The assumption that a high proportion of immature herd ungulates in an archaeological site indicates a domestic population implies (a) there is a 'natural' population structure common to all wild herds, and (b) predators will kill a representative sample of this population. We challenge these implications by demonstrating that wide variations occur in the population structures of non-domestic herds, in the vulnerability of age and sex classes to predation, and in the selective hunting of herd animals by predators, including man. We conclude that the initial assumption cannot be sustained.