4-NITROQUINOLINE N-oxide (NQO), which was already known to have antifungal and mutagenic activity, was first shown to be carcinogenic by Nakahara, Fukuoka and Sugimura (1957). These authors applied the compound to the skin of 20 mixed strain mice and obtained tumours on each of the 14 which survived more than 100 days. Rather unexpectedly, the tumours on 4 of these mice were found to be sarcomas, though some difficulty was experienced in determining their exact cell type. Differences in the degree of incidental surface erosion were suggested to account for thQ production of carcinomas on some mice and sarcomas on others.The carcinogenicity of NQO for mouse skin has also been investigated by Lacassagne, Buu-Hoi and Zajdela (1961). They reported a strain difference in response to NQO, papillomas and epitheliomas being produced on their C57B1/Z mice but not on XVIInc/Z mice. However, their ascribing this effect to possible differences in the SH-content of epidermal cells of the two strains appeared to us unjustified on the evidence available, since these workers also described severe toxic reactions to the NQO applications, which resulted in progressive emaciation, cachexia and death of many mice without the production of tumours. Nakahara et al. (1957) only applied very small amounts of NQO (about 0.02 ml. of a 0-25 per cent solution in benzene 3 times a week), but their loss of 6 out of 20 mice during the first 100 days of treatment is also suggestive of marked toxicity.The present paper summarises the results of a number of separate experiments with NQO carried out in our laboratories using both random-bred mice and pure-line mice of several strains. The first experiments (I-III), using random-bred stock albinos and C57B1 x IF F1 hybrids, demonstrated toxic effects of NQO similar to those described by the earlier workers, and we also obtained some sarcomas in addition to papillomas and carcinomas. Another series of tests was then carried out in which smaller amounts of NQO were applied to the skin of these mice and of four pure-line strains of mouse, with a view to elucidating the question of strain differences in response to NQO under conditions of reduced toxicity (Experiment IV). This experiment demonstrated marked strain differences in the toxic and carcinogenic effects of NQO. Moreover, it was unexpectedly found that in some strains of mouse, sarcomas were induced in larger numbers than epithelial tumours. A further experiment (V) with one of these strains is also included in this report.