Effect of Body Weight on Lethal Dose Estimates for the Western Spruce Budworm12

Abstract

Procedures for calculating lethal dose and confidence interval estimates for larvae of Choristoneura occidentalis Freeman of a given body weight based on multiple probit or logit regression techniques are described. Results of experiments with mexaearbate, pyrethrins, and DDT were similar for both probit and logit methods. Estimated LDs and their 95% confidence limits increased with increasing body weight. When the hypothesis of proportionality is rejected, use of the proportional model may produce estimates of LDs which are too low and 95% confidence limits which are too narrow, particularly in higher weight classes. These results suggest that topical application experiments should be done with a uniform application volume, and the hypothesis of porportional response should be tested. Since grouping tends to mask departures from proportionality in the data, the use of the basic model as a standard procedure for data analysis is recommended.