Table 16. Spearman-rank (rho, corrected for ties) correlations between dioxin equivalents deter- 

 mined in chemical analyses and dioxin equivalents (tcdd-eqs) determined in rat hepatoma bioas- 

 says of sediment extracts. F 1 fraction = whole extract. F 5 fraction = PAH fraction. F 7 

 fraction = bulk (>2 - ortho - chloro - substituted) PCB fraction. F 8 = mono - ortho - chloro - 

 substituted PCB fraction. F 9 = non - ortho - chloro - substituted PCB fraction. F 11 = com- 

 bined total PCB fraction. F 12 = PCDD/PCDF fraction. 



*p<0.05, **p<0.001, ***p<0.0001 



This relationship between the total cumulative tcdd equivalents from the chemical analyses and the 

 tcdd equivalents from the H4IIE rat hepatoma bioassays of the F12 fraction is illustrated in Figure 28. 

 The relationship is very strong and nearly linear. 



Also, there was a relatively strong correlation and relationship between the concentrations of total 

 PAHs in the sediments and the tcdd equivalents determined in the H4IIE bioassays of the F5 (PAHs) 

 fraction (Figure 29). This relationship was not as strong as that observed with the dioxins and would be 

 improved by deletion of the data from two samples. Nevertheless, a strong pattern was obvious be- 

 tween the chemical estimate and the bioassay estimate of PAH concentrations. 



Relationships Between Toxicity and Physical-Chemical Parameters: Phase 1. The relationships 

 between the four measures of toxicity and the concentrations of potential contaminants and other pa- 

 rameters were compared using non-parametric, Spearman-rank correlations (Tables 15-18). Although 

 the correlation analyses cannot be interpreted as evidence of cause-effect relationships, they can iden- 

 tify patterns in co-variance or concordance between dependent variables (i.e., toxicity) and indepen- 

 dent variables (i.e., potential toxicants). The correlation coefficients are accompanied by the level of 

 significance of the correlations. To account for Type 1 errors in the correlations, the significance level 

 (p=0.05) should be divided by the number of variables and the adjusted significance level used as the 

 critical p value. 



The total concentrations of most trace metals were not significantly correlated with amphipod survival, 

 bivalve survival, or bivalve normal development (Table 17). In contrast, most of the metals were 

 weakly (but significantly) correlated with the microbial bioluminescence EC50s. Only the concentra- 

 tions of mercury and tin were significantly negatively correlated with amphipod survival. None of the 

 metals or other parameters were correlated with bivalve survival and only the concentration of carbon- 

 ate was correlated with bivalve normal development. In the microbial bioluminescence test, all of the 

 potentially toxic trace metals (i.e., Ag, Cr, Cd, Cu, Hg, Pb, Sn, and Zn) were significantly (p<0.05) 

 correlated with toxicity. Also, the Microtox test results were correlated with the total organic carbon 

 content (% TOC). Only those correlations shown with "**" would remain significant if the number of 

 variables (18) were taken into account. 



76 



