Figure 3.

Temperature-dependent efficiencies of macronutrient conversion and digestion over the fifth stadium showing linear regression lines at 26, 32 and 38°C. An ANCOVA (table S1, electronic supplementary material) removed effects of covariates and factors; insets display resultant estimated marginal means and their standard errors at the three temperatures. The ordinate in each case has the same dimension as that of the parent plot. Insects at 26 and 38°C experience significantly lower utilization efficiency of (b) ingested protein, (c) carbohydrates or (a) both compared with insects at 32°C. This is due to (g–i) inefficiencies in the conversion of digested food rather than (d–f) inefficiencies in digestibility. See text for additional explanation. n = 30 for each temperature; insects failing to molt were excluded from further analysis. *p < 0.05; **p < 0.01; ***p < 0.001.