Dietary Restriction (DR), a reduction in nutrients without starvation, has been well known for nearly a century to reduce reproduction and extend lifespan. This surprising lifespan extension effect has been demonstrated in numerous species across diverse eukaryotic taxa, but we are still a long way from understanding what drives the response and what, if any, is its evolutionary significance. A prominent hypothesis suggests that the life-history responses to DR evolved as an adaptive strategy whereby resources are redirected from reproduction to somatic maintenance, to promote survival during resource scarcity. In line with this idea, it has been proposed that DR animals may adaptively reduce activity rates and mating behaviours, such that lifespan extension could be due in part to reduced wear and tear and avoidance of hazards, like predation and exposure to injury. However, evidence for this idea has been equivocal, and available data has come only from laboratory studies, in which hazards are largely absent and behaviours are likely to be constrained by the artificial environment. Here we present new data from a study on the water strider Gerris buenoi, in which we show that DR applied in a semi-wild environment, although reducing reproduction, fails to elicit the extended lifespan or reduced ageing responses that are well documented in the same species in the lab. This suggests that the lifespan extension effect of DR is likely to be a laboratory artefact, rather than an adaptive response to resource scarcity that evolved in early eukaryotes. Moreover, we show that while diet drives some behavioural differences in mating rates, foraging rates and refuge use, DR does not appear to reduce exposure to predation, as reinforced by autopsy data. These results suggest diet-mediated behavioural modification is unlikely to be driven by selection for extended survival during famine.