threshold that can be used to accurately classify sediment
samples as both toxic and not toxic. The results of this evaluation
were not substantially different when the PECs for the
individuals PAHs (i.e., instead of the PEC for total PAHs) were
used to calculate the mean PEC quotients (Table 7). Kemble et
al. (1999) reported similar results when the mean PEC quotients
were evaluated using the results of only 28-day toxicity
tests with H. azteca (n 5 149, 32% of the samples were toxic).
To examine further the relationship between the degree of
chemical contamination and probability of observing toxicity
in freshwater sediments, the incidence of toxicity within various
ranges of mean PEC quotients was calculated (e.g., , 0.1,
0.1–0.2, 0.2–0.3). Next, these data were plotted against the
midpoint of each range of mean PEC quotients (Figure 1).
Subsequent curve-fitting indicated that the mean PEC-quotient
is highly correlated with incidence of toxicity (r2 5 0.98), with
the relationship being an exponential function. The resultant