Table 35. Spearman rank correlations between measures of bioeffects in fish and 

 measures of length, weight, GSI, and HSI. 



Spearman rank correlation analysis among the various measures of effects also did not 

 show many high correlations (Table 36). AHH activity was most highly correlated with 

 EROD activity, as would be expected. Many of the suite of cytochrome P-450 measures (i.e., 

 total P-450, EROD, P-450E) were most highly correlated with each other. The counts of 

 total micronuclei were not very highly correlated with the other measures. Measures of 

 testosterone and estradiol 2-hydroxylase content were not highly correlated with the other 

 measures. 



Correlations Among Bioeffects Measures and Contaminant Concentrations in Fish. 



None of the biological measures was particularly highly correlated with any of the 

 chemical contaminants in the livers of the fish (Table 37). Total counts of micronuclei were 

 most highly correlated (negatively) with mirex and aldrin concentrations. Cytochrome P- 

 450 and P-450E content and EROD activities were weakly positively correlated with tPCB 

 concentrations. Estradiol 2-hydroxylase activity was relatively highly correlated with 

 dieldrin concentrations. No single chemical or chemical class stood out as consistently being 

 correlated with the biological measures. 



Within-Site Variability in Bioeffects Measures in Fish. 



Analytical variability, as evaluated with the sediment bioassays, could not be 

 evaluated with the bioeffects measures in fish, since replicate analyses of individual fish 

 were not performed on all the fish with all of the measures. However, within-site 

 variability and between-site discriminatory power can be compared among measures, since 

 all of the measures were performed on the same fish. Within-site variability among fish 

 can be viewed two (opposing) ways. First, if one assumes that individual fish sampled 

 within one area (site) had different histories of exposure to contaminants, then a bioeffects 

 measure that has high within-site variability may simply reflect that variability if the 

 measure is especially sensitive. If, on the other hand, one assumes that the fish sampled in 

 one area (site) are from a relatively homogeneous population and, therefore, all the 

 individuals have had a similar history of contaminant exposure, then a bioeffects measure 

 with high within-site variability may indicate relatively low analytical precision. 

 Unfortunately, Platichthys stellatus migrate in and out of San Francisco Bay annually and 

 little is known of the fidelity of the returning adults to specific areas of the estuary. 

 Although Spies et al. (1985, 1988) have demonstrated differences in measures of contaminants 



74 



