characteristic fluorescence and identified it as benzota) pyrene. the carcinogens now known. It is one of the most potent of all 55 TABLE 2.--Carcinogenic Polycyclic Compounds Isolated From Cigarette Smoke Compound 1. Benzo(s)pyrene 2. Dibenzo(s,i)pyrene 3. Dibenao(s,h)snthrscene 4. Benao(c)phenanthrene 5. Dibens(s,j)acridine 6, Dibenz(a,h)acridine 7. 7H-Dibenzo(c,g)carbszole structllre ' I / / (Id?? : 1; > w I> `I \ ' \ / Carcino- genicity Amount reported, rg/KMM cigarettes ++++ ++++ ++ 16 (ave. of 10 reports) 0.02-10 (2 reports) (1 retort) ot- not stated + 2.7 (1 report) 0.1 (1 report) 0.7 (1 report) 56 Since the discovery of carcinogenic hydrocarbons, a large number of polycyclic hydrocarbons and heterocyclic analogs have been tested for car- cinogenicity to mice and to rats in many laboratories, both by application to the skin and by subcutaneous injection. Bioassays in different labora- tories, often on independently prepared samples, are remarkably consistent and place a series of hydrocarbons in the same relative order of potency. A compilation (and its supplement) prepared by J. L. Hartwell (16) of the IKational Cancer Institute lists 2108 compounds of which 481 were reported to cause malignant tumors in animals. All but one of the polycyclic hydro- carbons listed in Table 2 as having been identified in tobacco smoke have already been documented in the Hartwell report and can be assigned a rating as very potent ( + + + + ), potent ( + + + ) , moderately carcino- genic I, + + ), or weakly carcinogenic ( + ) (31). Many other such com- pounds studied are reported in the Hartwell survey and in another by :Irthur D. Little, Inc. (31). The rating assigned to dibenzo (a,;) pyrene is based on experiments with over 10,000 inbred mice in which one subcutaneous injection in the groin of 0.5 mg. of hydrocarbon in tricaprylin produced 50 percent sarcomas at the injection site in 14 weeks and 98 percent tumors in 24 weeks (20). Benzo(a)pyrene is one of the two most potent of the seven carcinogens detected in tobacco smoke and it is present in much larger quantity than any of the other carcinogens listed. Two polycyclic hydro- carbons isolated from tobacco smoke but not yet adequately tested for carcinogenicity are: benzo (j ) Auoranthene and dibenzo (a,l) pyrene. Identification of benzo (a)pyrene is reported in 19 separate investiga- tions; the amount given in the table per 1000 cigarettes (70 mm. long, Neighing about 1.0 g. each) is the average of 10 values selected on the basis of the quality of criteria used for identification (31). Compounds 1, 2, 3, 4, and benzo (j) fluoranthene were identified in one laboratory over a period of years and are listed together in a review by Van Duuren (44). Isolation of the three heterocyclic carcinogens (5,6,7) is reported by Van Duuren (45). Because of losses in the process of fractionation and purification, the amount of carcinogens reported in a given investigation may be less than the amount actually present. Wy d n er and Hoffman (50) investigated this point by adding a known amount of radioactive C"-1abelled benzo(a)pyrene to a smoke condensate and applied the usual procedure for isolation of benzo(a)pyrene, which involved, in the last stages, chromatographing twice on silica gel and four times on paper. The activity of the benzo(a) pyrene finally isolated indicated a loss of 3540 percent of carcinogen during proc- essing. Th e amount of benzo(a) pyrene given in Table 2 thus should be multiplied by a factor of 1.5 to give the estimated true amount. Probably the amounts of the other carcinogens in smoke are also at least 1.5 times the reported amounts. Relatively little work has been done on the components of smoke produced with cigars and pipes. Table 3 summarizing a comparative study made in one laboratory (5) indicates that the amount of benzo(a)pyrene, the only carcinogen in the group studied, increases sharply from cigarettes to cigars to pipes. 57 TABLE 3.-Polycyclic hydrocarbons isolated from tobacco smoke [,,g. pm 1000 g. of tobxcm consumrd~ COCARCINOGENS Assays of tobacco smoke tars for carcinogenicity are done by applying a dilute solution of tar in an organic solvent with a camel's hair brush to the backs of mice beginning when the animals are about six weeks old. Applica. tion is repeated three times a week for a period of a year or more. The results of a number of such assays present a puzzling anomaly: the total tar from cigarettes has about 40 times the carcinogenic potency of the benzo( a) pyrene present in the tar. The other carcinogens known to be present in tobacco smoke are, with the exception of dibenzo(a,i) pyrene, much less potent than benzo (a) pyrene and they are present in smaller amounts. Apparently, there. fore, the whole is greater than the sum of the known parts (27, 33,49). One possible or partial explanation of the discrepancy is that the tar con. tains compounds which, although not themselves carcinogenic, can enhance the cancer-producing properties of the carcinogens. Berenblum and Shubik (3), reporting on cocarcinogenesis. described the potentiating effect of croton oil, which itself is noncarcinogenic except in certain strains of mice (4a), on the action of hydrocarbon carcinogens. Phenol is reported to have a similar potentiating effect (4. 50) and, as noted above. cigarette smoke contains considerable phenolic material. Long-chain fatty acid esters (39) and free fatty acids (19) have been shown to function as cocarcinogens, and sub. stances of both types occur abundantly in tobacco smoke. It is possible that the potentiatinp action of croton oil is due to the presence of fatty acids and their esters. A further observation of possible importance is that some poly cyclic hydrocarbons. though very weak or inactive as carcinogens, are capable of initiating malignant growth under the influence of a promoter. Thus henz (a) anthracene, identified in cigarette smoke, is verv weak or inactive in initiating malignant growth by itself. but initiates carcinogenesis under the influence of croton oil as promoter (15). If more were known about the possible cocarcinogenicity of the many inactive components of tobacco smoke, some of the apparent discrepancy between isolation and bioassay data might disappear. It is possible that some of the carcinogenicity of smoke is due to hydroperoxides formed from un- saturated smoke components and destroyed in the isolation procedures. Furthermore both sets of data are far from precise; for example, one esti- mate of the amount of the highly potent dibenzoi a,i)pyrene per 1000 cigarettes (Table 2) is 0.02~~. and another is 1Opg. However. it is not necessary to wait for an exact balance of the two sets of data to draw a conclusion from each. The isolation experiments, taken 58 alone, indicate that cigarette smoke contains a number of identified chemicals which are carcinogenic to mice. The bioassavs suggest that cigarette smoke probably contains components which. actin g in a manner as yet undescribed, are involved in the induction of tumors in mice. Assessment of all conceivable synergistic effects presents a gigantic problem for exploration. Tobacco smoke contains considerable amounts of phenols and fatty acids, both of which, as previously mentioned, enhance the activity of known carcinogens. Cellulose acetate filters now in use remove `XL80 percent of acidic constituents of tobacco smoke. MECHANISM OF THE FORMATION OF CARCINOGENS Most of the carcinogenic compounds identified in cigarette smoke tar are not present in the native tobacco leaf but are formed by pyrolysis at the high burning temperature of cigarettes. Van Duuren (4.4) reports formation of benzo(a) pyrene and pyrene on pyrolysis of stigmasterol, a smoke com- HO Stigma&sol Benro(a)pyrene Pyrene ponent. Similar pyrolysis of pyridine or of nicotine gives dibenzo( a,j) acridine and dibenzo (a,h) acridine, both of which are carcinogenic (Table 2). Pyrolysis of nontobacco cigarettes made from vegetable fibers and spinach resulted in formation of benzo( a jpyrene (50). Hurd and co-workers (22) by careful experimentation have elaborated plausible mechanisms for the formation of polycyclic aromatics by pyrolysis of materials of low molecular weight at temperatures in the range 800-900" C. Postulated radical intermediates are: (a) CHz=C=kH - CH~-C+ZH (b) EH-cH=I~H - ~H=cHGH tc) CH=CH~H=CH These radicals can arise from propylene, toluene, picoline, or pyridine. A variety of polycyclic hydrocarbons can be generated by reaction of these radicals with themselves or with other small radicals present in the heating zone. For example, dimerization of (b ) should give benzene. 59 Jt thus appears that the pyrolysis of many organic materials can lead to the formation of components carcinogenic to mice. Cigarette paper con. sists essentially of cellulose. Pyrolysis of cellulose has been shown to produce henzo(a)pyrene. The observation (2') that treatment of tobacco with copper nitrate decreases the benzo (a) pyrene content of the cigarette smoke suggests a possibility for improvement by the use of additives or catalysts. The fact that side-stream smoke contains three times more benzo (a) pyrene than main-stream smoke has been cited (50) as evidence that more efficient oxidation could conceivably lower the content of carcinogenic hydrocarbons. THE G.4S PHASE The gas phase accounts for 60 percent of total cigarette smoke. Hobbs et al. ( 34, 35`1 found that 98.9 mole percent of the gas phase is made up of the following seven components: Yitrogen ____------- --------- ________. 73 mole percent Oxygen---- --------_____------_______ 10 Carbon-dioxide ____- -- _______-__----_ - 9.5 Carbon-monoxide--------------------- 4.2 Hydrogen---------------------------~ 1. Argon------------------------------. 0.6 Methane----------------------------- 0.6 98. 9 The approximately one percent of the gas phase not accounted for by the seven major constituents contains numerous compounds, no less than 43 of I\ hirh have been identified as present in trace amounts. Some of these are listed in Table 4 (1). TABLE 4.-Some gases found in cigarette smoke (1 (PPm) 100 NX) 5. cinl 2d 0. 5 Unknown l-one Sonc sonr Irritant Irritant Irritant Irritnnt Irritant Irntant Irritant Irritant r."known i;;itant Repiratory enzyme poison Unknown EFFECTS ON CILIARY ACTIVITY* An important line of investigation was opened up by the report by Hilding (1s) that cigarette smoke is capable of inhibiting the transport activity of ciliated cells such as found in the respiratory tract. It has been suggested ( 10. 17 I that failure of ciliary function to provide a constantly moving stream of mucus enables environmental carcinogens to reach the epithelial cells. Kensler and Battista t 28) describe development of a method of bioassay for inhibition of ciliary transport activity involving exposure of the trachea of a rabbit to the test material. The smoke from a regular cigarette was found to inhibit transport activity by 50 percent after exposure to two or three puffs. Several commercial filter cigarettes gave essentially the same result. The fact that these filters lower the phenol content by 70 to SO percent and trap about 4.0 percent of the particulate phase suggested that neither phenolic nor particulate materials are responsible for the inhibi- tion noted. The next trial was with an absolute filter. that is, one which removes the entire particulate phase and gives nonvisible gas. The obser- vation that such treatment did not significantly alter the inhibitory effect of the puff established that components of the gas phase are responsible for inhibition of ciliary transport activity. Assays of known components of the gas phase showed the followin g compounds to possess such activity: hydrogen cyanide, formaldehyde. acetaldehyde. acrolein, and ammonia, al- though no one of these occurs at levels high enough to produce the effect noted for smoke. Activated carbons differ markedly in their adsorption characteristics. Carbon filters previously employed in cigarettes do not have the specific power to scrub the gas phase. It has been reported that a filter containing special carbon granules removes gaseous constituents which depress ciliary activity (28) . PESTICIDES AND ADDITIVES Before 1930 practically the only insecticides used in the growing of to- bacco were lead arsenate and paris green (the mixed acetate-arsenite salt of copper). Analysis of 6 brands of American cigarettes purchased in 1933 showed a range of 7.5-26.4 parts of As,O, per million, with an average value of 13.9 ppm. (6). Cogbill and Hobbs (S) found that main-stream smoke of Cigarettes containing 7.1 pg. of arsenic per cigarette contains 0.031 pg. per puff. This amount would be equivalent to 0.25 pg. of arsenic per cigarette (8 puffs), and hence a smoker consuming 2.5 packs of such cigarettes per day might inhale 12.5 pg. of arsenic per day. By comparison, analysis of the atmosphere of New York City over a 12-year period indicated an average content of 100-400 pg. of arsenic per 10 cubic meters, which is an approxi- mate daily intake per person (38). Extensive Federal efforts to discourage the use of arsenicals for the control of tobacco hornworms on the growing tobacco crop resulted in a sharp de- `This tapir is disrussel~ more fully in ~haptrr IO. 61 cline in the arsenic content of cigarettes after 1950. Thus, the average arsenic content of 17 brands of cigarettes analyzed in 1958 was 6.2 ppm. of As,O, (14). It seems unlikely that the amount of arsenic derived even from unfiltered cigarettes is sufficient to present a health hazard. Chemicals recommended by the Department of Agriculture for the control of tobacco insects are: malathion, parathion, Endosulfan, DDT, TDE, end&, dieldrin, Guthion, aldrin, heptachlor, Diazinon, Dylox, Sevin, and chlordane (42a). Trace amounts of TDE and endrin have been detected in commercial cigarettes and cigarette smoke. Guthion and Sevin residues were detected in main-stream cigarette smoke at levels approximating 0.3 percent and l percent of that added to cigarettes prior to smoking. Tobacco treated with Guthion and Sevin at the recommended levels showed no measurable con- tamination of main-stream cigarette smoke (4b). (For discussion of car- cinogenicity of tobacco pesticides, see Chapter 9.) Cigarette manufacture in the United States includes use of additives such as sugars, humectants, synthetic flavors, licorice, menthol, vanillin, and rum. Glycerol and methylglycerol are looked on with disfavor as humectants be- cause on pyrolysis they yield the irritants acrolein and methylyglyoxal. Additives have not been used in the manufacture of domestic British cigarettes since the Customs and Excise Act of 1952, Clause 176, and probably longer, inasmuch as Section 5 of the Tobacco Act of 1842 imposed a widespread prohibition on the use of additives in tobacco manufacture. SUMMARY Of the several hundred compounds isolated from the tobacco leaf, two groups are specific to tobacco. One of these groups includes the alkaloid nicotine and related substances. The other includes compounds described as isoprenoids. Cigarette smoke is an heterogeneous mixture of gases, uncon- densed vapors, and particulate matter. In investigating chemical composition and biological properties, it is necessary to deal separately with the particulate phase and gas phase of smoke. Components of the particulate phase other than the higher polycyclics include aliphatic and alicyclic hydrocarbons, terpenes and isoprenoid hydro- carbons, alcohols and esters, sterols, aldehydes and ketones, acids, phenols and polyphenols, alkaloids, nitrogen bases, heterocyclics, amino acids, and inorganic chemicals such as arsenic. potassium, and some metals. Seven polycyclic compounds isolated from cigarette smoke have been estahlished to be carcinogenic. They are shown in Table 2. The over-all carcinogenic potency of tobacco tar is many times the effect which can be attributed to substances isolated from it. The d'ff 1 erence may be associated in part with the presence in tobacco smoke of cocarcinopens, several of which have been identified as smoke components. Components of the gas phase of cigarette smoke have been shown to pro- duce various undesirable effects on test animals or organs, one of which is suppression of ciliary transport activity in trachea and bronchi. 62 REFERENCES 1. Albert, R. E., Nelson, N. Special report to the Surgeon General's Ad- visory Committee on Smoking and Health. 2. Alvord, E. T., Cardon, S. Z. The inhibition of formation of 3,4-benzpy- rene. Brit J Cancer 10: 49%506, 1956. 3. Berenblum. I.. Shubik, P. The role of croton oil applications, associated with a single painting of a carcinogen, in tumour induction of the mouse skin. Brit J Cancer 1: 379-82. 1947. 1. Boutwell, R., Bosch. D. K. The tumor-promoting action of phenol and related compounds for mouse skin. Cancer Res 19: 413-24, 1959. 4a. Boutwell, R., Bosch, D. K., and Rusch, H. P. On the role of croton oil in tumor formation. Cancer Res 17: 71, 1957. 4b. Bowery, T. G., Guthrie, F. E. Determination of insecticide residues on green and flue-cured tobacco and in main-stream cigarette smoke. Agriculture and Food Chem 9( 3) : 193-7, 1961. 5. Campbell, J. M., Lindsey, A. J. P o ycyclic hydrocarbons in cigar smoke. 1 Brit J Cancer 11: 192-5, 1957. 6. Carey, F. P.. Blodgett. G.: Satterlee, H. S. Preparation of samples for determination of arsenic. 0. xygen-bomb combustion method. Industr Eng Chem Anal Ed 6. 327-30. 193-1. 7. Clemo, G. R. Some aspects of the chemistry of cigarette smoke. Tetra- hedron 3: 168-74, 1958. 8. Copbill, E. C., Hobbs, M. F,. Transfer of metallic constituents of ciga- rets to the main-stream smoke. Tobacco Sci 1: 68-73: 1957. 9. Dauben, W. G., Thiessen, W. E., Resnick. P. R. Cembrene, A 14-mem- bered ring diterpene hydrocarbon. J Amer Chem Sot 84: 2015-6, 1962. 10. Falk, H. L.. Tremer, H. M., Kotin, P. Effect of cigarette smoke and its constituents on ciliated mucus-secreting epithelium. J Nat Cancer Inst 23: 999-1012, 1959. 11. Fieser, L. F., Fieser, M. Topics in organic chemistry. New York, Reinhold, 1963, p. 43-56. 12. Fieser, L. F., Greene, T. W., Bischoff, F., Lopez, G., Rupp, J. J. [Com- munication to the editor] A carcinogenic oxidation product of choles- terol. J Amer Chem Sot 77: 3928-9, 1955. 13. Grossman, J. D.. Deszyck, E. J., Ikeda, R. M., Bavley, A. A study of pyrolysis of solanesol. Chem Industr 1950-1962. la. Guthrie, F. E., McCants, C. B., Small, H. G., Jr, Arsenic content of commercial tobacco, 1917-1958. Tobacco Sci 3: 624, 1959. 15. Hadler, H. I., Darchun, V., Lee, K. Initiation and promotion activity of certain polynuclear hydrocarbons. J Nat Cancer Inst 23: 1383-7, 1959. 16. Hartwell, J. L. S urvey of compounds which have been tested for car- cinogenic activity, Federal Security .4gency, Public Health Service Pub No. 149, 1951. 583 p, 17. Hilding, A. C. On cigarette smoking, bronchial carcinoma and ciliary action. 3. Accumulation of cigarette tar upon artificially produced 714-422 O-64-6 63 deciliated islands in the respiratory epithelium. Ann Othol 65: 116 30, 1956. 18. Hilding. A. C. On cigarette smoking. bronchial carcinoma and ciliary action. 2. Experimental study on the filtering action of cow's lungs, the deposition of tar in the bronchial tree and removal by ciliary action, New Eng J Med 251: 1155-60, 1956. 19. Holsti. P. Tumor promoting effects of some long chain fatty acids in experimental skin carcinogenesis in the mouse. Acta Path Microbial Stand -16: 51-8, 1959. 20. Homburger, F., Tregier, A. Modifying factors in carcinopenesis. Progr Exp Tumor Res 1: 31 l-28, 1960. 21. Hurd, C. D., Macon, A. R. Pyrolytic formation of arenes. 4. Pyrolysis of benzene, toluene and radioactive toluene. J -\mer Chem Sot 84: 4524-6, 1962. 22. Hurd, C. D., Macon, A. R., Simon. J. I.. Levetan, R. V. Pyrolytic forma- tion of arenes. 1. Survey of general principles and findings. J Amer Chem Sot 84: 4509-15, 1962. 23. Hurd, C. D., Simon, J. I. Pyrolytic formation of arenes. 3. Pyrolysis of pyridine: picolines and methylpyrazine. J Amer Chem Sot 84: 4519-24, 1962. 24. Johnstone, R. A. W., Plimmer, J. R. The chemical constituents of to- bacco and tobacco smoke. Chcm Rev 59: 885-936, 1959. 25. Keith, C. H.: Newsome, J. R. Quantitative studies on cigarette smoke. 1. An automatic smoking machine. Tobacco 144: (13) 26-32. May 29, 1957. 26. Keith, C. H., Newsome, J. R. Quantitative studies on cigarette smoke, 2. The effect of physical variables on the weight of smoke. Tobacco 144 (14) : 26-31, Apr 5, 1957. 27. Kennaway, E., Lindsey, A. J. Some possible exogenous factors in the causation of lung cancer. Brit hIed Bull 14: 124-31, 1958. 28. Kensler, C. J., Battista, S. P. Components of cigarette smoke with ciliary- depressant activity. New Eng J Med 269: 1161-1166, 1963. 29. Kosak, A. I., Swinehart. J. S., Taber. D., Van Duuren, B. L. Stig- master01 in cigarette smoke. Science 125: 991-2, 1957. 30. Langer, G.. Fisher, N. A. Concentration and particle size of cigarette- smoke particles. AM&\ Arch Industr Health 13: 372-8: 1956. 31. Liggett & Myers Tobacco Co. Arthur D. Little, Inc. Special report to the Surgeon General's Advisory Committee on Smoking and Health. 32. Lindsey, A. J. Some observations upon the chemistry of tobacco smoke. In: James, G., Rosenthal. T.. eds. Tobacco and health. Springfield, Ill.. Thomas, 1962. Chapter 2: p. 21-32. 33. Orris, L., Van Duuren, B. L.. Kosak. A. I.. Nelson. N.: Schmitt, F. L. The carcinogenicity~ for mouse skin and the aromatic hydrocarbon content of cigarette-smoke condensates. J Nat Cancer Inst 21: 557-61, 1958. 34. Osborne, J. S., Adamek, S.. Hobbs: M. E. Some components of gas phase of cigaret smoke. Anal Chem 28: 211-5, 1956. 64. 35. Philippe, R. J.. Hobbs, M. E. Some components of the gas phase of cigaret smoke. Anal Chem 2::: 2002-6. 1936. 36. Roberts. D. L.. Rowland, R. I,. >Iacrccyclic diterpenes a and B-4, 8, 13.Duvatriene -I.:<-dials from tobacco. J Org Chem 27: 396%95, 1962. 37. Ro\$land, R. L.. Rodgman. A.. Schumacher. J. 5.. Roberts. D. L., Cook, L. 0.. Walker. W. E. 1063 I In press). ;