Despite the morphological differences of adults, larval bivalves are difficult to identify optically. Multidimensional scaling allowed us to see ontogenetic and species-specific developments in shape. With discriminant analysis on larger individuals (length over 150 micrometer), up to 74% of the larvae could be predicted correctly.

Diet quality (measured as polyunsaturated fatty acid (PUFA) content) of adult bivalves influenced breeding success, while the effect of enrichment of larval diet was minimal. Implications for field studies: monitoring chlorophyll content is not sufficient to predict reproductive output of bivalves and development of their larvae. Algal species differ in PUFA content and it is important which algae are available as food. No effect of PUFA supplementation on larval growth, development or mortality was found. This could be an artefact due to unsuitability for larval uptake of the fed lipid emulsion. However, food quantity for larvae may well be a determining factor for the survival of larvae in the field. Comparing maximal assimilation rate of an individual larva with the available palatable energy in the water showed that a larva never achieves maximal energy intake and thus can never grow at maximum growth rate.

The larval settlement process is difficult to study in the field. Laboratory (flume) studies are an alternative when similarity of scale is maintained. Apart from the Reynolds number a dimensionless scaling parameter of importance for settlement studies: the ratio between advection and turbulent mixing in the water column (Péclet number, or Rouse number). We manipulated this ratio by means of a grid throughout the water column and roughness structures on the substrate. Shear velocity and roughness height change when turbulence is artificially increased. Turbulence affects (lowers) the Péclet number. The Péclet number in the flume under normal circumstances is high, pointing at a bias towards sinking under laboratory conditions in the flume compared to similar flow velocities in the field. The settlement phase for weakly swimming bivalve larvae could be more dependent on active behavior in the field than presumed according to flume studies. In field circumstances turbulence and advection could be roughly balanced and (vertical) swimming or deliberate sinking could make a difference in settlement probability for small bivalve larvae. Resuspension possibility can be influenced by active behavior as well (byssus-threads) which gives the larvae even a larger window for active choice. Flume studies and modeling exercises confirmed the influence of turbulent mixing on settlement. At low mixing constants, there was a rapid increase in settlement up till the critical resuspension point, after which erosion caused loss of settled particles. If it takes a larva on average several minutes to travel through the water column, while the horizontal flow velocity is around 0.1 meters per second, then the horizontal scale of patch choice should be in the order of magnitude of tens of meters. Virtually any flume tank would be too small to investigate this, field experiments are needed. Small-scale patterns in larval settlement must be the result of a smaller-scale processes, e.g. in the viscous sublayer and secondary settlement. Particle Image Velocimetry (PIV) was evaluated against Acoustic Doppler Velocimeter (ADV) measurements. Close to the boundary, PIV measurements show lower flow velocities than expected from ADV profile extrapolations. Even under reasonably high flow velocities (0.15 meters per second) and increased turbulence levels, we find a viscous sublayer. Observed differences in turbulence intensity in the logarithmic part of the boundary layer measured with an ADV extinguish close to the substrate resulting in similar conditions with dominant viscous forces in near-bottom layers under different turbulent conditions in the water column. PIV measurements indicate reasonably high shear stresses present close to the bottom. This could be an argument for the entrapment of larvae in the viscous sublayer.

email address: irishendriks@irishendriks.nl