pod survival was highly variable (Figure 32). Relatively high survival occurred in many of the samples. 

 In the samples in which these compounds equalled or exceeded the ERM value, however, amphipod 

 survival was universally low and significantly different from controls. Amphipod survival was 0.0% in 

 one sample from the East River that had extremely high concentrations of PAHs. 



Amphipod survival was very high in all except two samples in which the total PAH concentrations 

 were below the ERL value of Long et al. ( 1 995). Amphipod survival was relatively low in many of the 

 samples with total PAH concentrations above the ERL. Amphipod survival ranged from 0.0% to 40% 

 in the four samples with total PAH concentrations above the ERM guideline. The data in Figure 30 

 illustrate a relatively consistent decrease in amphipod survival with increasing concentrations of total 

 PAHs, in agreement with the significant correlation (Rho = -0.603, p<0.001). 



The concentrations of both fluoranthene and phenanthrene normalized to TOC content were signifi- 

 cantly correlated with amphipod survival and the concentrations in many samples equalled or ex- 

 ceeded their respective proposed National SQC (U.S. EPA, 1994). The relationships between these 

 two compounds and amphipod survival are illustrated in Figures 34 and 35. In both cases amphipod 

 survival was relatively high in most samples with chemical levels below the SQC, and decreased steadily 

 as the concentrations exceeded the respective SQCs. 



In tables 22-24 the average concentrations of toxicants in the samples that were toxic to amphipod 

 survival are compared to those in the samples that were not toxic. Also, the average concentrations in 

 the toxic samples were divided by the average concentrations in the nontoxic samples and these ratios 

 were compared among chemicals. Finally, the average concentrations in the toxic samples were com- 

 pared with the sediment quality guidelines (SQG) of Long et al. (1995), or Long and Morgan (1990), or 

 the proposed National SQC (EPA, 1994). No SQG were available for substances such as aluminum 

 and iron. We assumed that chemicals that contributed substantially to the observed toxicity would be 

 correlated with toxicity and highly elevated in concentration in the toxic samples, and the average 

 concentrations in the highly toxic samples would exceed applicable ERM or SQC values. In the am- 

 phipod tests 17 samples analyzed for chemical substances were not significantly toxic (i.e., different 

 from controls), 2 were significantly different from controls (but survival exceeded 80% of controls), 

 and 19 samples caused amphipod survival in less than 80% of controls. Average amphipod survival 

 was 98.4% in the nontoxic samples and 30.1% in the highly toxic samples. 



The average concentrations of all the trace metals were very similar in the nontoxic, significantly toxic, 

 and highly toxic samples, based upon the results of the amphipod tests (Table 22). The ratios in aver- 

 age concentrations between the nontoxic samples and either the significantly toxic or highly toxic 

 samples ranged from 0. 1 to 2.2. Most ratios were 1 .0 or thereabouts. The concentrations of mercury in 

 the highly toxic samples were the most elevated of the metals, exceeding the concentrations in the 

 nontoxic samples by a factor of 2.2, and exceeding the ERM value of 0.71 ppm (Long et al., 1995) by 

 a factor of 4.5. The average concentrations of most metals exceeded the ERL values in both the 

 nontoxic and the toxic samples, illustrating the relative similarity in concentrations among the samples. 

 The mean total SEM concentrations exceeded the total AVS concentrations only in the nontoxic samples 

 (a result of two nontoxic, sandy samples). Most of the variability in the SEM/ AVS ratios was contrib- 

 uted by the concentrations of zinc in the samples. 



The average concentrations of chlorinated organic compounds (PCBs and pesticides) in the toxic samples 

 often were very similar (i.e., ratios of about 1 .0) to the concentrations in the significantly toxic samples 

 (Table 23). However, the ratios in chemical concentrations between nontoxic and highly toxic samples 

 often exceeded 2.0 and ranged upwards to 20.3 in the highly toxic samples. The average concentra- 



88 



