Summary: The importance of Freund's adjuvants for the production of experimental autoimmune nephrosis was studied. The frequency of disease production depended on the amounts as well as on the frequency with which they were used. Incomplete adjuvants or replacement of Mycobacterium tuberculosis by Mycobacterium butyricum, did not produce renal disease. The virulence of the organism or the method by which Mycobacterium tuberculosis was killed did not affect disease production. Production of a less severe, sterile peritonitis by replacement of Freund's adjuvants by talcum suspensions containing rat kidney antigen, failed to produce renal disease. Splenectomy and partial removal of the omentum before treatment did not affect incidence of ensuing renal disease. When rat kidney cortex suspension was used as antigenic material, the disease incidence was 81%. With material obtained from renal medulla it was 25%. ACTH and cortisone prevented the development of renal disease. Nitrogen mustard decreased the incidence from 87% to 30%. Guinea pigs did not tolerate the i.p. injections and died after 2 to 4 injections. Six dogs received up to 10 injections. The degree of resulting peritonitis was less severe and was only localized around injection sites. No renal disease was noted and precipitins to dog kidney antigen could not be detected in their sera. The severity of peritonitis was more marked in 16 treated rabbits. Bands of precipitation to rabbit kidney suspensions were detected in their sera, but no evidence of renal disease was obtained.