completed. Therefore, no post-disposal samples were available. 

 However, based on our previous experience, we could predict that 

 the elevated levels of trace metals observed during the disposal 

 period were generally of short duration. The elevated level of 

 copper during disposal at WLISrN was the only inconsistency 

 encountered in this analysis. Such an occurrence could be 

 interpreted as being coincidental with the disposal period or 

 attributable to other unknown environmental factors at the site; 

 further analyses by stepwise multiple regression favor the latter 

 interpretation . 



Stepwise multiple regression analyses (Table 5) show that 

 the volume of dredged materials disposed was correlated with the 

 concentrations of chromium and nickel at WLISc, of copper, iron 

 and nickel at 500MW. In contrast, for the reference populations 

 maintained at WLISrN and Rlr, dredged volume as an independent 

 variable was not entered into the regression models. The results 

 are consistent with our previous studies that show the intrinsic 

 variables, W/D and L, generally could account for the major 

 proportion of variance observed in the tissue trace metals. 

 There are two lines of evidence suggesting that dredged material 

 disposal played a minor role in the uptake of trace metals by the 

 mussels; these are (1) the dredged volume entered only ca. 28% of 

 the cases and (2) it was the third or fourth variable entered 

 into the model. The figure of 28% was derived from the 

 assumption that if the elevated levels of the nine trace metals 

 were associated with the dredged volume at WLISc and 500MW, one 

 would expect that the dredged volume entered all 18 cases (2 

 stations x 9 trace metals) of the stepwise multiple regression 

 analyses. In the present study, the dredged volume entered the 

 regression model as an independent variable only on 5 occasions 

 (5/18, or ca. 28%) . 



Polychlorinated Biphenyl fPCBs) Concentrations . The mean 

 concentrations of the Aroclors and the total PCBs from the four 

 mussel monitoring populations are summarized in Table 6. 

 Temporal variations in tissue Aroclors and PCBs are presented in 

 the Appendix (Tables 5-9) . The mean concentrations of Aroclor 

 1242, 1254 and 1260 fall in two distinct groups: the Rlr and the 

 three populations in western Long Island Sound. The latter has 

 nearly twice the Aroclor concentrations of the Rlr population. 

 When the data sets were subjected to two-way ANOVA, the null 

 hypothesis that there were neither spatial (station) nor temporal 

 (before-during disposal) differences in Aroclor concentrations 

 among the four mussel populations, was rejected. PCBs showed 

 significant between-station differences (Table 6, For Aroclors 

 1254+1260: F=6.92, d.f.=3,35, P=0.0009; For Total PCBs: F=4.43, 

 d.f.=3,35, P=0.0097), while the Aroclor 1242 and total PCBs 

 exhibited temporal or before-during disposal differences in 

 mussel populations held at WLISrN and WLISc (level of 

 significance see F values listed in Table 6B) . Further analyses 

 of the concentrations of Aroclors 1254+1260 and total PCBs using 



