Toxicity of carbaryl and malathion to two federally endangered fishes, as estimated by regression and Anova

Toxicity of technical carbaryl, Sevin®-4-Oil, and technical malathion to federally endangered Colorado squawfish (Ptychocheilus lucius) and bonytail (Gila elegans) was estimated by 4-d renewal-acute and 32-d early life-stage (ELS) tests. Median lethal concentrations were estimated by probit analysis. Effect concentrations for ELS tests were estimated by ANOVA and an alternative procedure using a linear-plateau regression model. Linear-plateau regression estimated a threshold concentration above which toxic effects began to occur. Median lethal concentrations and 95% confidence limits (in parentheses) for carbaryl, Sevin-4-Oil, and malathion were 1.31 (1.23, 1.40), 3.18 (2.87, 3.52), and 9.14 mg/L (8.36, 10.0) for Colorado squawfish and 2.02 (1.78, 2.25), 3.31 (3.06, 3.55), and 15.3 mg/L (14.4, 16.4) for bonytail. NOECs for Colorado squawfish were 445 μg/L carbaryl and 1,680 μg/L malathion. The NOECs for bonytail were 650 μg/L carbaryl and 990 μg/L malathion. Threshold concentrations for Colorado squawfish in ELS tests were 364 μg/L carbaryl and 455 μg/L malathion. Threshold concentrations for bonytail were 217 μg/L carbaryl and 521 μg/L malathion. Estimates of effect concentrations from linear-plateau regression were consistently lower than those estimated by hypothesis testing. Linear-plateau regression models accounted for a significant amount of total variation and adequately described the observed concentration-response relation.