In support of the theory of a toxic genesis of mental diseases arising from chronic infection, some authors have attributed a special rôle to the B. coli. As long ago as 1904 Johnsohn and Goodall (1) examined the agglutination of a series of psychotics to the B. coli. They found this reaction completely or weakly positive in 50% of their cases of insanity (mostly with a serum dilution of 1: 100), while only 15.5% of their control cases showed a positive reaction. These results were not confirmed by Puca (2). Later on Goodall, together with Barton White (3), also examined the stool flora of psychotics, and found no qualitative difference as to the coliform group between psychotics and normal persons. On the other hand, Stewart (4), in more recent experiments, claims to have found such differences; thus he emphasizes the much higher frequency of the B. para-coli and B. coli mutabile in mental patients, especially in recent cases, and the parallelism between their presence and the intensity of the symptoms. Further, he found a difference in the phenol formation in broth cultures obtained from stools-of psychotics and normal persons, the former producing on the average 0.02 to 0.03%, and the latter only 0.008 to 0.03% of phenol. This stronger production of phenol is attributed by Stewart to the high frequency of B. Morgan and other phenol-producing bacteria in psychotics. He discusses the possibility that aromatic bodies produced by these bacteria in the acid milieu of the large intestine may represent the toxic agent in mental diseases.