This article concerns the prehistory of Northern Italy from the end of the ninth to the beginning of the fourth millennium bc. The main aim in this work is to interpret the available archaeological evidence in an attempt to understand the cultural change which occurred through the period under study.

The gap existing in our knowledge between the Late Epigravettian and the Early Mesolithic has been partially rilled by recent discoveries of Final Epigravettian encampments on the Baldo (Bagolini and Nisi 1976) and Bondone (Bagolini 1976) mountains in the Veneto. and Trentino regions. Unfortunately nothing is known about the natural history of the area, because the Baldo and Bondone sites did not yield any faunal remains. Only a few charcoal samples were recovered from the peat bog of Viotte on Bondone, where Final Epigravettian hunters settled. These charcoal samples have not yet been identified. In my opinion, the Bondone encampment might be attributed to the Alleröd oscillation. Palynological evidence of this period is almost totally lacking except for the area south of Lake Garda. At the Castellaro Lagusello peat bog, Bertoldi (1965) identified both the Bölling and Alleröd oscillations from the same core at the depths of 6·50 and 6·0 m respectively. The presence of oak pollen at the depth of 6·50 m varies from 0·2 to 0·6%. This pollen disappears after this moment to reappear at the depth of 6·0 m when birch pollen increases and pine decreases abruptly. The Bölling oscillation has been dated by the radio-carbon method to 11250±250 bc, while no date is available for the Alleröd one.

The samples to be discussed in this paper derive from a series of late Mesolithic shell-midden sites on the island of Oronsay in the Inner Hebrides. Despite the small size of the island (less than six square kilometres) at least six separate middens have now been identified on Oronsay, of which all except one are located along the south-east-facing coast of the island (fig. 1). Radiocarbon dates ranging from 3695 ± 80 to 3200 ± 380 bc have been obtained on samples of charcoal from the middens, and point to a considerable measure of chronological overlap in the occupation of the different sites (Mellars 1978).

Archaeologically, the chief interest of these sites lies in the excellent conditions for the preservation of faunal remains, which provides a unique insight into the economic activities of coastal-living groups shortly before the appearance of formally ‘Neolithic’ communities in northern Britain. In addition to a wide range of molluscan, mammalian and bird remains, the sites have yielded exceptionally large numbers of fish bones. While the total range of fish exploited from the sites is impressive (comprising at least 12 species), analysis has shown that by far the major emphasis in the fishing activities was placed on the exploitation of a single species—the saithe or coalfish (Pollachius virens L.). In all the samples so far analysed this species accounts for at least 95 per cent of the total fish-bone material, and in many of the samples it is the only species represented.

Six of the twenty-five British implement petrology groups (Clough and Cummins 1979, 127)—Group I (Penzance, Cornwall), Group IV (Callington, Cornwall), Group VI (Langdale, Lake District), Group VII (Penmaenmawr, North Wales), Group VIII (South-west Wales), and Group XVI (Camborne, Cornwall)—account for almost half of all the stone axes so far examined from England and Wales. In every part for the country, one or other of these six is the most abundant individual group. On the basis of stone axe distribution studies (Cummins 1979), the country seems to fall naturally into three major provinces (fig. 1), which might possibly be interpreted as Neolithic tribal territories. Northern and Central England, the largest of these provinces, is dominated by Group VI axes which, though originating in the Lake District, seem to have been distributed from a secondary centre in Humberside. Wales, including Herefordshire and part of Shropshire, forms another province and is dominated by Welsh axes, Group VII in the north, and Group VIII in the south. Southern England, the third province, is dominated by Cornish ‘greenstone’ axes, mainly Group I but locally, in the south-west, Groups IV and XVI.

Cumulative percentages of all axes belonging to each group plotted against distance from its distribution centre (Cummins 1979, figs. 4–9) give an indication of absolute frequency variation in relation to that centre. The shape of the plots is controlled by two variables, (i) the cumulative increase of area with distance from the centre, and (ii) the variation in average frequency of the grouped axes (per unit area) with distance from the centre.

In their report of Durrington Walls, Wainwright and Longworth (1971) comment ‘It is noteworthy that not one certain grain impression has been recorded on any Grooved Ware sherds, nor is there any other evidence for the cultivation of cereals’. They conclude that the economy of Grooved Ware cultures is based on pastoralism and strandlooping rather than agriculture, and this conclusion has been generally accepted. Two points may be raised about this conclusion. First, the absence of grain impressions need not imply the absence of cereal cultivation. Hubbard (1975) has convincingly argued that the occurrence of pottery impressions does not provide the representative picture of crop economy suggested by Jessen and Helbaek (1944, 28). Such variables as the nature of the refuse deposited in the potter's working area and the degree of cleanliness required by the potter in his work will strongly influence the occurrence of impressions. For example, Hubbard has suggested that the sharp decrease in the frequency of impressions of emmer wheat in the Bronze Age may reflect its careful storage for human consumption, in contrast to barley which, as animal feed, would be more likely to get scattered around the site and incorporated into pottery.

The second point concerns the apparent absence of carbonised grain. This may be an illusory consequence of inadequate methods of retrieval in the field. Over the last decade, the use of water flotation has demonstrated that carbonised grain and seeds are many times more common than had previously been thought (Jones 1978).

The Shaugh Moor project is concerned with an area of moorland in south Dartmoor north-east and north respectively of the villages of Shaugh Prior and Wotter (fig. 1). The physical threat to the evidence for settlement and land-use caused by the operations of the China Clay industry involved the Central Excavation Unit of the Department of the Environment from 1976 in a programme of survey, excavation and environmental studies related to the settlements, land boundaries, burial mounds and ceremonial structures that are to be found on this piece of moorland. The background to the project and its preliminary research strategy have been outlined in Paper I (Wainwright et al. 1979) and this publication describes the investigation of a stone-walled enclosure surrounding houses and other structures that was totally excavated in 1977 and 1978. Subsequent papers will describe the related archaeological and scientific investigations into the past environment and land-use of this block of moorland and its adjacent region.

This barrow was excavated by the author for the Ancient Monuments Inspectorate of the then Ministry of Public Buildings and Works (now the Department of the Environment) in August and September 1965. The barrow was scheduled under the Ancient Monuments Acts by the Department of the Environment as Wiltshire 563, and all finds were accordingly thus marked at the time of excavation. In 1958 the barrow was observed by the Ancient Monuments Inspectorate as being 4½ ft high and the mound was spread by agricultural activity to a diameter of 75 ft. In 1965 this bowl barrow stood barely 18 in. high, having suffered much from the effects of ploughing. The barrow had been disturbed mainly in the centre of the mound by early, possibly nineteenth century, excavations. At least one slit trench had been dug into the side of the mound during the 1939–45 war and much other disturbance had been caused by rabbits.

Residual finds from this site caused a particular problem. Where they seem to be dateable they have been inserted in their appropriate section, otherwise where there is a possible alternative date they have been cross referenced. All ranging rods and scales shown in the photographs of the site are marked in feet and inches, except where otherwise obvious.

Excavated sites referable to the earlier first millennium B.C. are rare in the south-west peninsula. Although Dartmoor and Bodmin Moor exhibit well preserved Bronze Age landscapes, most of the upland settlements are attributed to the second millennium and there are few which on present evidence fall within the succeeding centuries (Silvester 1979). Beyond the granite moors, pottery from Kent's Cavern, Torquay and the Mount Batten peninsula, near Plymouth, suggest sporadic occupation, while radiocarbon dates from Killibury in Cornwall imply activity prior to the construction of the hillfort (Miles 1977, 111). Yet such is the imbalance of archaeological activity in the region that even in well worked areas such as west Cornwall, few sites of relevant date can be readily identified. In Devon, other than on Dartmoor and Exmoor, the limestone plateaux around Torbay provide the only surface evidence of open settlements. A small proportion of the modern fields and woods on the limestone preserve groups of clearance cairns and rubble banks of varying complexity; many of these are undoubtedly medieval in date, but in two locations at Dainton, also known as Miltor Mator Common (SX 859668), and Walls Hill, Torquay (SX 935651), there are the unobtrusive remains of rectangular field systems. The former consisted of a minimum of perhaps ten fields, averaging just over 0·4 ha. in area, and randomly spaced cairns, most of which are presumably contemporary (fig. 1). Part of Walls Hill is covered by dense ashwood scrub, a recurring feature of the limestone plateaux, but on the cliff top are the vestiges of nine fields which are generally of smaller size than those at Dainton. That these sites survive is fortuitous but perhaps not surprising; the soil on the plateau is well-drained but normally less than 0·3 m in depth (Clayden, 1971, 122). Loose rock occurs frequently and occasional outcrops add to the problems of agricultural use. Consequently the limestone hills tend to be given over to permanent pasture and are ploughed only rarely.

One of the major problems of British prehistory has been the contrast between the mass of Late Bronze Age metalwork and the rarity of contemporary settlements. The Berkshire river gravels are one area in which a high proportion of bronze objects is recorded in apparent isolation. With the increasing recognition of Late Bronze Age pottery, however, it has been possible to identify domestic finds of this period among the artefacts from gravel pits around Reading. Part of the gap in the settlement record has also been closed by the excavation of two sites on the Berkshire Downs, the earthwork enclosure at Rams Hill, and an open site at Beedon Manor Farm (Bradley and Ellison 1975; Richards in press). But it was not until 1974 that a Bronze Age settlement on the gravels could be examined in situ, and since the formation of the Berkshire Archaeological Unit a series of five sites have been sampled or more extensively investigated. This paper is concerned with the two most extensive sites, those at Aldermaston Wharf and Knight's Farm, Burghfield, but will make cross reference to the other work where necessary, in particular to a more recently recorded site at Pingewood.

It is now clear why this evidence was so difficult to find. The pottery is extremely friable and would not survive on the surface; and the gravel sites contain very few worked flints. The main features are small pits under 50 cm deep, and for this reason the sites cannot be detected from the air; and, even if they could be recognized, there would be nothing to distinguish them from Iron Age open sites, like those in the Upper Thames Basin.

This paper discusses the development of pottery styles in southern and eastern England during the first half of the first millennium B.C. The region discussed is Hawkes' Southern Province (1959, fig. 1), and excursions will also be made into the Eastern Province (fig. 1).

The discussion of settlement sequences, and the dating of individual sites, is still largely dependent upon ceramic refuse derived from such sites. The analysis of settlement patterns for the earlier part of the first millennium B.C. thus rests upon our understanding of the ceramic traditions of that period, and this has led to considerable confusion. Two problems must be isolated at the outset. The first results from the re-evaluation of British Bronze Age chronology which took place in the 1950's, leading to Hawkes' 1960 scheme and then to Burgess' rather rigid reading of the evidence nine years later (Butler and Smith 1956; Smith, M. A. 1959; Smith, I. F. 1961; Hawkes 1960; Burgess 1969). The main effect, if we are to follow Burgess, was to draw Beakers, Food Vessels and most Urn forms back into an early Bronze Age the end of which, around 1400 B.C., was marked by the end of the Wessex grave series. As for the Middle and Late Bronze Age Burgess concluded that ‘over much of the British Isles there are no settlements, burials, defended sites, pottery, or other non-metallic cultural material which can safely be assigned to the Middle or Late Bronze Age. There are a few localized exceptions such as the Deverel-Rimbury culture and Flat-rim ware’ (Burgess 1969, 29).

In recent years increasing attention has been focused on the palaeoeconomic interpretation of prehistoric sites, and a number of projects have been carried out both in Britain and abroad on plant and animal remains from archaeological sites with a view to interpreting subsistence patterns and activities of early communities. Though large-scale, systematic sampling and recovery of palaeoethnobotanical data has been carried out on several sites in the last ten years, the practice is still not widespread and has yet to be judged by its results. Fundamental issues relating to sampling, recovery and data presentation still require attention. Although studies of palaeoeconomic material have been done on Iron age sites in southern Britain, for example at Portway (Murphy 1977a) and Abingdon (Parrington 1978), as yet little evaluation of this evidence has taken place. The aim of this paper is to present, assess and compare the evidence from two such studies of different types of Iron Age site.