Larvae of coastal gastropods sink in turbulence and may use nearshore turbulence as an initial settlement cue. Our objective was to quantify the relationship between turbulence and the proportion of sinking larvae for competent mud snail veligers (Ilyanassa obsoleta). We exposed larvae to a range of field-relevant turbulence conditions ( ϭ 8.1 ϫ 10 Ϫ3 to 2.7 ϫ 10 0 cm 2 s Ϫ3 ) in a grid-stirred tank, holding other factors constant. We used a video plankton recorder to record larval movements in still water and in turbulence. Larval trajectories and velocity measurements were extracted using video-image analysis. We also measured turbulent flow velocities independently, using laser Doppler velocimetry. To interpret empirical measurements in terms of larval behavior, we developed a threecomponent, normal mixture model for vertical velocity distributions of larvae in turbulence. The model was fitted to observed larval velocities by maximum likelihood, to estimate the proportions of sinking, hovering, and swimming larvae. Over the range of turbulence intensities found in typical coastal habitats, the proportion of sinking larvae increased exponentially (r 2 ϭ 0.89) with the log of the turbulence dissipation rate. The net mean behavioral velocity of the larvae shifted from positive to negative when the dissipation rate reached ϳ10 Ϫ1 cm 2 s Ϫ3 . By sinking when they enter turbulent, shallow water, competent larvae could improve their chances of settling in favorable coastal habitats.Very little is known about how larval behavior in the plankton affects patterns of larval supply and settlement of benthic invertebrates. Much work has been done to describe larval behavior during the exploration of substrates, when larvae can sometimes select settlement sites over small scales (millimeters to centimeters). Less progress has been made on understanding the behavioral contribution while larvae are transported through the water column to benthic habitats. Under some hydrodynamic conditions, larvae could settle more successfully if they responded to waterborne cues by sinking toward the benthos. If the swimming velocity and gravitational sinking velocity differ by a factor of two or more, behavioral changes can significantly affect larval sink-1 Corresponding author (hfuchs@whoi.edu).
AcknowledgmentsY. Yamashita helped collect and culture the larvae. We are grateful to J. Sisson for assistance with LDV measurements and to J. H. Trowbridge for guidance on spectral analysis. B. Raubenheimer and E. A. Terray also gave advice on flow data analysis. We thank C. DiBacco for generously sharing his culturing expertise and supplies, V. R. Starczak for advising us on the experimental design, and S. P. McKenna for introducing us to the turbulence tank. S. M. Gallager provided video equipment, software, and advice on particle tracking.