Proceedings of the Prehistoric Society 37 (1)

The first report by CBA Group 4 covers 350 implements (including axe-hammers, battle-axes, mace-heads and adzes). Of the 192 groupable examples Group VI has the highest total; Groups XVIII, VII and I are also well represented. In Yorkshire the large numbers of glacial erratics make identification of "local" versus "foreign" rocks very difficult. Search in Upper Teesdale for a likely factory-site for Group XVIII, Whin Sill (which provides a very wide range of implements) was unrewarded. On all the distribution maps the Flamborough-Rudston area stands out for its major concentration of tools, as also of ritual monuments. Few of the axes have associations, and there is insufficient evidence to suggest trade routes.

A philosophical re-appraisal of both the objectives of archaeological study and the standards of explanation found acceptable. Up to now, practical problems of fieldwork have dominated our methodology and archaeologists have tended to avoid "explanation" for fear of being caught out in "speculation". Participants in the current debate about quantification are almost all primarily concerned with material culture, but it is likely that more testable hypotheses can be derived from the study of population and ecology in site catchment areas than from anything connected with vague notions of "culture process". The preoccupation with reconstructing culture history, in fact, tends to impede alternative explanations of the data.

Statistical comparison of observed and random distributions of finds confirms the importance, deduced on other grounds, of natural physiographic regions. Even allowing for differential recovery, certain regions stand out as particularly poor or particularly rich in finds. The upland areas show a reduced density compared with earlier periods, as if some shift of population to lower ground was already in progress. The "Jurassic routeway" does not show an above-average share of finds, whereas the valleys of Trent, Ancholme and Witham have very marked concentrations, particularly at important river crossings. The preponderance of fine weaponry in the Witham valley argues for votive deposition. Distributional analyses of this type may be more useful than chronological and typological discussions.

Analysis of barrow densities in Wiltshire and Dorset brings out certain groupings which seem to reflect distinct territories. Two areas stand out for their high proportion of "fancy barrows" distributed densely over fairly narrow linear areas. For various reasons an explanation in differential living densities is unsatisfactory, though there could be a chronological element. The theory of "Wessex undertakers" has severe practical difficulties in its simple form; however, it may be time to revive the concept of Wessex pastoralists. If these visited their traditional grazing grounds in Wessex at certain seasons they could also bring with them any cremations for burial on home ground. Apart from a few ?cultivators in permanent residence, the population would disperse for the rest of the year. The incidence of leatherworking equipment appropriate to pastoralists appears significantly high in the Stonehenge area. Ethnographic parallels for "home base" transhuming populations are plentiful, and their grazing patterns (especially in relation to co-existing cultivators) are well documented. The rise of Wessex power could be due to the phenomenon, common elsewhere, of dominance by pastoralists over cultivators.

Ethnographic parallels suggest alternative explanations of "storage pits", "drying racks" and "granaries" of the earlier pre-Roman Iron Age. Evidence is known for pits holding not only grain but water, sling-stones, manure or burials. Four-post structures can be watch-towers or fighting platforms; they may also double as stores for grain or any other foodstuff or equipment requiring protection from damp. Storage of foodstuff both above and below ground insures against loss of an entire crop. Exposure-burials on two- or four-post structures (cf Mississippi Indians) could account for the occasional scatters of human skeletal fragments on EIA domestic sites as well as for the general absence of formal EIA burials.

In response to Hodson's paper on Cluster Analysis (abstract 70/750) the problem of the logical correlation of variables is raised, and an attempt is made to defend "matrix ordination (seriation) and score contouring to delimit clusters" against rebuke. J G