Many hypotheses have been proposed to explain multiple mating in females. One of them is bet hedging, that is avoiding having no or very few offspring in any given generation, rather than maximizing the expected number of offspring. However, within-generation bet hedging is generally believed to be an unimportant evolutionary force, except in very small populations. In this study, we derive predictions of the bet-hedging hypothesis for a case in which local insect populations are often small, offspring performance varies, for example, due to inbreeding depression, and the groups of gregarious larvae have to exceed a threshold size before they are likely to survive throughout the larval stage. These conditions exist for populations of the Glanville fritillary butterfly (Melitaea cinxia), potentially making bet-hedging benefits larger than usual. We observed matings in a field cage, which allowed detailed observations under practically natural conditions, and analyzed genetic paternity of egg clutches laid by females under direct observation. The egg-laying and survival patterns are in line with the predictions, supporting the hypothesis that multiple mating in M. cinxia presents a rare case of within-generation bet hedging.