TESTING FOR LONG-TERM TOXICITY 827 formation. An example is that of urethan (ethyl carbamate) which is excreted more slowly in infant mice (10) and is a more active agent in in- fants (9). Despite the apparent enhanced responsiveness of newborn animals to carcinogens, there are still too many unknown factors involved in this technique. Therefore, in routine tests for the safety of new compounds, it is considered best to start with weanling animals or those a few weeks postweanling (9, 11). Starting with animals which are already adult may not leave enough of the animal's life span for tumors to develop, especi- ally in the case of weaker carcinogens. A second factor which influences the results of tests for chronic toxicity is the sex of the animals. Bates (12) reported that male mice had more skin tumors from painting a typical carcinogen 7,12-dimethylbenz(a)- anthracene on the skin than did females (Table II). Some explanations for these differences are given by Bock (13). In many cases male ani- mals also responded to a greater degree to other carcinogens which affect the internal organs. Table II Skin Tumors in Mice after 7,12-Dimethylbenz(a)anthracene• Sex Tumor Incidence (%) Tumors per Mouse Male 79 5.90 Female 18 0.26 • Data from Bates (12). The tumors were induced by a single dose of 16tag of dimethylbenz- anthracene applied to the back, followed by application of 0.2 ml of 1% croton oil in acetone weekly. The experiment was terminated after 23 weeks. Controls treated with acetone alone had 0.06 tumors per mouse. An illustration of the part played by hormones comes from the report of Morris and Firminger (14). Male ACI rats are much more suscepti- ble to the hepatocarcinogenic effects of 2-diacetylaminofluorene than are females. However, castration of males and administration of estrogen led to a tumor response more nearly like that of females. Conversely, females which had been ovariectomized and then given testosterone re- sponded like intact males. Because of these differing responses, test systems should include both male and female animals. Likewise, species and strain play an important role in the response. Thus, it was noted in some of the first work on cutaneous response to the pure aromatic hydrocarbons that there was a great range in the reactivity of various species. For example, with benzo(a)pyrene, a hydrocarbon

828 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table III Species Difference in Response to Benzo(a)pyrene• Tumor Incidence Duration Species (Range, %) (Range) Mouse 60-100 75-200 days Rabbit 2-15 8-55 months Rat ... 2 years Guinea pig ... 2 years Fowl ... 2 years Monkey ... Up to 10 years • Data from Hartwell (15). The benzo(a)pyrene was applied to the skin by painting an acetone or benzene solution (0.1-0.5 % ) two or three times weekly. Table IV Mouse Strain and Response to Repeated 3-Methylcholanthrene a Papilloma Incidence Average Time to Autopsy Strain (%) (Weeks) A 100 30 BALB/c 100 31 C57BL 97 36 C3H 100 28 DBA/2 100 35 I 100 27 R III 100 43 • Data from Andervont and Edgcomb (16). A 0.25% solution of 3-methylcholanthrene was applied weekly to the skin of the interscapular area by a single brush stroke. often used as a standard carcinogen, the mouse was most sensitive, the rab- bit next, while some species did not respond within a useful time period (15) (Table III). Within a given species, the response to any given carcinogen may vary greatly. Although repeated skin painting with $-methylcholanthrene presents a sizable carcinogenic dose, mice of some strains such as the G57BL or RIII either took a longer time period to die from the tumors induced or else did not develop tumors. On the other hand, a high per- centage of strain I mice developed multiple papillomas and died early (1õ) (Table IV). A study with the hepatocarcinogen ethionine exhibits clearly the wide variation in response. Holtzman male rats fed ethionine for 5 months developed no liver cancer under identical conditions Carworth Wistars had an 86% incidence of these tumors (17) (Table V).