Sparse gypsy moth,Lymantria dispar(L.), populations studied in the vicinity of Glenville, N. Y., between 1958 and 1964 tended to increase rapidly to outbreak levels. Conversely, equally sparse populations studied in the vicinity of Eastford, CT between 1965 and 1971, tended to remain sparse. Major mechanisms determining numerical differences between the two areas acted primarily during three age-intervals. First, the survival rate among instars I–III was higher in Glenville than in Eastford, although this difference decreased as density increased. These results were interpreted as a consequence primarily of airborne dispersal processes that occur largely during instar I. Second, the survival rate among instars IV–VI was higher in Glenville than in Eastford, and this difference increased as density increased. Although predation processes were probably responsible for most of the instar IV–VI mortality among the sparse populations in both areas, these results suggest that predation during this stage was effective in maintaining sparse populations at innocuous levels only in Eastford. Third, pupae were more likely to survive in Glenville than in Eastford, and this difference was relatively constant across the entire density range that was common to the 2 areas. Since vertebrate predators, especiallyPeromyscus leucopus, were known to have killed most of the pupae in the Eastford populations, these results imply that these predators played a major role in maintaining the Eastford populations at innocuous levels.