THE TOXICOLOGY OF ARTIFICIAL COLOURING MATERIALS 429 is that there is no basis for the argument that, in choosing safe colourings, less hazard is likely to be associated with the use of colourings that do not produce sarcomas than with the use of those that do. Superficially reason- able though it may be, this attitude simply evades the issue and perpetuates current misconceptions. Undoubtedly some true carcinogens do give rise to local tumours on subcutaneous injection but in such cases their carcino- genicity has invariably been established by other means. Effects on reproduction, toxicity to the foetus and teratogenicity Very little information is available in this area of toxicology, except for acid disazo dyes such as Try_pan blue, Evans blue and various types of Niagara blue. An excellent review has been published (34). While this group of dyes finds no application in foodstuffs, and probably none in cosmetics, the possibility exists that action on foetal development is an attribute of some part of the molecule of the disazo dyes. Beaudoin and Pickering (35) suggested that 1-napththylamine sulphonated in two positions, preferably 3 and 6, is a basis for teratogenicity. Accordingly Christie (36) tested the effect of 1,7diamino-8-naphthol-3,6-disulphonic acid given by the subcutaneous route to rats on day 8 of pregnancy. The maternal deaths, resorptions and developmental retardations observed were attributed to renal toxicity of the compound in the mothers rather than to any direct effect on the embryo. The present position, therefore, appears to be that foetal toxicity and teratogenic activity are confined to a special class of dyes given by injection in a variety of species. The relevance of these results to cosmetics colourings is doubtful. Mutagenesis The observation by German workers that Erythrosine had a slight but nevertheless genuine mutagenic effect on E. coli was followed up by Lfick et al (37) who tested a variety of xanthene and other colourings. Xanthene itself, eosine, eosine BNX and erythrosine showed a very slight but statistically significant mutagenic effect on the bacteria when tested at a concentration of 10-4M, supposedly the same order of magnitude as that used in foodstuffs. Erythrosine at 1%, and Rhodamine B at a 0.5% concentration had distinct mutagenic effects. Induction of mutation in yeast has been reported by Nagai (38), using basic triphenylmethane dyes (such as methyl violet and p-rosaniline) and xanthenes (Pyronines Y and B). Although concentrations as low as 1 ppm or less produced effects, it is questionable whether these results have any significance from the standpoint of mammalian toxicology.

430 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Dose and effect The use of colourings in food is regarded as self-limiting and, subject to good manufacturing practice, no attempt is usually made to prescribe upper limits for the levels of application in particular foods. This factor, taken together with the possible presence of colourings in many articles of diet, consumed by both the very young and the very old, makes the task of safety evaluation much more exacting than it need be with cosmetics colours, where maximum levels of use can be prescribed for various appli- cations. Thus Davis and Fitzhugh (39) were able to establish 0.05øfo as a satisfactory no-effect level for Lithol red (D & C Red No. 10) in rats. Accepting that in such circumstances a reasonable safety factor (i.e. ratio between the maximum tolerated dose producing no ill-effects in animals and the likely total daily intake in man) such as 100 can be agreed upon, the levels of testing in animals are readily arrived at. Provided the maximum no-effect level is thus established in long-term lifespan studies in the animals, what happens at higher levels is of less importance as long as carcinogenesis is not involved. The colouring matters testing programme initiated by the Toilet Goods Association U.S.A., has been based on this approach (40). Finally, there exists the need to discriminate between colourings in- corporated in preparations whose use involves the likelihood of ingestion, and those products intended solely for external application. (Received: 14th April 1966) REFERENCES (1) Golberg, L. J. Soc. Cosmetic Chemists 15 177 (1964). (2) Cainan, C. D. J. Soc. Cosmetic Chemists 18 3 (1967). (3) Hansen, W. H., Davis, K. J., Fitzhugh, O. G. and Nelson, A. A. Toxicol. Appl. Pharmacol. 5 105 (1963). (4) Radomski, J. L. and Mellinger, T. J. J. Pharmacol. Exptl. Therap. 130 259 (1962). (5) Scheline, R. R. and Longberg, B. Acta Pharmacol. Toxicol. 9,11 1 (1965). (6) Ryan, A. J. and Wright, S. E. Nature 195 1009 (1962). (7) Daniel, J. W. Toxicol. Appl. Pharmacol. 4 572 (1962). (8) Ryan, A. J. and Wright, S. E. J. Phar•n. Pharmacol. 13 492 (1961). (9) Priestly, B. G. and O'Reilly, W. J. ,[. Pharm. Pharmacol. 18 41 (1965). (10) Webb, J. M., Fonda, Margaret and Brouwer, E. A. J. Pharmacol. Exptl. Therap. 137 141 (1962). (11) Webb, J. M. and Brouwer, E. A. Federation Proc. 21 449 (1962). (12) Andersen, C. J., Keiding, N. R. and Nielsen, A. B. Scan& J. Clin. Lab. Invest. 16 249 (1964). (13) Hess, S. M. and Fitzhugh, O. G. J. Pharmacol. Exptl. Therap. 11{ 38 (1955). (14) Fours, J. R., Kamm, J. J. and Brodie, B. B. J. Pharmacol. Exptl. Therap. 19,0 291 (1957). 1.q• Lindstrom, H. V., Hansen, W. H., Nelson, A. A. and Fitzhugh, 0. G. J. Pharmacol. Exptl. Therap. 1{9, 257 (1963).