Almost 50 years ago Neel proposed a hypothesis to explain the prevalence of obesity and diabetes in modern society--the 'thrifty gene' hypothesis. The fundamental basis of the hypothesis was that, in our early evolutionary history, genes, that promoted efficient fat deposition would have been advantageous because they allowed their holders to survive at periods of famine. In modern society, such genes are disadvantageous because they promote fat deposition in preparation for a famine that never comes, and the result is widespread obesity and diabetes. In recent years I, and others, have questioned some of the fundamental assumptions of this hypothesis--particularly focusing on whether differential survival of lean against obese in famines provides sufficient selective pressure for the spread of so-called 'thrifty genes'. These arguments have been criticized because famines not only affect survival but also fecundity, and obese people would be expected to sustain fecundity longer in the face of food shortages. In this paper, I show that the reduced fecundity argument is flawed because famines are almost universally followed by periods of enhanced fecundity, which offsets the decline observed during the famine itself. The net effect of famines on fecundity is consequently insufficient to rescue the thrifty gene idea. Elsewhere, I have suggested an alternative scenario that subsections of the population have a genetic predisposition to obesity due to an absence of selection, combined with genetic drift. The scenario presented earlier was based on evidence from prehistory concerning the release of our ancestors from heavy predation pressure around 2 million years ago. I suggest here that this is one of a number of potential scenarios based on random genetic drift that may explain the specific aetiology of the obesity epidemic. Together, these alternatives, based on central notion that genetic drift rather than positive selection was a dominant factor, may be called the 'drifty gene' hypothesis.