Two recent trials of chemotherapy in relatively large numbers of patients with lepromatous leprosy generated data that permitted analysis of the effects of treatment regimens and of various pretreatment characteristics of the patients. The results of treatment were measured by inoculation of mice with Mycobacterium leprae recovered from skin biopsy specimens obtained from the patients at intervals during the trials. The pretreatment variables—sex, age, histopathological and clinical classifications, logarithmic biopsy index, bacterial index, and the common logarithm of the number of M. leprae per skin biopsy specimen (LAFB)—were found to be uniformly distributed among the 36 patients treated by regimens 1, 2, 4 and 5 of trial I. The ten patients treated by regimen 3 were excluded from this analysis. These pretreatment variables were also found to be uniformly distributed among all 21 patients treated by the two regimens of trial II. The effects of the pretreatment variables and of the treatment regimens on the response to treatment were considered in 31 patients treated by regimens 1, 2, 4 and 5 of trial I (5 patients were excluded from this analysis because of evidence of unauthorized dapsone intake) and in the 21 patients of trial II. The treatment regimen itself was the most important determinant of the rate of loss of M. leprae infective for mice in both trials. In addition, the pretreatment value of the LAFB was found to be significantly related to the response to treatment in trial I. Had the random allocation of patients to each regimen not resulted in a uniform distribution of the pretreatment LAFB, the results of the trial might have been confounded. Stratification of patients by the number of M. leprae in the skin, as indicated by the LAFB, is recommended for future short-term trials of antimicrobial therapy of lepromatous leprosy.