In the cytological studies of the geographic races of Lymantria dispar, GOLDSCHMIDT ('17, '20, '32) pointed out racial size difference of chromosomes indicating that the amount of chromatin is in an inverse proportion to the strength of sex-races. Close examination of his data revealed some questionable points worthy of study. The reinvestigation was here undertaken by the authors with material coming from twelve localities. The races coming under study are treated as follows: the weak race from Sapporo, the neutral race from Kyoto, Ayabe and Kumamoto, and finally the strong race from Aomori, Tokyo, Matumoto, Misima, Nagoya, Gifu, ogaki and Maibara.The haploid number, 31, was obtained in the study of the primary and secondary spermatocyte throughout the geographic races under observation. The diploid number, 62, was observed in the spermatogonial metaphase of one of the races. The chromosomes of the haploid group are round or oval in outline and vary in size very slightly forming a well graded series in the diminution of size. It was found to be practically impossible to group the chromosomes as large, medium and small ones, as done by GOLDSCHMIDT.The comparative study of chromosomes of the geographic races leads to the conclusion that the chromosomes of the studied races are fairly identical in their general appearance, especially in regard to the size of chromosomes, and that there is detected no essential morphological difference in the chromosomes among the races obtained from the different localities and also among the different sex-races.In the light of these studies it must be concluded that GOLDSCHMIDT'S contention that the chromosomes of the weak, neutral, and strong sex-races of Lymantria dispar show perceptible size differences resulted from mistaken observation and irregular technique. The error apparently resulted from: (1) a confusion between the secondary and. primary spermatocytes; and (2) varying degrees of decolorization in the staining process, producing the variation of chromosome sizes.