The primary object of this study, as we have already stated, was to look for evidence of an epidemic strain among cocci of this type. With regard to this group of bacteria, perhaps more than any other, sweeping statements have been made concerning the relationship of strains to epidemics. This is largely due to their ubiquity and the difficulty of demonstrating clear cut proofs of such relationship and also to the ease with which deductions are made from insufficient evidence. The most important question relating to this subject about which we have little accurate knowledge, is the question of the permanency of those characteristics which are supposed to be indicative of species, of specific pathogenicity and of related characteristics which might be used in diagnosis. We know a little of the limits of change in vitro but essential changes in the animal have not yet been demonstrated with controls sufficiently free from error.