Within populations of the seed beetle Callosobruchus maculatus (F.) (Coleoptera: Bruchidae), female longevity is positively correlated with fecundity, both phenotypically and genetically. Yet manipulations of egg-laying rate (through deprivation of seeds or mates) consistently suggest a trade-off between these traits. In this study, females were exposed to four levels of seed availability and the relationship between fecundity and longevity was examined both across and within environments. The expected trade-off was observed across environments, as females with few or no seeds laid fewer eggs and lived longer than females with many seeds. Within environments, however, the relationship was inconsistent; individual longevity and fecundity were positively correlated when seeds were abundant, uncorrelated when seeds were few and negatively correlated when seeds were absent. Body mass at adult emergence was positively correlated with fecundity, but its effect was stronger when seeds were present than when they were absent. After the effects of mass were statistically controlled, longevity and fecundity remained positively correlated among females with many seeds and negatively correlated among those without seeds. Thus, even a single technique (phenotypic correlation) can produce divergent conclusions about the ‘cost of reproduction’ in insects. The reversal in the correlation appears to depend on variation in reproductive allocation; when seeds are scarce or absent, differences in reproductive effort among individuals may be large enough to overwhelm differences in resource acquisition.