Early Man: Quaternary geology and Palaeolithic Man

The quaternary geology of Oxfordshire with reference to Palaeolithic Man (fn. 1) : Introduction

In the last hundred years the superficial deposits of Oxfordshire and the flint implements that occur in some of them have received a great deal of attention. Buckland, Phillips, Prestwich, and Professor Sollas in turn have investigated them during their tenure of office in Oxford; in Reading Professor Hawkins and Mr. Ll. Treacher have studied them for many years, while at the other end of the county Miss M. E. Tomlinson of the University of Birmingham has added to our knowledge. The Geological Survey has made important contributions in its Memoirs, both of the Old and New Series, which are supplemented by occasional short reports in the annual Summary of Progress. Independent researches have been carried out by geologists too numerous to mention here, but many useful contributions are mentioned below and in the bibliographic references of the Geological Survey memoirs. In the space available in this volume it is impossible to include an exhaustive bibliography. A recent publication of the Geological Survey, namely, London and the Thames Valley (1935), attempts to view the Pleistocene deposits of the valley as a whole: much, however, remains to be done. The same task is now to be attempted for the Upper Thames and part of the Middle Thames, and it must be clearly understood that finality has not been reached. The following pages are no more than an attempt by the writer to review as a whole the extremely varied, widespread, confused, and confusing superficial deposits of a county that embraces regions with widely dissimilar post-Tertiary histories.

The writer is indebted to Professor H. L. Hawkins for reading the final draft, and for suggestions, which have been adopted.

The superfical deposits

These will be grouped as follows:

I. High-level deposits and the surfaces on which they rest.

A. Plateau Drifts:

(1) Deposits of uncertain origin, in part, at least, residuals of older strata.

(2) Glacial deposits.

B. Clay-with-flints and analogous accumulations.

II. Boulder Clays, with gravels and alluvial deposits.

III. Trail, Coombe Rock, alluvial fans, and dry valleys.

From the above it emerges that within the county are boulder clays and glacial beds at high and low levels, non-glacial deposits of streams and rivers, and superficial accumulations formed before, during, and after the Ice Age. Parts of the county were overwhelmed by land ice; much of it escaped but endured a rigorous climate. There is evidence of two genial climatic oscillations; Palaeolithic implements and the varied mammalian life of Quaternary time are fairly well represented.

I. High-level Deposits and the Surfaces on which they rest.

This part of England has been a land area almost without interruption since early Tertiary times: a vast amount of denudation has been effected and unfossiliferous detrital residues have been formed, eroded, and redeposited. It is often impossible, therefore, to discover the age of a deposit or of a surface upon which it rests, and the two may be of widely dissimilar age. Much has been written on the subject in the London Basin, but in the writer's opinion the results achieved there cannot readily be applied on the west side of the gorge through which the Thames passes near Goring, since that gap in the chalk escarpment may be a geologically recent feature. The sub-Eocene surfaces in the Chiltern Hills can be mapped in detail to the top of the scarp; there are regions of similar altitude that are either bare chalk or covered by residual deposits: these surfaces may be assumed to be stripped portions of the sub-Eocene floor, but their cover may be of later date.

In the great clay lowlands west of the Chiltern escarpment and on the flanks of the Cotswolds beyond it is even more difficult to assess the age of surfaces or residual beds upon them, but we can fairly assume that most of this country was moulded after Eocene times. It is also clear that the northern affluents of the Upper Thames Basin at one time had longer courses than they now occupy, and that some of them, notably the Evenlode and Cherwell, have lost or are losing their original watersheds to the Severn Basin. Whether or not the original headwaters of the Thames Basin were in Wales and the Welsh border is a matter of opinion.

In the country around Oxford some early Pleistocene deposits are preserved in the form of a thin drift over the plateau-like divides of Jurassic limestones between the rivers or above bends in their courses: the surfaces upon which they lie are older and may represent within reasonable limits the configuration of the country at the close of Tertiary times. If this is so, the higher parts of the county have not suffered severe denudation in Quaternary times, and locally the old valley surfaces may have been within 100 ft. or less of their present level. On the whole the river valleys and adjacent lowlands have probably been lowered 200–300 ft. This is at any rate the order of magnitude. The question is discussed in more detail in the second edition of the Geological Survey Memoir, 'The Geology of the Country around Oxford': (fn. 2)

A. Plateau Drifts.

Most of the superficial deposits outlined above come under the indefinite heading of 'plateau drift', which is honoured by long and unsatisfactory usage in south-central and southern England. Especially on the south side of the Thames, downstream of Goring Gap, they are complicated by the presence of 'southern drift', the product of the denudation of the folded Cretaceous beds. North of the Thames the following are included under the title of 'Plateau Drift':

(1) Deposits of Uncertain Origin, in part at least Residuals of Older Strata. On the Chiltern Hills there are patches and long trains of mixed deposits, sometimes disordered, sometimes rudely bedded or false-bedded, sands, gravels, pebbly drifts, wisps of clay, nodules of flint, battered pebbles, yellow and white quartz, and other durable substances. Some of the patches, often marked by trees and heath, are at the highest levels, to the very crest of the Chiltern scarp, and the gravels of similar composition are frequent between bare, dry valleys cut in the Chalk.

The banded arrangement may therefore be due chiefly to denudation, and the sheets may have been more or less continuous. Probably the whole of this material has been derived from the Eocene cover, or the Chalk, but the nature of its redeposition has not yet been fully explained.

In the lowlands of the Thame and Thames many of the higher hills of the clay-land are capped by sandy gravel, which bears no obvious relation to the present drainage. In the Thame Basin flint pebbles have been brought considerable distances from the Chiltern scarp, elsewhere quartz and quartzite become the more prominent constituents, with local material from the Jurassic rocks; lime-bearing rocks are usually absent. The siliceous rocks are generally considered to have been brought from the Midlands in early stages of the drainage system. Quartz sand and pebbles are available, especially in the Greensands.

On the higher parts of the Cotswolds scattered pebbles of quartz and quartzite lie on the surface and in solution-pipes in the limestone. Their origin is obscure, but again is often attributed to the early stages of the drainage system.

(2) Glacial Deposits. These are among the most interesting and perplexing in the county. Their typical development, in the country around Oxford, has been discussed in the Survey Memoirs (both editions) and in a paper published in The Quarterly Journal of the Geological Society of London, 1929, lxxxv, 359–88. Briefly it may be stated that they consist of pebbles and boulders in a matrix of sand or clay, usually unstratified; they are devoid of limestone or soluble constituents and their vertical distribution is approximately between 100 and 350 ft. above the nearest main river (e.g. Thames, Cherwell, Evenlode), and below about 600 ft. O.D., above which scattered pebbles and small pockets occur. The pebbles and boulders are almost without exception foreign to the district in which they occur, and striated boulders are common: in other words, they are erratics and the products of glaciation. To what extent glaciers entered the district and to what height is by no means clear; in fact, the sources of some of the material are so widespread that it is difficult to credit land-ice alone with its transportation. It has been suggested that a considerable submergence took place in early Pleistocene times and that some of the erratic material accomplished part of its journey in floating glacier and shore-ice. Stiff clay recognized as Boulder Clay occurs in some places, but the greater part of the mass has been deposited at all altitudes between the stated limits and seems to have been subsequently re-sorted by the rivers and deposited in terraces or benches. (fn. 3)

Analogous deposits occur in the gorge of the Thames between Goring Heath and Reading, and have been claimed by Hawkins (see pp. 334–5) to be of glacial origin. He also considers that a part of the excavation of the gorge may have been effected by glacial overflow. The detail of this argument is of interest, and readers are strongly recommended to consult it. Harmer's hypothesis of 'Lake Oxford' will be familiar to most readers, one of many attempts to account for the creation of Goring Gap, but Hawkins's thesis and the fact that the Plateau Drift is redeposited on graded courses from Oxford to Reading both tend to displace the theory.

Dines has also considered the problems of this undoubtedly glacial drift in the north-western Cotswolds, (fn. 4) and concludes that there is no suggestion of their glaciation having been severe. He attributes the absence of thick or large spreads of Boulder Clay on the uplands to the 'probability that only the higher portions of the ice-sheet over-rode the hills'. Denudation undoubtedly reduced these patches, and it remains to be proved that Boulder Clay was in fact laid down on the higher parts of the uplands.

B. Clay-with-flints and Analogous Accumulations.

In contrast to the drifts enumerated above we have to consider superficial accumulations which have been formed in place. Of these the most obvious is the clay-with-flints of the Chiltern Hills, a residue not uniformly spread over the chalk. It is widespread and thick in the Princes Risborough district, where the county boundary traverses the Chilterns, and it supports the magnificent beech-woods of this part of the hills. With other superficial deposits it covers the surface in such thickly wooded districts as those of Nettlebed and Henley. It is much reduced on the bare slopes between the Thames near Wallingford and the Oxford-Henley road, and north-eastwards along the face of the scarp, where the barren slopes are in contrast to the dense woods of the crest. It is evidently a residue of chalk that has been removed by solution, flint nodules, often much shattered, thus being concentrated. Locally, however, fine residues of Eocene sands and clay may be involved in a red flinty earth. Special reference should be made to H. J. Osborne White's well-known contributions to the study of the clay-with-flints and other superficial accumulations of the chalk downs. (fn. 5)

On the Jurassic limestones a somewhat similar reddish-brown earth is formed on plateau-sites, and is frequently seen as a filling of joints and solution-cavities.

Locally, deep subsoils of similar red earth are produced on gravels principally composed of Jurassic limestone, and pipes penetrate the gravel below: they are well known on the terrace gravels of Handborough and Kirtlington, and the writer has seen them also on much younger gravels near Yarnton. These occurrences show that the material is still 'growing'; at Yarnton the process can be watched in detail.

II. Boulder Clays, River Gravels, and Alluvium.

Apart from the Glacial Plateau Drift of the Upper Thames Basin and of the Goring gorge, Boulder Clay of more recent age encroaches on the region, but barely enters it. The best-known district is in the head of the Evenlode Valley in the Moreton-in-Marsh area, where Miss Tomlinson has devoted attention to it. In the Cherwell Valley remarkable Boulder Clays are described in the Memoir (of the Old Series) from Banbury, (fn. 6) but much yet remains to be done in these areas. Farther east rewashed Boulder Clays in which the writer found Scandinavian and North British erratics occur on the Tingewick-Buckingham ridge, but they do not seem to have advanced farther towards Bicester. On the eastern boundary glacial deposits probably of the same age are known in the country between Buckingham and Aylesbury, but again they do not seem to have penetrated much farther west into Oxfordshire. There is no evidence that the glaciers passed through the Cotswolds to the Oxford Clay plain, and they therefore did not reach the Upper Thames, where river gravels are characteristically developed. The gravels are composed essentially of local materials with an admixture of erratics derived from the Glacial Plateau Drift.

Throughout the valleys the older gravels are disposed in marked terraces, locally obliterated or masked, described below (pp. 228 et seqq.). Successive lowerings of the valley floors led to the abandonment of the gravels in terraces, and to the cutting of magnificent incised valleys and meanders through the limestones.

Before attaining their present level the rivers cut some 30 ft. deeper, and the channels are now filled with fine gravel and sand. The last event in the activity of all the rivers has been a slackening of flow and consequent deposition of fine alluvium. Some branches of the rivers have been abandoned and have become congested with alluvium, e.g. the old Cherwell Valley from Thrup to Oxey Mead.

III. Trail, Coombe Rock, Alluvial Fans, and Dry Valleys.

Although the Pleistocene glaciers scarcely entered the county after the period of the Glacial Plateau Drift, the surface of the non-glaciated country was profoundly affected by the presence of ground-ice, snow slopes, and heavy runoff of mud, stones, and water from slopes during periods of melting, probably seasonal. Signs of such movement are ubiquitous. Trail is a disturbed layer, usually of river gravel, the bedding of the original deposit being bent into festoons, sometimes with indications of flow or creep, so that the material is folded over itself. It is typically developed where there is some surface slope, but even on extensive flats signs of disturbance, usually without lateral shift, may be observed. The pebbles typically occur with their longer axes steeply inclined or vertical instead of horizontal. In the writer's opinion this festoon-arrangement is due to the former presence of ground-ice and to the seasonal thawing of the uppermost few feet. At Wolvercote near Oxford a local variety of this trail is known as warp; it demonstrates the removal of much of the fine clay of an underlying sandy silt and the concentration of the sand to a depth of 4—6 ft.: this material has been kneaded into the underlying unaltered silt, contorting its upper layers, and driving balls of sand into it. In the same neighbourhood slices of Oxford Clay have been incorporated with gravels which have flowed down hill-sides, the resulting confused deposit simulating Boulder Clay. Locally, some clay slopes are still not stabilized, and in winter movement takes place, the turf being bulged like a wrinkled carpet.

These flows of mud and gravel recall the transgressive movement of chalk and flint surface-rubble in the Chilterns, a type of flow which has now ceased, but is attributed to excessive surface creep with the aid of freezing, thawing, and strong scouring of steep slopes by water. A tumultuous deposit of Coombe Rock may be seen in many declivities in the chalk country, hollows were filled and upland surfaces stripped to bare rock. The flint nodules were usually shattered and the chalky clay reduced to a paste which cements flints and lumps of chalk: some of the flints are striated as a result of this type of flow, and again a superficial similarity to Boulder Clay may be noted. Many of the Coombe Rocks are charged with material derived from the Eocene or from superficial drifts of the Chilterns.

The intense erosion of the limestone and chalk uplands, probably at a time when underground drainage was prevented by ground-ice, led to the creation of far-reaching systems of deep valleys; they are not obvious in the clay lands. These valleys are now dry; some may contain winter bournes in their lower reaches or carry off surface water in periods of exceptional rain; all of them possess the smooth curves, meanders, and interlocking spurs of typical river valleys. Here and there the absence of these features may denote a valley formed by subsidence of solution-cavities in the limestones, but in the writer's experience this type of valley is difficult to substantiate in the county, though common in more massive limestone terrain. The dip-slope dry valleys of the Chilterns and Cotswolds are of unusual depth and grandeur.

The faces of the chalk escarpment, and of the gaps in it, are scored by steep, short, dry valleys, and their heads usually lie in steep-sided coombes; the runoff has been severe, though now nil, and broad alluvial fans splay out in the lower ground in front of them. Examples may be seen in Goring Gap, and along the escarpment of the Chilterns north-eastward, e.g. the long slopes above Ewelme and Benson. Hawkins has described some of these coombes, and contrasts the deeply scored dip-slope between Wycombe and Nettlebed with the smooth section from Nettlebed to Whitchurch.

The stratigraphical order of the superficial deposits

In the previous pages is given an outline of the varied deposits and surface features, for the greater part produced in Quaternary time. Such of the deposits as have some chronological value will now be arranged, as far as is possible at present, in their order of formation with reference also to Palaeolithic man.

I. The Thames Basin above Witney and Newbridge.

In this region of headwater tributaries we may differentiate the dip-slope streams and dry valleys from the sluggish river of the clay lowlands. With the former the Windrush may be included. All are deeply cut in the Cotswold limestones; all have incised meanders, especially in the central and lower parts of the courses, and terraces of gravel above their present valley-floors may be entirely absent. Bare benches which may have been cut by the rivers occur at intervals, as in the Windrush Valley, where also a meandering course has been partly abandoned near Asthall. All these dip-slope streams seem to have had simple histories of downward erosion in more or less meandering courses, and probably all of them, and the lower parts of some of the valleys now dry, were deepened below their present floors. Rather fine lime gravels were then deposited and the valleys were aggraded to their present levels. It is common to find shallow gravel-pits along their courses, and the broad V-section is interrupted in its lower part by a horizontal line of gravel and thin alluvium. This is especially well marked downstream from Burford: much of the water is conveyed underground, and the stream has an acutely meandering course in the alluvial floor.

The failure of run-off along the dry valleys of the limestone slopes and the continued bed-erosion of the main river courses have converted some of the dry tributaries into hanging valleys. In the past, high winter or spring floods and extreme low levels in late summer probably had much to do with the great width of the main river and the production of extensive gravel-flats, abandoned as the river lowered its course, or shifted sideways. When Lechlade is reached these flats are fully developed in a succession of terraces, and there all the main stages that will be described from the Oxford district may be identified. Some years ago the writer mapped the terraces in outline from Lechlade downstream, but the area is now being officially surveyed by Messrs. Richardson and Dines, and a full account will be found in their forthcoming Memoir. To the best of the writer's knowledge no Palaeolithic implements have been found in situ in these gravels; there is a single Lower Palaeolithic specimen from Broadwell in the Ashmolean Museum.

Reference should be made to 'The Country around Cirencester', (fn. 7) in which the gravels of the Oxford Clay plain are divided into higher and lower groups with an Older Alluvium locally covering the latter.

II. The Valley of the Evenlode.

The head of the valley lies outside the county boundary above Moreton-inMarsh, but it is necessary to stress the fact that the valley is cut through the Cotswold scarp and is wide open to the Severn Basin, the original headwaters having evidently been captured, the Evenlode consequently truncated, and the volume of its waters severely reduced. (fn. 8) Miss Tomlinson's identification of Boulder Clays in this area is important, since some of the gravels of the Evenlode farther downstream are considered to be outwash from the terminal moraine. (fn. 9) To a point near Charlbury the river flows in a wide and open valley of Liassic rocks crowned by Jurassic limestones. Downstream from Charlbury the Lias disappears under the thick oolite limestones through which the river has cut a gorge: the valley widens again on the Oxford Clay plain. Although gravels occur in and near the valley between Kingham, Ascot-under-Wychwood, Chadlington, and Spelsbury, their mutual relations are perhaps best known downstream of Charlbury. The highest terrace is at about 100 ft. above the river, typically developed at Long Handborough, and its gravel is composed of local limestone, Liassic debris, and much material derived from the Glacial Plateau Drift, than which clearly it is younger. Remains of Pleistocene mammals are found here, archaic forms of Elephas antiquus and E. trogontherii, Rhinoceros leptorhinus or megarhinus, with red deer, horse, and bos sp. (fn. 10) (Fig. 1).

No flint implements have been found in situ, but a rolled Chellean implement from the district is in the Ashmolean Museum.

The river had evidently settled in a meandering course in these gravels when it began to deepen its bed: rejuvenation trapped it in its meanders, which were incised ultimately through about 100 ft. of limestones and clays. Some part of the meander belt was abandoned, and the old course may be identified between the river, Wilcote, and North Leigh. The 100-ft. gravels may be traced also over the slopes to Church Handborough: but when they were abandoned the river took a course on their east side, west of Begbroke Common, running due south to join the Thames at Cassington. A great deal of erosion took place throughout the valley, and the next great gravel spreads occur about 15 ft. above the river, their base locally descending in deeper channels to present river level (Fig. 1). Near Spelsbury and Chadlington there are gravels at about 50 ft. above the river.

III. The Valleys of the Cherwell and the Ray.

The Cherwell rises a long way north of the county boundary and flows due south to Banbury: its present course cuts through Boulder Clays as far south as this point, but their relation to any terrace gravels south of Banbury remains to be determined. North of the town the country is low and of subdued relief, but a few miles to the south the river cuts through the Lias and the Oolite escarpment. Thence to the Oxford Clay plain its course across the strike of the beds closely parallels that of the Evenlode: a deep, wide, meandering gorge has been cut which is best seen in the neighbourhood of Tackley, where an abandoned part of the meandering river course may be observed. The meanders are flanked by gravels, more particularly on the east (Buckinghamshire) side, which are especially prominent at, and north of, Kirtlington. No implements or bones have yet been found in the gravels; in their lower parts slightly abraded large slabs of local limestone are a feature, an indication of severe erosion before the main gravels were laid down. A few miles below the passage of the river on to the Oxford Clay it is joined by the Ray. The old course of the Cherwell was more westerly, to join the Thames near King's Weir Lock, while the Ray probably flowed on the east of the site of Oxford to the Thames near Iffley. The region between these two older courses is thickly encumbered with gravels (p. 232). The Ray drains a peculiar region consisting almost exclusively of clays. It passes through a faulted limestone belt at Islip by a narrow channel, the clay land on the upstream side now being appreciably lower than the limestone. At one time the Ray escaped by a channel a little to the north, now blocked by gravels. Within its basin perhaps the most interesting feature is Ot Moor, a low-lying area, often water-logged, of alluvium on clays. From Ot Moor a valley passes directly to the Thame, but no water from the Ray flows down it, nor appears ever to have done so. Similarly, on the western side through-valleys lead from Ot Moor to the present Cherwell Valley, but no water drains through them. These and other valleys, notably that from Swinford Bridge to Botley, were probably formed by headwater erosion of small brooks; they had no effect on the main system of drainage.

IV. The Oxford District: Summary of the Upper Thames Succession.

The Thames above Newbridge, the Windrush, Evenlode, Cherwell, and Ray are directed towards the clay plain around Oxford like the spokes of a wheel, each of them discharging in the past large quantities of gravel, and latterly of fine silt. It is to be expected, then, that much can be deduced here of the history of the river system, fauna, and man himself. The following is a summary (see Fig. 1).

The 100-ft. terrace', resting on Plateau Drift, is seen in the Upper Thames Valley especially near Bampton, in the Evenlode near Handborough, and in the Cherwell Valley near Kirtlington: Handborough is the type area and has given its name to the terrace, which is associated with a 'warm' early Pleistocene fauna and probably with Chellean and perhaps older Acheulean implements. At Wolvercote near Oxford, near Spelsbury in the Evenlode Valley, at Campsfield Farm south of Woodstock (old Cherwell Valley), and elsewhere a terrace gravel occurs between 40 and 50 ft. above the adjacent rivers: it contains rolled Chellean implements (at Pear Tree Hill and at Wolvercote, within a mile of one another), but no fauna is known with certainty to occur in it, except perhaps horse at Spelsbury. It has been associated by Miss Tomlinson with the moraines in the upper Evenlode Valley, of which it is possiby the out-wash. (fn. 11) Throughout most of the river courses it was almost entirely destroyed, and a further period of downcutting ceased at about 15 ft. above present river levels, with deeper channels. Enormous volumes of gravel were then deposited in the Thames above and below Newbridge, from Eynsham through Cassington to Yarnton, Summertown, and Marston, through Oxford to Iffley, at Radley on the Berkshire side, and near Dorchester at the Thames-Thame confluence. The gravel is also extensively developed on the Berkshire bank through Abingdon, Drayton, and Didcot. Wherever the stratigraphy of this great terrace has been studied it is found to consist of two gravels separated by an eroded surface: locally one or the other may have been destroyed. Throughout the area the geologically older gravel is exceedingly rich in remains of mammoth (Elephas primigenius sibericus) and woolly rhinoceros (Rhinoceros tichorhinus), with bison and other mammals, (fn. 12) while an older form of elephant (the 'Ilford form') also occurs. The upper gravels contain a temperate molluscan fauna with the form Corbicula fluminalis which no longer lives in the country and is associated with a warm climate, Elephas antiquus of evolved type, abundant remains of hippopotamus, horse, red deer, Cervus megaceros, and others. Lower and Middle Acheulean implements have been found in some quantity, notably in north Oxford (Summertown, &c.), and one late Acheulean implement (at Eynsham). Thus a gravel with a markedly 'warm' fauna lies unconformably upon deposits with a Siberian fauna. From its richest sites this has been termed the Summertown-Radley terrace (Fig. 2).

The sequence is complicated by the presence of Upper Acheulean and later implements, and virtually the same 'warm' fauna, in a channel cut in the next higher terrace at Wolvercote (Wolvercote Channel). (fn. 13) The sequence of fauna and industry therefore goes to higher levels rather than to lower. Moreover, the deposits of the channel in which these occur indicate a strong current of water; they were not formed in a pond, and the ends of the channel were open. The superficial deposits at the top of the Wolvercote Channel are continued on the surface of the 'warm' gravels of the Summertown level. So it has been suggested that in the Oxford district (but not necessarily throughout the whole Thames Basin) the 'warm' gravels were accumulated above their present level till the rivers were flowing at the height of the Wolvercote Channel. The channel itself contains in its lower part gravels with the 'warm' fauna (without hippopotamus), Upper Acheulean and Micoque implements. These are covered by sands with a temperate molluscan fauna and by a thin seam of peat (in which a few Alpine and sub-Alpine species occur with a majority that are still living in the district) (Fig. 3).

The channel was then filled with fine silt: the current had thus slackened or almost ceased; a reindeer antler (unique in the district) was found in the silt some years ago, also teeth of a small horse and a few flakes, which are not typical of an industry but on general grounds have been considered to be Middle Palaeolithic. At the top is the 'warp' (p. 227), indicating nonfluviatile conditions and probably the presence of ground-ice, which, with 'trail', is continued over the Summertown gravels. The trail of the Summertown gravels is covered by a foot or two of bedded gravel, and the whole was truncated by river erosion. The flood plain and the submerged deposits of the buried channel follow: these bear no sign of trail in their visible upper part; from their lower parts teeth of Siberian mammoths, some in perfect condition and not water-worn, have been obtained (Fig. 4).

In the waterlogged gravels, therefore, are hidden the remains of the whole of the Upper Palaeolithic period.

V. The Thame Valley and the Gault Lowlands.

Only the lower part of the course of the Thame lies within the county, and throughout the greater part of its lower valley the river flows over Kimeridge Clay, among Portland Beds, and passes on to the Gault near its mouth. In spite of its present relations to the geological formations, however, it is a river of the Gault lowlands which has slipped towards the Chilterns. In part of its course its tendency to move down the dip of the beds has been arrested by resistant Portland Beds, which remain as ridges, some still capped with Gault. The northern tributaries rise in the Oxford Clay: the river's function as a drainage for the Chiltern scarp seems to be of secondary importance, and the composition of its gravels shows that it has destroyed completely large areas of Plateau Drift which probably lay on Gault, Kimeridge, Ampthill, and Oxford Clays. Its junction with the Thames has also migrated in a southerly direction, since oolitic limestones of the Thames occur in the gravels around Dorchester. (fn. 14) In this great clay lowland the positions of towns and villages are sure indications of lowlying gravel patches, as in the Thames Valley above Oxford. There is also a wide spread of gravel between Ascot and Chalgrove, consisting essentially of flint rubble from the Chilterns. T. I. Pocock (fn. 15) considered that this extensive deposit might mark an old channel of the Thame, rather than of the adjacent brooks.

The terraces of the Thame may be grouped in a fourfold series analogous to that of the Upper Thames and its tributaries, but the essential fauna and implements are still lacking. (fn. 16)

Below the confluence of the Thame and Thames the combined waters flow over a wide Gault plain to Benson, where they pass on to the Upper Greensand and, near Wallingford Bridge, on to the Chalk, through which they pass by Goring Gap. South of Benson the wide expanses of gravel, probably to be correlated with the Summertown-Radley terrace, contract to a narrow ribbon and locally disappear. The valley and its deposits downstream from Dorchester is described by A. J. Jukes-Browne and H. J. Osborne White, (fn. 17) special attention being given to the fossil molluscan fauna.

VI. The Chiltern Escarpment.

The superficial deposits of the escarpment and dip-slope to the Thames near Henley have been reviewed (pp. 225–8); here only one deposit calls for attention, namely, the prominent spread of flint-gravel which stretches from Cuxham to Ipsden, Britwell Salome and Ewelme being situated on its margins. Pits have revealed as much as 20 ft. of angular and rolled flint gravel, with seams of sand, and no constituents beyond those available from the Chalk and the Eocene deposits of the crest of the Chilterns seem to occur. Locally the upper parts of the gravels are much disturbed (cf. trail). The deposits thus seem to be the products of streams flowing down the escarpment under exceptional conditions rather than ancient gravels of the Thames. Osborne White associates them with the coombes and valleys of the uplands, now dry. They are of special interest on account of the Lower Palaeolithic implements which have been found in them, notably near Ewelme: several implements from Rumbold's and Painter's pits near Ewelme are in the Ashmolean Museum. In view of the nature of the deposits these implements have little stratigraphical importance. Remains of mammoth and red deer have also been found. (fn. 18) In the same district similar implements have been found in the gravels near Benson and Wallingford (Turner's Court, Gould's Heath and Grove, Lonesome Farm).

VII. The Thames between Benson and Henley.

Any co-ordinated description of the superficial deposits of this part of the Thames Valley is fraught with difficulty: on the west is the succession of the Upper Thames; on the east, between Henley and Maidenhead, the succession of the Lower Thames—Boyn Hill, Taplow, and other gravels—is recognized. Much work has been done in the Goring-Reading-Maidenhead part of the valley, but as yet the Upper, Middle, and Lower Thames successions are not welded into a single chronological sequence. The difficulties of establishing such a sequence are increased if we admit Professor Hawkins's deductions (fn. 19) of late and post-Tertiary folding in the Reading sector of the London Basin, causing the river to traverse southward, and of the part played by glacial outwash streams flowing from land ice lying along the north-western parts of Goring Gap. In so far as it is possible to discuss effectively in the present publication unfinished work, the following may be offered as a general statement.

1. The Glacial Plateau Drift of the Oxford district is recognized along the valley within the Chilterns: it is well described by Hawkins (loc. cit.) from Coldharbour at 541 ft. O.D., about 2 miles north of Whitchurch. Thence it spreads out and occupies a wide belt between Reading and Henley on the north side of the river, where it is grouped in gravel terraces, the heights of which above the Thames recall the series identified by the Rev. Charles Overy in the country around Oxford. (fn. 20)

2. Many of the patches of drift marked as 'Plateau Drift' on the 1-in. geological maps, together with the Pebble Gravel, are to be distinguished from the Glacial Plateau Drift. Some contain Lower Palaeolithic implements, some might with good reason be included among the higher river terraces. Thus the gravels at Caversham, (fn. 21) in which primitive Chellean implements occur, (fn. 22) suggest a Thames terrace at about 140 ft. above the present level of the river.

3. A 100-ft. terrace, so prominent in the Thames upstream of Dorchester, is lacking in the low ground near the foot of the Chilterns, and in much of the valley through the chalk: near the eastern boundary of the county the Boyn Hill terrace occurs, and, within broad limits, it has some features in common with the Handborough terrace of Oxford: it is not suggested that these two were contemporary. The 140-ft. terrace at Caversham may be older than either.

4. Terrace gravels 30–40 ft. and 20–30 ft. above the river are noted in the country between Dorchester and Cleeve, and the valley is burdened with gravel which passes beneath the alluvium and tongues into the mouths of the combes and dry valleys of the chalk. The 20–30-ft. gravel seems, in the field, to be the continuation of the Summertown-Radley terrace of Dorchester and Drayton, and the higher gravels to 40 ft. recall the Wolvercote terrace, but with little more than altitude to support the analogy.

5. Between Cleeve and Whitchurch river gravels are virtually limited to the flood plain, and in the vicinity of Coombe End Farm, about a mile east of Basildon, are the interesting false-bedded glauconitic sands described by Hawkins (loc. cit.), which have led him to suppose that the Goring gorge was blocked by a tongue of more or less motionless ice at the west end, the waters of which did much to erode the river bed between Goring and Reading: some of the coombes were left as hanging valleys, and the river still shows signs of a period of rapid bed-erosion.

6. From Whitchurch to Shiplake brown loam masks the gravels with but little interruption on the north (Oxfordshire) side of the river. The interesting gravels of Tilehurst, Reading, Sonning, and Wargrave lie outside the county boundary. Palaeolithic implements and mammalian remains (fn. 23) have long been known from them, especially from the work of Ll. Treacher, who has made an extensive collection from the whole area, including the pits at Caversham mentioned above. While these gravels cannot be effectively discussed in this volume, perhaps it is permissible to add that they probably hold the key to the correlation of later Acheulean and Middle Palaeolithic deposits west of Cleeve and east of Henley.

7. The gravels mentioned above occur again on the left side of the river from a point opposite Wargrave to Henley, at about 20 ft. above river level.

8. The mollusca and stratification of the alluvium are of special interest in the western part of this region. (fn. 24)

The superficial deposits and Palaeolithic Man: summary

Excepting certain accumulations of indefinite age located especially on the Chilterns, the oldest Quaternary deposit of the county is the Glacial Plateau Drift, characteristically displayed on the flanks of the Cotswolds and around Oxford: it contains no mammalian remains or flint implements.

The oldest river gravel that contains primitive Lower Palaeolithic implements is probably that at about 140 ft. above the Thames at Caversham. The greater part of the Glacial Plateau Drift had already been redeposited by the rivers in terraces at greater heights than this above their present courses.

The oldest river gravels containing mammalian remains so far as yet known are those of the Handborough terrace near Oxford, at about 100 ft. above the adjacent rivers: the fauna indicates a 'warm' climate, and is of rather early Pleistocene type. (fn. 25) Unfortunately no implements are definitely associated with this stage; it may be Chellean or early Acheulean.

A '50-ft. terrace' succeeded that just described: in the Oxford district and Evenlode Valley it is believed to be associated with a boulder clay; (fn. 26) rolled Chellean implements have been found in it at Pear Tree Hill, Wolvercote. It is almost unfossiliferous, though it is widely recognized in the basins of the Upper Thames and Thame. (fn. 27)

In all the river valleys upstream from Goring Gap a prominent terracegravel occurs at and about 15–20 ft. above the rivers: in the Oxford district it has been called the Summertown-Radley terrace. Wherever its stratigraphy has been studied in detail in this region two distinct gravels have been identified. The lower is rich in mammalian remains, indicating a 'cold' or steppe climate: the upper contains a 'warm' fauna. (fn. 28) Rolled Chellean and Lower Acheulean, Middle Acheulean, and rare Upper Acheulean implements occur in these gravels, especially in numerous places between Wolvercote and the Thames at Folly Bridge. Collections and records will be found in the Ashmolean, University, and Pitt-Rivers Museums in Oxford.

Similar implements are found in abundance in the river gravels of the Reading district (fn. 29) and farther downstream, as at Sonning and Shiplake. They occur also in the patches of superficial flint-gravel between Ewelme and Benson. Between Wallingford, (fn. 30) Reading, (fn. 31) and Henley much of the above mammalian and molluscan fauna has been found in low-lying gravels, but their relation to the Upper Thames sequence is not yet certain.

In the vicinity of Oxford the next deposits, in order of age, are the gravels and sands of the channel (fn. 32) which cuts across the '50-ft. terrace' at Wolvercote.

The basement gravels contain almost the same mammalian remains as the upper gravels of the Summertown-Radley terrace, with many Acheulean implements, including late Acheulean specimens of great size and beauty. The largest collections from this site are in the Ashmolean (Pl. I), University, and PittRivers Museums, many of them from the collections of Manning and A. M. Bell. The basement gravels are covered by fine sands and gravels with temperate molluscan fauna which were much eroded before silt and clay were deposited. At the eroded junction vegetable remains were found by Bell: they included Alpine and sub-Alpine species and elytra of beetles, nearly all referable to existing species. (fn. 33) The thick overlying silt and clay are virtually barren, but a part of an antler of reindeer from them is in the Manning collection in the Museum of Manchester University. The few implements referable to sands and silts (other than derived specimens) suggest Middle Palaeolithic technique. The top of the clay and silts are strongly disturbed and masses of sand have been driven into them from 'trail' or 'warp' above: the conditions suggest movement by freezing and thawing of the superficial layers. Similar disturbance is seen in the upper part of the Summertown-Radley gravels, locally covered by undisturbed river gravel, and it is supposed that severe climatic conditions prevailed while the rivers were falling below their present levels by some 20–30 ft. Unrolled teeth of mammoth and rolled Lower Palaeolithic implements have been dredged from the gravels of this buried channel, notably at North and South Hinksey, Iffley, and Folly Bridge, Oxford. Although the sequence of events from the bottom of the Wolvercote Channel to the gravels of the flood plain and buried channel cannot yet be harmonized with the succession of the Thames gravels from Wallingford to Henley, the flood plain gravels and the filling of the deep channel with silt, mud, and peaty beds, capped with alluvium, can be recognized throughout. The fauna of these beds is Holocene, and it appears probable that the deposits represent the greater part of post-Palaeolithic time. The position of the Upper Palaeolithic industries in the sequence probably lies in the flood plain gravels and their submerged counterparts.

The distribution of Lower Palaeolithic implements is interesting: on the Chilterns and in the London Basin they are found locally in astonishing numbers: in the clay plains of the Thame and Upper Thames they are almost unknown, except in the vicinity of Oxford (including Marston, Iffley, and Bagley Wood); the writer also knows of a few found near Abingdon. On the Cotswold slopes again they are extremely scarce, becoming fairly common outside the county boundary towards the east. A few implements have been found on the surface, notably at Broadwell (Manning coll., Ashmolean Museum) and in the Evenlode Valley near Kingham, Charlbury, Handborough, and Cassington (the last dredged from the river). (fn. 34)

Middle Palaeolithic implements are found on the Chilterns, a few can be attributed to the Wolvercote Channel near Oxford, and, so far as the writer knows, none has been found elsewhere in the clay plains or on the Cotswolds, while Upper Palaeolithic implements are not recorded.

It is probable that continued search will produce more implements in the lower areas, but the excessive scarcity compared with the abundance of the London Basin is nevertheless remarkable.

The distribution of the mammals is also uneven. Of the 'warm' fauna the abundance of hippopotamus and Elephas antiquus is outstanding, while the Siberian mammoth is also very common. Musk ox, on the other hand, has not been found, and of reindeer only one occurrence is known. Cervus megaceros makes a single and remarkably early appearance. A unique ramus of a new species of bear, Ursus anglicus, has also to be recorded. (fn. 35)

It is tempting to suppose that the clay plain offered some difficulties to human and mammalian progress: flint had to be imported for the making of implements, but, since the hard quartzites from the Glacial Plateau Drift were abundant and were used for this purpose, (fn. 36) lack of material was evidently no real deterrent to Palaeolithic Man's entry into the clay and limestone country.

Note. The names of parishes in which implements have been found are underlined on the distribution-map of the Prehistoric period. In a few cases it has been possible to underline other localities that are included among the names on the map. It must be emphasized that the underlining of a name does not necessarily indicate quantity. While implements are abundant on the south slopes of the Chilterns, and have been found in some numbers round Ewelme and Benson, and at Wolvercote, from other localities only isolated specimens have been recorded, e.g. Broadwell and Handborough (one apiece).