1. Three fractionation schemes are described for the concentration of the polycyclic aromatic hydrocarbons (PArH) of cigarette smoke condensate. 2. Schemes 1 and 2 have their first two stages in common which result in the separation of cyclohexane soluble material (Fraction G), representing 24% by weight of the original smoke condensate, from water soluble and methanol soluble constituents. 3. The mouse painting tests show that by direct comparison Fraction G contains a very high proportion of the tumorigenic components of stored nonvolatile whole smoke condensate (SWSC). Neutral fraction (33-35% w/w SWSC) at 600 mg./wk (equiv. dose) gives a similar number of both tumour bearing animals and carcinoma bearing animals as Fraction G (24% w/w SWSC) at 480 mg./ wk (equiv. dose). Fraction G is considered to be the more satisfactory material for further attempts to isolate the mouse skin carcinogenic factors in whole smoke condensate. 4. Fractions N, GI and HC1, the materials containing the PArH in most concentrated forms so far tested for biological activity, produced similar numbers of animals with tumours in each case. 5. In fractionation Scheme 1 considerable losses of carcinogenic material, which are not reflected in losses of PArH, occur during the last three stages of the separation. Recent work indicates that this loss occurs by irreversible adsorption or catalytic destruction of biologically active material on silica gel. Similar, but less pronounced, losses were encountered in the preparation of hydrocarbon fraction from neutral fraction (Scheme 3) by the partially irreversible adsorption or destruction of biologically active material on alumina. 6. Scheme 2 gave some loss of biologically active material but this was not appreciable. Whereas Fraction GI, containing all the PArH, proved to be as active as Fractions N and HC1, its activity was only one-quarter of that of Fraction G2, the recovered PArH-free material and one-sixth of the activity of whole smoke condensate. 7. There was no effect on the incidence of spontaneously occurring tumours or disease processes in any of the treated groups as compared with untreated control mice housed under identical conditions. 8. The general conclusions drawn from this work are that solvent partition methods for the fractionation of SWSC are highly satisfactory in minimising losses of biologically active material and that the unsubstituted polycylic aromatic hydrocarbons are probably not such important carcinogenic constituents of cigarette smoke condensate as has been thought hitherto. The authors thank the Tobacco Research Council for permission to publish this paper, Doctors T. D. Day and R. F. Davies for supervising all the biological experiments and post-mortem examinations and for carrying out the histological diagnoses, Mr. P. N. Lee for statistical assistance, Mr. T. Smith for supervising the large scale preparation of fractions and Mr. D. V. D. Thorowgood who was responsible for the animal husbandiy. © 1969, The British Empire Cancer Campaign for Research. All rights reserved.
