It is generally assumed that cryptic prey, such as birds' nests, should be over-dispersed as a defence against predation. This suggestion is based on the idea that predators intensify their searching effort in an area around prey items they have found and on frequent observations of predation rate going up with increasing prey density. However, for this model to be acceptable, one has to show that predation risk for a given item is related to the distance to its nearest neighbour. Contrary to this prediction, predation risk on randomly dispersed nests of mallard, Anas platyrhynchos, was unrelated to nearest neighbour distance. The selective pressure for over-dispersion of nests due to predation will be related to the probability of having neighbours too close, which depends on both nest density and the searching habits of the predators. The disadvantage of random dispersion therefore decreases at decreasing density. Thus predation may have been overrated as a factor causing over-dispersion of nests, especially at low nest density; several other factors may have a greater influence on the dispersion patterns of nests. © 1991.