JAHN ET AL.: FOOD-SEEKING LARVAL WHITE CROAKER 



Accordingly, the microplankton data set, repre- 

 senting six vertical profiles separated in time, 

 was reduced to principal components for exami- 

 nation of possible time effects. 



Twenty-four taxonomic/size categories of mi- 

 crozooplankton were used to compute principal 

 component scores for the 24 samples. The first 

 four principal components accounted for 60% of 

 the variance. No clear separation of midwater 

 from near-bottom samples was seen. The first 

 component, which accounted for 22% of the vari- 

 ance, separated the near-bottom samples into two 

 groups, morning to midday and afternoon (Fig. 1), 

 leaving the midwater samples at intermediate 

 projections. The midwater samples were in turn 

 separated by the fourth component (11% of the 

 variance) into time groups corresponding to those 

 of the near-bottom set. No stratification by sam- 

 pling height was seen within the near-bottom 

 samples, and none of the other axes provided sep- 

 aration by time. The highest loading variables on 

 components 1 and 4 (Table 1) were various sizes of 

 rotifer and, for component 1, three genera of 

 tintinnids (Favella, Acanthocystis , and Da- 

 dayella ). Much of the time-correlated variance 

 structure depicted in Figure 1 thus appears to be 

 due to change in the size composition of rotifers, 

 described in a later section, and a decrease in 

 these three tintinnids near the bottom in late af- 

 ternoon (Table 2). An identical analysis of the 



Table 1. — Loadings of important variables on tfie first and fourth 

 principal components of microplankton data. 



Component 1 



Component 4 



phytoplankton data found no trends in time or 

 depth. 



Larval Fish Abundance 



Of 1,125 total fish larvae taken in the six pump 

 samples, 666 (59%) were white croaker, a deep- 

 bodied, robust larva (Watson 1982). More than 

 half (338) of these had absorbed the yolk sac and 

 were thus of feeding size. The second most abun- 

 dant feeding-stage larva was an unidentified gob- 

 iid type (84 specimens), but this taxon was not 

 taken above 100 cm of the seabed and so was 

 excluded from the gut analysis. Feeding-stage 

 California sardine, northern anchovy, and Cali- 

 fornia halibut — all relatively abundant O0.2 

 m""^) in the area three weeks earlier — each repre- 

 sented <1% of the catch. Although the earlier 

 survey employed oblique bongo net tows, past 

 comparison of the Nielsen pump with bongo tows 

 found no significant differences in diversity or 

 abundance estimates based on similar-volume 

 samples (R. Schlotterbeck^). We therefore think 

 the differences between the February survey and 

 our March samples were due mainly to a real 

 change in the ichthyoplankton, from a typical 

 late winter assemblage (McGowen 1987; Walker 

 et al. 1987) to a more depauperate one. 



Vertical Distribution and Feeding 

 Incidence of Larval White Croaker 



White croaker free embryos ranged in abun- 

 dance from <0.1 m"'^ at 0.5 m to ~1 m"-^ at 1 m 

 to >2 m-3 at 6.7 m above the bottom. Of 61 free 

 embryos dissected, none had gut contents. 



Feeding-stage larvae of white croaker were 

 only slightly more abundant at 6.7 m (1.9-2.2 

 m"3) than at 1 m and 0.5 m (1.1-1.6 m'^), but 



3R. Schlotterbeck, Robert Schlotterbeck, Inc., 18842 

 Ridgeview Cr., Villa Park, CA 92667, pers. commun. April 

 1986. 



Table 2. — Density (cells per liter) of three tintinnids as a function of time and 

 sampling height. Each set of three numbers gives the density of Favella spp. (F), 

 Acanthocystis spp. (A), and Dadayella spp. (D). 



255 



