Estimation of a single probit line from multiple toxicity test data

Abstract

LC50 values for the same compound and the same species may vary considerably within a laboratory and between laboratories. These differences are usually attributed to variable test conditions and response of the test organisms to the toxicant. Furthermore, the lack of standardization for aquatic toxicity testing may contribute to the variability in LC50 values.To employ toxicity data for regulatory purposes, it may be useful to report one single LC50 value and its associated confidence interval, instead of several LC50 values for each test substance. To accomplish this, a procedure for combining probit lines from several toxicity tests was developed by modifying the maximum likelihood probit method with inclusion of the technique for parallel line probit analysis. The resulting single probit line from this method is referred to as a "grand probit line' and takes into account separate test results. To ease the calculation a BASIC program was developed for an IBM PC.