That sex determining systems ever change is paradoxical but can be explained by noting that conflict between selfish elements and their modifiers will often cause a shift in sex determining strategy. The evolution of the novel sex determining system of moles (Talpa europaea and T. occidentalis) may, we argue, be an example of just such a process. Three different models for the evolution of female intersexuality are presented. These all attempt to account for (1) the fact that a few years ago populations of moles had high frequencies of sterile XX individuals that were either morphologically male or intersex (other XX individuals were normal females) and (2) that presently, the XX individuals in the same population are exclusively fertile intersexes that are functionally female; i.e. have follicle producing ovotestes. This case history is compared to that of the wood lemming and two similarities are discussed. First, in both cases it is noted that one end product could be approached from different routes. Second, selfish elements may be involved in the evolution of both systems. In general, it is suggested that XY sex determination, far from being resilient to evolutionary change, is vulnerable to take-over by selfish elements. This is particularly the case in mammals in which transplacental interactions could allow manipulation of sex determination in one foetus by another. This, we also suggest, is a good candidate explanation for the evolution of novel sex determination in Talpa.