Interactive effects of haloclines and food patches on the vertical distribution of 3 species of temperate invertebrate larvae

In laboratory experiments, we examined the effect of haloclines and determined whether the presence of food patches overrides this effect on larval vertical distribution of the sea star Asterias rubens, the sea urchin Strongylocentrotus droebachiensis and the mussel Mytilus edulis. We experimentally constructed haloclines in which the salinity of the bottom water layer was 35 and that of the top layer was 21,24,27, and 30 (21/35,24/35, 27/35, and 30/35) for A. rubens and S. droebachiensis, and 24, 27,30 and 32 (24/35, 27/35, 30/35, and 32/35) for M. edulis. For each species and stage, additional halocline treatments (A. rubens: 24/32 and 27/32: 4-arm S. droebachiensis: 21/29 and 24/32; 6-arm S. droebachiensis: 24/29 and 24/32: M. edulis: 27/32 and 30/32) were used to determine whether the larval response to inhibitory salinity gradients was due to the absolute salinity of the top layer or the relative salinity difference between the two layers. Also, we measured the density of A. rubens and M. edulis to determine whether the specific gravity of larvae can explain the observed vertical distributions. Larvae aggregated at and below the halocline and these aggregations were more pronounced with increasing strength of the vertical salinity gradient. Threshold salinities in the top layer which inhibited similar to 100% of the larvae from crossing the halocline were 24 for A. rubens and M. edulis, and 21 for S. droebachiensis. These distributional patterns were not the result of larval density, which was greater than all treatment water densities for M. edulis and S. droebachiensis and lower for A. rubens. The effect of the presence of a food patch at inhibitory haloclines (A. rubens: 24/35 and 27/35; 4-arm S. droebachiensis: 21/34 and 24/34; M. edulis: 27/35) was determined by using three algal densities: 0, 5000 or 10 000 cells ml(-1) Thalassiosira pseudonana in either the top or the bottom water layer. For both A. rubens and M. edulis, the number of larvae at the halocline increased in the presence of a food patch, but this effect did not depend on algal density in the patch. For 4-arm S. droebachiensis, there was no effect of the presence of a food patch on larval vertical distribution. Our results suggest that low salinity may act as a barrier to vertical movement and that the presence of food patches above the halocline may strengthen the larval aggregation response to inhibitory haloclines. (C) 2008 Elsevier B.V. All rights reserved.