Factors influencing allocation of resources to male and female offspring continue to be of great interest to evolutionary biologists. A simultaneous hermaphrodite is capable of functioning in both male and female mode at the same time, and such a life-history strategy is adopted by most flowering plants and by many sessile aquatic animals. In this paper, we focus on hermaphrodites that nourish post-zygotic stages, e.g. flowering plants and internally fertilising invertebrates, and consider how their sex allocation should respond to an environmental stress that reduces prospects of survival but does not affect all individuals equally, rather acting only on a subset of the population. Whereas dissemination of pollen and sperm can begin at sexual maturation, release of seeds and larvae is delayed by embryonic development. We find that the evolutionarily stable strategy for allocation between male and female functions will be critically dependent on the effect of stress on the trade-off between the costs of male and female reproduction, (i.e. of sperm and embryos). Thus, we identify evaluation of this factor as an important challenge to empiricists interested in the effects of stress on sex allocation. When only a small fraction of the population is stressed, we predict that stressed individuals will allocate their resources entirely to male function and unstressed individuals will increase their allocation to female function. Conversely, when the fraction of stress-affected individuals is high, stressed individuals should respond to this stressor by increasing investment in sperm and unstressed individuals should invest solely in embryos. A further prediction of the model is that we would not expect to find populations in the natural world where both stressed and unstressed individuals are both hermaphrodite. (C) 2012 Elsevier Ltd. All rights reserved.