In an attempt to confirm the concept of Evans and Freeman that bacteria of ordinary size may occur as filterable forms in encephalitis of man and rabbits, 550 cultures were made from brain emulsions or filtrates (Berkefeld V) of 80 rabbits infected with 4 different strains of encephalitic virus. The inoculated rabbits were either killed while showing typical symptoms of the disease, or were allowed to die. Positive cultures were obtained as follows: staphylococ-cus, 4 times; B. coli, 3; diphtheroids, 3; M. cremoides, once; Gram-positive rods, 4 times; and Streptococcus viridans, once. The pathogenicities of the Gram-positive rods, diphtheroids, and streptococcus were tested in rabbits by intracerebral injections. Only negative results were obtained. The streptococcus rabbits were further tested for immunity to virus but they succumbed typically, thus showing that previous inoculation with the organism conferred no immunity to the virus. The shorter the interval between death and autopsy, the smaller was the number of organisms recovered from brain, suggesting the positive cultural results were due either to non-specific infection of the brain tissue before or just after death, or to technical error. The experiments failed to show any causal relationship between the ordinary visible bacteria and encephalitis, and furnished no evidence to support the suggestion that the filterable encephalitis virus may change into a true bacterial form.