The large-scale (km) distributional patterns of juvenile bivalves are established by larval settlement and subsequently changed due to actively initiated postlarval migrations (byssus-drifting), resuspension during sediment disturbance, and local differences in mortality. Repeated mapping of 0-group bivalve distribution during two summers was combined with simultaneous registrations of the numbers of drifting specimens. Species differed in their susceptibility to passive resuspension and their activity in byssus-drifting, but the two ways of entry into the water column were independent of each other. As a result, the relative magnitude of byssus-drifting and passive resuspension varied with the species. While hydrographical conditions always determined the changes in the distributional patterns of passively eroded species, this was not the case in active migrations of byssus-drifting bivalves. A comparison of the distributional patterns of juveniles over two consecutive years showed similar patterns in the tellinid clam Macoma balthica and the razor clam Ensis americanus. In these two species spatfall mainly occurred around mean low-tide level. Subsequently, juvenile M. balthica rapidly accumulated in the upper intertidal, whereas juvenile E. americanus accumulated subtidally. In mussels Mytilus edulis the distribution of juveniles only changed in a longshore direction, not in the tidal level occupied. The spatial pattern of cockles Cerastoderma edule changed from aggregation of the early spat in the mid and lower intertidal towards a more uniform distribution of recruits over these tidal flats. There were only a few patches of high abundance left at the end of summer. Finally, in juvenile clams Mya arenaria the spatial patterns of both spatfall and redistribution were unpredictable. On a km-scale, abundance of some species correlated with sediment granulometry. Presumably, this does not reflect a causal relationship but is a consequence of the correlation between grain size distribution and tidal level within the area studied. On a scale of hundreds of metres, there were no consistent correlations between juvenile abundance and sediment granulometry in any species. In summary, it seems that hydrography ruled the initial settlement of larvae to the sediment and strongly influenced the subsequent redistribution of juveniles caused by passive resuspension. The outcome of actively initiated migrations, on the other hand, was only weakly influenced by hydrography. Therefore it is suggested that hydrography plays a dominant role in the initial development of km-scale distributional patterns of just-settled bivalves in the Wadden Sea, while habitat selection is delayed to the byssus-drifting postlarvae phase.