In the Proceedings of the Prehistoric Society for 1964, the writer presented (Roe, 1964) an interim report on his programme of research on the British Lower and Middle Palaeolithic material, describing methods of studying assemblages of handaxes by metrical and statistical analysis, and outlining the first results obtained. The present paper offers a summary of the final results obtained for certain aspects of the same research project, in particular the discerning of various groupings among the 38 handaxe assemblages considered. The Groups which emerge are clusters of sites whose industries are similar in terms of the morphological range and usual degree of refinement of their handaxes.

The Thesis for which the research was undertaken was completed in March, 1967 (Roe, 1967), and is available for consultation. It is of course quite impossible to quote in the present paper more than a small part of the evidence upon which the final conclusions depend.

During the first quarter of this century, not much was known about the Indian Stone Ages and only two phases, Palaeolithic and Neolithic (including microliths), were recognized. Some of the earlier discoveries recorded by Robert Bruce Foote, Coggin Brown, Cammiade, Miles Burkitt, De Terra, Paterson and others gave an impetus for the later researchers and the Stone Age cultural sequence began to take a clear shape. The most significant among the later discoveries is the identification of a distinct culture by H. D. Sankalia, which is now designated as the Middle Stone Age, and the extent of its distribution is now well established. Prehistorians who launched a systematic survey unearthed the remnants of the stone using populations from different parts of the country, viz., the Panjab, Rajasthan, Uttar Pradesh, Bihar, West Bengal, Orissa, Madhya Pradesh, Maharashtra, Gujarat, Mysore, Andhra Pradesh and Madras recognizing three phases, The Early Stone Age, the Middle Stone Age and the Late Stone Age, occupying the time span of Pleistocene and Early Holocene times.

The material culture that represents the Stone Age Communities in various parts of the country displays a homogeneous pattern of development. The Early Stone Age is followed by the Middle Stone Age which in most of the regions gave way to the Late Stone Age; however, a blade and burin industry with a discontinuous horizontal distribution is known in a few regions between the Middle and Late Stone Ages.

The site whose excavation is here recorded is a small kite-shaped enclosure all but obliterated by ploughing. It is situated (ST (179) 942197; 6-inch sheet ST 91 NW) at the southern tip of a spur known as Berwick Down 1 mile north of the village of Tollard Royal on the borders of Wiltshire and Dorset. It is surrounded on its south or downhill side by a semi-circular bank and ditch. The locality has been recently described briefly by H. C. Bowen and P. Fowler whose plan (1966, 46–8, fig. 2) is here reproduced (fig. 2). The other two sites occupying the 16 acres of the spur comprise:

(1) An Iron Age settlement to the north consisting of a concentration of unenclosed pits, a large round house demarcated by a pennanular palisade groove and two cross-dykes.

(2) A circular enclosure 2½ acres in extent containing Romano-British hut platforms and crossed in its southern sector by a modern fence. To the north of this fence the earthworks have never been ploughed and are in a state of preservation, only too rarely found in southern England. To the south of the fence the downland has been heavily ploughed over a number of years.

The earthworks of the kite-shaped enclosure had become so degraded that in 1962 the Ministry of Public Building and Works initiated a trial excavation under the direction of Mr E. Greenfield. With the assistance of Miss V. Russell, Mr Greenfield covered the area with a 10 foot grid of test-holes which were expanded into trenches when required. In 1965 the site was put down to grass and the earthworks planned in the spring by members of the Royal Commission on Historical Monuments, by which time the bank and ditch of the kite-shaped enclosure were virtually invisible. In August and September of that year the interior of the enclosure was completely stripped by the author on behalf of the Ministry of Public Building and Works.

The Iron Age hill-fort at Grimthorpe (Grid reference SE.816535) in the parish of Millington, East Riding of Yorkshire, is on the western edge of the Yorkshire Wolds, with a commanding position over the Vale of York. There is an uninterrupted view to the White Horse on the Hambleton Hills, 25 miles to the north-west; beyond York, 13 miles to the west, to the Pennines; and to the south 25 miles to the chimneys of Keadby and Scunthorpe. To the west and south the land slopes away to the Vale of York, and to the north and east there is a sharper fall to Given Dale and Whitekeld Dale. The hill-fort defences follow the 520 feet contour, and enclose an approximately circular area of eight acres (fig. 1).

A traditional reference may be preserved in the field-name—Bruffs—perhaps a variation of ‘Brough’, which ‘refers in all cases to ancient camps, usually Roman ones’. But all surface indications have now been obliterated by ploughing, and even a century ago there was little more to be seen. John Phillips in 1853 noticed ‘unmistakable traces of ancient but unascertainable occupation’, and in 1871 an excavation by J. R. Mortimer located ‘the filled up inner ditch of a supposed camp’. But Mortimer was not concerned with the settlement; his interest had been aroused by the discovery, in 1868, of a burial with rich grave-goods, including metalwork with La Tène ornament, in a chalk-pit within the south-west sector of the hill-fort.

The site at Eldon's Seat (fig. 1) lies on the side of a shallow valley cut into the Kimmeridge clay bedrock. Occupation began probably in the 7th century B.C. or earlier, at a time when the ceramic assemblage was based upon Late Bronze Age traditions, and continued with interruptions until the 2nd century A.D. Three major periods have been isolated: during period I (Late Bronze Age style pottery) and period II (haematite-coated ware) a series of circular huts were built against the downhill side of a sporadically growing lynchet. The huts were occupied by farming communities who were also engaged in making armlets from the local Kimmeridge shale. Period III buildings have not been identified in the excavated area, but Durotrigan and Roman pottery, which typifies this phase, has been recovered from the uppermost layers of the lynchet.

The finds from the excavations will be deposited in the Dorset County Museum, Dorchester, Dorset.

The peat deposits in the Somerset Levels have yielded many finds of prehistoric and early historic age, ranging from flint and metal artefacts to roadways and other structures. These finds have been a source of interest to archaeologists for many years, and the work here described represents only one aspect of the early occupation of the region. The Levels themselves lie between the major uplands of the Quantock Hills and the Mendip Hills (fig. 1), but within the Levels are extensive ridges of limestone (including the Polden Hills and the Wedmore Ridge), smaller and lower areas of limestone (such as Westhay and Meare), and patches of sands (Burtle Beds) all of which provided dry ground for occupation or other activities. We are concerned in this paper with the earliest wooden roads and tracks of Neolithic age which were built to allow communication between the Westhay ridge and the largest Burtle Bed north of the Polden Hills (pl. XIV).

Some objects defy archaeologists to understand their function. Two gold rod-like fragments from the Kerivoa hoard, Côtes-du-Nord, and one from Plouarzel, Finistère (Briard, 1965, fig. 21, 2), are the only gold objects on the Continent with broadened terminals similar to the ‘neck-ring’ from Arlon, Belgium. Three similar rings in copper exist in Britain. These six objects, their function, and their relationship to other prehistoric objects, form the nucleus of this article.

The two Kerivoa fragments (pl. XVI, B), with rod-like body and broadened ends, fit together to form a complete object like that from Arlon. They were a part of the Kerivoa hoard which also contained three gold lunulae with a fragment of sheet gold broken into four pieces. These were found near Bourbriac, Côte-du-Nord, France, in 1932, surrounded by a dark mass, which has been interpreted as a disintegrated wooden box. The fragments with broadened terminals and their association with lunulae are critical to this paper.

In these Proceedings over 15 years ago Gordon Childe made a classic survey of the evidence for the earliest wagons and carts in the Near East and Europe, following it three years later with a similar treatment using some additional material which had become available by that time (Childe, 1951, 1954a, b). Since then much new primary archaeological evidence has been discovered or recognized for what it is, and there have been two recent important works of synthesis, that of Boná (1960) on vehicle models in the Middle Danube, and that of van der Waals (1964) of finds of actual disc wheels in the Netherlands. From 1956 a large number of actual vehicle finds have been made in excavations on the shores of Lake Sevan in Armenian SSR; an additional second find comes from Trialeti in Georgian SSR, and new vehicle-burials found near Elista in the Kalmyk steppe, so that we now have a total of nearly 30 wagons, carts and chariots surviving wholly or in part from southern Russia, mainly of the 2nd millennium B.C. In the last 10 years or so, too, new work on the Neolithic and earliest metal-using cultures of East Europe, and on the Kura-Araxes Culture of Transcaucasia and its counterpart in eastern Anatolia has thrown new light on early wheeled transport in such contexts. Despite the penetrating and far-reaching conclusions pursued by van der Waals only four years ago there still seems room for a survey reviewing these in connection with other aspects of this cardinal problem in Old World prehistory.

The very early development and extent of the obsidian trade in the Near East is being increasingly documented by the discovery and excavation of Early Neolithic sites throughout the area. Since the publication of our last paper (Renfrew, Cann and Dixon, 1966), obsidians from five aceramic Neolithic sites have been analysed, and the natural source of the material determined in each case. All the results fall into the framework of groups already established. In addition, one new obsidian source has been discovered.

This confirmation of the characterizations hitherto achieved is satisfactory, but perhaps more important is the new evidence for the reconstruction of the pattern of the trade or traffic in obsidian in Early Neolithic times. Regularities in this pattern are becoming apparent, which allow a greater insight into the trading mechanisms involved.

In this paper, written by Colin Renfrew on the basis of analyses and group divisions by J. E. Dixon and J. R. Cann, new information concerning the Anatolian sources is first considered, and new results for the Early Neolithic and later obsidians discussed. In the last section, an attempt is made at the more precise definition of the obsidian trading pattern in the 7th and 6th millennia B.C. on the basis of a more detailed analysis of the prehistoric trade statistics.

The post-Pleistocene archaeology of southern Italy is a rapidly developing study and for this reason, while many facets of it are still little known, others are undergoing fundamental changes as new discoveries are made. Particularly notable contributions have been made in the last 20 years by, among others, L. Cardini, P. Graziosi, A. Palma di Cesnola, E. Borzatti von Löwenstern, O. Cornaggia Castiglioni and F. Zorzi in the palaeolithic and mesolithic field, R. B. K. Stevenson, L. Bernabò Brea, D. Trump and S. Tinè in the Neolithic sphere and D. Trump, S. M. Puglisi, R.Peroni, F. G. Lo Porto and H. Muller-Karpe for the bronze age. As a result of this work it is now possible for us to reconstruct in outline the later prehistory of southern Italy. Most of the work that has been done has been on individual topics, and, while some attempts at synthesis have been made (most notably those of A. M. Radmilli, D. Trump and R. Peroni), these have taken the form of a culture-by-culture description, as it were a commentary upon a vast composite stratigraphy of the area. No attempt, however, has yet been made to produce a history of the human occupation of the land and it is towards this aim that my paper is directed.

The emphasis in this paper is on settlement and economy in relation to environment. It is through the techniques of his subsistence economy that man deals with (or in evolutionary terms, adapts to) his environment, and in this sense it can be regarded as the most important factor contributing to the nature and development of society.

In recent discussions of the origins and process of animal domestication (Reed, 1961, Zeuner, 1963), both authors rely on two kinds of evidence: on the one hand, the present distributions and characteristics of the different breeds of whatever animal is being discussed, together with its feral and wild relatives, and, on the other hand, the past record, given by literary and pictorial sources and the bones from archaeological and geological sites. Increased recognition of the limitations of the past record, whether in the accuracy of the information it appears to give (as in the case of pictorial sources), or in the certainty of the deductions we are at present capable of drawing from it (this applies especially to the osteological record), has led these authors to argue mainly from the present situation, using the past record to confirm or amplify the existing picture.

Arguing from the present, many hypotheses about the origins and process of domestication are available. The only test we have, when attempting to choose between these, lies in the direct evidence of the past record. The past record, it is freely admitted, is very fragmentary: the information provided by the present situation is more exact, ranges over a much wider field, and is more open to test and control. Nevertheless, the past record, however imperfect it is, is the only direct evidence we have about the process of domestication.

In the 1964 Proceedings a preliminary report was published of an initial sounding at this site, discovered and named (after a nearby village) in 1962. The main excavation, in the summer of 1964, was undertaken with the kind permission and co-operation of the Iranian Government and authorities, and with the financial assistance of grants from the British Academy and the Crowther Beynon Fund of the University of Cambridge. Subsequent laboratory analysis has been carried out mainly in the Department of Archaeology and Anthropology of the University of Cambridge.

Since the geographical situation of the site forms an essential factor in its interpretation, the main features may be repeated here for convenience. The cave is eroded in the base of an escarpment which follows the foot of the Elburz range where it bounds the coastal plain along the southern shores of the present Caspian. The mountains rise abruptly in an impressive series of ridges to over 10,000 ft; the plain extends to the north to the modern sea-shore now some 8 miles away from the site and rapidly retreating. It is known however that in the recent geological past this situation has varied greatly. At times the sea lapped the base of the hills leaving traces in the form of escarpments and raised beaches, while at others it retreated scores of miles to the north and may even have disappeared altogether.

The Early Iron Age pottery of the Herefordshire-Cotswold region is marked by a predominance of the well-known duck-stamped ware associated with a thinner spread of pottery ornamented with tooled lines. It has been customary to equate the former with the presence of an intrusive culture now generally labelled Western Third B, while the linear-tooled pottery has been referred to the Western Second B culture and is held to represent the indigeneous population. However, this is based on the unwritten assumption that the pottery was produced in or near the home and hence can be used as a cultural type-fossil. Petrological examination suggests that this is unwarranted since it appears that the pottery was the object of a well organized trade involving specialist potters based in the Malvern district. The distribution could thus be due to purely commercial factors, and this necessitates a reassessment of the cultural interpretation. The object of this paper is to present the new evidence and to discuss briefly its implications.

Although concerned mainly with the stamp ornamented and linear tooled ware of the west Midland and Cotswold regions, this paper deals occasionally with undecorated pieces. The ornamented vessel from Pen Dinas, Aberystwyth has been included because of its obvious resemblance to some of the pots from Bredon Hill, Worcestershire, but other stamped sherds such as that from Methyr Mawr, Glamorgan or the Cornish duck-stamped pottery have been excluded on typological grounds.