The hypothesis that larval herring Clupea harenyus select food by type as well as size was tested in laboratory experiments. Herring larvae were reared at 7 to 9'C on wild zooplankton. The percentage of herring larvae with food at day's end increased from 4 % (4 d) to 68 % (9 d) and averaged 83 % for larvae 25 to 74 d old. Larval herring selected particles according to type as well as size; this selectivity varied with larval age and hence size. Copepod nauplii and copepodites were preferred by larvae of all sizes. Among copepodites, Pseudocalanus sp, and Oithona sp. were preferred by smaller and larger larvae, respectively, while Acartia sp. was rarely ingested, even when of acceptable size. Mollusc veligers comprised a significant portion of the diet of young (4 to 24 d) larvae but were actively rejected by older larvae even when perceived and of acceptable size. Particles smaller than the largest acceptable size were consistently preferred, atypical of predation by vertebrates. These results indicate that larval herring select prey according to type as well as size and that this behavior is acquired through experience.

A deep aggregation of fifth copepodid (C5) Calanuspacificus in the Santa Barbara Basln (SBB) was mapped over yr w~t h zooplankton net tows, a n optical plankton counter, and a moored acoushc Doppler current profller Hlgh concentrations of diapausing C5 C pacificus bullt up in the deep waters of the SBB dunng the summer and into the fall Dunng the buddup the deep aggregation moved up from the bottom as oxygen became depleted In the basin S deep waters The deep aggregatlon apparently bullds u p due to the basin trapping C5s that migrate below the sill depth from water advected over the basin C5s are retalned within the basln until they swim or are forced above the sill depth Possible mechanisms responsible for the dispersal of the deep aggregation are mlgratlon to the surface waters with subsequent dispersal by surface currents advection over the basin s sill due to the buildup of oxygen-deficient water or flushing of the basin's deep water

Vertical mixing induced by storms is hypothesized to modify the spatial and temporal availability of food to fish larvae and thus influence their feeding and growth. We investigated the effects of storms on sagitta growth rates of 2 age classes (3 to 15 d and 16 to 50 d, post-hatching) of larval Atlantic menhaden Brevoortia tyrannus during winter 1986 in Onslow Bay, North Carolina, USA.We tested the null hypothesis that sagitta growth rate of 3 to 15 d and 16 to 50 d old larvae is independent of the timing and intensity of storms by using transfer functions to investigate relationships among daily time series of sagitta growth rate and meteorological and oceanographic variables. Variation in sagitta increment width was greatest during the first 1 to 2 wk after hatching. Reduction in sagitta growth rate coincided with storms and corroborated laboratory results that growth increments are formed daily and stressful events are manifest in sagitta microstructure. Age-related trends in sagitta growth rate were observed for both age classes and were removed from the analysis by a standardization procedure. Fluctuation in sagitta growth rate for 3 to 15 d old larvae was inversely crosscorrelated with time series of wind speed. Fluctuation in sagitta growth rate for 16 to 50 d old larvae was inversely cross-correlated with tune series of wind speed and heat flux The pattern of crosscorrelations for 3 to 15 d old larvae indicated an immediate response of sagitta growth rate to periods of strong winds while sagitta growth rate in 16 to 50 d old larvae lagged strong winds and heat fluxes by 2 to 5 d. Transfer function models incorporating wind speed and heat flux accounted for 40 to 54 % of sagitta growth rate variation. Reduction in larval Atlantic menhaden growth rate during early life is consistent with the critical period concept and may be related to dispersion and aggregation mechanisms coupled to wind-induced vertical mixing and its effect on food concentration and availability.

Interannual and regime (decadal) scale changes in climate affect the spatial distribution and productivity of marine fish species in numerous ecosystems. We analyzed a historical simulation (1965-2000) from an end-to-end ecosystem model of anchovy population dynamics for the California Current System to untangle the effects of warm versus cool conditions on recruitment. A�3-dimensional coupled hydrodynamic-NPZD (nitrogen-phytoplankton-zooplankton-detritus) model (ROMS-NEMURO) provided the physical conditions (circulation, temperature) and 3 zooplankton concentrations as inputs to an anchovy full life cycle individual-based model (IBM). Our analysis was focused on isolating the effects of the well-documented El Niño Southern Oscillation signal and 3 climate regimes on spawning habitat, development, and survival of eggs and yolk-sac larvae, growth and survival of larvae and juveniles, and ultimately recruitment of anchovy. The major drivers of lowered recruitment success in warm years and in warmer regimes were reduced survival and growth rates of eggs and larvae that resulted from the poleward shift of adults in response to warmer temperatures prior to spawning. Three model-data comparisons showed the model deviated from empirically derived values of annual recruitment success but agreed with data for annual mean latitude of egg distributions and predicted larval consumption rates versus measured zooplankton concentrations. More effort is needed to improve certain biological aspects of the IBM so that it can replicate empirically estimated recruitment fluctuations. Overall, the altered responses of anchovy to changing climate in the California Current domain illustrate the benefit of the present mechanistic approach to infer how anchovy may respond under future ecosystem conditions.