eggs laid by each female). The patterns for potential fecundity and total reproductive investment (data not shown) were identical to that demonstrated by realized fecundity. I used stepwise multiple linear regression to identify the most significant determinants of realized fecundity (Table 5-6). Cumulative intake during the reproductive lifespan was the primary variable selected by the model, explaining 82.8% of the variance in fecundity (Fig. 5-9a). When potential fecundity (number of eggs laid + number of eggs remaining in the ovaries at death) was regressed against cumulative intake during the reproductive lifespan, the same relationship existed with an adjusted R2 value of 0.847 (data not shown). In addition, growth rate during adulthood prior to first oviposition and cumulative intake during all juvenile stages were also selected as variables in a model that explained 92.8% of the variance in realized fecundity (model 3, F3,54 = 245.46, p < 0.0001). Stepwise multiple linear regression identified body mass at first oviposition, age at the adult molt, mass-specific intake during adulthood prior to first oviposition, and cumulative intake during adulthood prior to first oviposition as significant independent variables in a model that explained 70.6% of the variance in initial oviposition rate (model 7, F4,53 = 35.20, p < 0.0001). The data do not support the contention that decreased longevity is a cost of reproduction, at least in C. morosus. On the contrary, fecundity was significantly and positively related to adult lifespan when data for all treatments were combined (F1,56 = 25.67, p < 0.0001, R = 0.302, Fig. 5-9b). When potential fecundity (number of eggs laid + number of eggs remaining in the ovaries at death) was regressed against adult lifespan, the same relationship existed with an adjusted R2 value of 0.343 (data not shown). However, analysis of covariance indicated that adult lifespan did not have a significant effect on reproductive output (F1,52 = 3.284, p = 0.076) whereas treatment did have a significant effect (F4,52 = 32.463, p < 0.0001). This result was confirmed by