This is a study of the robustness, in estimating population size, of the usual DeLury Estimator and three modifications of it derived from the generalized exact deterministic catch equation of Beverton and Holt (1956, 1957). These estimators are called the Generalized DeLury Estimator, the Weighted Generalized DeLury Estimator, and the Effort-corrected DeLury Estimator.A computer was used to simulate varying environmental conditions, which included combinations of several vectors of catchability and effort and several types of random error. The results indicate that the generalized estimator is overall the best, at least for these simulated data. The effort-corrected estimator is just as effective, but requires effort correction. The usual DeLury estimator underestimates on the average. The weighted estimator has high variance.Catchability behavior has a marked effect on accuracy of any of these estimators. An increase in catchability from a previously constant level results in a subsequent increase in the estimate of population size. A decrease in catchability results in a decrease in the estimate. If catchability cycles, the estimate cycles with the same crest and period.Effort has little effect, at least for the vectors studied in this model.