Abstract: Anopheles sinensis were fed through membrane on blood from five patients naturally infected with vivax malaria in the southern pfert of Yunnan Province. After feeding at 8-hour intervals for 48 hours., all unfed mo'squitoes were removed from cages and the engorged ones were incubated at 26±1℃ and 70-80% RH on 10% glucose. 13 to 100 mosquitoes from each batch were dissected according to the level of infection 6-7 days later and the number of oocysts were recorded. The percentage of mosquitoes infected and the mean number of oocysts per infected gut were taken as critaria of the viability of the gametocytes. Both thick and thin, blood films were taken at the time when mosquitoes fed, and the number of asexual parasites and gametocytes were counted by the routine method.The general conclusions drawn from our work are: (1) The viability period of the gametocytes was demonstrated, but the nocturnal periodicity of morphologically mature gametocytes, as previously noted by Hawking et al., was not regularly observed. It is doubtful whether Hawking's work on monkey infections could be applied to human Plasmodium vivax. (2) There is no marked and definite correlation between the infecti-vity of mosquitoes and the total parasite count or gametocyte density. (3) The infectivity of gametocytes for mosquitoes reached a peak at the interschizogony period and fell to low level at the time when schizogony occurred. (4) The time required for the rise and decline of gametocyte viability is little longer than the time required for an schizogony cycle.