Few would doubt that prehistorians have come to regard the formulating of models and the testing of hypotheses as among their most important concerns in recent years. This trend towards explaining, rather than merely accumulating data can be regarded partially as reflecting a desire to maintain some degree of command over a rapidly expanding body of information, but also the feeling that prehistoric archaeology has tended to lag behind related behavioural disciplines such as ecology or geography. What has become a highly conspicuous feature of these developments is that prehistorians have tended to undertake two radically different kinds of studies, particularly on prehistoric subsistence. The first relies upon published archaeological data, often accumulated over several years, from regions as large as countries or even continents; the second type, however, tends to be more modest in scope, and based upon a small body of evidence that has been specifically collected from either a small number of sites, or at most, a small region such as a single valley system. Recent studies by Murray (1970) on European neolithic animal husbandry, or by Ammerman & Cavalli-Sforza (1971) on European neolithic dispersal rates are convenient examples of the first type of study: both may be contrasted with the more detailed, localised types of studies on the same topics by Payne (1972) and Barker (1975). Whilst these two approaches may be complementary, it is nevertheless noticeable that they often reach strikingly different conclusions that frequently result from fundamental divergences of opinion about the quality of the data itself, and the most appropriate way to examine it.