In this experiment, a 3 molting (Tester-3) and a 5 molting (Tester-5) strain of silkworm were used as material. In both strains molting characteristic is not strictly fixed, but varies according to environmental conditions, i.e. in one and the same batch of worms there occured 3 molters and 4 molters in the former and 4 molters and 5 molters in the latter. The author attempted to investigate the relationship between the moltinism revealed in the selected silkworms by means of starvation and the resistance of the silkworms to various pathogenous factor. First instar silkworms after 70% of the total feeding period had passed stage were starved and molted worms were separated from non-molted ones. The molted worms were collected and used for further investigation. The number of moltings in the later stages and in those of the next generation of test animals were compared with those of the untreated controls.<BR>The results obtained were as follows:<BR>1) In Tester-3, after starvation in the late first instar, the ratio of worms which molted 4 times after starvation to those which molted 3 times was higher than in the controls, while in Tester-5, the ratio of 5 molters to 4 molters was higher than in the controls.<BR>2) The next generation obtained from the treated silkworms showed a higher proportion of 4 molters in Tester-3 and 5 molters in Tester-5, respectively, than those in the next gene ration of controls.<BR>3) As regards to the resistance which was measured by moltinism and the rate of pupation (each of which is regarded as one of the criteria for judging normal health) 3 molters were found to be more healthy than 4 molters in Tester-3, and 4 molters were more healthy than 5 molters in Tester-5.<BR>It was concluded from the above mentioned results that:<BR>a) Selection by starvation in the late first instar results in the increase of the number of molting.<BR>b) The reason for the above is considered to be that the 4 molters separated out from Tester-3 and the 5 molters from Tester-5 have a shorter duration of the first stage than the 3 and 4 molters of each strain. That is to say, 4 molters can pass the critical molting poinr earlier than 3 molters, and 5 molters earlier than 4 molters. When the starvation began at the same time in 3 molters and 4 molters in Tester-3, a larger number of 4 molters entet the first molt than 3 molters and more 5 molters than 4 molters in Tester-5. Thus more 4 molters in Tester-3 and 5 molters in Tester-5 are picked up.<BR>c) The selection for resistance to starvation gave a negative result, i. e. obtaining of a weak train instead of healthy, because the selection by means of starvation led to the increase of the number of molting and subsequently weakness.