422 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS It is my intention as far as possible to avoid covering ground that was very adequately dealt with at the 1963 Symposium on "Toxicology of Cosmetic Matehals," and in particular my own contribution on "The assessment of safety-in-use: just how much is contributed by feeding tests in animals ?" (1). I am fortunate also in having Prof. Cainan dis- cussing reactions to colouring materials (2) and need not therefore elaborate on this aspect of their toxicology. This discussion will be principally concerned with food colourings, for that is the field in which there is most knowledge and experience. It does not follow automatically that acceptance for food use renders such colours safe for use in cosmetics. Some mention will be made of the additional tests involved to establish suitability for incorporation into formulations intended for external application. Specifications and impurities I need hardly stress the importance of knowing what we are testing when we undertake to study the toxicology of a material. While speci- fications are available for many synthetic colourings, particularly those used in food, they usually fall far short of what is desirable from the standpoint of the toxicologist. Not for him the pathetic charade of limits for lead, arsenic or copper, archaic devices intended to ensure good manu- facturing practice, and now quite wrongly interpreted as safeguards against toxic hazard. How adequate is the emphasis on limits for total amines or total ether extractable matehals ? This is the age of chromatography, and there is no reason why it should not be applied to separate, identify and ultimately to standardize the by-products that are present, so that their role in the production of biological effects may be accurately assessed. The time has long since passed when toxicological evaluation could be carried out "blind." If we are to make full use of recent advances in biochemical pharmacology we need to know all we can about the chemical composition of the matehal under investigation. A striking instance of the discrepancies between apparently authentic specimens of Ponceau 3R, recently reported by Hansen, et al (3), underlines the need to characterize most fully any colouring undergoing safety evaluation. METABOLIC CHANGES In the intestinal tract Colouring matters entering the gut are subjected to the action of acid, digestive enzymes and the gut flora. The degradation undergone as a

THE TOXICOLOGY OF ARTIFICIAL COLOURING MATERIALS 423 result of the red'uctive fission of azo linkages in water-soluble colourings has been studied by Radomski and Mellinger (4) who showed that amines so formed in the rat were absorbed, metabolized and excreted in the same manner as the identical amino compounds given in free form by stomach tube. The most characteristic compound split off in this way is sulphanilic acid, which is absorbed and emerges in the urine as the free acid and its N-acetyl derivative (5). By administering the colouring by the oral route in some rats and intraperitoneally in others, one can judge the extent to which intestinal degradation takes place. Thus when Tartrazine is given by mouth, sulphanilic acid appears in the urine but no free colour- ing is excreted in urine or faeces. When the compound is given parenterally the animal is dyed bright yellow and free colouring appears in the urine, but without sulphanilic acid (6). Whatever the extent to which a colouring is absorbed from the intestine, a complicating factor is the proportion of biliary excretion, which provides a direct route from the liver back into the intestine. Daniel (7) and Ryan and Wright (8) have shown that some water-soluble azo colourings are excreted almost quantitatively in bile. In a study of the relation of protein binding to biliary excretion, Priestly and O'Reilly (9) concluded that preferential binding to liver proteins, as against plasma proteins, was the determining factor - at least in the case of the colours studied by them. Biliary excretion has the effect of recycling the intact colouring, or the products derived from it, through the intestine, a process termed "enterohepatic circulation." The stability of halogenated derivatives of fluorescein, given by mouth to rats, has been studied by Webb et al (10). Of the di-, tri- and tetra- halogenated colours, only the 4-iodo and 4-bromo derivatives were dehalo- genated to fluorescein. Recoveries of unchanged material in the faeces were almost quantitative with tetrahalogenated derivatives, but fluorescein and its dihalogenated derivatives were cleared more slowly than the others. Webb and Brouwer (11) found that increasing halogenation of fluorescein diverts excretion from the urine to bile and hence the faeces, but also increases the total excretion. Elevation of serum protein-bound iodine in man has been quoted as evidence that erythrosine is deiodinated after ingestion (12). How- ever, since no effort was made to distinguish between protein-bound erythrosine-iodine and protein-bound iodine, this conclusion is unconvinc- ing. Sulphonated colourings that are not subject to attack by the