FISHERY BULLETIN: VOL. 81, NO. 3 



1956), available in the IMSL Library's 3 statistical 

 package. Variances of catches were heteroscedastic 

 among depth blocks and between diel periods, which 

 precluded use of parametric analysis of variance. 

 Wilson's three-way ANOVA's (either with unequal 

 replication or without replication) were used, 

 depending on the category (e.g., total individuals or 

 species counts) or species being considered. 



Quantitative clustering of species was carried out 

 using the Ecological Analysis Package (E AP) 4 . Mean 

 abundance of species by cruise date, depth, and diel 

 period was transformed by its square-root in order to 

 counter the tendency of the Bray- Curtis Index (Clif- 

 ford and Stephenson 1975) to overemphasize abun- 

 dant species. Flexible sorting was used to maximize 

 the separation between groups. Only species with a 

 minimum total occurrence of 20 were considered in 

 this analysis. 



Spearman's rank correlations (rj were calculated to 

 examine the relationships between diel period, water 

 depth, and surface water temperature and CPUE for 

 1) total individuals, 2) total individuals minusE. mor- 

 dax, 3) certain major species, and 4) species counts. 

 CPUE data for taxonomic categories 1-3 were 

 related to water clarity by using Spearman's rank test 

 within each depth block; in addition, a parametric 

 analysis of covariance (ANCOVA), with water clarity 

 as covariate of CPUE and depth block as the treat- 

 ment effect, was performed to estimate the general 

 magnitude of the potential influence of depth- 

 specific differences in water clarity on daytime 

 catches. Diel catch efficiency data for Seriphus 

 politics were related to water clarity by parametric or 

 Spearman rank correlation, as appropriate. 



Day-Night Comparison of 

 Foregut Fullness 



Preliminary examination of the variance: mean 

 ratios of CPUE data indicated that fishes were dis- 

 persed more at night than during the day. In order to 

 investigate the possible role of feeding behavior in 

 this nocturnal dispersal pattern, we examined 

 archived stomachs from paired day (1000-1300 h) 

 and night (2100-0100 h) samples of fishes of compa- 

 rable sizes for the five most abundant taxa (E. mor- 

 dax; Seriphus politus; white croaker, Genyonemus 

 lineatus; Pacific butterfish, Peprilus simillimus; and 

 atherinid spp., the latter represented by jacksmelt, 



J IMSL Library, Sixth Floor, NBC Building, 7500 Bellaire Blvd., 

 Houston, TX 77036. 



4 Robert W. Smith (author of EAP), Ecological Data Analysts, 1151 

 Avila Drive, Ojai, CA 93023. 



Atherinopsis californiensis). Contents of the foregut 

 were removed, either dried at 40°C (for E. mordax, S. 

 politus, and G. lineatus) or blotted dry (P. simillimus 

 and A. californiensis), and weighed. A contents index 

 (CI) was then calculated for each specimen as 

 follows: 



CI = (weight contents/weight of fish) X 10 5 . 



Diel overlap in gut evacuation did not create a major 

 problem except that it tended to make the analysis 

 more conservative (i.e., more difficult to detect day- 

 night differences). 



The CI's for day-caught versus night-caught fishes 

 were compared by either Wilcoxon signed-ranks test 

 for paired comparisons or Wilcoxon two sample test, 

 depending on the number and temporal distribution 

 of samples. 



RESULTS 

 Species Composition 



Sixty-two taxa representing 33 families of teleost 

 and elasmobranch fishes were collected in 643 net- 

 hauls partitioned among the three depth blocks and 

 two diel periods made during the 19-mo period, Sep- 

 tember 1979-March 1981 (Table 1). The catch was 

 overwhelmingly dominated by E. mordax. Seriphus 

 politus, G. lineatus, P. simillimus, and a species com- 

 plex of silversides (atherinid spp.) were also abun- 

 dant in the catch. These top five taxa accounted for 

 >98% of the numbers of total individuals sampled 

 (Table 1). The atherinid species complex was a com- 

 posite of three species (Atherinopsis californiensis; 

 topsmelt, Atherinops affinis; California grunion, 

 Leuresthes tenuis) that were not readily identifiable in 

 the field. Subsamples of "atherinid spp." field 

 catches were about 48% Atherinopsis californiensis, 

 42% L. tenuis, and 10/J Atherinops affinis. 



Location Comparison 



Location differences were insignificant (P > 0.05) 

 for most categories and species within depth blocks 

 and diel periods based on Mests for paired com- 

 parisons. Differences were detected in the following 

 cases: 1) Atherinids were more abundant at night in 

 the upcoast area (P < 0.01); 2) Pacific barracuda, 

 Sphyraena argentea, occurred in greater numbers at 

 night upcoast (P < 0.01); 3) T. symmetricus was more 

 abundant both day and night upcoast (P< 0.01); and 

 4) Scomber japonicus was also caught in greater num- 

 bers upcoast (P < 0.05), but only at night during the 



572 



