N1(-)[tris(hydroxymethyl)]methyl-4-nitro-o-phenylenediamine was fed in the diet to groups of 30 male and 55 female Sprague-Dawley rats at levels of 0.2, 0.6 and 2.0% for up to 6 months. One mid-dose and two high-dose females developed palpable mammary masses that were subsequently diagnosed as mammary adenocarcinomas at a 13-wk interim kill involving 10 rats/sex/group. After 14 wk, 25 females per group with no apparent masses were mated in a reproduction/teratology study. Mammary tumours developed in a dose-related fashion both in the pregnant rats and in the remaining 20 females/group that continued on treatment for 6 months. On gestation day 20 (wk 17-18) the final incidences of mammary adenocarcinomas in the low-, mid- and high-dose mated dose groups were 20, 60 and 84%, respectively, while the corresponding incidences in the non-mated females at 6 months were 5, 40 and 85%. Most mammary tumours were encapsulated but, at 6 months, lung metastases were noted in four rats, and four females also had Zymbal's gland tumours. Non-neoplastic changes in male and female rats considered to be related to treatment included increases in thyroid follicular cell size accompanied by an accumulation of golden-brown pigment, multifocal hepatic necrosis with non-suppurative inflammation, and renal tubular pigmentation. Increases in foetal variations in the mid- and high-dose groups were considered to be related non-specifically to retarded growth. Malformations observed in the high-dose group were found primarily in single foetuses and were not considered to be treatment related. Although the mean numbers of micronucleated polychromatic erythrocytes in bone marrow obtained from high-dose treated females after 13 wk slightly exceeded historical negative control values, the data were not considered indicative of a genotoxic effect because of the absence of either a dose relationship or a substantive increase in the frequency of micronucleated cells.