A study of the data presented leads us to feel that in no case did the fungus have a pathognomonic relationship to the diagnoses involved. It is possible, but difficult of proof, that the fungus in any case modified or influenced the symptomatology. The course of therapy ignored the fungus findings and in all cases the response to therapy was characteristic and average for the diagnoses concerned. The surprisingly even balance between positives and negatives in the various diagnoses listed in table 3 strengthens this feeling. It is further confirmed by the wide variety of fungi present under the same clinical diagnosis, and the wide distribution of each fungus type among different diagnoses, as shown in table 4.
It is probable that further survey studies would be of value in elucidating the problem of pathogenicity of intestinal fungi. Such studies should use a uniform method of culture and classification. They should be made so far as possible on individuals with a single type of pathology, of localized residential history, with repeated cultures, associated cultures of the mouth and of gastric and duodenal contents, and close correlative study of the chemical and bacterial features of the specimens cultured.