We used imprinted ducklings and wing-clipped female mallards with broods as a complement to observational field studies. Because mallard ducklings search for food by themselves, the food searching efficiency and behaviour of imprinted ducklings and ducklings together with a wing-clipped female ought to reflect the general conditions in the lake they search for food, if not in a strict quantitative way, at least it should reflect the relative differences in conditions for the lakes used in the experiments. We were interested in why mallards start breeding so early and how climate change might influence early and late breeding ducks. In a cross over experiment conducted 2004 and 2005, we studied the result of releasing two sets of female mallards with ducklings. The first one was done close to the same time as wild mallards normally started their breeding season in the region. The other set of broods was delayed 12 days. The eggs for the late release of ducklings were kept at low temperature and 12 days later they were put under sitting mallard females. The hens and broods were subsequently released with an interval of 12 days (one brood in each of 10 lakes in the first release, and the late release in ten other lakes). The next year the lakes were reversed. Survival of ducklings was followed every day during the first 12 days, and later every 4 days up till 24 days. Invertebrates in the lakes were sampled as well (of both benthic and pelagic origins). The data are analyzed in relation to food abundance during the early and late part of the two breeding seasons.