The long history of investigations at Hoxne, Suffolk, beginning when John Frere discovered Palaeolithic implements there in the last decade of the eighteenth century, has been described by Moir. Moir himself worked on the deposits at Hoxne, and the results of his investigations, together with those of Reid, have formed the basis for our knowledge of the geology and archaeology of the deposits. From these investigations it is clear that Hoxne is an important site, for it is one of the rare places where there are interglacial lake deposits and Palaeolithic implements in direct association with ground moraines of the older glaciations. Moreover, the deposits occupy a well-known position in the East Anglian Quaternary succession.

This paper deals with a period of transition from prehistory to history. This age of transition is one which the Prehistoric Society cannot overlook, for its very title implies an interest in the historical method. But the transition raises special problems. History, starting from the written record, is largely occupied with political questions and political entities. Man, as we learnt from Aristotle, is a political animal and no-one, who studies the archaeology of the historical ages, would wish to deny the profound influence of political factors on the development of cultures. But prehistory, based on the natural sciences and on archaeology, focuses attention primarily on material phenomena, though the growing interest in the sociological implications of these phenomena points the way to a better appreciation of the political factors, which were operative in prehistoric as in later times.

This paper describes and discusses a group of five hoards of metal-work and coins found in 1948–50 by ploughing supplemented by excavation at Ken Hill, Snettisham, in north-west Norfolk. On account of the large quantity of precious metals included the find has been termed collectively ‘The Snettisham Treasure.’ Its principal contents are as follows:—Hoard A (p. 36) contained the remains of four gold tubular torcs. Hoard B contained three staters, four quarter-staters of the Bellovaci; four staters and one quarter-stater of the Gaulish Atrebates, all of gold (p. 59): Hoard C contained at least 145 speculum (tin) coins of Allen's Class I (p. 72) and three buffer terminal bronze torcs (p. 52). From Hoards B and C came at least 48 loop terminal torcs of gold alloy, bronze and tin (p. 46); 17 ‘ingot-bracelets’ of bronze and tin (p. 52) 14 rings of gold alloy, bronze and tin (p. 54); U-shaped bronze binding (p. 56); fragment of a bronze bridle bit (p. 57); two dome-shaped bronze rivets (p. 57); 10 iron nails (p. 58) and miscellaneous fragments of sheet-bronze and ‘cake’ of gold alloy and tin (pp. 57–58). Hoard D consisted of a gold loop terminal torc (p. 46) with securing ring of gold (p. 54). Hoard E contained the following gold objects: ring terminal torc (p. 63) with quarter-stater of Gaulish Atrebates (p. 59); bracelet (p. 66) and large buffer terminal torc (p. 67).

The cave known as the Mugharet Kebara opens at the western foot of Mount Carmel, 13 km. to the south of the Wady el-Mughara, and 1½ km. south of the railway station of Zikhron Yaaqov. It faces north-west on the 60 m. contour line, at 2½ km. from the sea, from which it is separated by a strip of marshy land and a wider belt of sand dunes.

The first sounding on this site was made in 1927 by Dr M. Stekelis, who was unable to continue owing to lack of funds. In 1930, unaware that Dr Stekelis had already tested the cave, I went over from the Wady el-Mughara with Dr Theodore McCown and spent one day in digging a trial trench (Garrod 1932), which revealed the presence of a Natufian layer with an underlying level containing an industry then unfamiliar, but since described under the name Kebaran (Turville-Petre, 1932). At the bottom of the trench we found a number of typical Aurignacian implements. In the summer of 1931 Francis Turville-Petre and Mrs C. A. Baynes undertook a three months' season at Kebara on behalf of the British School of Archaeology in Jerusalem and the American School of Prehistoric Research. During this time they camped with us at the Wady el-Mughara, so the two expeditions were in close touch and able to compare results throughout the campaign. The work at Kebara revealed a very rich Lower Natufian horizon (Layer B) with a thin underlying level (Layer C) containing the new Kebaran industry. This culture, which has proved to be fairly widespread in Palestine, Lebanon and Syria, is Neuville's Upper Palaeolithic VI (Neuville, 1951). Turville-Petre published the material from Layers B and C in the year following the excavation (Turville-Petre, 1932), but the Aurignacian industries of Layers D and E which were dug to contact with an underlying Levalloiso-Mousterian, have for various reasons never been studied or described in print.

It is convenient to begin this study of Azilian harpoons by first considering the Magdalenian harpoons of northern Spain. The two Basque sites of Aitzbitarte, a few miles within the Spanish side of the frontier, and Isturitz, not many miles within the French side of the frontier, are useful in delimiting the Spanish area from the point of view of the fauna. At Aitzbitarte, Harlé (1908) found a great predominance of red deer with one or two odd fragments of reindeer antler, whereas the well-known excavations by St. Périer (1936) at Isturitz showed a great preponderance of reindeer in conformity with other sites in the Pyrenees. The other occurrences in the three east Spanish provinces agree with that at Aitzbitarte (for instance, at Valle, Lorenzo Sierra, 1908, gives the ratio of antler fragments as red deer to reindeer: 73–1). Judged by the fauna, therefore, the modern political frontier corresponds roughly to the limit of the effective penetration of the reindeer westward, and consequently to the line of demarcation between harpoon-heads made respectively from red deer and reindeer antler.

The subdivision of the Magdalenian deposits in north Spain has never been satısfactorily resolved, and they have obstinately refused to conform to the French sequence. The main difficulty has been the lack of stratified sites; only Cueto de la Mina has yielded three layers; and two or three other caves have given two layers. Nevertheless two stages, an earlier one without and a later one with harpoons have been clearly distinguished (Breuil and Obermaier, 1936, 165–70, 192–4). Although no stratigraphical division of this later stage is possible, a number of interesting points arise from a typological study of the known harpoons.

The granite upland of Dartmoor has attracted settlers coming into the south-west of Britain from the early Bronze Age (Beaker phase) onwards. The long outline of the moor is visible from the South Devon coast, which with its series of good harbours invites intrusion. The navigable lower reaches of the south-flowing rivers—the Plym, Yealm, Erme, Avon, Dart and Teign—and the ridgeways along their watersheds provide convenient thoroughfares across the intervening coastal belt, the South Hams and the hinterland of Torbay, as far as the foot of the main escarpment. Above this steep ascent, where the granite overlies the Devonian rock formations, and where the rivers run in deep wooded gorges, the moorland mass forms a rolling tableland (1000-1400 feet high), broken only by the profile of the numerous upland valleys or by the fantastic eroded outlines of the tors. It is in this zone (light shading, fig. 1) that the evidence for most of the prehistoric settlement and early cultivation is to be found. Between 1400 and 1500 feet, the ground rises markedly again to the uninhabited northern heights (Yestor 2039 feet) and the southern heights (Ryder's Hill 1692 feet), in which the rivers have their sources and which are now blanketed in peat bog (dark shading, fig. 1). The heights apart, Dartmoor contains a broad tract of elevated open country, well-watered grass and heather moor, ideally suited to sustain a primitive pastoral economy. In many places on the marginal slopes and in the upland valleys a sufficient depth of sandy soil has accumulated on top of the rock to make arable cultivation possible; at present, cultivation ceases at about 1000 feet, which is the normal tres line.

The pioneer study of seed and plant-finds from prehistoric and Roman sites in this country was made by Helbaek and Jessen in 1944, and it was with the assistance of their findings that the present paper was originally written. Subsequently Helbaek carried the same research further, publishing his results in the Proceedings of the Prehistoric Society for 1952. These new results have considerably modified our notions of the cereal crops grown in both the prehistoric and Roman periods in Britain, partly owing to the examination of hitherto unanalysed finds and partly to the accumulation of experience in identification, which made it possible to correct previous errors. Thus, it has now been established that Triticum vulgare (Breadwheat) and Triticum compactum (Clubwheat) are much rarer among prehistoric finds than previously thought; and many specimens previously identified as Emmer were found by Helbaek to be Spelt. These modifications have necessitated a reconsideration of the bases of the present paper, but seem to the writer to confirm its main thesis.

Sir Lindsay Scott carried out excavations on the barrow on Whiteleaf Hill, near Princes Risborough, Bucks., with meticulous care from 1934 till the outbreak of war in 1939 interrupted operations. After the war he had no opportunity to resume excavations beyond making a 2 ft. section across the ditch on the west, and he was not able to complete the work he had planned or to prepare the material for publication before his untimely death in 1952. He gave four short and explicitly provisional reports in P.P.S. 1935, p. 132; 1936, p. 312; 1937, p. 440, and in the Records of Bucks. 1941–6, p. 298, but insisted that ‘this description must be taken as entirely provisional.’ His papers comprised a master plan contoured at 6 inch intervals, and showing the grid over the excavated area, small detailed plans of individual features—post holes, pits, and the ‘peristalith trench’—and a series of sections taken across the whole excavated area at 2 ft. intervals on both X and Y axes, but no complete plan of the excavated area nor any detailed description of the observations made in the course of the excavation. The relics are, however, all numbered with three co-ordinates, so that the position of each can be exactly located on the plans and sections, and these, especially the Neolithic pottery, are of such exceptional importance that it is an obvious duty to attempt a provisional account of the results achieved, without waiting for the completion of the excavation.

In the course of enquiries into the nature of the primary vegetation of Britain and its associated soils, the soils buried under prehistoric earthworks have provided valuable material. Such work is usually carried out on barrows which can be dated by the archaeological finds in them, but occasionally samples have been examined from undated barrows. Pollen analysis of the old land surface under an earthwork gives a broad picture of the vegetation extant on the site up to the time that earthwork was built, and it has been found that this picture varies according to the impact of prehistoric man upon the vegetation. This effect is, of course, progressive and it is possible to show that the later the earthwork the further removed is the flora from the forest conditions which once prevailed. Once such a sequence has been established, it should be possible, by pollen analysis alone, to allot an undated construction to its approximate position in the sequence.