Each treatment mean in the data set was subsequently categorized by region (coastal or inland), harvest age (one month increments ranging from 9 to 26 months), cultivar, and eldana resistance rating of each cultivar (resistant or susceptible). For eachregion, mean TERC was fitted as a function of HA using curvilinear regression, and optimal HA was interpreted from the peaks of the fitted curves. A similar method was used to fit individual cultivar curves, as well as curves for eldana resistant and susceptible cultivar groups. Within each region, the fitted curves of three cultivars representing different time periods of cultivar release were compared with each other to determine if optimal HA changed with time of release. The%IB was also plotted against HA for resistant and susceptible cultivar groups using linear regression. These cultivar groups consisted of five resistant and seven susceptible cultivars in the coastal region, and six resistant and seven susceptible cultivarsfor the inland region. In all cases, linear and polynomial fits were tested for significance, and where necessary, significant differences between curves were evaluated using F-tests.