ABSTRACT
Several distribution types are applied to species sensitivity data including the log-logistic (Kooijman, 1987; Aldenberg and Slob, 1993), Gompertz (Newman et al., 2000), and lognormal (Wagner and Løkke, 1991) distributions; however, method formalization in North America increasingly favors the lognormal distribution (e.g., Baker et al., 1994; U.S. EPA, 1998c). The predicted mean value for 5% (HC
) or 10% (HC
) is frequently the inferred protective concentration. Safety factors have been included in some cases because application of the HC
without such adjustment assures the protection of 100 –
p
percentile of species in only 50% of all cases. As an early example, Van Straalen and Denneman (1989) described conservative adjustments to predictions from a log-logistic model. The lower 95% confidence or tolerance limits for the HC
were also promoted as conservative estimators of a protective concentration (Wagner and Løkke, 1991; Jagoe and Newman, 1997; Newman et al., 2000).
