The choice of a mussel bed as a settling locality by conspecific mussel larvae is a trade-off between reduced fitness due to an increased risk of larval predation and post-settling food competition with the filtering adults and the benefit from a reduced post-settling mortality. This reduced post-settling mortality may be due to a reduced benthic predation in habitats with high complexity. In a field experiment, the larval settling of blue mussels, Mytilus edulis, was recorded on an artificial substrate 0.25, 1 and 2 m above the bottom during six periods in spring 1999. The experiment was conducted at four stations with dense mussel beds located at two of these stations. Two of the four stations, one with mussel beds and one without, were located in a wind-exposed area and the two other stations in a sheltered area. It was demonstrated that the larval settling was reduced up to 1 m above the mussel beds. A size analysis of the settled mussels indicated that the reduction in settling intensity close to mussel beds was due to a pre-settling process: the larvae were predated by the filtering adult mussels. Settling was significantly affected by wind stress. During periods with a high mean wind velocity and a turbulent water column, the larvae showed a reduced settlement 1 m above a mussel bed relative to 2 m above, whereas the same patterns not could be observed outside a mussel bed. The importance of the filtration activity of the adult mussels and the behaviour of the larvae is discussed. The recruitment of blue mussels into a mussel bed was investigated. Here, significant positive correlations were observed between the density of recruits and density of adult mussels and weights of empty shells. This indicates that the filtering mussels not only increase the mortality of larvae, but also serve as an important substrate reducing post-settling predation from benthic predators.