In the production of the infective cercariae of trematodes, the terminal investment hypothesis of life-history theory predicts that the rate of host exploitation and cercarial production should increase during the period of cercarial shedding since the reproductive value of the parasite decreases during this period. In contrast, a bet hedging hypothesis that focuses on the success of transmission when host contact rate is variable predicts that cercarial production should decrease in an attempt to keep the host alive for longer and thus would increase the probability of successful transmission. We examined these two hypotheses under laboratory conditions and recorded the production of Diplostomum spathaceum cercariae from naturally infected snail hosts, Lymnaea stagnalis. The average number of cercariae produced per day decreased as the snail host approached death counter to the terminal investment hypothesis. The finding supports the prediction of the bet hedging hypothesis and implies that the pattern of cercarial production may be explained by reduced virulence of the parasite within the snails to ensure extended total production time of cercariae. Nevertheless, survival of infected snails was still lower than uninfected snails suggesting that ultimately the infection still increased snail mortality rate. Cercarial production varied between days but was not cyclic, probably because of the physiology of the sporocysts within snails. Fewer cercariae were released at night, which may increase transmission efficiency to diurnally-active fish hosts. The mechanisms associated with daily cercarial production are discussed.