For the last approximately 10 yr, the Gypsy Moth Life Stage (GLS) model has been used by pest managers to predict when important events in the gypsy moth, Lymantria dispar L., life cycle will occur (e.g., peak second larval instar population and male moth flight). Although the GLS model has been shown to outperform other gypsy moth phenology models, its predictions have not always been as accurate as desired. Differences between predicted and observed egg hatch phenology prompted a re-examination of the original experimental data that were used in the construction of the egg hatch submodels of the original GLS model, and a data processing error was discovered to have truncated the postdiapause experimental data. Analysis of the complete data set confirmed that developmental rates in the postdiapause phase were age and temperature dependent but that the developmental response to temperature is distinctly nonlinear at postdiapause initiation, in contrast to the indeterminate response previously reported. By incorporating the new estimates of developmental rate patterns and parameters into the GLS model, errors in the GLS-simulated egg hatch period were reduced by 33-71% and error in date of 50% cumulative egg hatch by 25-100%.