Table 15. Minimum detectable distance between sites, assuming n=2, n=3, n=4, and n=5 

 stations per site, based upon ANOVA power analysis, where a = 0.05 and power = 90%. 



n = 2 n = 3 n = 4 n = 5 



R. abronius 



survival 

 avoidance 



A. abdita 



survival 

 avoidance 



M. edulis 



percent abnormal 

 percent survival 



S. purpuratus 



percent abnormal 

 echinochrome 



D. gyrociliatus 



egg production 0.76 0.62 0.53 0.48 



A variety of calculations were performed to compare the relative analytical precision 

 and discriminatory power of 15 of the toxicity end-points (Tables 16 and 17). For some end- 

 points {e.g., R. abronius avoidance) the degree of within-sample variance differed greatly 

 between samples taken at some sites (e.g., YB samples) (Table 16). For other end-points there 

 was relatively high homogeneity (e.g., M. edulis percent normal development in VA 

 sediments). The averages of the SDs and CVs for each toxicity test are compared in Table 10, 

 the latter as an index of precision. Given that the SDs were largely influenced by the units 

 in which the end-points were reported, the CVs are a better basis for comparison. Among all 

 of the end-points, that of percent normal development in M. edulis had the lowest average 

 CV (3.9%). Among the other end-points measured in all 15 samples, those of A. abdita 

 survival and S. purpuratus percent normal development and echinochrome content also had 

 relatively low average CVs. The CVs for the avoidance end-points of R. abronius and A. 

 abdita were the highest. Among the end-points in the tests with S. purpuratus measured in 

 five samples, the incidence of micronuclei was highly variable, and the number of mitoses 

 per embryo and egg fertilization success were the least variable. 



The quotients obtained by dividing the total range in mean values by the average SD for 

 each end-point are compared in Table 10. This quotient, the discriminatory power of the test, 

 is intended to identify those end-points with the widest range in response and the lowest 

 analytical variability independent of the use of controls. The discriminatory power was 

 highest with the M. edulis percent normal development end-point and the R. abronius percent 

 survival end-point. They had 6.5 and 6.1 SDs within the range in response, respectively. 

 Among the other end-points measured in all 15 samples, those of avoidance of sediments by R. 

 abronius and A. abdita had the lowest discriminatory power and those of M. edulis survival, 

 S. purpuratus percent normal development and echinochrome content, and D. gyrociliatus egg 

 production were intermediate. Among the S. purpuratus end-points measured in five samples, 

 that of cytological abnormalities had a relatively high discriminatory power, while the 

 others had relatively low values. 



44 



