Stress is here defined as an environmental condition that reduces Darwinian fitness when first applied. Optimal stress responses (i.e. those that maximize Darwinian fitness) are calculated for different levels of growth and mortality stress, and are found to depend critically on the shape of the trade-off curve relating mortality to growth rate. If the trade-off does not change shape when stress is applied, then the optimal strategy is to spend less on personal defence for both mortality and growth stresses. However, if stress does change the shape of the trade-off the predictions may be modified, or reversed. This optimality analysis is rigorous and easy to apply. What is more difficult, is to establish the shapes and positions of trade-off curves in particular cases. This problem is discussed and some suggestions are made. The theory's predictions are applied speculatively to biogeographical data on marine animals and are found to be qualitatively successful, although some of the needed data are lacking. The applications and testability of the theory in the study of ageing and a variety of other processes are considered.