Most models of parental investment in offspring assume a trade-off between propagule size and number, and an increasing "concave down" function relating offspring fitness to propagule size. In this study, we test these two fundamental assumptions, using three closely related species of crickets, Gryllus firmus, G. veletis, and G. pennsylvanicus. Egg weight, 35-day fecundity and 35-day egg biomass were estimated in a population of each species, and the relationships between these reproductive traits and date of egg laying and body size were estimated. The relationships between egg weight and offspring survival were also sought for eggs buried at different depths, soil moistures, and soil types (G. firmus and G. veletis), as well as in the field (G. pennsylvanicus). A trade-off between egg weight and 35-day fecundity was revealed in a multivariate analysis taking into account among-species variation in egg weight and body size. Independent of the environmental conditions affecting the eggs, a positive correlation existed between the number of larvae that emerged from the soil and propagule weight in each species. Therefore, these findings provide partial support for the assumptions considered in the models mentioned above. A single optimal egg size was favored in two out of the three sets of conditions in which the functions relating egg weight to larval survival could be derived. The conditions encountered by the eggs, however, influenced the average survival of the larvae, as well as the shape of the relationship between egg weight and offspring survival. This suggests that cricket eggs frequently face heterogeneous environments with respect to egg and hatchling survival; the implication of habitat heterogeneity on the evolution of an optimal egg size is considered. The relationships between the reproductive components and female age and size, as well as between egg size and variation in cricket life-history, are discussed in an ecological and evolutionary context.