One of the key tenets of life-history theory is that reproduction and survival are linked and that they trade-off with each other. When dietary resources are limited, reduced reproduction with a concomitant increase in survival is commonly observed. It is often hypothesized that this dietary restriction effect results from strategically reduced investment in reproduction in favor of somatic maintenance to survive starvation periods until resources become plentiful again. We used experimental evolution to test this "waiting-for-the-good-times" hypothesis, which predicts that selection under sustained dietary restriction will favor increased investment in reproduction at the cost of survival because "good-times" never come. We assayed fecundity and survival of female Drosophila melanogaster fruit flies that had evolved for 50 generations on three different diets varying in protein content-low (classic dietary restriction diet), standard, and high-in a full-factorial design. High-diet females evolved overall increased fecundity but showed reduced survival on low and standard diets. Low-diet females evolved reduced survival on low diet without corresponding increase in reproduction. In general, there was little correspondence between the evolution of survival and fecundity across all dietary regimes. Our results contradict the hypothesis that resource reallocation between fecundity and somatic maintenance underpins life span extension under dietary restriction.