A study was made of the intracellular localization of Type 2 poliomyelitis virus, using the technique of Mirsky and Pollister (23) for cellular fractionation. After isotonic saline homogenization of central nervous system tissue from infected mice, and subsequent centrifugation of the suspension, the virus present in the supernatant fluid was held to be of cytoplasmic origin. Upon serial washings of the sediment with physiological saline, the resulting supernates contained progressively less virus until by the seventh washing, virtually none was present. At this point extraction of the washed sediment with molar NaCl, which lyses the nuclei, yielded substantial amounts of virus, and this was assumed to be from nuclear sources. The possibility has not been excluded however that the "nuclear" sediment was contaminated by cytoplasmic particles too large to remain in the supernate. Experiments on the increase of virus during the incubation and acute stages of infection have revealed that it was first detectable in the "cytoplasmic" fraction and subsequently in the "nuclear" fraction. Virus in the "nuclear" fraction from paralyzed mice sometimes reached titers almost as high as those found in the "cytoplasm." Adsorption experiments indicated that the "nuclear" fraction of CNS tissue from normal, uninoculated mice did not adsorb added Type 2 poliomyelitis virus, nor did such fractions adsorb virus procured from the "cytoplasm" or "nuclei" of infected cells. Although individual mice varied in their response after virus injection, the "cytoplasmic" fraction of paralytic mice was found to contain virus regularly, whereas little more than half of the non-paralytic mice yielded it. When virus was present in the "cytoplasm," it could be found in the "nuclear" fraction of paralytic mice with much greater regularity than in that of non-paralytic mice. A comparison between the lines of the MEF1 strain of poliomyelitis virus, "adapted" and "non-adapted" to newborn mice, and the Lansing strain, revealed no differences in their intracellular increase. In both infant and adult mice, the chief difference in the findings with non-paralyzed and paralyzed mice lay in the greater concentration of virus in the "nuclear" fractions of the latter group.