Fig. 2. Feeding over time as related to different combinations of juvenile hormone (JH) and temperature. (A) Mean (± s.d.) feeding for four treatments over time. Groups of bars within days are not significantly different (NS), as determined by non-parametric Kruskal–Wallis tests (P>0.05, N=3). (B) Linear regression results (mean ± s.d.) for feeding over time among different treatments. All regressions were significant (P<0.05, N=3). Slope values shown at the right with the same letter are not significantly different by linear regression analysis (GLM; P>0.05). Note that feeding rates were highest for the 27°C untreated controls, which have exposure to only endogenous JH.