When we attempt to carry out an experimental study of human behaviour we are often frustrated because we can manipulate only a few unimportant variables, and because the period over which we can observe their effects is so short. To escape from such limitations we have recourse to animal studies, with a consequent reduction in the range and complexity of the phenomena which are accessible to us. There is an alternative whose value is becoming more and more apparent, namely, to exploit the variables which nature has provided in differences of ecology, culture, and language. This approach is not new, but a more adequate specification of the relevant variables has become possible through the growth of anthropology and linguistics. We can now undertake the study of basic psychological processes, which is our real concern, much better equipped than were our predecessors to assess the long-term effects of complex environmental conditions upon many aspects of behaviour.