Physiology of ectothermic organisms depends on microclimate temperature. In some insect–plant relationships, the herbivore physically manipulates its proximate environ-ment (i.e., plant tissues). However, little is known about the effects of this manipulation on the insect microclimate. We studied the thermal environment of the leaf-mining insect Phyllonorycter blancardella (Lepidoptera: Gracillariidae). This herbivore modifies both morphology and physiology of attacked apple leaf tissues to construct a mine inside which the entire larval development occurs. Spectral measurements showed that absorbance of mined leaf tissues differed from that of intact leaf tissues. Gas exchange measurements in mined leaf tissues demonstrated that responses of stomata to changes in climatic parameters were modified compared to intact leaf tissues. We built an energy budget model to predict the temperature within a mine given climatic variables, and measured parameters related to radiative absorption properties and to the ecophysiology of stomata. Model predictions were verified using experimental measurements made under a large range of climatic conditions. Radiation level was the most influential variable on both mine and leaf temperatures, and a mine was always warmer than a leaf. Mine temperature was predicted to be up to 108C above ambient air and 58C warmer than intact leaf tissues at high radiation levels. The decoupling between mine temperature and leaf temperature was significant. The model was manipulated to quantify the separate effects of altered absorbance and modified stomatal behavior. Both effects contributed almost equally to the temperature excess. A mine gained more radiative heat than a leaf and the observed stomatal closure limited latent heat losses. We suggest that this warm microclimate allows larvae to develop faster, leading to a reduced risk of attack by parasitoids. The model, which is the most complete one to date for any herbivorous insect, shows that the second trophic level manages and partially controls the first one, even to the point of one trophic partner co-opting the physiology of the other. These reciprocal influences imply that heat budgets cannot be simply built for each trophic level independently.