Six-week-old male CF-1 mice were fed the AIN-76 diet, given eight weekly s.c. injections of either the colon carcinogen 1,2-dimethylhydrazine or saline, and killed 24 weeks after the last injection. Parameters measured in the large bowel included the incidence and locations of all intestine (gut)-associated lymphoid tissue (GALT) sites; the locations, incidence, and sizes of all aberrant crypt foci (ACF); and the incidence, locations, and types of all overt tumors. In saline-treated mice the distribution of GALT along the length of the large bowel was bimodal, with a majority peak of lymphoid nodules occurring in the distal large bowel and a minority peak occurring in the proximal large bowel. No ACF or tumors were present in the large bowel of the saline-treated mice. In 1,2-dimethylhydrazine-treated mice the majority of ACF were present in the middle third of the colon, between the two peaks of GALT, but the majority of the tumors were found over the GALT in the distal colon. There was a significant positive linear regression relationship between the numerical distribution of GALT and the numerical distribution of tumors along the length of the large bowel. There was no significant relationship between the distribution of ACF and the distribution of (a) tumors or (b) GALT along the length of the large bowel. Thus the numerical density of lymphoid nodules, not the numbers or distribution of ACF, was the significant predictor of the distribution of tumors in the large bowel of 1,2-dimethylhydrazine-treated mice. It is proposed that lymphoid nodules in the distal large bowel play a promotional role following initiation of colon carcinogenesis and that ACF have little if any malignant potential in the mouse.