Evolution and the variation of mammalian sex ratios at birth: reflections on Trivers and Willard (1973).

The Trivers-Willard (TW) argument may be interpreted to yield a large class of true predictions and a small class of false predictions. This would explain the relative numbers of empirical successes and failures of TW. It is argued here that in mammals (including man), one cause of these failures is a constraint imposed by a small minority of females in the breeding population. These are females that are stressed and consequently have high testosterone (T) levels which in turn cause (a) reproductive impairment and (b) (contrary to TW, but in conformity with my hypothesis) a tendency to produce sons. TW reasoning applies to both parents, not just the mother. Indeed, the TW prediction seems fulfilled more generally by fathers than by mothers. My hormonal hypothesis can apparently explain both the successes and the failures of TW.