FISHERY BULLETIN: VOL. 84, NO. 4 



The precision of the duplicate tows was evaluated 

 by comparing numbers of eggs, larvae, and zoo- 

 plankton settled volumes from the May, June, and 

 July tows. No statistical difference was detected 

 between first and second tows (2-tailed P = 0.407, 

 Wilcoxon Matched Pairs test, Hull and Nie 1981: 

 228). The mean coefficient of variation for these 

 paired tows was 0.22. Because there were no statis- 

 tical differences between these duplicates, only one 

 of each pair of the remaining samples was sorted. 



Data Analysis 



Eggs, larvae, zooplankton, microzooplankton, and 

 plankton volume per 1,000 m 3 were calculated 

 based on flowmeter readings. Temperature and 

 salinity stratification variables were created by 

 taking the difference between surface and bottom 

 values. Salinity stratification represented the inten- 

 sity of estuarine circulation or freshwater runoff; 

 temperature stratification represented water 

 column stability and revealed atmospheric tempera- 

 ture extremes. 



Distributions of the variables were examined for 

 skewness, kurtosis, and unreasonable range limits 

 indicative of keypunch errors. Normality of the 

 original variables and of \og(X + 1) transformations 

 was tested (Kolmogorov-Smirnov test; Hull and Nie 

 1981:224). Variances of the transformed variables 

 were not heteroscedastic. Biological and environ- 

 mental variables were plotted against month, sta- 

 tion, and each other to look for spatial patterns, 

 seasonal trends, and nonliner relationships (espe- 

 cially nonmonotonicity) between pairs of variables. 



Analysis of variance (ANOVA) was used to assess 

 the effects of month of the year and station loca- 

 tion on numbers of eggs and numbers of larvae. 

 Stepwise multiple linear regression was used to ex- 

 amine which of the other variables could account 

 statistically for the variability in numbers of eggs 

 and larvae. Logarithmic transformations of stan- 

 dardized numbers of eggs, larvae, zooplankton, and 

 microzooplankton were used in the regressions and 

 in the ANOVA' s. 



Ichthyoplankton abundance is often expressed as 

 numbers of ichthyoplankton under an area of sea 

 surface by multiplying density per cubic meter times 

 water depth (Smith and Richardson 1977). In deep 

 water tows are made below the depth range of most 

 eggs and larvae, so the tow depth is used as the ef- 

 fective water depth. Standardizing a unit of sea sur- 

 face area allows comparisons of total numbers of 

 eggs and larvae in the water column from areas with 

 different water depths. Abundance standardized to 



area of sea surface was used to estimate total egg 

 production. However, larvae that were relatively 

 uncommon in deep water could be as abundant as 

 more concentrated larvae in shallow water, but ex- 

 posed to different concentrations of predators and 

 prey; therefore, densities of larvae and plankton 

 were used to examine relationships between ich- 

 thyoplankton, other plankton, and environmental 

 variables. 



The method used to estimate spawning stock bio- 

 mass was a direct estimate because it incorporated 

 batch fecundity from histological data (Hunter and 

 Goldberg 1980) and daily egg production estimates 

 from ichthyoplankton surveys (Parker 1980). 

 Parker's equation for the direct estimate of biomass 

 from egg abundance is 



5 = P{ab'c)- l d 



Parker (1980) estimated the coefficient of varia- 

 tion of the estimate of spawning stock to be 0.614. 

 Most of this statistical error was due to error in the 

 estimate of egg production. Daily egg production 

 was estimated in my study by dividing the egg abun- 

 dance by the number of days needed to hatch at the 

 ambient temperature (interpolated from Zweifel and 

 Lasker 1976, fig. 7). 



Numbers per square meter of Bay surface were 

 calculated by multiplying density per cubic meter 

 times water depth at the station. The areas repre- 

 sented by the stations were estimated from the chart 

 of the Bay in Conomos and Peterson (1977). Total 

 numbers of eggs and larvae were calculated from 

 estimates per square meter times the area repre- 

 sented by the sample. 



RESULTS 



Eggs and Larvae 



Either eggs, larvae, or both were present every 

 month of the year. Eggs were present every month 

 except December and January. Only one egg was 



882 



