Food availability modulates survival, reproduction and thereby population size. In addition to direct effects, food availability has indirect effects through density of conspecifics and predators. We tested the prediction that food availability in isolation affects reproductive success by experimentally manipulating food availability continuously for 3 years in zebra finches (Taeniopygia guttata) housed in outdoor aviaries. To this end, we applied a technique that mimics natural variation in food availability: increasing the effort required per food reward without affecting diet. Lower food availability resulted in a slight delay of start of laying and fewer clutches per season, but did not affect clutch size or number of offspring reared per annum. However, increasing foraging costs substantially reduced offspring growth. Thus, food availability in isolation did not impact the quantity of offspring reared, at the expense of offspring quality. Growth declined strongly with brood size, and we interpret the lack of response with respect to offspring number as an adaptation to environments with low predictability, at the time of egg laying, of food availability during the period of peak food demand, typically weeks later. Manipulated natal brood size of the parents did not affect reproductive success. Individuals that were more successful reproducers were more likely to survive to the next breeding season, as frequently found in natural populations. We conclude that the causal mechanisms underlying associations between food availability and reproductive success in natural conditions may be more complex than usually assumed. Experiments in semi-natural meso-populations can contribute to further unravelling these mechanisms.