Predation regulates the survival of mallard nests as revealed by an experiment in two landscape types

Ecological theory frequently postulates that most animal species are subject to density-dependent mechanisms in some stage of the year or life. However, few avian studies have succeeded to give evidence for regulatory mechanisms by replicated experiments in natural habitats. Breeding success is to a great extent determined by conditions during the nesting period and the major cause of nesting failure is nest predation. In addition, predators may possibly regulate fluctuating populations and to test the hypothesis that nest survival is negatively density-dependent we conducted an experiment with manipulated densities of semi-natural Mallard nests and adult pairs using 32 small to medium sized lakes (~3-5 hectares) in southern Sweden during two nesting seasons (2003-2004). Predicting that predation rates are higher in open compared to more vegetated habitats, half of the lakes were in agricultural and the other half in forest landscapes. Using real Mallard eggs, nests were constructed in two densities; either 2 nests per lake (low density) or 8 nests per lake (high density). Model fitting in program MARK revealed that nest survival was negatively affected by nest density but not by pair density. Further, predation rates were much higher in agricultural landscapes than in forested. The effects of nest density and habitat were consistent in the two years. Covariates added to the model matrix revealed a negative effect of other waterfowl present on the lakes and a positive effect of study day. However, the number of avian predators observed at the lakes did not affect survival. Our experiment is the first to demonstrate density-dependent nest predation in Mallard. The consistency of the density effect between years and habitats shows that regulatory mechanisms may be truly significant for Mallard, possibly affecting population dynamics of the species.