The diets of larval gulf menhaden Brevoortia patronus, spot Leiostomus x d n t h u~s , and Atlantic croaker Micropogonias undulatus, collected in the northern Gulf of Mexico in December 1979, February and December 1980, and February 1981, were significantly (P<0.001) different when compared with respect to length by step-wise discriminant analysis, and there was little overlap when the diets of size-specific larvae that CO-occurred in discrete collections were compared by percentsimilarity. Gulf menhaden larvae had a more diverse diet that included phytoplankters (diatoms and dinoflagellates) as well as zooplankters (tintinnids, pelecypods, pteropods, and all stages of copepods). The diets of larval spot and Atlantic croaker were largely restricted to zooplankton. Overall diet distinctiveness, the lack of diet overlap, and the lack of small-scale CO-occurrence indicate that the larvae of these three species do not compete for food.

The distribution of larval fishes is shaped by the Mississippi River plume at both coarse (kilometers) and fine (tens to hundreds of meters) spatial scales. Density estimates of larval fishes (No. m-3), based upon ichthyoplankton samples collected about the Mississippi River plume in February 1982, December 1982, and November 1983, were often greater by a factor of 10, and sometimes by several orders of magnitude, at the plume front than they were inside (within) or outside of the plume. Greater densities at the plume front were most apparent for larval Brevoortia patronus (a clupeid) and unidentified clupeid larvae. Frontal convergence apparently affects the accumulation of larval fishes at the frontal interface, but high densities were not evident everywhere along the front. An index of potential convergent velocity indicates maximal velocities of the order of 0. lo-0.30 m s-l, but this index does not explain variation in the accumulation of larval fishes. Frontal convergence may be localized and affected by secondary circulations.Planktonic organisms are typically aggregated in the vertical and horizontal dimensions on a variety of temporal and spatial scales. This aggregation is influenced by biological characteristics of these organisms and by physical fc:atures of their habitat (Haury et al. 1978; Longhurst 198 1; Owen 198 1;Mackas et al. 1985). The propagation of plankters, their behavior, and their buoyancy interact with the physical hydrographic structure and hydrodynamic motion of the ocean to affect aggregation. Propagation contributes to aggregation on most temporal and spatial scales; e.g. the cell division of phytoplankters, the reproduction of zooplankters, and the spawning of fishes produce aggregates of propagules over various scales that depend on the taxa. Interaction of buoyancy and behavior of plankters with hydrographic structure is effective more often in shaping spatial distribution at miAcknowledgments We acknowledge and appreciate the cooperation of S. Cummings for MOCNESS operations, the officers of the Oregon II for their navigation in the cluttered and busy waters of the northern Gulf of Mexico, and a squad of technicians for plankton sorting. J. M. Brubaker offered valuable counsel on the hydrodynamics of frontal convergence. P. B. Ortner reviewed the manuscript.T. Leming supplied the satellite images.

No abstract

Herring (Clupea harengus L.) show a characteristic 'startle' response when subjected to vibrational stimuli from a diaphragm in the wall of their tank. Threshold measurements on fish from 2-8 to 17 cm total length tested at frequencies from 70 to 200 Hz showed that the response was elicited by sound pressures between 2 and 18 Pa, the most sensitive fish being in the length range from 8-11 cm. Intermediate-sized fish of 12-13 cm also responded to sounds from a loudspeaker in air above the tank, the mean threshold being about 5 Pa. The stimulus was thought to be the sound pressure rather than the particle velocity component of the stimulus, the gas-filled pro-otic bulla acting as part of the pressure detecting system. Pre-treatment of fish by shock pressure increases of 4 atm to burst the bulla membrane increased the threshold about ten times. The bulla membrane is coupled hydrodynamically with the head lateral line system via a lateral recess membrane. Decoupling of this link by cutting a window over the lateral recess did not affect the response much, nor did treatment of the lateral line by methylene blue or streptomycin as blocking agents.