Thermal tolerance is an important factor influencing the distribution of ectotherms, but we still have limited understanding of the ability of species to evolve different thermal limits. Recent studies suggest that species may have limited capacity to evolve higher thermal limits in response to slower, more ecologically relevant rates of warming. However, these conclusions are based on univariate estimates of adaptive capacity. To test these findings within an explicitly multivariate context, we used a paternal half-sibling breeding design to estimate the multivariate evolutionary potential for upper thermal limits in Drosophila melanogaster. We assessed heat tolerance using static (basal and hardened) and ramping assays. Additive genetic variances were significantly different from zero only for the static measures of heat tolerance. Our G: matrix analysis revealed that any response to selection for increased heat tolerance will largely be driven by static basal and hardened heat tolerance, with minimal contribution from ramping heat tolerance. These results suggest that the capacity to evolve upper thermal limits in nature may depend on the type of thermal stress experienced.