Fig. 1. (A) Mass-independent critical thermal minimum (CT_min) onset and recovery (least-square means ± 95% confidence intervals) of P. marioni larvae repeatedly exposed to cold (filled squares, broken lines) or handling controls (open triangles, solid lines). Number of exposures on x axis includes 5 days recovery after final exposure (r-5). (B) Mass at dissection (g) and (C) gutless mass (g; mass at dissection – gut + contents) of Pringleophaga marioni larvae exposed to –5°C on up to 5 occasions (filled squares, broken line) and handling controls (open triangles and solid lines). Number of exposures on x axis, includes 5 days (r-5) and 30 days (r-30) recovery after 5 exposures. Values are least-square means with initial mass as covariate ± 95% confidence intervals. All analyses were performed as ANCOVA on log₁₀-transformed data. Asterisks indicate masses that are significantly different from initial mass (0 exposures), on the basis of non-overlapping confidence intervals. Control-treatment pairs whose confidence intervals do not overlap are significantly different (P<0.05). Mass at dissection: main effect F_1,134=80.99, P<0.0001; treatment x exposure interaction F_7,134=4.721, P<0.0001. Gutless mass: main effect F_1,133=21.61, P<0.0001; treatment x exposure interaction F_7,133=1.71, P=0.112.