The recognition by archaeologists over the last fifteen years or so that even a discipline as down to earth as prehistory requires a theoretical basis has led to a diversity of research objectives. Essentially this has been a move away from the delineation of chronology and culture narrative towards an understanding of the behaviour of prehistoric man, requiring the development of theories relating to his ecological adaptation, his social organization, and his cognitive behaviour. The reassessment of archaeology has largely been at an explanatory or interpretative level. To some extent this has of course been paralleled by the increasingly sophisticated scientific and analytical techniques that are now available to aid in the interpretation of archaeological data. However, whilst archaeologists are more rigorous than ever before, and have developed a whole range of concepts for tackling the problems of prehistory, the actual data base employed by archaeologists has remained remarkably static. Field theory and methodology is essentially that of the earliest archaeologists. As Clarke (1973, 10) has commented, ‘harnessing powerful new methodological horses to rickety old conceptual carts has proved to be a powerful but drastic way of improving archaeological constructs by elimination’. Equally, though, the gleaming, streamlined conceptual chariots of the 1980s will go no further nor faster unless some attempt is made to update the roads over which they are expected to travel. What is necessary is a reassessment of what archaeological field data represent, i.e. how is the archaeological record formed?

Stratigraphic and typological analysis tends to confirm the two-fold division of the Zarzian suggested by Dorothy Garrod, while palaeobotanical data indicate that a very dry steppe prevailed throughout the sequence. Both the low carrying capacity of the contemporary vegetation then and at present in winter, and the resultant winter migration to warmer areas of wild game at present (and presumably in the past) suggest cyclical short distance movements by hunter-gatherers and fishers within two ecological zones: 1, the mountainous area which must have been occupied from spring to autumn; 2, the winter camping grounds in the upper plains along the foothills region of the Zagros Mountains.

When and how were the Mediterranean islands first settled? Has insularity itself—the special characteristics of islands everywhere—acted as a constraint on the manner and rate of their colonization by man? If so, is it possible for archaeologists to make use of ecological and biogeographical models which have been developed to account for the abundance and diversity of animals and plants on islands of varying size and remoteness? This paper offers a brief review of the available data on the first of these important questions, seen in the light of the second and third, and it proposes some modifications to the scenarios of colonization to be found in most current accounts of early island prehistory in the Mediterranean. As a reflection of personal research interests, I emphasize the east Mediterranean evidence, but there are useful insights to be gleaned, I believe, by comparing what we find there with the pattern for the islands of the west.

Summary. Excavations at Bury Hill dated a small, continuously-ditched enclosure to the early Neolithic. The interior of the enclosure, c. 1 ha in extent, was sampled by two trial cuttings, 9 m wide, running across the enclosure at right angles to one another. No Neolithic features were found. Sampling the ditch produced early Neolithic pottery, flintwork and animal bone. Though continuous, the ditch had a ‘beaded’ appearance, with a single entrance. No corresponding bank survived, though aerial photographs indicated its existence on the inside of the ditch.

The complex of sites that forms the subject of this account lies in the extreme north-east corner of England, about 10 km south-east of the river Tweed in the parishes of Kirknewton, Ewart, Akeld and Milfield, district of Berwick-upon-Tweed, county of Northumberland (fig. 1). The sites have mainly been discovered by means of air photography, but some were recorded as earthworks in the last century while two standing stones survive to this day. The modern discovery of the complex can be attributed to St Joseph during aerial surveys in the late 1940s, when a henge monument 1 km south-south-east of Milfield was located; this was subsequently included by Atkinson (1949–50, 64–5; 1951, 105–6) in his list of henges. It was recorded as a class II monument, and the presence of a pair of roughly parallel ditches (a ‘droveway’) passing through the site was noted. Further aerial survey work revealed that this site was one of several in the area around Milfield. In 1971 McCord and Jobey published further sites of henge type at West Akeld Steads, Ewart Park, East Marleyknowe and Milfield North (1971, 123–4), bringing the total of hengiform sites to five, and recording the full known extent of the ‘droveway’. The remaining sites were located by St Joseph and McCord but not published; they were noticed by the present author (among others) in the course of examination of all air photographs of the area, and appear on the plots of air photographs in the archives of the Northern Air Photography Committee, held in the Department of Archaeology, University of Newcastle upon Tyne (fig. 2).

The pit alignment at Milfield North, discussed in the preceding paper, is one of six alignments revealed in the Milfield Basin within the last decade. All discoveries have so far been confined to the sands and gravels of the former delta surface, and whereas additional examples, as well as their extension onto the heavier loams overlying the sandstone series, remain to be found, so crisp has been the definition by means of intensive aerial survey in this valley, that only a limited enlargement of the present distribution pattern by this means is anticipated. The characteristics of the individual alignments are as follows:

1. Ewart 1. NT95343209 to NT96103162

An irregular line of closely spaced pits that for most of its observable course follows the crest of a gentle but marked elevation in the sand and gravel-terrace. Two interruptions are visible towards its eastern end before it angles southwards to disappear into a wood. Observed over 1100 m.

Source: Cambridge University Collection. BDE35, BDE36, BKC40.

Museum of Antiquities, Newcastle upon Tyne. A1656–1662.

2. Ewart 2. NT95353165 to NT95803159

Lying to the south of Ewart 1 and running broadly parallel to it. The pit sizes and spacing is similar to Ewart 1 although their inter-relationship is obscured at their eastern end by a plantation. Ewart 2 fades away at its western extremity. Observed over 300 m.

Source: Cambridge University Collection. BDE35, BDE36.

Summary. An extensive urn cemetery associated with a complex flint platform, excavated by Max Dacre between 1966 and 1970, included burials of late Neolithic, Early Bronze Age, Middle Bronze Age (‘Deverel-Rimbury’) and Late Bronze Age date. The cemetery developed organically from a late Neolithic/Early Bronze Age focus which had evolved around one or more large sarsen stones.

The pottery sequence is of particular interest. The chronological precedence of all the barrel urn types of Central Wessex has been demonstrated for the first time and the Deverel-Rimbury phases contain pottery which relates both to the local Wessex sequence and to the Lower Thames Valley assemblages. The later Deverel-Rimbury phases also include vessels of the post-Deverel-Rimbury tradition and the final burials were interred in Late Bronze Age jars.

Analysis of the associated cremations gives some indication of the age and sex structure of each phase of burials, although identification proved difficult owing to the post-incineration process of pulverization to which the remains had been subjected in all phases. The existence of a range of age groups and both sexes in each phase serves to confirm the hypothesis that modular units within such later Bronze Age cemeteries represented the burial places of individual small social groups.

The urn cemetery developed gradually over a period of 1500 years (from c. 2100 to 600 BC).

The third phase of the Shaugh Moor Project comprised survey and excavation work on the Saddlesborough Main reave and Wotter Common (fig. 1). Following the decision of the china clay companies not to develop the area immediately around Saddlesborough (Area Y), Watts, Blake and Bearne & Co Ltd and English China Clays Ltd embarked upon a joint tipping programme which will eventually destroy the terminal reave east of Saddlesborough and most of the monuments on Wotter Common (fig. 2).

The area to the north of Saddlesborough was surveyed in 1976 and limited excavation was undertaken early in 1977 (Site 10, Wainwright et al 1979). An enclosure, 1 km to the north-east of Site 10, was totally excavated in 1977 and 1978 (Site 15, Wainwright and Smith 1980). In 1979 and 1980 the area threatened by the quarry extensions was surveyed and a programme of limited excavation was undertaken on the Saddlesborough Main reave and Wotter Common (Site 208).

This paper is derived from chapters of a PhD dissertation by Keith Muckelroy which was partially completed at the time of his death in September 1980. Publication of sections relating to work on two Bronze Age wreck sites was thought appropriate since the maritime context of the material and Muckelroy's assessment of its significance related to Bronze Age studies in general as well as the more specific focus of Middle Bronze Age maritime archaeology which formed the core of the research. In these chapters Muckelroy discussed the individual artefact types and their continental relationship and assessed the implications of the Langdon Bay and Moor Sand cargoes in terms of Bronze Age seafaring and the seaborne transportation of goods. His detailed discussion of the bronze forms has been summarized here; and it is upon the maritime aspect that this paper concentrates, this being the area in which Muckelroy's work at the two sites was making its unique contribution.

The two sites under examination, Langdon Bay and Moor Sand, are well-enough known through various interim statements (Baker and Brannigan 1978; Coombs 1976; Muckelroy and Baker 1979; Muckelroy 1980) to require little detailed introduction other than a short description here. Both are important not so much for the particular objects found, but for the implications of the associations and contexts and as a result of the briefest consideration it can be asserted that in the later second millennium BC there existed vessels of size and quality sufficient to transport bulk materials across the channel, probably at regular intervals, and certainly guaranteed of receptions which warranted the organization of Bronze Age shipping by a community or segment of a community.

It is now more than a decade since Avery & Close-Brooks (1969) and Musson (1970) recognized that some roundhouses which, at first sight, appeared in the archaeological record to have been single-ring buildings were most probably of double-ring construction. Using the excavation-plans of timber roundhouses at the Bronze Age settlements on Shearplace Hill in Dorset and Itford Hill in Sussex as the mainstays of their arguments, these writers gave reasons for believing that the ring of postholes, which constituted the principal archaeological evidence for the buildings in question, had held the uprights of an internal roof-supporting structure, while a separate, concentric, external wall, set some distance outside that post-ring, had received a more meagre earth-cut foundation, which in some cases had either ‘disappeared’ or ‘escaped detection’. These particular roundhouses had each stood on a terraced platform, and it is likely that this had afforded their ground-plans a measure of protection against at least mechanical weathering, with the result that ‘tenuous traces’ of an outer wall-groove survived in the chalk-bedrock of the House A platform at Shearplace Hill, while at Itford Hill the extent of the ‘trodden chalk’ areas of some of the floors supplied a vital clue.