Home » Recreation » Books

The Stone Age Cultures of South Africa

A.J.H. Goodwin & C. van Riet Lowe

Preface

So much progress has been made in the study of South African prehistory in the past three or four years, and so large a measure of agreement has been reached among the various workers as to the grouping and sequence and significance of the stone implements and other cultural remains, that the time appears ripe for a statement of the conclusions which have been reached and a fairly comprehensive description of the material and occurrences on which they are based.

Many pioneers in this field have contributed towards laying the foundations of our present-day knowledge; most of them are referred to largely in these pages. But as far as published work is concerned, and particularly in point of illustrating by means of abundant figures the types of objects under discussion, there can be little question but that the late Dr. Péringuey’s “Stone Ages of South Africa,” appearing as Vol. VIII of these Annals, was the most important of the earlier publications. It is thus appropriate that the present volume should also be issued by the South African Museum in the same series of its Annals. The interval since Dr. Péringuey’s work was prepared is eighteen years. In European archaeology such a period nowadays brings no very large change in general conceptions; in South African archaeology the interval has brought something like a revolution. Current and future investigations are likely in time to produce more or less considerable modifications of the system here adopted, but its main lines seem now to be sufficiently firmly established to warrant the production of such a work as this, which it is hoped will serve as a new basis and starting-point for archaeological work in South Africa.

Our indebtedness to the two authors is very great. Mr. C. van Riet Lowe’s brilliant work in the Orange Free State has been a leading factor in clearing up the confusion of ideas that prevailed until very recently, and his careful and finely illustrated account of his material and deductions is an extremely valuable contribution to this volume. To Mr. Goodwin we are even more indebted, for, in addition to his very substantial share in preparing the volume, the Museum has profited to an extent that is difficult to acknowledge at all adequately by his systematic study and arrangement of its immense collection of stone implements.

It remains to acknowledge the financial help that has been received from the Committee on African Life and Languages of the University of Cape Town. Our commitments for publication at the present time are particularly heavy, owing to the amount of systematic work that is being done on the collections; and the monetary support that has been afforded to us by Professor T.T. Barnard and his Committee has made it possible to be more generous both in text and illustration than would otherwise have been possible.

Leonard Gill.
South African Museum,
August 1928.

An Introductory Survey of the Geographical and Archaeological Conditions in South Africa.

The science of archaeology depends basically upon our knowledge of the geographical, climatic, and geological conditions present at various phases of the Quaternary period. The chronology supplied by geological evidence must necessarily be accepted by the archaeologist. The distributions of the various types of culture can only be accounted for by a wide knowledge of the geographical and climatic conditions of the various periods.

Archaeology is a human science, and therefore cannot be treated as we would treat a purely physical science. We are not dealing with a science of a type similar to chemistry or geology which are necessarily bound by ever-present factors; we must thus take into account the climatic and geographical causes necessarily affecting or confining individual groups of the human type. We are dealing primarily with a migrant and evolving animal, an animal which, within certain limits, is capable of inventing and perpetuating his various reactions to environment, or modifying those reactions to meet certain needs. The only tangible results of those cultural reactions we have to hand are those implements and objects which early man has left us, and which have survived to our day. Our knowledge of mankind is thus an ever increasing ratio: of the earliest periods we know relatively nothing; of the following periods we know more, but, even so, far less than we know of modern man.

Our conception of early geography within the human period in South Africa is hampered by a variety of factors. Foremost is our lack of knowledge. Second only to this is our lack of glaciations of any extent or of other chronological landmarks during the Quaternary period. Our only great glaciation (apparently from the north, peculiarly enough) was long previous to human evolution.1 It is possible that knowledge we may in time acquire of minor glaciations spreading from the Basutoland mountain-masses may help us to date our prehistoric period with some degree of accuracy. A further difficulty lies in the fact that whereas Europe is split up geographically into a variety of regions, alpine areas, broken littorals, temperate regions, and so on, from which prehistoric conditions can be judged, as yet hardly sufiicient material can be accumulated with our present knowledge of Africa to justify such regioning. We know little of our prehistoric river system, and as little of the climatic conditions presented at any one period; while the extreme shortness and the isolation of our coastline from other countries does not allow of our using the presence or subsidence of land bridges as an aid in defining the date of any migration.

What were the barriers hindering man’s movements in Africa during the prehistoric period? What were the highways which invited his migrations? What were the causes enforcing great game movements, making it necessary for man, the hunter, to follow? We cannot yet tell with any certainty, but the subject is an important one. It would be safest to deal with the geographical conditions now present in Africa, and to work back to the past conditions so far as we do know them.

Africa is bounded to the north by the Mediterranean Sea, which is surrounded by a climatic area that is at present a single entity: the northernmost strip of Africa is not truly African from the human point of view. The Mediterranean forms a barrier which to-day is far less strong than the Sahara. This was probably more true during the Capsio-Aurignacian period of Europe, though even at that period the land bridges giving easier access to Europe were compensated for by a more fertile Sahara region. Since pre-Roman times the Sahara has become more and more of a barrier, until now it acts as a filter. Higher cultures can pass from north to south and survive, but lower cultures passing from south to north are immediately subdued and assimilated by the higher, better organised folks of the Mediterranean. As a result, although “something new” is always expected out of Africa, these things are only “new” from their very age. Africa is a pocket from which nothing tangible returns.

South of the Sahara Africa divides into two regions. The Congo basin, with the forested regions of West Africa on the western half, forms an area attractive to a hunting people, and is fairly abundant in edible game; in the eastern part of the continent lies the mixed mountain and lake region of Africa, a definite highway to the south, with a great variety of climatic variation within a limited area. Still further south we reach that section of Africa which has most greatly affected the present distribution of South African tribes, and which most certainly affected the Later Stone Age peoples of this area.

At the present time we know that South Africa can be roughly divided into three main sections, running from north to south. The western desert, practically useless as a highway, not attractive to a hunting people, but definitely a barrier. This desert originally cut across into the heart of Rhodesia, but has decreased, leaving a passably good though dry climate; but this extension appears to have belonged to a period long previous to man. East of this the central highroad lies between this desert and the great mountain masses of the eastward and southern fringe of the African plateau. These mountains form an “eddy” between the central highway and the fertile eastern highway lying eastward of the main mountain masses, and running from north to south across the myriad short rivers of the east coast (see diagram).

Cutting across Africa in the opposite direction are the rivers. First comes the Zambesi, passing across the central and eastern highroads, while further south the Limpopo does likewise. Neither of these two barriers is impassable, though both in all probability constitute lines of lateral movement across the great highroads. Further south we have westward-flowing rivers. The Kunene cuts off the fertile territory of Angola from the less fertile Kaokoveldt, which in turn gives way to the semi-arid land of South-West Africa. South of the Kalahari, but cutting also across the central highroad, is the Orange River, together with its great tributaries. These rivers did not form barriers, but seem to have enclosed a great area of attraction within their basin. The desirability of this area was increased a hundred-fold by the presence of large amounts of indurated shale or lydianite, a material of a tractable type with an even fracture and of a hardness eminently desirable to the worker in stone. We will see how the Fauresmith implement makers were either attracted on their migrations by the presence of these delights, or else evolved a variation of their original culture during their sojourn in this region. In the Later Stone Age, too, we will see how Smithfield man apparently began his industry and evolved it here.

We can presume quite fairly that much the same conditions extended right back to the beginning of the Later Stone Age. During the Middle Stone Age a somewhat similar state of affairs seems to have existed. There still seems to have been a western desert region, though whether it extended over its present area we do not know.

The Kalahari seems to be moving southward, leaving Southern Rhodesia, and increasing in the south. Probably in time it will link with the semi-arid Karroo region. This desiccation seems to have been general in Africa, and dates back in its beginnings, so far as can be judged, to some period after the beginning of the Middle Stone Age.2 Certain investigators regard this desiccation as slower than has hitherto been believed, and would date it as previous to the human period; others regard it as part of a rhythm. The fact remains that during the Earlier Stone Age the vegetation and animal life supported by climatic conditions in various parts of South Africa was capable of supporting peoples where water does not at present exist. Similarly, in semi-arid parts of South Africa, relatively thick population by hunting peoples was possible, together with the animal life concomitant.3

Much of this aridity appears to be due to the breaking down of barriers previously supporting large lakes, and also to the gouging out of deep passages by certain rivers,4 resulting in a failure to spread and too great fixity of course. Much aridity has similarly been caused by lack of vegetation of a type capable of preventing erosion and of building up surface soil. This has been largely due, in all probability, not so much to a decrease in rainfall, as to the alternating dry periods and seasons of torrential rains. This has resulted in a loss of ability to assimilate the rain which does fall, and in the resultant removal of lighter surface soil.

The Industries. — The work, of which these papers is the first report of a detailed type, was begun in 1924 and is still continuing. The survey was started at the South African Museum, and at first consisted in the sorting of all theimplements into types, and the subsequent association of each implement with others into industrial groups. This process is still going on, but sufficient is now known to allow of our giving an introductory series of papers on South African prehistory, owing not only to my museum research and my inadequate field-work, but mostly to Mr. C. van Riet Lowe’s excellent, detailed, and untiring work at various sites, more especially in the Free State. His work has proved to us very conclusively that the lydianite area of the Orange River basin is an archaeological unit in itself and must be treated as such. This beautiful dull-black material was practically a new medium, and as a result the Free State was the mother of variations in industries which can only be judged when a thorough archaeological survey of the Free State is done. He is doing it, and while this present series of papers is to be regarded as general, very much has been left unsaid about the Free State which he alone is in a position to say. It is his exact and superior knowledge of the Free State too which has led me to invite his co-operation in the Fauresmith paper, and later in the Smithfield paper. These two papers are to be regarded in the light of additional to his forthcoming work on the Orange Free State archaeology.

As a result of our joint work, and with the help given me by the Museums of the Union and by various field-workers, the following general scheme has been arrived at. How final it is remains to be seen, and especially is this true of the Free State, as we shall see.

Earlier Stone Age:	Fauresmith, Victoria West, Stellenbosch.
Middle Stone Age:	Still Bay, Glen Grey, and a variety of industries showing Mousterian origins.
Later Stone Age:	Wilton, Smithfield, “A,” “B,” and “C.”

Of these we may say that the Stellenbosch is a relatively pure offshoot from the same source as the Lower Palaeolithic of North Africa; the Victoria West is an apparent local evolution.

The Fauresmith is of great interest as it appears to be the first of the South African industries to evolve within the lydianite area of the Free State. It is basically of a Lower Palaeolithic origin with possible Mousterian additions. The coup-de-poing is augmented by a strong flake industry, trimmed flakes and true cores appearing. Within the Free State, or at least in the region of the Modder, Riet, and Orange rivers, this industry seems to shade directly into the Smithfield “A” period of the Later Stone Age. Elsewhere normal elements of Middle Stone Age type intervene between the Earlier Stone Age and Smithfield “B.”

The Middle Stone Age shows a number of industries, not yet sorted fully, showing a common origin with the Mousterian of North Africa. They have little in common with one another except their obvious origin and the fact that they show varieties of points, with types paralleling the Mousterian, proto-Solutrean, and even the Lower Solutrean of Europe. The Middle Stone Age seems actually to have been either an evolution, or more likely a series of offshoots from an evolution much the same as that appearing in the east of Europe during those three periods.

The Later Stone Age is Neo-anthropic in character, and is probably basically an offshoot of a late Mesolithic or early Neolithic stock of North African origin. The Smithfield “A,” as we have seen, is very much the same as the Fauresmith without the coup-de-poing, and perhaps marks Neo-anthropic influences on the already mixed Fauresmith makers. Smithfield “B” is even more definitely Neo-anthropic. Pottery and polished stones appear, though the unpolished implements are all of Capsio-Aurignacian type. It is an advance upon the Smithield “A,” and appears to follow it in time. Smithfield “C” is similarly Neo-anthropic, and is generally to be regarded as a cave culture.

The Wilton seems to have been of late Capsian origin with Neolithic additions. It is a cave culture primarily, though open sites do appear. An almost exact parallel to this industry5 is visible in North Africa at the Djebel Redeyef rock-shelter.

The Smithield “A” and “B” are now definitely associable with the rock engravings of the dolerite areas, while Smithfield “C” and Wilton are as definitely associable with the cave paintings of the Union. The Later Stone Age folk were thus the artist race of South Africa, and physically can be regarded as belonging to the “San” or so-called Bushman race.

The Earlier Stone Age in South Africa.

The Stellenbosch Industry.

In 1912 the late Dr. Péringuey, then Director of the South African Museum at Cape Town, wrote his monograph “The Stone Ages of South Africa.” In this work he drew world-wide attention to the archaeological remains of the Union, thus making the first great attempt to lay a firm foundation for this science in South Africa, though similar work had been done previously on a smaller scale by J.P. Johnson. South Africa had not been behind in her archaeology. In 1866 Sir Langham Dale found and recognised the first stone implement in South Africa. By 1872, papers of extreme archaeological interest were appearing in the Cape Monthly Magazine. Péringuey, however, attempted a new nomenclature, realising the difficulties which had been encountered by previous writers. Johnson, for instance, had been forced to speak of “Acheulic” and “Solutric” in attempting to draw parallels without implying exact identity between South Africa and Europe.

It was Dr. Péringuey who first made use of the term “Stellenbosch type,” but his use of the term was apparently much narrower than that which is now in use. He says:6

I find no difficulty in dividing the South African bouchers into several types, owing to their appearance or facies, or to the material of which they are made.

On the following page he says of his Stellenbosch implements:

They are of a type so numerously illustrated in all the South African districts of the Cape Colony and also beyond (Cape, Stellenbosch, Paarl, Worcester, Tulbagh, Ceres, Clanwilliam, Malmesbury, Piquetberg, Caledon, Mossel Bay, Knysna, Port Elizabeth) that they may well be ranked under the name “Stellenbosch type.”

He speaks also of the “Orange River type”:

These palaeoliths are mostly made of banded jasper, brown or yellow, and occasionally white with bluish veins .... The implements thus produced rival the best Acheulean flints in finish. They do not usually attain the great size and heavy weight of the Stellenbosch type examples.

He also gives a list of districts in which this type is found: Alice, Bedford, East London, Carnarvon, Kenhardt, Prieska, Warrenton, Pniel, Vryburg, Modder River, Smithfield, Transkei, Pretoria, Witwatersrand, Potchefstroom, Vereeniging, Swaziland. Further on he states:

On the whole, and in spite of the differences mentioned, the facies or general appearance of these two, or perhaps three, types is astonishingly alike.

Since his death some four years ago a vast amount of new material has been accumulated at various museums in South Africa, most notably at the South African Museum, Cape Town, the McGregor Museum, Kimberley, and the Albany Museum at Grahamstown. The archaeological side of anthropology is also being stressed at the National Museum, Bloemfontein, the Port Elizabeth Museum, and the Transvaal Museum, Pretoria All these collections I have been privileged to visit, and permitted to make notes and drawings of the various implements represented. As a result of these various finds it has been deemed best to regard all Péringuey’s types as falling under the single heading of “Stellenbosch.” Péringuey himself has admitted (vide supra) that his differentiation was based largely upon differences of material, and that the general appearance of his types is often astonishingly alike. For these and other reasons it has been thought safer to regard the Stellenbosch type as a single industry, while recognising local variations depending more or less entirely upon the differences of material.

Two main geographical areas are immediately discernible, and these fall roughly into Péringuey’s two original groupings. The first is the Cape System, or Southern Mountains area. This lies roughly between Cape Town and Port Elizabeth, or even East London, and for about a hundred miles inland. The second is the Vaal River area, lying between the Transvaal border (near Warrenton) and the junction of the Vaal and Harts Rivers. But these can only be regarded as geographical regions, the industries only showing variations due to differences in the material used. The first of these two areas has never been completely written up, though Péringuey touched upon various sites at Stellenbosch, and in the neighbourhood of Knysna.

Typology.

Throughout the Stellenbosch industry, the most typical and obvious implement is the coup-de-poing or boucher. This implement is probably the best known of any type, occurring as it does throughout the entire Old World. There is little difference in the implements appearing from India, South Africa, North Africa, or Europe. The one basic difference is that of material. While in Europe and North Africa flint is used, in India and in Central and Southern Africa the most convenient local material is made use of. This implement definitely proves a cultural unity between the Lower Palaeolithic of Europe and our Stellenbosch industry in South Africa, though the original home of this culture has not yet been ascertained.

The coup-de-poing is best described as a stone some 6 inches long, shaped by human agency, normally from a water-worn boulder, by means of flakes removed about the perimeter. The resultant shape is much that of a shelled almond, a stone with lenticular cross-section, and bounded by an amygdaloid edge. While this shape is the most common, two or three varieties appear; a pear-shape (ficron) similar to the almond but with a more tapering point; a leaf-shape (limande) tending to a point at either end; an oval coup-de-poing; and rarely, a circular type. In the last three the greatest width of face is across the mid-length of the specimen, while in the two former it is about one-third of the way down the mid-line. In all cases there appears to be a desire for symmetry across this line, except in the leaf-shape, where a few specimens appear showing a tendency towards a lunate or half-moon shape (see text-fig. 1).

All the sites known appear to have been factory or workshop sites, and hence a high proportion of well-made implements is hardly to be expected, and in fact the percentage is often very low; but in all cases in which large numbers of specimens have been sent in from sites, a strong tendency is noticeable towards a well-shaped, evenly worked implement. The whole body of the implement is surrounded by a sharp edge, often further sharpened by step-flaking, or resolved-flaking. This edge extends round the butt, and right round to the point. The thickest part of the implement (usually from an inch to an inch-and-a-half) is in the centre of its greatest width, the point (or points) tapering from here. The point end of the implement always shows the most delicate working. The resulting tool is neat and well made, denoting the early presence of artistic pride in South Africa. In the best specimens the ratio between length : breadth : thickness is about 5 : 2¹/₂ : 1.

The second artefact worthy of note is the cleaver, or biseau. This would appear to me not to have been an implement, but rather a crucial stage in the manufacture of the coup-de-poing, a stage at which it could very readily be seen whether or not the resulting implement could have been successfully completed (see Plate V). However this may be, this artefact must in fairness be included as a type-specimen of the Stellenbosch industry. The implement is exactly similar to the coup-de-poing, except that in place of the point there is a straight unworked working-edge at right angles to the length of the implement. In a few instances, from the Vaal River only, so far, the point has been completed and the wedge-shaped edge takes the place of the butt. The similarity to the modern axe is immediately apparent, and unsafe conclusions as to hafting, uses, and so on can be jumped at with the utmost ease. The size is slightly greater, generally speaking, than that of the coup-de-poing (see text-fig. 2).

Lastly come a heterogeneous group of hammer stones, discoidal artefacts, and the like. These are formed by their use, and thus fall into the class of instruments rather than that of implements. The commonest shape, and that to which the term “discoidal artefact” has been applied, is a rough polyhedral disc, often slightly truncated so as to leave a small platform in the centre of each face. This type has obviously been formed by the use of the perimeter of a circular pebble as a hammer stone (see text-fig. 3).

Péringuey attempted to show that a knowledge of a flake technique was to be found with the Stellenbosch material. There does not seem to be sufficient evidence that this was usually so. All the examples he brings forward are of little use as proofs. For instance, the specimens he describes from Smithield may yet prove to belong to the Fauresmith Industry. The Robberg specimen is apparently not a human-made flake, and certainly not an implement. The East London specimen is similarly doubtful. The Nooitgedacht specimens appear to be flakes struck from the coup-de-poing in the making. This, however, is negative evidence. It seems certain, from Mr. Lowe’s finds at Knysna, that flakes struck off in the making of the coup-de-poing were occasionally used as rough scrapers.

We may thus state that there is not sufficient evidence that the makers of the Stellenbosch implements normally used a pure flake technique (implying the removal of fluting flakes or longitudinal trimming flakes), nor is there evidence that this industry contains conventionalised flake implements. We can, however, presume the use of spalls or flakes struck from the coup-de-poing, as scrapers, etc. The action of water, pressure, and similar natural forces, upon the edges of likely-looking flakes must be taken into consideration in this connection. On the other hand, however, Mr. Lowe found at Knysna, stratified at depth and lying cheek by jowl with Stellenbosch material, large flake cores which he asserts belong to the assemblage; these come from the Knysna-Concordia road. He regards this as evidence of a knowledge of a crude flake-technique. This discovery will be mentioned again later in dealing with Mr. Lowe’s Knysna finds.

Southern Mountain Area.

The Southern Mountain or Cape System area appears at some time either to have been very thickly populated by the bearers of the Stellenbosch culture, or else populated over a considerable period of time.

The Cape System consists of a group of sandstones and subordinate shales, and comprises a three-fold grouping, viz., the Table Mountain, the Bokkeveld, and the Witteberg Series, distinguishable from one another only by their fossil contents. The Table Mountain consist of thick unfossilised grits and quartzitic sandstones. The Bokkeveld consist of similar sandstones but with shales and Hagstones plus marine fossils of Devonian affinities, and on the whole associable with the Gondwana of India, with the Devonian of America, and with occurrences in the Sahara. The Witteberg sandstones are also quartzitic, but the fossils are carboniferous, and the shales subordinate. In the western portion of the Cape System granites of pre-Cape Age underlie the Table Mountain Series and outcrop at the Cape, Paarl, and various neighbouring sites.

The material used by the makers of the Stellenbosch implements in this area was invariably a quartzitic sandstone, either from the Table Mountain Series or from the Bokkeveld. The most easily obtainable form was the larger talus material from the mountains, the talus consisting of quartzitic sandstones of both types. It is thus not possible to differentiate between the two sources of material.

The normal processes of denudation have formed much of the sandstone into river gravels, and further denudation has left these gravels capping the foothills below the mountain ranges. It is from these water-worn gravels that the Earlier Stone Age folk obtained their material, and worked their implements on the spot. As a result most of the implement sites lie on the higher lands, well above (often 100 feet or more) the present river levels. Such is the mode of occurrence in and about Stellenbosch, Paarl, and Wellington, and in fact all the towns lying below the mountains in this part of the country.

Paarl.

Péringuey7 describes various sites of this type. Speaking of occurrences at Paarl, he says that in an ironstone gravel quarry he found several coups-de-poing of large size.

Some 400 or 500 yards away on the higher slope of the hill, and about 1¹/₂ or 2 miles from the [Berg] River, a piece of ground had been delved for establishing a vineyard. Alongside were two heaps of palaeoliths and nuclei, thrown aside by the workmen .... Not a single scraper or small spall was to be had. These palaeoliths and nuclei were found at a depth of 2¹/₂ or 3 feet — the depth of the delving.

Unluckily he gives no exact locality for this particular site, and thus it has not since been searched for further material. Dr. Péringuey presumes that the material for manufacturing the coups-de-poing was brought up from the river below for the purposes of manufacture. Judging from other exactly similar sites this would appear to be an unnecessary presumption; it would seem more likely that high gravels were also used here, and that these gravels, containing the sandstone material, have since been left as broken terraces capping the foothills. The Berg River has out down very considerably since these gravels were laid, which is not surprising when it is remembered that this point is only thirty miles in a direct line from the sea.

The Rev. Breedt of Wellington very kindly motored me to a number of other sites in the Paarl vicinity. One of these lay some three miles north of Huguenot station on the Paarl-Wellington road, which runs west of the Berg River, and at the fork of the Klipheuvel-Wellington road. Here the circumstances are similar to those obtaining at Péringuey’s site. Implements occur to a depth of two feet in a fine gravel formed (apparently) of a fine ironstone and kaolin from the Paarl granites. An exactly similar occurrence is to be seen about 50 yards from the crossroads of the Malmesbury-Paarl road and the Wellington-Klipheuvel road, towards Wellington. At both these sites, Mr. Breedt and myself found implements of normal Stellenbosch type in the slight roadside diggings excavated to obtain road metal (see Plate I).

Wellington.

Mr. Homan, until lately the owner of Zoutendal farm some two miles north-west of Wellington station, invited me to visit the estate as he had come across a number of implements in the gravels capping the hill which forms the main part of the orchard.

The highest point of the orchard lies about 80 feet above the present river level. The river terrace in which the trees now stand is composed of a large number of boulders of quartzitic sandstone which have aggregated at this point and cover an area of over 300 yards by 500 yards. Here worked pebbles formed a very small part of the gravels, perhaps 2%, and although huge piles of stones had been collected and cast out, implements were relatively scarce. However, sufficient was found to show that the normal Stellenbosch types were again represented from this site. The finest specimen is still in the keeping of Mr. Homan.

He also brought out a point of interest regarding the soil. Before the planting of fruit trees began the whole area was covered by the South African Olive (Olea verrucosa), a remarkably long-lived tree. These trees had been standing sufficiently long to make a considerable change in the structure of the soil, which is clayey and fertile where olive trees once stood. The height above the present river level argues great age too, and the implements are in some cases considerably water-worn, pointing to the fact that these gravels were in process of formation when the implements were being made. Others show no signs of wear and may be of slightly later age, or perhaps escaped rolling. The gravels here, though actually lower than the other Wellington gravels, are not really lower in relation to the present river levels; the Zoutendal gravels do, however, appear to be newer than the gravels lying immediately below the mountain range, if we may judge from the relative wear of the implements enclosed.

Above Wellington and immediately below the mountain are a number of sites similar to those to be studied at Stellenbosch (see later). Mr. Breedt kindly took Mr. Burkitt and myself to a site on Mr. Brink’s farm a few miles out of Wellington. The occurrence here is very much the same as that found elsewhere in this district, a gravel left high and dry to cap a hillock. The implements were again of normal Stellenbosch type.

Later Mr. Breedt took me to a further farm Bovlei set similarly upon a rise. The circumstances of the find were identical with those obtaining at Péringuey’s Paarl site. A field had been broken up for purposes of tree planting, and all stones likely to impede the plough had been thrown out and piled. In these piles were large numbers of implements in excellent condition. No flakes were represented, these, of course, not being thrown out. Time would not allow of a further search, but flakes should most certainly be found turned up by the plough, as the larger implements are all from the upper two feet of soil.

Simondium.

Returning further south, some seven miles due south of Paarl lies the station of Simondium towards the French Hoek Valley. Dr. Péringuey notes implements from here and gives a large scale map of the district.8

West of Simondium towers Simonsberg, forming the western flank of the Drakenstein Valley. This mountain falls abruptly, then levels out towards the railway station, leaving a succession of small hills. One of these, Pontac Hill, behind the homestead of the Pomona estate, proved to contain gravels, exactly similar to those occurring commonly below these mountains. A donga, or erosion gulley some 10 feet deep, showed that some of the material from the gravels above had dropped to a lower level. Dr. Péringuey traced this site for about a mile up the ascending talus of Pontac Hill and about 250 feet above the present level of the Berg River. Two miles below this highest point, in a cutting at Simondium station, he found further evidences, probably the lower end of the same gravels. The implements were not in a rolled condition, and hence must have been made from the gravels when these were lying exposed.

Stellenbosch.

Stellenbosch lies some 10 miles south-east of Simondium and 15 south of Paarl. The situation of the town is very much that of Wellington, nestling in the Stellenbosch mountains which form part of the main massive range running northwards to form part of the escarpment of the main plateau. The mountains here also drop steeply and form rounded hillocks as they near the river level and the plain. Here again the gravels lie at the tops of these hillocks, while the river system has cut down leaving them as a series of graded terraces, all being of probably much the same age. Implements of Earlier Stone Age type are scarce in the valleys, though they do occur, having been apparently washed down from the higher levels.

A site of this type was discovered by Mr. Burkitt and myself near Stellenbosch station. The road leads across the railway line past a wine-factory, and thence crosses a small stream. Near the road, and on the far side of the stream from Stellenbosch, is the cemetery. Nothing appears to have been dug up in excavating this ground in the process of burial, but towards, and on both sides of, the road, some 20 feet above the level of the stream, a considerable amount of material has been excavated for brick- or road-making purposes. In these quarries and near the surface appear a small number of implements. The percentage here is slight compared with the unworked material, but there is sufficient to show a workshop site, apparently in two layers.

This site forms a part of Dr. Péringuey’s Bosman’s Crossing site, which has been cut in two by the stream. A spur of the Papagaiberg abuts on a tributary of the Eerste River, and had been intersected by a road on one side and a railway cutting on the other. The space between these had been used for a considerable period as a brickfield. Dr. Péringuey found two layers of worked implements. These layers ran right through the entire spur, reappearing on both sides of the road, and also of the railway cutting. The lower layer apparently lay directly on the granite formation which it followed. The brick clay lying above these implements amounted, in places, to 20 feet. This suggests a very considerable lapse of time since the implements were deposited, even allowing for the fact that the overlying clay formed part of the Papagaiberg talus and was probably quick in forming. This part of Stellenbosch has changed considerably during the past fifteen years and “Bosman’s Crossing” does not now appear under that name, but it seems that it stood a little south of the site now taken up by Stellenbosch railway station.

A somewhat similar site has been revealed some 200 yards west of Du Toit station; this, however, is not very rich, and the material appears to have gravitated down from a higher level. The ground here has been excavated for yellow clay, and as a result the surface soil for some distance above the lip of the quarry has been denuded, carrying implements into the otherwise sterile quarry.

Higher up the slopes towards the Stellenbosch mountains are various sites. The late Hon. John X. Merriman submitted implements of this Stellenbosch type from his farm Schoongezicht on the Rustenberg road. Similar finds are reported from Rustenberg farm itself (see Plate II).

At Lorraine, a farm some 3 miles from Stellenbosch on the French Hoek road, Mr. Lowe discovered a number of implements on the crest of a hill which had been ploughed and dug up as an orchard. None of the implements were in situ, but all had been found in the clayey surface soil of the field, to a depth of some 3 feet. There is a considerable amount of gravel material, rounded boulders, etc., with the implements, but the latter were not in a rolled state, and appear to have been made from the gravels.

While this site does, not appear to extend to the top of the hill, it is quite thick on the face overlooking Stellenbosch. Very little well-finished material appears, though Mr. Hill, the owner of the farm, has collected a few specimens which show that the desired standard was similar to that expected elsewhere in the same industry. The site is extremely rich, some 20% of the gravel material showing signs — however slight — of working; perhaps a quarter of this material is sufficiently advanced to show the general morphology. From these it is seen that the only implements made (excluding the possibility of a minor flake industry) were the coup-de-poing, the biseau, and the usual fabricating artefacts normal to the industry. The shape of the coup-de-poing (as is usual in the Stellenbosch district) is the almond, thc size being normally in the region of 7 inches long by 3¹/₂ inches wide by 2 inches thick in well-advanced specimens.

At Blaauwklip Spruit (or Blouklipspruit) some 4 miles out of Stellenbosch on the old Somerset road, the road crosses a small low-level concrete bridge over the spruit (rivulet), and a hundred yards up the hillside is the farm entrance. A little way in front of the homestead is a “land,” ploughed up and worked as a vegetable garden. From the top 3 or 4 feet, a number of implements in varying stages have been thrown out. These show that here again the implement-bearing gravels do not quite crest the hill. The site is almost exactly the same as that at Lorraine, and as rich. Mr. Lowe discovered the site in the stream below, where he found that rolled implements and refuse occurred actually in the gravels of the stream; this led him to search on the banks above the stream, and the main site was revealed. On a later visit Professor Dart discovered similar material in these gravels, including a few worked flakes. One such flake appears to be a scraper of a fairly definite type, and is thus of interest in this connection. The almond-shape again appears to be the normal, the size being about 6 inches by 3 inches by 2 or 1¹/₂ inches thick in the more finished implements. One or two roughly blocked-out specimens reach a length of 10 inches, a width of 5 inches, and a thickness of 3 inches. The types are identical with those occurring at other sites in the neighbourhood.

Almost exactly similar implements have been sent in from various sites of like type in the neighbourhood. Mr. Lowe reports other sites at Ida’s Valley, Rustenberg, on the Stellenbosch-French Hoek road, on Hell’s Hoogte (in a gravel pit immediately adjacent to the road right on the Hoogte), and also on the slopes of the Drakenstein beyond, notably at Pniel (Stellenbosch).

Villiersdorp.

Some 30 miles east of Stellenbosch, and behind the mountains overlooking this town, is Villiersdorp. Mr. C.O. Payne has sent in a large number of implements from here. These occur in exactly similar fashion in the gravel pits cresting the foothills below the mountains, and apparently lie in the gravels above the town. They are the same in every way as the Stellenbosch and Wellington specimens, the size and shape agreeing, as also the mode of occurrence.

General.

The various sites so far mentioned form a group of sites lying in an “L” 20 miles by 30. They are of interest as they show that the area below the mountains was comparatively thickly populated at this period. While we cannot regard the implements found on the higher-level sites, immediately below the mountains, as of an age equal to that of the gravels containing them, owing to the unrolled condition of the artefacts, yet the accumulation of soil above the gravels — usually only a couple of feet — is considerable when we realise that these gravels are situated on the tops of foothills, which have since been left standing as terraces by rivers which have carved out valleys below. In the case of Mr. Homan’s farm, Zoutendal, the whole hill stands well away from, and well below, the hills above Wellington, and the gravels were obviously formed by the Berg River when running at a height of about 100 feet above its present level. The implements in this gravel are rolled, and while identical in every way with those appearing from the higher terraces, are obviously older than this lower terrace, and in fact may have gravitated or been washed down from the higher terraces at a time when these newer gravels were in process of formation.

It would appear, then, that the whole area between Wellington, Paarl, and Stellenbosch at one time formed a plain below the mountains, some 250 to 300 feet above its present level. Gravels formed of mountain talus were left in this plain, and implements were made from the river material ready to hand. Subsequently the river level fell by some 150 feet, at which time the gravels on Zoutendal farm and similar sites were left, containing water-worn implements. Since this period the river has dropped a further 100 feet.

Later Stone Age material appears to be confined to the surface of the Stellenbosch Flats, or plains below the mountain ranges, and thus belong to a far later period than the material accumulated in the gravels.

Cape Peninsula.

Between the area discussed and the Cape Peninsula lies a plain some 30 miles across and generally not over 100 feet above sea-level. The surface is sandy, and rests directly upon the granite. The Cape Peninsula is a high-standing promontory rising directly from the Cape Flats and bounding them at their southern end, and reaching a height of 3600 feet in places, then falling almost directly to the sea. Formerly this mountain range was presumably linked up with the Stellenbosch mountains which bound the Flats to the north, and the formation, Table Mountain Sandstone upon granite, is identical.

Throughout the Peninsula sporadic finds of Stellenbosch implements have occurred. No site has yet proved really rich, but many show a fair number of specimens. One of the most interesting is the site which appears to extend from Bishopscourt, through the “Hen and Chickens” (a rock formation) and on to Kirstenbosch. Sporadic finds have appeared from here at different times. No really good specimens appear, but the rejects show that the desired implement was here the same as in the Stellenbosch area. Various single specimens appear from sites on the Cape Flats; in fact, as Péringuey mentions, the first implement of this type found in South Africa was discovered on the Koeberg Road by the Hon. W. F. Lyttelton in 1880.9 Further finds have been reported from the Wynberg and Constantia valley, but exact localities are not known. Dr. Péringuey found coups-de-poing at Fish Hoek,10 and Colonel Hardy has found similar implements at the opposite end of the same valley, some 4 miles distant, at Noord Hoek.

Probably one of the best Cape Peninsula sites was discovered by Mr. Lowe and myself almost half way down the slope below Cape Point lighthouse towards the Cape of Good Hope. This was a surface site, though the implements were found embedded firmly in a soft rock, and had been covered by shifting sand which had moved with the prevalent winds and left the site bare. The soft rock had obviously been formed from such wind-blown sand. Here again the implements are of the usual Stellenbosch type, and show no variation in shape or size.

Clanwilliam.

Before leaving the western end of the Cape System, it is worth noting that implements have been sent in from the Clanwilliam district and various other sites by Mr. J.M. Bain, an ardent collector, now resident near Stellenbosch. Somerset West and Gordon’s Bay have also yielded a little, and it has been stated that a site occurs on the Steenbras Mountains behind Gordon’s Bay. Péringuey11 describes a site at Beukesfontein, Ceres, and another at Maatjesfontein in the Laingsburg district, but both of these lie on the northern edge of the Cape System, and the material used in both cases is the bluish-grey Dwyka chert. This is of interest, as it probably shows that the Stellenbosch folk were proceeding southward and had not yet accepted the Table Mountain Sandstone as a possible material.

Pringle Bay, Hangklip.

Some of the most beautifully finished implements from the Southern Mountains area were found by the writer at Pringle Bay, collected on a hard, dry, turfy loess formed below an old spring on a mountain slope. Together with the material of Stellenbosch type were found implements of Still Bay facies, the material was all on the surface, and relative age was not discernible (see Plate III).

Montagu.

Farther east, the South African Museum excavated a cave some few miles from Montagu. This site may be regarded as the most important hitherto found in the Union, in that it is the only large cave which has so far been excavated with anything like full scientiic care and thoughtfulness. This cave was first discovered as an archaeological site by Mr. F. Jansen (now of East London), who has since achieved some little fame in the matter of the Victoria West implements. As a result of his discoveries, Dr. Haughton and Dr. K.H. Barnard of the South African Museum excavated the cave as far as the fallen rock would allow.

Some 6 miles east of Montagu a valley cuts southward into the Wittebergen and branches out into a “Y.” Almost opposite the fork, but on the right-hand side of the Western branch of the valley, appears a cave, some 200 feet up the face. A constant water supply exists in the valley below, and the cave is eminently suitable for bats and men. The main cave extends some 50 feet into the mountain, then at a height of 20 feet a narrow passage-way leads far into the mountain. Prehistoric man did not make use of this upper cave — probably having no means of access — but for some little time it has been used for obtaining supplies of bat guano.

The deposit was found to consist of beautifully stratified layers to a depth of about 16 feet12 If we leave out of consideration various rises and falls in the strata, we may broadly summarise the deposits thus:

1. A 6-inch layer of surface soil.
2. A 6-inch layer of Wilton material.
3. A 1-foot sterile layer of ferruginous sand.
4. 2 to 3-foot layer of stone implements of Stellenbosch type.
5. 3 feet of sterile ferruginous sand.
6. 3 to 4-foot layer of stone implements of Stellenbosch type.
7. A 3-foot sterile layer of ferruginous sand.
8. A layer of Stellenbosch implements, up to a foot in depth.

The cave proved a double sequence. Layer H was difficult of access owing to fallen material, and only two good coups-de-poing were recovered. These were of the pear-shaped variety of the Stellenbosch industry. Layer F, the next inhabited layer after (ie. above) H, contained similar pear-shaped specimens with a very small percentage of almond shapes (15 out of 77). In the next inhabited layer, layer D, the percentage of almonds immediately rises (70 out of 89), while in addition four are almost round, or ovate. So, judging only from these better specimens, we may express the sequence thus:

	Pear	Almond	Ovate	Total
Layer D	15	70	4	89
Layer F	62	15	0	77
Layer H	2	0	0	2

There was thus a change in fashion between layers F and D, which resulted in the pear-shape being displaced by the almond, and, to a lesser extent, by the ovate. It would be interesting to discover whether or not the ovate finally displaced both the earlier types. The second point of interest is the appearance of atypical Wilton types in layer B, again separated from the Earlier Stone Age material by a sterile layer.

Half-way between Montagu hospital and the road-bridge a number of coups-de-poing were found in a slight cutting made by a rain-water gully. These are not well made, and are of little interest.

Riversdale.

Across the Langebergen to the south-east lies Riversdale. Mr. C.H. Heese has done a large amount of indefatigable work here in the few years of his residence. He has discovered among other things a scattered Stellenbosch site at Riversdale about the showground. The implements, which are often well made, lie about a foot beneath the surface of the ground in ironstone gravels. They are of the usual quartzitic sandstone. The site extends across to the railway line where it is again revealed in a cutting. This site appears to be of the same foothill gravel type as those occurring in and about Stellenbosch district.

At Still Bay on the coast 20 miles south-south-east of Riversdale, coups-de-poing are also found, embedded in a limy concretion. This site will be referred to in a later paper dealing with the Still Bay industry.

Mossel Bay.

Probably one of the richest sites in this part of the Cape System is that discovered by Mr. J.H. Power at Mossel Bay. The town of Mossel Bay is situated on a promontory ending in Cape St. Blaize. The whole of this promontory forms an extremely rich archaeological field, and has been inhabited for some thousands of years by various peoples. Along the central keel of this projection of land runs the main road to Cape Town; a little south of this, and about 6 miles from the Mossel Bay station, are a number of shifting sand-dunes, situated high up on the shore. Beneath these are a large number of implements in various stages of manufacture. Mr. Power made several visits to the site and succeeded in taking back a vast collection to the McGregor Museum, Kimberley. This collection is of extreme interest, as Mr. Power, not satisfied with obtaining the best implements only, also obtained a huge variety of unfinished specimens, which indicate to the student the technology and methods employed by these folk in the manufacture of their implements, which will be referred to later. Any attempt at dating the finds was impossible in this case, as the sand overlay the material and kept it from being buried under any more stable deposit. The constant rolling and shifting of the sanddunes have alternately revealed and covered different portions of the site, making it difficult to judge the full extent of the deposit. Mr. M.C. Burkitt visited this site while at Mossel Bay, and took a considerable amount of material with him to Cambridge, leaving the greater volume behind at the South African Museum. There is reason to believe that Mossel Bay will yield a vast amount of valuable information if carefully investigated by the right workers.

Oudtshoorn.

Above the Kammanassie Mountains stands Oudtshoorn, where the 1925 meeting of the S.A.A.A.S. was held. While attending this meeting, I came across an extraordinary site on a small hill a mile or so north of the town. The coups-de-poing are small (4 inches long), and all show a convex cleavage face along the two edges and the butt. The fact that it seems impossible to remove flakes of this type from a stone with any consistency has led me to the belief that these were of purely natural origin, but further research may throw more light on the subject. The “technique” (if such a term may be employed) reminds one immediately of the scaling or shelling normally found in the spheroidal weathering of dolerite, but occurring also in cases of similar weathering in certain quartzitic sandstones. Until further knowledge is obtained these finds will have to be ignored.

George.

George is a little forest town below the mountains, some 40 miles south of Oudtshoorn. Visiting here for a few hours I was surprised to find a well-made coup-de-poing directly under a street lamp in the exact centre of the town. This find was providential, as it was made late at night. The following morning I was able to locate four implement-bearing sites. The first lay a little way north of the town in a donga running beneath the railway line due east of a small reservoir.

The town is built on a single street, which is joined at the north end, and at the middle, by two branches of an entrant road. The southern branch passes the golf course and the Industrial School, the northern runs directly from this road to the Dutch Reformed Church. The second site would appear to lie between these two roads on either side of a stream, as I found two indications — one opposite the golf course between the roads, the second in a small cutting on the northern branch, lying 3 feet deep in the talus material here prevalent. The golf course and the brickfields yielded nothing, nor did the commonage, except for a single specimen some 2 miles south of the town, and a little east of the continuation of the main street. There is reason to believe that a large amount of material could be obtained from this part of the country directly below the mountains if the local foresters could be made keen on archaeology. One of the difficulties at present is the fact that stones are returned to the ground in the process of tree planting, with the result that surface indications are rare, and that the stratification is upset in the afforested areas. Foresters might find themselves in the position to note stratification and also to collect some of the better specimens.

Knysna.

About 40 miles east of George stands the forest-harbour of Knysna situated below the Outeniqua mountains. The place-names in this area (Gouwkamma, Tzitzikamma, Kamnassie, Knysna, Outeniqua, etc.) immediately remind one that it was inhabited until comparatively recent times by pastoral tribes of Hottentots. The country here is extremely fertile, and has probably supported a considerable population for some thousands of years. The cave- and open-sites show an extraordinary richness which supports this view.

In 1922, Mr. C. van Riet Lowe found himself in a position to study the Stellenbosch industry of this area with remarkable results.13 During this period he kept in constant touch with the anthropological department at the University of Cape Town, and the greater part of his collection is housed there. I quote freely from his paper and letters to the department.

Situated about 200 yards on the left of the road from Knysna to the Concordia Plantations and about 2¹/₂ miles from the village of Knysna is a small rounded hill, separated from the main ridge by a slight saddle. The entire surface of this hill and of those in its immediate neighbourhood is composed of a dark loamy soil with slight impregnations of ine ferruginous gravel in the form of “iron pan” — the whole overlying Table Mountain Sandstone.

As a result of his excavations here, Mr. Lowe discovered a number of stone implements of Stellenbosch type. These seldom appeared on the surface, but a large number of artefacts were discovered in the soil and sub-soil of this hill at depths varying from 18 to 80 inches, through a layer of sand and in the “iron pan.” All these, and in fact all the implements found, were of Table Mountain sandstone.

A number of coups-de-poing varying in size from 5 pounds (9 inches) to ¹/₄ lb. (4 inches) were found. In association with these were a number of scrapers, knives, etc. Most interesting of all was the association of a single true flake-core with the flakes and coups-de-poing. This would imply a further evolution in stone technique than is normally presumed for this industry. A number of detaching hammers and discoidal artefacts also appear from this site.

In all, Mr. Lowe has recorded 10 areas in which coups-de-poing and associated artefacts were found. He gives a locality sketch of Knysna and the immediate surroundings, marking the sites and lettering them. Areas A, B, and C appear to be at or near workshop sites (see map).

1. Rich in yields, described above.
2. The surface and sub-soil here (overlying Table Mountain sandstone) are similar to those of site A, but the ground has been cut up by dongas, revealing implements which appear to have gravitated from a position further east.
3. The formation here is typical of the Enon Conglomerates, but is denuded. “Iron pan” is noticeable. Many of the numerous boulders from these beds have been worked, so a few advanced specimens were found.
4. Isolated coups-de-poing, flake knives, and scrapers appear from the neighbourhood of “The Hill” (C.W. Thesen), on the western slopes and near the crest of the hill on the west of the re-entrant shown on his map, and near the conjunction of the Enon Conglomerates and the Table Mountain sandstone.
5. This area is about 4 miles from Knysna on the Plettenberg Bay road. It contains recent gravel pits. This site is in or near a workshop.
6. Stands on a spur of the hill overlooking the Knysna-Concordia road, about a mile north-east of Knysna. The formation here is exactly similar to A, B, D, and E.
7. Lies about a quarter of a mile east of F, and is similar.
8. Lies in Block C, of the Concordia Plantation.
11. Lies about 60 yards east of the 14^th milestone from Knysna on the Plettenberg Bay road.
12. Is a quarry some 200 yards east of the 16^th milestone.

He mentions three other isolated finds.

From all his sites the implements are of precisely the same nature — coups-de-poing, biseaux, scrapers, and flakes, the implements being in all stages of manufacture, though completed artefacts are rare. All show signs of having been made from water-worn boulders.

The country-side along the Knysna to Plettenberg Bay road is flat or slightly undulating. It contains many pans (fens) and the sub-soil is impregnated throughout with the hydrated ferric oxide known as “iron pan,” the result of badly drained territory. This iron pan is extensively dug and used for road-metal. The small quarries at the 14^th (K) and 16^th (L) milestones were opened for this purpose and revealed a large number of coups-de-poing and flakes at a depth of about 3 feet.

Mr. Lowe showed that the ridge constituting the “back-bone” along which the Concordia road runs, including almost the entire length of the present Eucalyptus Avenue, was at one time a large Stellenbosch industry settlement. The implements are identical with those appearing from elsewhere in the Cape, both in technique and shape and also in material. The only addition appears to be the presence of worked flakes and a single true flake-core. One specimen of particular interest consists of an elliptical scraper about 89mm by 63·5mm by 25·4mm (maximum), with “heavy” secondary trimming along the major axis edge opposite the bulb; it is in Mr. Lowe’s collection. However, further research will be necessary before any conclusions can be arrived at on this point.

Other Sites.

Various other known sites in the Eastern Province of the Cape fall within the Cape System. Finds have been sent in from Kingwilliamstown, some of which are in Dr. Bruce-Bayes’ collection, East London, and some in Kingwilliamstown Museum ; others are from Stone’s Hill, Grahamstown, and from Healdtown, both in the Albany Museum collection, and others from Alice, the last being represented in the Kingwilliamstown Museum.

Middledrift.

The most interesting groups from this part of the Southern Mountains System are those to be found at Middledrift, a village set below the Amatola Mountains near the source of the Keiskama River. Here the brothers Wilson, Mr. Gladwin, and others have made some very fine discoveries.

The Wilsons regard the coups-de-poing as contemporaneous with the points usually associable with the Middle Stone Age, but their evidence is from alluvial sites, and is therefore not strong. Evidence of perhaps greater value lies in the fact that these two types appear at a depth of from 4 to 12 feet below the surface, showing that the Wilton and Smithfield implements to be found on the surface are of a later date.

Father Stapleton of Grahamstown differentiates two river terraces at Middledrift. The lower lies at about 40 feet above the present river bed, and appears about 50 yards east of the suspension bridge which crosses the river below the hotel. The earlier (upper) terrace is visible some 400 yards further east, and stands about 80 feet above the river bed. The road cuts through both these terraces, but on visiting the gravels I found no implements whatsoever, nor artificially worked stones in the upper terraces (negative evidence), while discovering two industries at the lower terrace. The first was of Stellenbosch type, much rolled and obviously deposited as part of the gravel. The second was Middle Stone Age, consisting of a well-worked point and some flakes, quite unrolled, and obviously of a date later than that of the gravels.

In the Euphorbia Donga, about a mile from Middledrift, were found a large number of coups-de-poing, at a depth of about 6 to 10 feet in the yellow clayey earth which had been cut into by the donga and its tributaries. Further from the town in a field overlooking the valley similar finds were made, but at no great depth. These sites were discovered by the local workers, and Mr. Burkitt and myself were enabled to visit there through their kindness.

Methods of Manufacture.

It would be as well to note here a few variations in the methods of manufacture employed in the various sites within the Cape System. Primarily the Stellenbosch industry is a pure core industry. Péringuey pointed out that in many instances a large portion of the butt of the coup-de-poing was unworked, and the surface was composed of the original boulder-face. He seemed to regard this as typical of the Stellenbosch industry, but actually it only occurs in artefacts in an early stage of manufacture — all well-finished coups-de-poing show none, or in a few cases only a minute portion, of the original boulder-face.

This brings us to the method very often employed. An ovate river pebble some ten inches long was taken, and four or five flakes removed from different directions, but all from about the same point. The result was much that obtained by sharpening a pencil. (Fig. 4a.) One end was pointed, the rest was left intact. The next step was the normal one of removing flakes from about the perimeter.

Another method (found by Lowe at Knysna) was to take a large pebble, split off a reasonable flake, and trim this from the boulder-face, working round the perimeter. (See Plate IV.) Here the boulder-face took up the whole of one face of the coup-de-poing, the work would be continued in the normal way from now on.

A third method employed was to remove a large flake, or perhaps better, a large lump, then to remove from this a flake. The result was a large flake having two convex faces, one forming each face. The object can best be described as a lens from which two opposite edges have been cut off square. (Fig. 4b.) Trimmed flakes were next removed from along each edge, from alternate faces, and if successful the work would be continued in the normal way. One point is of importance, as it will be reverted to in speaking of the Fauresmith coup-de-poing; the two convex faces each have a normal bulb of percussion, but each is at the side of the face it governs, i.e. at the edge of the implement, not at either end. Actually a few instances occur where the bulb is at the butt end, but these are negligible. This point may seem unimportant, but it is of extreme interest in any study of comparative material from Fauresmith, the Vaal River, and the Cape System.

Vaal River Area.

This area has already been written up by the writer14 and little more need be said about it in this paper. Major Collins, writing for the Journal of the Royal Anthropological Instutute,15 mentions a number of sites on the Vaal River, but most of these are outside the “Vaal area” as defined above (lying between Warrenton and the junction of the Vaal and Harts Rivers). He mentions sites between Meyerton and Vereeniging, at Vereeniging, near Klerksdorp, and so on. Johnson16 mentions similar sites. Mr. Leslie of Vereeniging sent some implements from the most interesting site, and Dr. Péringuey speaks of it in his work.17

Vereeniging.

The Vaal River has left extensive river terraces east and west of Vereeniging. East of the town is a small pit in which chipped quartzite flakes and various pebbles appear. West of the town, Mr. T.N. Leslie discovered unfinished coups-de-poing of diabase, smoothed by the action of water; with these he found quartzite implements of rougher manufacture, and showing less signs of wear than the diabase specimens. Now diabase is a tough material with a relatively clean fracture, and tends to be abraded under the action of rolling; quartzite, on the other hand, is a hard, brittle material, tending to shatter when struck, but only yielding to extreme rolling. The fact that these quartzite artefacts show less signs of wear, does not (as Péringuey and Leslie would like us to believe) necessarily imply two different periods and types of workmanship. The diabase implements all down the Vaal are well made, proving a good and tractable material; on the other hand the quartzite implements throughout the country show a greater percentage of rejects, and a less even fracture.

Bloemhof.

Professor Dart of the Witwatersrand University18 has described implements from the gravels at Bloemhof. The Vaal River here forms an island during the rainy season, but leaves one channel high and dry during the dryer season. It was in this dry river channel, which is 4 feet above the normal water level (i.e. 7-10 feet from bed to bed), and at a depth of 4-5 feet in the gravel, that implements and two right upper molars of an extinct elephant (Archidiskodon) were found. The implements are of the usual type to be found along the Vaal, so far as one can judge from the photograph. Professor Dart speaks of this gravel as the “mammoth gravel” and presumes that it “must reach back to a rather early phase of the Pleistocene.”

Dr. R. Broom in replying to this article suggests that19 “it may be regarded as almost certain that they (the elephant remains) are very much older than the lowest gravels of the Vaal, which cannot be of great geological antiquity.”

The remains described by Dr. Haughton20 as Loxodonta griqua and Griquatherium cingulatum, are all from the middle (60-70-foot) terrace of the Vaal, and date this terrace as of about Middle Pleistocene period. It seems more than probable that the lowest terrace at Bloemhof (if, of course, it is equivalent to the lowest terrace at Windsorton, etc.) is of middle to late Pleistocene date, and the implements of a similar age.

From Windsorton southward to the Vaal and Harts River junction, the Vaal is lined by three gravel terraces; the highest lies about 200-400 feet above the Vaal, and at a great distance (often 6 miles) from the river. The next gravel lies between 30 and 70 feet above the river, while the third series lies in the present river bed, or immediately above it, and is relatively modern. The highest gravel contains no human remains, and is apparently of Pliocene age. The next gravels are all Pleistocene, so far as fossil content proves, and contain implements of Stellenbosch type.

Windsorton.

From Windsorton a number of implements have reached the Kimberley Museum. Here, as always, it is a little difficult to assign the implements with any certainty to their correct terrace. The depth from the surface to the finds is given, but neither the height of the gravel above the Vaal, nor the height of the ground-level above the Vaal. However, the gravels here are of great age; they are composed of a mixture of fine water-borne materials and boulders, and have an overburden of sand often 24 feet thick (see Plate V).

Riverton.21

On Tipperary farm, Riverton, a small diamondiferous gravel was worked at the Rapids; a dam was built across the river and the working carried on below this. The gravel lies in the present river bed, and the dam was originally constructed of boulders from these gravels, and contains a large number of coups-de-poing, biseaux, and discoidal artefacts, all of which, while crude and rough, fall within the variation usual to the rejects of the Vaal River area. The material used is diabase, the shapes are those typical of the Vaal River. Capt. Shortridge of the Kingwilliamstown Museum obtained better specimens than the writer, from Mr. S. Tapscott who owns the farm. These show a tendency to a small coup-de-poing, 50% of the specimens being under 5 inches in length, 32% being between 5 and 6 inches in length, and the remaining 18% being over 6 inches. The better finished implements are all of a neat almond shape.

Pniel Mission Station.

This site is probably the richest Stellenbosch site in the Union, so far as the high standard of the rejects is concerned. It is also of interest as it forms the type of site of Dr. van Hoepen’s “Pniel Culture.”22 The subject of this culture will have to be reverted to when we have dealt with the Pniel finds.

The alluvium about 100 yards from the old Mission building lies in the present bed of the river, and in order to work it a breakwater or dam was built enclosing part of the river. This wall was constructed of river gravels taken from the river bed which just here skirts the foot of a dolerite kopje. In the gravels were found large numbers of coups-de-poing, biseaux, and discoidal artefacts these have all been thrown up to form a breakwater. Two shapes are present: an almond shape and a narrow pear shape. The latter may prove (judging only from the Montagu Cave analogy) to be earlier, but owing to the fact that the implements are always in an already disturbed position when found, no proof is available from this site. One point is of slight interest, the pear-shaped coups-de-poing tend to be slightly larger (up to 29 cm. in length) than the almonds. The normal variation in the two types lies between 21 cm. and 12 cm., a fairly wide range.

A peculiar type of triangular biseau is represented from this site. In the more usual biseau the implement is similar to the coup-de-poing, but the end which would normally be taken up by the point is composed in the biseau of the intersection of two cleavage faces, leaving a sharp edge more or less at right angles to the length of the implement. The biseau from Pniel, however, is often of a type which can only be regarded as a reversal of this; the point is present, but the portion, which in the coup-de-poing would form the butt here consists of the axe-like edge.23 With these normal Stellenbosch types appear a few unconventionalised flake-implements, of doubtful association.

Method of Manufacture.

It would be as well here to enter into the method of manufacture usual in this area. The material employed is generally amygdaloidal diabase, and has apparently been obtained from the river gravels themselves. The makers appear to have struck off a piece of material from a stone, then to have removed a large flake from the lump taken. The resultant flake was usually the size of the whole hand, or a little longer, and the two faces consisted of positive cleavages covering the whole of each face, the two ends often being formed by the intersection of these two curved surfaces, while the two edges consisted of the outside of the original stone. The shape can best be described as that of a lens which has had two opposite edges removed by having been out off square, leaving two parallel edges revealing the section of the lens in each case. This shape once obtained, the two edges were trimmed alternately, leaving a section (taken through the width of the artefact) similar to a rough parallelogram. If the work was successful so far, the two edges would again be worked, but in each case from the face opposite to that chosen before. This would reduce the size of the artefact and the shape would tend to be that of a coup-de-poing: or in some instances, one end or the other would remain unworked, leaving the axe-like edge resulting from the intersection of the two original faces untouched. The former shape would represent the coup-de-poing, the latter the two types of biseau.

In view of this technique, which is very consistent at Pniel, it appears impossible to associate these implements, as Dr. van Hoepen has done, with the Victoria West types, or with the single Le Vallois flake found by him at Thaba Nchu, about 150 miles east. It seems to be this association (forced and unproved as it is) which forms the basis of Dr. van Hoepen’s Pniel culture.

It is necessary here to turn back to the methods of manufacture employed in the Southern Cape System area. Here we saw two methods: one was the method noted by Péringuey, the taking of a river pebble of a likely shape and size, and the removal of four or five flakes struck from a single point at one end, the resulting artefact being rounded at the opposite end where the original pebble surface appeared, and tapering at the end, from whence the flakes were struck, the flake-scars giving an impression exactly similar to the chip-scars left when sharpening a pencil. It is obvious that such a method of initial blocking out could only be employed where rounded river pebbles were abundant. From Mossel Bay the evidence brought by Mr. Powell shows, on the other hand, that the method of manufacture employed there was in most instances exactly that employed at Pniel and all down the Vaal area. In other instances the usual Stellenbosch technique was employed. At Knysna, Mr. Lowe’s material showed that a variety of methods was employed in the manufacture of the coup-de-poing; the most interesting being the occasional utilisation of a large lenticular flake struck from a heavy water-worn boulder, and worked in such a way as to leave the outside of the boulder unworked and the whole of the cleavage face worked away by the removal of trimming flakes from the surrounding edge; the resulting artefact is aptly described by Mr. Lowe as a “turtle back.” The evidence thus leads one directly to the conclusion that the technique employed varied with the material used and the mode of occurrence of that material in each individual instance. The variations occur consistently right through the Stellenbosch industry, and it is therefore quite impossible to differentiate between the Vaal River industries and those found in the Cape System area on the grounds of technique employed.

Barkly West.

At Barkly West the river has apparently changed its course a number of times, leaving a series of gravels on the south side (Old Pniel) and on the north (Barkly West).

There are three terraces represented here. The lowest lies a little above the present river level, and is best seen (though sadly disturbed by diamond diggers) on the southern side of the river. The second is some thirty feet above this, and appears at a depth of up to 24 feet in Canteen Kop on the northern side of the river. The highest terrace stands about 100 feet further up and contains no human implements or remains. It is not possible to differentiate between the types of implement found, as the two lower gravels (if two occur on the southern side) are completely mixed up at Old Pniel, and only one appears at Canteen Kop. The implements found at Canteen Kop are of the same material and shape as those from Pniel and other Vaal sites, but the rolled condition precluded any attempt at discovery to what extent the technique was the same. The implements are generally larger than those from Pniel.24

Waldeck’s Plant and Niekerk’s Rush.

The material from these sites is similar in every way to that found at Pniel, and no further remark is here called for. Mr Hodkinson25 pointed out that the implements occur at a depth of from 10 to 30 feet below the surface in gravels over which lie 10 feet of red sand. These gravels may be of an age comparable with the lowest or the second gravels at Barkly West, but cannot be regarded as of an age comparable with the highest terrace at Barkly or the Klipdam gravel. It is possible that the Pleistocene fossils described by Beck26 (Mastodon or Bunolophodon) are from this terrace, as also Fraas’ discoveries27 of Equus zebra, Damaliscus sp., Hippopotamus, and Iridina.

Mr. Hodkinson also showed that the gravels at this part of the river prove a sequence between the Smithfield (Later Stone Age) and Stellenbosch industries, the former lying on the surface, the latter being included in the gravels directly beneath.

Droogeveld, Rooipoort, and Middelplaats North also show implements of similar types.28

Owing to the fact that the Vaal gravels are diamondiferous they have been thoroughly explored, with two main results. First, access to the gravel materials is easy, owing to the fact that the pitting and working necessary has been done by the diggers. On the other hand the disturbance of the gravels is extreme, and many gravels have been gone through a second time (as at Old Pniel) with more scientific methods. The result is that we have accumulated a considerable amount of show-material for museum purposes, but very little fact on which to base our deductions. The palaeontologist finds himself at a similar loss. Dr. Haughton29 can only state of the specimens he examined that all “probably came from the 60-foot terrace,” — apparently the second terrace above the river.

It seems that an entirely new geological survey of this part of the river will be necessary to discover which gravels can be correlated, and also to discover whether du Toit’s original hypothesis of a series of barriers holding back the gravels30 is correct, or whether these gravels grade with the river — a point brought out by Mr. Burkitt on his recent visit to this country. It would seem necessary also, in such a survey, to give the total depth of gravel at each point along the river, and the depth of the gravel below the present surface. Only by this means can we arrive at an immediate recognition of the comparative age of a ind from the meagre evidence (which almost always takes the form of depth below surface) supplied by donors of implements.

One point may prove of extreme importance. At a time when the highest (oldest) gravel was being deposited the Vaal River seems to have had a course very different from that at present followed. It seems to have turned westward from Windsorton, along the latitude 28° 20″, then to have turned southward through Klipdam, thence following the present tributary of the Vaal due south to join the Vaal at a point 24° 37″ east by 28° 30″ south. The gravels along this line show no evidence whatsoever, so far, of stone implements, nor do the highest gravels, which seem to follow the river westward from this point. If we could obtain a fairly representative group of fossils from the uppermost terrace at Klipdam, and compare it with a group from the middle terrace at Barkly West, or further down the river, we would be in a position to gauge the geological moment at which man appeared on the Vaal River. The evidence is available, and becomes increasingly so as diamond working continues.

General Distribution and Facts.

While these two main areas, that of the Vaal and the Cape System area, might appear to form two separate culture-centres, there is no reason to presume this to be so. They are very much more likely to be geographical areas which are better known to us at the present day owing to the drying up of the western and central portion of the Union. That this desiccation has occurred is proved by Penning’s discovery of stone implements of Stellenbosch type in what is now desert,31 and also Dr. K.H. Barnard’s discovery of a large implement site on the surface and under similar desert conditions.

Other Sites.

A few interesting sites outside the two areas noted are worthy of mention. Mr. Bacon of Randfontein, Transvaal, sent a number of perfect coups-de-poing of quartzitic sandstone to the South African Museum and the Port Elizabeth Museum. These were found at a depth of 6 feet in a natural clay-puddled pan in Randfontein, which was being worked for brickmaking purposes. The shape is amygdaloid and the implements are very well made, showing a high standard of workmanship.

Swaziland.

In August 1921, the Government Secretary, Mbabane, Swaziland, submitted a number of specimens to Dr. Péringuey at the South African Museum. These had been discovered by the mining engineer at Ezulweni, Swaziland Tin Mines, Mr. Pote. The finds consisted of (a) coups-de-poing, (b) lance-heads, (c) iron bangles or rings. The finding is described thus:

Mr. Pote was present when the “Monitor” (a powerful jet used for excavating and washing the tin-bearing gravels) was at work, and personally took out what appeared to be a mass of concrete. On breaking it open the rings were found, as if they had been worn or tied together. They were found a good 20 feet below the present surface on the top of what is known as bedrock (a soft substance underlying the tin-bearing gravels). In this gravel, “flint implements” and weapons are frequently found in great profusion. The “flints” sent with the rings were found within 6 feet of the rings and at the same level. The red soil which forms the overburden has never been found to contain “flints” … Everything goes to show that the “flints” in the tin-bearing gravels are of immense antiquity, and the question arises whether it is possible (if these rings are of the same date as the “flints”) for iron to have been preserved for such a lengthy period?

The idea that the iron bangles are of a date contemporaneous with the implements (of indurated shale, not flint) must, of course, be discarded as impossible. Erosion gullies (dongas) have apparently been the means whereby the iron and stone objects have fallen into association; Dr. Péringuey was in complete agreement with this view. Mr. F.H. Dutton in his reply objects, and states that “it seems certain that the rings were not buried in a deep hole, and that they were in man’s possession before that great mass 25 feet deep was deposited there.”

Dr. Péringuey also notes that two types of stone implement are present.

At Mbabane there have been found among the small objects usually called “arrow-heads,” five lance-head-like narrow points worked on both faces, of remarkable workmanship. The contiguity of these small spalls with the large bouchers is very surprising, and requires still a good deal of explaining.

The three types are most certainly in complete “Glozellian” association, and further evidence is necessary before we can definitely regard the Stellenbosch types of coup-de-poing, the typically Middle Stone Age lance-heads, and the iron bangles as contemporaneous.

Pretoria.

From the Aapies River gravels in Pretoria, Mr. Leith collected a number of extremely bad coups-de-poing, and formless implements of this group. These are of quartzite and ill made and uncouth as a result. It has been suggested32 that these implements should form the basis of a subdivision of the Stellenbosch, a Pretoria industry or phase. Mr. Leith in his publication33 spoke of these Pretoria finds as “Eoliths”; Péringuey points out34 that coups-de-poing were found with these in the same gravels. It seems unnecessary to make a dividing line between these Pretoria implements and those from other sites, the fact that the material is intractable and fractures unevenly accounts for the difference in standard, as may be seen at a large variety of sites in different parts of the country, e.g., along the Orange River and in the Barberton and Lake Chrissie districts.

The factor of material is a basic and controlling one in South Africa. In Europe, flint and sometimes other tractable materials were used; in South Africa owing to the complete absence of true flint we have a great variety of materials, but it is obvious after a thorough survey of sites that the industry is the same wherever the material allows, but is markedly and consistently different directly the material is of a bad type. This can at once be seen in Péringuey’s series of plates.

An example may be taken from the Orange River. From the junction of the Orange and Vaal rivers westward, quartzites abound, while here and there banded jaspers appear. The implements from Griquatown, some few miles north of the Orange, are of jasper, and show the peculiar flat cleavage with clean edging so typical of jasper; the resulting implements are generally flatter than those of diabase, dolerite, and quartzite from further up the river, owing to the banded structure of the jasper. It is at once apparent that the material with its peculiar structure has affected the implement; in much the same way objects made of pitch-pine and of ebony in our own technique vary considerably from each other. Whereas we have suited our material to our needs — concrete, cast iron, wood, wrought iron, goldfoil, each supplying an individual need — the maker of the various Stellenbosch implements conversely had to suit his implement to whatever material came to his hand. Directly quartzites are the best obtainable materials the workmanship falls off to a great extent. The size of the implements drops from 6 inches to 4 or even 3 inches. The length : width : thickness ratio changes from the fairly constant 5 : 2¹/₂ : 1 found in the normal Stellenbosch to 6 : 4 : 3 in quartzite implements. Whether we shall have to make a separate industry of this quartzite group it is difficult to say; such a course might be legitimate, but seems unnecessary in view of the fact that the difference is largely one of material. As a possible instance of reversion to type, Dr. Haughton discovered a typical Stellenbosch coup-de-poing of perfect shape, some 7 inches long, 4 across and about 1 inch thick in Gordonia district. The implement is of a brown quartzitic sandstone, identical with some of the better Table Mountain Series material in fracturing qualities, and it was possibly the chef-d’œuvre of one of the wandering peoples who had passed across the Orange River, leaving behind the jaspers and quartzites of that region (Plate I.)

It is exactly this difficulty of material that has worried Rev. Neville Jones in his Rhodesian work.35 On p. 23 he says:

The early Rhodesian had to make use of hard intrusive rocks, silcrete, and any other stone of local occurrence and sufficiently hard to withstand wear and tear. While some of these rocks could be easily fractured, others, while being suitably tough, were refractory, and the most skilled artificer could not make a presentable tool from them. The result is that in some places, where the worker was evidently torn between a natural preference for producing a shapely implement and the obvious duty of producing a strong and tough one, there is often a wide variation in technique, and it would be easily possible, in more than one locality, to make a series covering whole ranges of European cultures. This is somewhat confusing, and would tend to lead one to the conclusion that any attempt to decide the relative antiquity of the implements in any given locality must be made on the evidence of the best examples only.

In spite of the hesitancy necessarily arising from the realisation of this state of affairs, Rev. Neville Jones still feels that at certain sites he is justified in presuming a sequence paralleling the Chelleo-Acheulean sequence of Europe; on p. 25 he continues:

Some years ago I spent part of a holiday in examining an area to the south of Vryburg, and at Tyger Kloof36 I found patches of gravel on high ground on the top of the Kaap Plateau and occupying depressions in the Campbell-Rand Dolomite. These were evidently the remnants of an ancient river terrace, which it was difficult to relate to any existing stream, unless it be to a deep kloof, generally more or less dry in the winter months, and half a mile or so away. In this ancient gravel I found a number of bouchers of typical Chellean type, but nothing comparable with those of Acheulean culture. At Taungs, a few miles to the south, there is, however, a gravel of obviously later date over which the present river flows, and there the Acheulean type is abundant. These two localities combine to furnish evidence that is, I think, sufficient of itself to indicate that there were at least two culture periods in the South African Lower Palaeolithic. I therefore hold that Dr. Péringuey’s original divisions should stand, though I would prefer to call them “South African Chellean” and “South African Acheulean”

Mr. E.G. Bryant has lately presented the South African Museum with a single coup-de-poing recalling the Chellean of Europe. It appears to be from a high residual gravel, capping a kopje in Khosi’s Native Reserve, British Bechuanaland.

Unluckily we have very little definite evidence of such a sequence as yet in the Union, though that does not preclude the possibility of future evidence forcing us to differentiate. Dr. van Hoepen drew attention to the vast differences between the well-made implements of the Vaal River and the uncouth specimens of the Aapies River, Pretoria — Leith even speaks of these latter as “Eoliths” — but even here the material used on the Vaal is far more workable than the quartzites of the Aapies. Too little definite evidence of a true sequence of a bad early type followed by a good later type, both in similar material, has yet been discovered. One local sequence in shapes has been proved at Montagu Cave by Dr. K.H. Barnard, but even this does not constitute a cultural change, and more evidence from widely distributed sites would be necessary before any generalisation is possible.

Rev. Neville Jones gives a table of his finds at Taungs and Tyger Kloof, which is given below with slight alterations.

Chronology	Finds	Locality
6	Later Bushman implements of crude workmanship, surface.	Taungs and Tyger Kloof.
5	Earlier Bushman implements from subsoil.	Tyger Kloof.
4	Well-worked flake implements from top 6 inches of pebble bed, apparently Mousterian in type, merging into …	Taungs.
3	Earlier flake implements, found throughout pebble bed, possibly some coups-de-poing.	Taungs.
2	Coups-de-poing, scrapers, and probably some flakes from pebble bed. (Péringuey’s “Orange River type”).	Taungs.
1	Cruder coups-de-poing, scrapers, and flakes from high terrace gravels. (Péringuey’s “Stellenbosch”).	Tyger Kloof.

Chronological Dating.

The chronology of the Stellenbosch industry relative to Middle Stone Age material is perfectly clear. The discovery by the writer of Middle Stone Age material at Middledrift lying unrolled and on a gravel containing rolled material of Stellenbosch type should be mentioned here. The discovery of Middle Stone Age material together with Stellenbosch coups-de-poing at Mbabane is of importance, but the evidence is not sufficiently definite for us to draw conclusions. The same may be said of the Pringle Bay finds.

The Vaal River gravels have also produced evidence that the Middle Stone Age material there is of later age than the Stellenbosch material. A considerable amount of other data of a similar nature exists, but so far we have discovered no definite cave-sequence of these two ages.

Montagu Cave gives the best proof of a sequence between Stellenbosch and Later Stone Age types, but the Vaal River gravels give a similar sequence with great consistency; Later Stone Age material being found on the banks, while the gravels directly below contain Stellenbosch material.

Dr. Lebzelter while on a visit from Austria did a considerable amount of research work in the Natal Highlands south of Vryheid. The result of this work is perhaps a little startling in one respect, and is certainly worthy of note here. He states that wind has built up the earth into four loess-like terraces, composed of sand and small stones. Subsequently the sand has blown out from the surface of these terraces, leaving a “gravel” of small stones only, lying directly upon the normal loess. The lowest gravel is thus the earliest, in his estimation, while the highest is the latest.37

“The surface finds,” he says, “embrace two kinds of artefacts, viz., small implements similar to those of the rock-shelters and of the first gravelly layer, … and also large hand-axes of earlier palaeolithic character. Our observations have brought us to the conclusion that these massive hand-axes of early palaeolithic type are here younger, or later, than the culture of the first gravel layer, which, however, itself bears a similarity to the late palaeolithic, European Capsian.”

These hand-axes or coups-de-poing appear on the surface of the humus which has accumulated over the highest “gravel,” and their position is indeed extraordinary unless they have gravitated from a still earlier river terrace situated at a higher level, but since entirely denuded.

His further finds consist of:

II	“Inyatian”	(Smithfield types).
III	“Mangenian”	(probably also Smithfield).
IV	“Isikwenenian”	(similar to Middle Stone Age specimens).
V	“Inxobongoan”	(definitely Middle Stone Age and recalling Glen Grey Falls material, and Swaziland lance-heads).
VI	“Ingeleduan”	(apparently similar to IV, but on finer flakes, Middle Stone Age).

This series (VI to II) is perfectly normal, Middle Stone Age followed by Later, but he believes the coup-de-poing types to be later than even the Later Stone Age material, whereas they should precede the entire series. The evidence which has already accumulated elsewhere to prove the sequence Earlier-Middle-Later is so overwhelming and so consistent that it seems improbable that the Earlier Stone Age could have survived in this one area right up to such comparatively modern times. The findings, while suggestive and worthy of note, appear to be based upon some flaw in research which it is impossible to discover from the preliminary report here quoted.

Conclusions.

This paper, while giving a fair general survey of the sites represented at the various museums in the Union, is not by any means an exhaustive study. An exhaustive study of sites, or indeed a distribution map, would be misleading. From Cape Town the main railway line runs north-east to Johannesburg. East of this line the country is covered by a network of railroads, westward hardly any such lines exist. The obvious cause of this distribution of railroads is the uneven density of population, while the chief result is that the archaeology of the Union is only fairly represented from the south-eastern half of the sub-continent; our distribution map follows the railways. How misleading such a map would be is immediately evident when we realise that wherever geologists have been in the north-western districts, implements of Stellenbosch facies have been discovered.

A further point must also be brought out: in this paper it has more or less been presumed that the Stellenbosch Industry is a single unity. Within limits this is true, but it is necessary to bear in mind the fact that variation within comparatively small limits does exist. These variations, however, all appear to centre round a norm, and they are of two types — those due to variation in material and source of supply, and those due to an individualism in time or place which we might term “fashion.” Under the first heading come the extreme variations existing between the finest implements of quartzitic sandstone (the Pringle Bay specimen and the Gordonia specimen being typical) and the ill-made specimens of white quartz. It seems enough to state that so far as my knowledge goes (judging always from the best implements from representative sites) where good stone was used good implements appear, and that no really well-made implements of quartz or granite are known, though one or two promising specimens have been found.38 So far as the distribution is concerned no great variation is known; we have seen that in the various well-known areas the mode of manufacture and the implements made are identical if we consider that some were made from rock fragments and some from water-worn boulders, the proportion varying in the different districts.

As to fashion, we do not know much about the chronological order in which the various shapes occurred. Montagu Cave gave us a clue for the Southern Mountain area, but no additional evidence of a similar or of a further sequence has yet appeared either in this area or from elsewhere.

These facts do not preclude the possibility that in time we may have to divide this phase into a number of either periods or areas, or even both.39 It seems unlikely that South Africa was the field of a single invasion of Lower Palaeolithic culture which ceased from evolving within the Union, and also failed to receive successive suggestions and impacts from more northerly evolving peoples (if such they were).

In this connection also it must be remembered that a differentiation has already been made between the Stellenbosch Industry and the Fauresmith, and this in itself is an acceptance of the possibility of some change greater than a mere change of fashion. Rev. Neville Jones’s Tyger Kloof finds and similar evidence from this area should most certainly evoke a division into a Lower and an Upper Stellenbosch, even if no evidence is forthcoming from Stellenbosch itself. But it seems a pity to multiply our terminology to include evolved or degenerated forms.

The Victoria West Industry.

Before discussing this industry it would be as well to note that the neighbourhood of Victoria West provides examples of three, or perhaps even four, different industries; these, however, fall under their own various headings and do not affect the actual Victoria West Industry.

The village of Victoria West is situated some seven miles west of the main Cape-De Aar line at Hutchinson, and at about 31°24″ south by 23°5″ east. The village lies, enclosed on three sides, in a funnel-shaped valley between mountains of a tafelberg type, composed of a soft blue shale capped with dolerite.

It would appear that at some very remote period the whole of this portion of the Karroo was capped with a dolerite sheet, at least 20 feet thick. Subsequently the entire sheet has been broken up by the normal processes of denudation, and as a result the whole country has been left as a plain from which rise hills, perhaps 500 feet high, all exactly level at their tops, and reaching to the remains of the original dolerite sheet, which is still retained in the form of a cap. The hills are of two types, directly resulting from this process of denudation. The first is the “tafelberg” type, consisting of a large dolerite cap, and falling away at the edges to form a sloping hillside. The second type is the same, but the dolerite cap is merely a small circle of the original sheet, and the resulting hill takes the form of a slightly truncated cone with a dolerite crown, pyramidal, and known as the “spitzkop” type.

The dolerite is slowly weathering away with the passage of time, but these caps still stand as “built-up” blocks of dolerite, or large unrolled but heavily weathered blocks of rock. The sides of the mountains are composed of talus and broken-down material from the level of the dolerite and below it, and blocks of dolerite thus form an integral part of the mountain slope, mixed in with an earthy rubble of disintegrated dolerites and shales.

The surface of the talus is composed of dolerite boulders, varying enormously in size, which extend from the cap to the edge of the hill slope, and there stop in a very marked line. The reason for this demarcation appears to lie in the fact that the action of water is slowly building up sandy alluvial material from the slopes to form a plain out of which the hills rise. The dolerite talus thus disappears and extends, in all probability, well below the present surface, but has slowly been covered by the sandier soil gathering round the feet of the hills. Here and there slight rises in the plain point to the subsided remains of old spitzkops, still held together by the broken-down cap of dolerite, but slowly conforming to the plain.

The hills lying north and south of Victoria West are of the tafelberg type40, and they run together above the town to form a narrow pass, opening out almost immediately into a wide basin. In the middle of this pass and overlooking the town from its western end stands a small spitzkop, attached by a saddle to the southernmost hill. Across this poort has been constructed a dam, which holds back the water from which the town is supplied. The dam is small, but the water extends back from the wall for a distance of two miles, and is a mile wide at its greatest width. This reservoir forms a small corner of what must once have been a vast lake or pan extending over an area of a hundred square miles, being held back by the hills above the town, and it has subsequently broken through and carved the poort. This lake would cover the present farms of Blaauwkrantz, Ganskraal, part of Gemsbokfontein, and portions of the Victoria West Commonage.

Either owing to the presence of the lake, or to the breaking through of the natural dam holding back this vast mass of water, the whole area in the valley covered by Victoria West appears at one time to have formed the bed of a stream, either wide and short-lived as a result of the bursting dam, or perhaps perennial and meandering over the enclosed plain. As a result it is noticeable that at the height of the present native location on the northern hillside, and at a corresponding level on the opposite slope, there is a deposit of lime, while some of the softer rocks near the native location on the northern side have been considerably undercut by the action of water.

History.

Mr. F. Jansen, while Resident Magistrate at Victoria West, had his attention drawn to rough coups-de-poing behind the present hospital, and as a result he spent a considerable amount of time and energy in searching for further sites in the vicinity and elsewhere. He was able in time to locate various sites at Victoria West itself. These will be referred to as:

• Hospital,
• Golf Course,
• Station,
• Western Hillside (south-west), and
• Eastern Hillside (north-east).

He has also located four sites outside the immediate vicinity of the town:

• Zuurkop (Wolvefontein),
• Vingerfontein,
• Melton Wold, and
• Loxton.

All named from various farms on which they appear. Besides these there are a variety of minor sites from which individual specimens have been collected. At all these sites he discovered that the coup-de-poing, roughly made in dolerite, was present, together with a homogeneous group of implements of a type not normally associable with the coup-de-poing.

Description of Implements.

These “new” implements41 are in their general conception similar. They are more or less high-backed objects showing a core technique. The high back is formed by the removal of a number of flakes from a circular edge which bounds the implement. The opposite face is the more interesting and consists of a single negative cleavage scar forming a depression over the greater part of that face, and edged for perhaps a third of the circumference by a narrow lip formed of the scars left by the removal of a number of minor flakes: in many instances this lip has been worn smooth by weathering, or perhaps from use. (See Plates VI and VII.)

The size and the general shape of the artefacts vary, the length varies (at Victoria West) from 4 inches to 9 inches, the flake scar of the main flake removed is relative, of course, to the size of the implement, but always maintains a fairly constant ratio to the whole of the face-generally about 3 : 4.

There are some three or four recognisably different norms about which recognisable types may be grouped. The “hoenderbek” (fowl beak) shape comes to a point at the one end, the main flake is struck off from one side of the under face. The “horse-hoof” type is round, and the back tends to be flatter, the under face shows the normal cleavage scar, but very often it appears that a second and smaller flake was struck from the same point as the main scar, and covers a large part of the main scar. The “skilpad” (tortoise shape) type is similarly round, but has a higher back, while only a single flake has been removed from the under face. So much can be gathered from the implements found; the extreme weathering of the dolerite in all instances makes it difficult in many cases to see from what point, relative to the core, the main flake has been struck off. Similarly it is always difficult to see from where the smaller flakes were removed. The flaking is generally heavy, and no light secondary trimming is visible.

Material.

The dolerite used is always fine (though coarse material was ready to hand) and the texture may best be described by stating that the consistency and timbre were very similar to those of cast-iron. While the fracture of this material is of course of the normal conchoidal type, the bulb of percussion is often small, and hardly visible, while the secondary ridge, encircling the bulb at a radius of about an inch and a half, is excessively pronounced. (See Plate VIII, 1.) From experiments conducted with fresh fractures it was ascertained that the great saucer-like depth of a negative scar left by the removal of a flake was due to the cast of this marked secondary ridge, and also to a tendency towards a very sudden upward turn taken by the cleavage on nearing the release at the front end of the flake.

One other fact is of interest, the note given off when the dolerite is struck is musical, pointing to considerable vibration within the stone. Much of this vibration appears to be across the direction of the stone, giving one or two interesting results. If a long piece of stone is held by the middle, and an attempt is made to remove one end by a blow, a portion from the opposite end is often removed; the cleavage in such an instance occurs at a point symmetrical about the part held to the point of percussion, and at the opposite side. Again, if an attempt is made to remove a flake from a piece of material, the blow will very often fracture the stone into two halves, the fracture in this instance being at right angles to the plane of the desired flake cleavage.

Argument.

When Mr. Jansen first announced his discovery, he met with a cool reception. Individual artefacts of this type which had weathered to any degree, especially in dolerite, which is a difficult material to work, were regarded as more than probably natural forms. Professor Schwartz of Grahamstown, from specimens submitted to him, declared that their form was definitely the result of insolation — the effects of alternating extremes of heat and cold.

This view was based upon three hypotheses. First, that the greater number of specimens lie on the concave face when found, the pyramidally worked face being uppermost. Secondly, it was doubted whether flaking by the percussion method would produce the saucer-like depression forming the under side of these specimens. Finally, he and Mr. J. Hewitt doubt whether it is possible to remove a flake 9 inches in width from this type of dolerite without a manched hammer.

Mr. Hewitt has dealt with these objections in a paper,42 and the views expressed are worth quoting. He gives three possible views:

Against the objection that all the Victoria West implements are the results of insolation, it may be urged that owing to the rolling of these artefacts, rolling which is evidence in all surface specimens, they would naturally stabilise themselves upon their widest and flattest base — the concave face. If this part were shielded from the sun in this way, it would not be influenced by the extremes of heat and cold, and hence the main flake could not have been removed by insolation. Conversely, if the object lay upon its back, and it were possible to remove the main flake by insolation, then the flake-scars on the shielded face could not have been removed by insolation. We are in fact dealing with an artefact which has both faces trimmed: one face by the removal of a number of flakes leaving a pyramidal back; the other by the removal of a single large face, bounded on two sides by a number of small flake-scars, giving a concave face.

It has already been shown that dolerite cleaves without a marked bulb of percussion, but with a well-defined secondary ripple and a marked “kick-up” at the release of the flake; Both of these factors would go to produce the saucer-like depression on the one face. Thermal fracture, or fracture by insolation, on the other hand, shows, in all the specimens I examined at Victoria West, that a curve is presented in only one plane.

Finally, both Mr. Jansen and myself have proved it possible to remove a flake comparable in size with any removed on Victoria West specimens, with a manched hammer or with a properly peined hammerstone, though it is admitted that the fracture is precarious, and governed largely by the lateral vibrations set up in the stone.

Quite apart from the quality and character of the material, the mode of occurrence in the various sites precludes any possibility of thermal fracture. The chief points worthy of note in this connection are that the distribution is not even, sites are defined, and each is confined to an area usually about 100 yards square. These sites do not occur on the same faces of hills, thus they do not get sunshine at the same period of the day. The sites do not in all instances form part of talus material, the station and the golf-course sites stand out in the plain, and the material appears to have been actually brought there.

The Sites.

The most easily accessible site is that upon which the railway station has been built. This site consists of a large number of stones, all well under a cubic foot in size, scattered over an area of perhaps half a mile by a quarter of a mile. The station stands in the centre of this site, and the goods yard has been entirely cleared of artefacts and other stones. On the north-east side of the station this site extends on the flat, while on the south-west it rises to a small eminence. Over the whole surface hundreds of worked stones appear in all stages of manufacture. To the north-east the site stops at a long low rise, composed of a large variety of stones, but containing little dolerite. The site appears to continue to the east of this hill on a scattered stony surface.

The golf course lies north of the village, a little way round the foot of the hill bounding the north-eastern side of the town. Here the site extends in disconnected patches for a distance of perhaps two miles, lying below the mountain slope which it follows round as far as Marseilles farm. This series of sites does not form part of the talus, but between the talus-edge and the string of sites lies a perfectly bare strip perhaps two hundred yards wide, while on these patches in the plain almost every stone shows some signs of having been fractured by human agency. In the actual talus, stones have also been worked, but here the percentage of worked material is low, the greater number of stones being pure talus material.

The school site lies on and in the talus of the opposite hill, which bounds the Victoria West valley on its south-western side. The greater part of the material appears to lie on the surface, but a certain quantity appears in a trench dug 3 feet deep into a limy concretion below the school. This site is probably the continuation of the site called by us the “West Bank Site,” which lies on and in the talus of the same hill, but about a mile nearer the dam. At this West Bank site we have the most interesting collection of finds. On the surface a considerable number of worked artefacts were to be found, and when an avenue of pepper trees was planted here some years ago, Mr. Jansen found artefacts at a depth of 4 feet beneath the present surface of the soil and in a limy deposit. Appreciating the significance of this fact, Mr, Jansen asked for the help of eight native prisoners, and permission was granted by the Department of Justice at Pretoria. With their help we were able to discover a considerable amount about this industry. Below the surface was an apparently sterile layer of 2 feet, below this again was an aggregation of implements, about 2 feet thick. (See Plate IX, 2.) The same types reappeared at depths of 8 to 10 feet, each layer showing relatively less weathering than the one above, until at a depth of 10 feet no appreciable weathering was visible.

This lowest layer rested upon a hard natural concrete composed of lime, shale, and talus from the mountain; this concretion in turn rested upon the original soft shale bed which here rises at an angle steeper than that of the talus, and crops out a few hundred feet up the slope.

It would appear from the relative weathering of the artefacts that there is no definite reason to regard these deposits as successive occupational layers, they would seem rather to mark successive falls of rock higher up the mountain slope, forcing surface material to slide down and cover other surface material until the talus has accumulated, together with the contained implements, to a depth of 10 feet. Mr. Jansen regards these layers as occupational layers divided by considerable intermediate periods. In any attempt to date the finds, however, we must regard the period as dating from an age previous to or contemporaneous with the lowest deposit of the series. An examination of the implements in the lowest layer in their unweathered state forces one to the conclusion that the stones are the work of man, artefacts in the true sense, and that we are in the presence of a definable and recognisable industry of a specialised type. The methods employed in the manufacture become more apparent. At this level, as on the surface, the implements are of the types described above, and are mixed with coups-de-poing of a rough type, compatible with the fracturing qualities of the dolerite from which they are made, and obviously suffering in finish as a result of the intractable nature of the material employed.

One implement from this site is of supreme interest; it consists of a coup-de-poing in what would appear to be a grey quartzitic sandstone. It was originally heart-shaped, some 4 inches long, by 3 across, and extremely well made. The edges are straight, the workmanship fine, and the implement appeared at a depth of 8 feet from the surface in absolute association with the Victoria West types. The differences between the two types of coup-de-poing are extreme, and would probably be more so but for the fact that this finer specimen was damaged in the recovery. (The main drawback in the use of convict labour in excavating.) This shows immediately that either the Victoria West workers were capable of producing a finer implement, given the right material; or else that they were sufficiently friendly with the Stellenbosch industry groups to trade with them, and to appreciate the neater implements so obtained. One small fact is of interest, the heart-shape of this single implement is not the shape usually favoured by the Victoria West workers, nor for that matter has it any direct affinities with the shapes used by the Stellenbosch workers. It falls into a shape-class of its own. Mr. Jansen suggests that the isolated well-made coup-de-poing found here, and the isolated specimen to be referred to from Cofimvaba, both imply contemporaneity of the Stellenbosch and Victoria West implements, and a certain amount of exchange between the groups. (See Plate VIII, 3.)

Other Sites.

A little way past the dam enclosing the end of the valley, and along the Carnarvon road, the hills are split up, to the north, by peculiar valleys running up from the roadway. In most of these are to be found artefacts of the usual Victoria West types. This is of interest as the lake (now most inadequately represented by the dam) which at one time covered this basin must have entered the mouths of these various valleys, and it is possible that the makers of the implements date from a period when this lake was present, and that they lived beside the coves about its edge. After passing those valleys which show signs of the relatively recent presence of water, the implements cease until further sites at Zuurkop (Wolvefontein), Vingerfontein, and Melton Wold are reached.

The site at Zuurkop (Wolvefontein farm) lies on the slope of the conical hill bearing that name. The road cuts through the site, which extends to the edge of the bed of an intermittent river in the valley below. The whole site might be some two miles long and half a mile across. The material used is the same fine-grained dolerite we get at all sites, and implements appear in great profusion over the surface of the ground. The commonest type from this site is the high-backed tortoise-shape. In spite of the fact that no excavation has been made at this site, it appears probable that the site is of purely surface nature, and that it does not descend to any depth.

Melton Wold, the next site, is still further from Victoria West in the same line. It is not so rich as the Zuurkop sites, but is similar in other ways. The road cuts through the surface site, which appears scattered.

A little further along the same road, and round the curve of the hill overlooking Melton Wold, access can be obtained to the site on Vingerfontein. This site lies perhaps a mile eastward from the road almost directly below the mountain. The material is again similar, and bouchers are once more to be associated with the Victoria West types, which do not vary at all considerably from those to be found at other sites.

Mr. Jansen also discovered a site at Loxton, 52 miles west of Victoria West. The implements from here are the same in type and form as the other Victoria West implements, but heavily weathered. The position is very similar to the Victoria West sites, a narrow poort with a plain beyond.

Further Distribution.

Quite apart from these sites in the more immediate neighbourhood of Victoria West, it has now been found that this type of implement has a very much wider distribution than was at first supposed. The writer has found specimens at Nakob (Southwest Africa) and at Cofimvaba (Transkei), which so far appear to be the western and eastern extremes of distribution. Mr. C. van Riet Lowe has also found a specimen at De Put, Edenburg district, and another at Spitzkop, in the Fauresmith district, O.F.S., while Mr. M.C. Burkitt, on his visit to the Union in 1927, discovered a specimen at the Half-way House on the Kimberley to Barkly West Road.

Mr. C.H. Heese spent a short time at Britstown (some 50 miles north of Victoria West) after his visit to the Victoria West sites, and here made an extremely interesting discovery. Writing apropos of his finds here, Mr. Heese says:45

Concerning other sites also visited, Mr. Heese says:

It is worth noting that the Britstown lydianite specimens are the only implements of Victoria West type so far known in any material except dolerite. They are of extreme interest as they are slightly different from the normal shapes.

Manufacture.

To understand the following remarks on the manufacture of the Victoria West implements it is necessary to state that this paper is the direct outcome of the wish of the South African Association for the Advancement of Science, Section E, that a small commission should visit Victoria West with the intention of clearing up two points: first, whether the objects here discussed were capable of being regarded as of human origin; and, secondly, whether the objects collected were core-implements, and hence of interest in themselves, or conversely, whether they were prepared cores from which a needed flake had been struck. As a result, the various sites were visited by Mr. F. Jansen, the discoverer of the implements, Mr. C.H. Heese of Riversdale, and the writer [A.J.H. Goodwin].

In spite of the acknowledged fact that all three of us were probably biased in our preconceived ideas on the first point, in favour of accepting the objects as of definitely human origin; yet I feel that we have made out a reasonably strong case in favour of their acceptance as artefacts, a verdict which seems to have been supported by the Nakob, De Put, and Cofimvaba finds, and perhaps also by Dr. van Hoepen’s discoveries at Bloemhof.47

The second point, however, proved the centre of heated arguments, and has by no means been cleared up yet. In an argument of this type, it must be remembered that absolute proof is not easily obtainable, probability and circumstantial evidence play a subtile but prejudicial part in any attempt at demonstration. Exactly the same state of affairs is to be seen in the European discussion of Eoliths; it was proved conclusively that stones identical with Eolithic types could be made with the aid of concrete-mixers, but this did not necessarily show that Eoliths were not the product of a human or humanoid being. Points of this kind cannot be regarded as proven, an open mind must be kept on the subject, and one may not with any fairness “believe”; one can only “incline to the view,” until further evidence is forthcoming. However, for the sake of euphony, the word “believe” will often stand for the less definite phrase in order to avoid repetition in the following argument.

The question under discussion is as to whether these high-backed artefacts showing a core technique were the desired implements or whether the core was merely a well-prepared nucleus, shaped to produce a flake of a definite type. In other words, is the core a variety of the coup-de-poing; or is it a “tortoise-core” similar in type to those discovered at Northfleet (England)?48

Mr. Jansen inclines to the view that the large core would be the desired implement, and the flake thrown away; he bases this view largely upon the method he believes to have been used in the manufacture of the implement. He infers that the main flake (covering about two-thirds or more of the under side of the finished implement) was struck off first from a “raw” block of dolerite. The next step would be the removal of flakes all round, striking the blows about the edge of the main flake-scar, in fact using the sudden upward turn at the release of the flake as a striking platform. From these flake-scars were next struck a series of flakes at right angles to them. The result would be, according to Mr. Jansen, the high-backed upper face of the implement.

This method of manufacture does not appear possible in view of our knowledge of the fracture of stone, nor does it agree with the evidence of other observers. Dr. van Hoepen, Professor Radcliffe Brown, myself [A.J.H Goodwin], and others have noted that wherever the weathering did not interfere with our inspection of the small flake-scars bounding the main flake-scar, they were always struck off from the outside edge (i.e. not from the edge of the flake-scar). Dr. van Hoepen says:49

There is no difference in type between the smallest material from Bloemhof and the largest from Victoria West. From the Bloemhof material it is obvious that the stones were prepared in the first place in order that a great wide flake might be struck off. It thence follows that the large stones from Victoria West were also shaped with this intention. In the greater number of stones from Bloemhof, it is obvious that the main flake was struck off last, and this after the smaller flakes had been removed. In the Victoria West stones it would appear that the case is the same.

On the other hand, Mr. Heese says:50

If Mr. Goodwin has “never found that the flakes trimming the core were struck off from the edge left by the removal of the main flake,” some of my finds show such instances. Mr. Goodwin’s theory of the manufacture seems to hinge on this matter, as also his idea as to the whole purpose of the core — the “final flake.”

My own view is that the implement was made originally in the shape of a wide coup-de-poing, one face high-backed, the other less markedly convex. Almost the whole of this flatter face was removed by a single blow, using a point about an inch or less away from the bounding edge on the more pyramidal face as the point of percussion. The resultant core would be exactly of the type we now have, and the flake would be of a shape and size which we can generally presume was useful.

Mr. Jansen objects:

I don’t remember finding a single boucher as wide as the cores. On the contrary they were all rather flat, and narrow compared with their length.

Actually the South African Museum has two specimens collected by me at Victoria West, which are identical, except for the removal of the final flake, which occurs in one only. The strongest argument against the “desired flake” view is that in many of the horseshoe types, after the removal of the main flake, a second flake appears to have been struck of from within the first flake-scar. This most certainly does not fit in with my view expressed above, and I do not attempt to explain it, yet in spite of this second flake, the “split” or flaked coup-de-poing theory appears to me to be by far the more rational explanation.

The importance of agreement on the method of manufacture is obvious. If the main flake were removed first, then it would be untrimmed, and the core would thus be proved to be the implement wanted. If, on the other hand, the main flake were proved to have been struck off after the trimming of the core we would be left with three alternatives: either the makers needed the core, or the flake, or both.

It has been argued that Dr. van Hoepen’s specimens show a core some two inches long, from which a flake, perhaps an inch in diameter, had been removed, and that this “proves” that the core was needed, such a small flake being useless to the maker. But, conversely, we may argue that many cores were of such vast proportions that they were useless except to a race of giants, while even the flakes removed would prove cumbersome.

In attempting to judge from the large flakes removed, the difficulty lies in the fact that it is impossible to produce any authentic flakes. All the flakes which might have been produced by this means are too worn and weathered to prove anything, as Mr. Heese points out:

Dolerite, with its coarse grain, disintegrates at so fast a rate that small or thin implements soon become unrecognisable as such, if they do not break up altogether.

We are thus baulked at every turn, mainly by the degree of weathering present on the implements observed. If a sufficient number of unweathered specimens were produced from depth, a far saner view of the whole industry would result. For the present it is necessary both to reserve final judgments and to regard the cores as possible true implements, while still searching for worked flakes likely to have been struck Off and used as tools.

Chronological Sequence.

The finding of the heart-shaped coup-de-poing in the series of deposits on the south-western side of the Victoria West valley in complete association with the Victoria West implements does not tell us very much as to relative chronology. The heart-shape does not yet imply any one particular phase of the Stellenbosch Industry, and thus does not give us any conception as to whether the local industry is an early or a late offshoot. We do not know what the exact relationship existing between the two actually is.

The limy deposit in which this association was proved continues without a break along the lower part of this south-western hillside towards the dam. Some two or three hundred yards along it skirts the edge of Moonlight Kop. Here the deposit is exposed by the road-cutting, and is seen to be upwards of six feet in depth. Over the deposit talus has accumulated to a depth of one or two feet. We can safely presume from the continuity of the deposit, and the fact that throughout the distance between the western hillside site and the Moonlight Kop sporadic implements of Victoria West type occur in this deposit, that the whole layer is of a single age, and is all contemporaneous with the Victoria West implements. Lying directly on the talus material which covers this extension of the Victoria West deposit the writer discovered a small but very rich site of the Smithfield “C” phase of the Later Stone Age. This site proves that a very considerable period had elapsed between the Victoria West and Smithfield “C” periods. (See Plate IX, 1.)

Some years ago Rev. Perold of Victoria West submitted a number of Middle Stone Age implements to the South African Museum. These came from the surface near the golf-course site, and the circumstances of the discovery point to their being of a date somewhat later than that of the Victoria West implements.

These facts justify us in regarding the Victoria West Industry as of the Earlier Stone Age and partly or entirely contemporary with the Stellenbosch Industry.

The Fauresmith Industry.

The Fauresmith is an Industry closely allied to the Stellenbosch. It is difficult as yet to say whether it is an evolved or specialised branch of the Stellenbosch, due partly or entirely to the presence of a useful material as Lowe is inclined to believe, or whether we are here presented for the first time with an infiltration of that racial or cultural impetus which, coming into South Africa, was to give us the industries grouped together as the Middle Stone Age. It is sufficient to say that the differences between this industry and the Stellenbosch have been regarded by both the writers of this paper, and by many others, as sufficiently deinite to necessitate the invention of a new category.

Association.

As in the Stellenbosch industry, so in the Fauresmith, the coup-de-poing is the most typical implement (see Plates XI and XII). The workmanship is fine, which may be due to a superior standard having been desired and attained, or purely to the fact that the indurated shale (lydianite) invariably used is the most tractable and amenable material of any used in South Africa. The coup-de-poing are of a type which Mr. M.C. Burkitt, while visiting the Union, stated to be very similar to those of the La Micoque period or variation in Europe. The shape is generally a neat almond, sometimes an ovate; the limande is rare, and the triangular coup-de-poing exceedingly scarce; the pear-shaped variety, so far as is known, does not appear at all. The size is generally small, and the implements are of a length and weight which make them eminently suitable for use in the hand (Plate X).

The following table gives a comparison between the average length, breadth, and thickness of implements in millimetres from twelve Stellenbosch sites, the largest implement and the smallest implement of the group. The same is done for six Fauresmith sites.

	Stellenbosch			Fauresmith
	Length	Breadth	Thickness	Length	Breadth	Thickness
Greatest individual specimen:	260	102	40	164	87	47
Highest single-site average:	230	102	38	126	73	35
Average of all specimens:	170	88	40	108	64	28
Lowest single-site average:	150	76	35	100	59	27
Smallest individual specimen:	101	66	30	74	53	21
Average ratio:	4·25 : 2·2 : 1			4 : 2·3 : 1

It will at once be seen that the greatest individual Fauresmith specimen falls below the average Stellenbosch, while some Fauresmith sites show an average lower than the smallest individual Stellenbosch specimen. The sites were chosen quite at random, and further examples would show much the same result. Thirty advanced Fauresmith coup-de-poing were measured, and fifty advanced Stellenbosch implements. These include the implements measured by Lowe.51

Associated with these coup-de-poing are discs, scrapers (text-figs. 1, 2, 3, 4), slightly trimmed fiake points, and occasionally biseaux. Faceted hammer stones or polyhedral stones, three or four inches in diameter, are also found. The biseaux associated are ill made and of coarser material, most commonly dolerite, but deeply patinated and almost unrecognisable. This fact forces Goodwin to the conclusion that the association is a chance one, and that either they mark the presence of Stellenbosch implements which have slipped in amongst the Fauresmith, or that they are the result of trials of new material. The fact remains that these biseaux show execrable workmanship, bad choice of material, and a complete lack of finish, and yet they are nominally associated with well-made coups-de-poing of eminently suitable material, whence it may appear after further investigation that the Fauresmith lndustry lacks the biseau type completely. On the other hand, Lowe considers the biseau to be merely aberrant.

Technique.

Before enlarging upon the methods employed in the manufacture of the coup-de-poing, it would perhaps be best to speak of the flake technique as it is found in the smaller implements of this industry.

The most typical element in the Mousterian technique of North Africa and of Europe is the presence of a flake industry in which the flakes are trimmed on one face only and show a faceted butt. In this technique either the striking platform has been trimmed by the removal of various small flakes before the needed flake was removed, or else the necessary flake has had all signs of the striking platform removed by chipping over that end. This peculiar method of butting is very typical of most Fauresmith flakes, and would thus appear to prove the presence of what we may call (for want of a more exact term) Mousterian influence. This term is thus meant to imply the presence of a culturally acquired knowledge of the Mousterian technique, and does not in any way presume the presence of Neanderthal man in South Africa.

The Mousterians also show a lack of skill in completely mastering the fluting technique, or longitudinal trimming, which is so typical of the later Neo-anthropic industries — this failure is again present in the Fauresmith Industry.

The flakes do not show much sign of any conventionalised tools unless the side-, end-, and hooked-scrapers are to be regarded as such. The end-scrapers are coarser than those we will come to later in the Smithfield industry, but they fall quite within the limits of the Smithfield types, and might perhaps have been classed as such if picked up unpatinated, unassociated, and singly. The side-scrapers suggest the concavo-convex of “Smithfield A” (Later Stone Age), and Lowe has found three particularly well-made specimens (text-figs. 2, 3, 4). The hooked-scrapers are very typical, not appearing in any other South African industries. A few other scrapers are known and one or two worked points also appear. Considering the number of well-made flakes and also the regular presence of true flake cores, the number of flake tools is disappointing. Both the coup-de-poing and the flake tools lie side by side at depth in various deposits and show identical patination and oxidisation. The patina is interesting; Lowe states that the commoner dark grey or blackish lydianite turns to a reddish-brown or terra cotta when exposed to the elements away from the action of water. This discoloration is due to the oxidation of the iron content, and depends largely, if not entirely, upon the amount of iron impregnation and the length of exposure. The incrustation may even amount to one-eighth of an inch in depth. Water — hydration and oxidation combined — turns the stone blue-grey, and the material tends to be water-Worn under the action of running water.

The Coup-de-poing.

It has been stated in an earlier paper52 that the type of Stellenbosch implement peculiar to the two lower Vaal River gravels shows a similarity to Fauresmith types. This is obvious when a large study collection is available, and Lowe regards this as being a probable key to the origin of Fauresmith Industry.

It was shown that the prevalent method of manufacture along the lower Vaal terraces consisted of the utilisation of a laterally removed flake as the block or core from which the coup-de-poing was made. In these gravels was also found a possible association, never yet proved with any certainty, of coups-de-poing of a fine type and flake implements. It is quite within reason to suppose that a flake technique was already present in an embryonic form in the Stellenbosch industry on the Vaal, on the Suikerboschrand, and at Knysna where the presence of a single flake core has raised a multitude of questions and surmises. The discovery of worked flakes at Blaauwklip-spruit near Stellenbosch itself, in the gravel terraces of the Suikerboschrand, near Heidelberg (Transvaal), and in similar terraces of the Klip River, near Henley, is also significant. This point, it will be seen, is of extreme importance to us in any attempt we may make to obtain a sane view of the Fauresmith Industry. The point which still remains to be proved is the presence of a true flake technique in the Stellenbosch. Péringuey long ago recognised the fact that chance spalls were trimmed for use by the Stellenbosch workers. It is also possible that much of the material we have hitherto regarded as of Stellenbosch type from the Free State belongs either to a transition phase or to the Fauresmith proper.

The Stellenbosch and Fauresmith coups-de-poing do show definite differences. One of these would appear to be fairly basic, the utilisation of a longitudinal flake as the material from which the Fauresmith coup-de-poing was made. This technique is quite obviously allied to the flake technique employed in the making of the smaller flake implements. As a result of this method the butt end of the coup-de-poing often shows a part of the striking platform left as a small eye-shaped flattening at this end. In many instances this “eye” has been dexterously removed by a single well-placed blow, leaving a peculiar twist very like an eye-lid and giving a very typical distortion at this point. In many unfinished specimens, the whole or a large part of the cleavage plane is visible, taking up the greater part of that face, the outer face being worked all over in all instances (Plate XII).

The process of manufacture thus consisted of two stages, the striking of a large and suitable flake from a core, and the subsequent trimming of this to a coup-de-poing shape. The large flake thus becomes a core in a coup-de-poing technique, which is peculiar in another way also. In one of the normal methods employed by the Stellenbosch makers the block was roughly hacked to shape, then reduced, the working becoming more and more delicate as the desired size was reached. This method was never used by the makers of the Fauresmith implement, but apparently the whole of the outer face was first worked, and the cleavage face worked later if the necessity arose and the shape proved suitable. The Stellenbosch makers also had a dehnite preference for water-worn boulders where these were obtainable, but the workers in lydianite preferred rock fragments of local origin.

The coups-de-poing from the Fauresmith sites also have a straight edge, differing from the zig-zag so often apparent in the Stellenbosch industry, and very often too the Fauresmith coup-de-poing shows a decided S-shaped twist or screw, reminding one at once of the typical twist of the European Acheulean. This is extremely rare or lacking in Stellenbosch implements (Plate XI).

The shape is often very typical. The Stellenbosch coup-de-poing is generally a true almond when viewed in plan, and if marked across its greatest width, the length of this line would only slightly exceed the distance between itself and the extreme butt of the implement, giving a shape much that of a slightly side-iiattened semicircle. If a shorter, stockier Fauresmith specimen were so marked it would show in most instances an almost exact half-circle.

So far as we now know, the Fauresmith Industry is confined approximately to the southern and south-western part of the Orange Free State and to the neighbouring districts of Kimberley and a portion of Herbert. It is typical of the Beaufort and Ecca shale areas of the Karroo system. These shales are intensely cut up and metamorphosed by plutonic intrusions (chiefly dolerite), and the result of the metamorphosis is the indurated shale or lydianite, as typical of this industry as it is of the later Smithield industries. It is a flinty material with a clean, dependable fracture leaving a sharp edge.

History

The history of our knowledge of this industry is interesting. Some years ago Mr. Leviseur of Bloemfontein submitted a number of implements to Dr. Péringuey of the South African Museum. These implements (S.A.M. 3240) were recognised by Goodwin as of interest in 1923 and shown to Professor Radcliffe Brown, then head of the Department of Social Anthropology at Cape Town University. It was agreed that the specimens showed a great variation from the Stellenbosch type and necessitated a more extended nomenclature. The points of greatest interest were the twisted coup-de-poing, the presence of the longitudinal flake-scar on one face of some specimens, the fine workmanship, the flake implements, and a worked- out core.

At apparently very much the same time Lowe stumbled on this industry himself, and his subsequent researches are therefore of extreme interest. Writing to Goodwin in 1923, he says:

Writing in July 1926, he says further:

This last point, the inter-relation of the Stellenbosch, Fauresmith, and Stilbaai types, will be spoken of in a later paper dealing with the Middle Stone Age periods, and more especially with the Glen Grey industry.

Writing in September of the same year (1926), Lowe says:

It will interest you to know that I have in two different localities found well-marked stratification. In the Kromellenboog spruit, and in the Fauresmith town spruit on the town commonage, we have states of affairs which are remarkably similar. Smithfield types are found on the natural ground surface, while below and separated by seven feet or so of sterile strata, we find in a 3-foot gravel bed (water-borne) implements of Fauresmith type.

Lowe was thus able to prove conclusively that the Fauresmith Industry was considerably earlier than the Smithfield “B” phase of the Later Stone Age. Hitherto the only proof we had of this fact was that Fauresmith implements are invariably heavily patinated, while Smithfield implements are lightly patinated or entirely free from oxidation.

Distribution.

Lowe gives the following factory sites in the O.F.S., the names of district and farm being given in each instance:

• Fauresmith district.
  • Blaauwheuvel
  • Brakfontein (Type)
  • Dwarsvlei
  • Erfdeel
  • Fauresmith Town Lands
  • Koffiefontein
  • Leeuwarden
  • Petrusberg
  • Rorich’s Hoop
  • Rooidraai
  • Spitzkop I
  • Spitzkop II
  • Valschfontein
  • Zuurfontein
• Bloemfontein district.
  • Honing Kopje
• Boshof district.
  • Damplaats
  • Elandsput
  • Meerlandsvlei
  • Schaapfontein
• Edenburg district.
  • Springfontein
• Luckhoff district.
  • Luckhoff
• Philippolis district.
  • Onder Dwars Rivier

Besides these a further site is represented in the South African Museum, from Fauresmith, and another in the Herbert district, while Kimberley shows several specimens, one from the site of the present library, and housed in the McGregor Museum. A single coup-de-poing from Brandfort has also been given to the Department of Social Anthropology at Cape Town University by Dr. Schonken of Klerksdorp.

Sites.

Brakfonteim (No. 231), District Fauresmith.

This site is situated on the right of the road from Koffiefontein to Fauresmith, and about twelve miles from the former.

In a small re-entrant into what appears to be an old consolidated sand-dune (now overgrown) immediately adjacent to the road, implements were first found by Lowe in 1926. Closer examination revealed the fact that the dune had encroached over the original site and completely buried it until erosion set in. At first only a few implements could be recovered, but after each rain and as opportunity offered, Lowe revisited the site and found more and more specimens being exposed. So that in course of time, the finest and most representative collection from any single Fauresmith site was made.

Save for a few biseau types in dolerite, the material used throughout was lydianite or indurated shale. Several extremely neat and “finished” coups-de-poing were found, a few side-scrapers (one of which is concave-convex), end-, and hooked-scrapers.

On account of the long burial and the consequent protection from the elements, the implements are in a rare and excellent state of preservation; they have weathered to various degrees, from a khaki to a rich light chocolate colour. Owing to this state of preservation and the richness of the site, Lowe has constituted this as the main type-site of the industry.

Onder Dwars Rivier

On the slopes of a kopje immediately adjacent to the Philippolis Road Station, and opposite the goods-yard, is an excellent and instructive Fauresmith site. Coups-de-poing and crude scrapers were first found here by Lowe in 1923. All the implements are on the surface, and the exceedingly heavy incrustation on the specimens suggests very long exposure. In cases the weathered surface consists of a terra-cotta coloured crust one-eighth of an inch in thickness; the angular edges appear rounded, and one entire implement looks as though it had been rolled under running water, which, however, is not the case, No biseaux were found, and the material used throughout is lydianite or indurated shale.

In the course of time, the makers of the Smithfield “B” implements came to the same locality and similarly established a factory site. They used Fauresmith coups-de-poing, etc., as fabricators, and they retrimmed Fauresmith flakes. These later implements show either very light patination or a complete lack of it, and where a Fauresmith flake has been retrimmed the body of the resultant implement is heavily patinated or incrusted, while the surfaces exposed by later secondary trimming are entirely or almost free from any signs of such oxidation.

Fauresmith.

Immediately above bed-rock, which varies from six to sixteen feet below the natural ground surface, the Fauresmith Town Spruit has, through and in the immediate vicinity of the village, exposed a bed of water-borne gravel that contains vast quantities of Fauresmith Industry remains. All the implements are somewhat water-worn, and the surfaces have weathered to a blue-grey colour, but wherever a fresh chip has been removed the original very dark grey to black of the unweathered lydianite can be seen.

Immediately south of the village and near the bridge, where the implementiferous gravels are about three feet in depth, and lie some seven feet below the natural ground level of the banks, the remains of a Smithfield “B” settlement appear on the bank. The separating stratum of earthy material is sterile and contains no remains of human handiwork.

Lockshoek.

On the farm Lockshoek, 14 miles north-east of Jagersfontein, and 21 miles east of Fauresmith, Lowe again found stratification, for here is an excellent Smithfield “B” surface site on the banks of the Kromelleboog Spruit, while beneath and separated from the surface by a seven-foot sterile stratum, occur water-borne gravels containing Fauresmith remains. Although these gravels have yielded many coups-de-poing and flakes — all slightly rolled, and of lydianite — no biseaux have yet been found. The indurated shale has weathered to a light blue-grey colour on the surfaces of the specimens, but has remained the usual dark-grey to black inside.

Chronology.

It is difficult to give any estimate as to what chronological relationship exists between the Stellenbosch and Fauresmith industries. From a purely evolutionary point of view, the Fauresmith is the “later” or more advanced of the two, but no evidence as to the actual sequence or contemporaneity of these two industries has yet appeared. We do know that the Fauresmith is very considerably earlier than either the Smithfield “B,” as has been shown, or the Wilton Industry, and the technique and type implements ally it sufficiently to the Stellenbosch to justify its inclusion in the Earlier Stone Age of South Africa. The fact that a new influence — called Mousterian — may be represented here for the first time would imply that the Fauresmith stands late in the Earlier Stone Age, and probably marks the arrival and first signs of the Mousterian influences which appear to have given us the Middle Stone Age.

Addendum.

One or two other sites are of interest while dealing with the Fauresmith Industry:

Cofimvaba.

The implements represented at the Albany and South African Museums from this site are of very considerable interest. The site implies an association rather similar to that occurring in the Fauresmith Industry throughout the Free State, though it lies about 130 miles south of the nearest Fauresmith Industry site. It shows the contiguity of coups-de-poing with a group of flake implements and trimmed flakes in the material sent to the Albany Museum, though the specimens in the South African Museum consist of flake implements only.

The administrative village of Cofimvaba lies in the Transkeian native territories on the Queenstown-Butterworth road. In 1925 Mr. C.W. Wilmot, then postmaster of Cofimvaba, presented a number of implements to the Albany Museum, Grahamstown, and to the South African Museum, Cape Town.54 These came from the river upon which the village is built, and from a stratum traceable for a distance of up to four miles along the river. The implements appear to have been covered or sealed in by a layer of river boulders and by thick deposits of sandy loam. Mr. P.H. Walker55 regards this layer as having been wind-deposited and to be composed of cave sandstone, implying a dry windy climate. Mr. Hewitt regards the deposits as consisting of a series of home or factory sites, though it appeared to Goodwin more probable that the stratum had been fed at some time from a factory site or series of workshops a little higher up, or bordering upon the stream. Neither of these points need worry us at all considerably here, the presence of the implements implies the presence of man; the subsequent accumulation of sandy material and boulders implies a considerable passage of time since the deposition of the implements.

Mr. Hewitt describes three groups of implements found, and these agree (except for the presence of the coups-de-poing) with the material sent to the South African Museum by Mr. Wilmot.

1. A number of long, slender, parallel-sided flakes (up to 5 inches in length), many quite untrimmed, some trimmed on the inner face, some on the outer face; triangular points with edge-trimming.
2. An end-scraper (one is also present in the S.A. Museum collection), possibly intrusive. A single large flattish scraper, oval in outline, 3¹/₄ in. by 2³/₄ in.
3. Coups-de-poing, not well made but “typical” and discoidal artefacts.

All these implements are of shale, more or less indurated.

Whether these Cofimvaba implements form a single series, or whether two groups have been mixed together, it is difficult to say. If two groups have been so mixed, then one belongs to the Stellenbosch Industry, and the other shows affinities with the Middle Stone Age material from elsewhere. If, on the other hand, they form a single group, then we may have a site representing the extreme south-eastern evolution of the Fauresmith.

Thabu Nchu.

Dr. van Hoepen of the National Museum, Bloemfontein,56 reports the presence of a well-made Levallois flake in his Museum from Thabu Nchu, some 45 miles east of Bloemfontein. The flake measures about 120 mm. × 90 mm. × 12 mm. (max.), and is worked over the whole of the outer face with primary chipping; no secondary or trimming flakes appear, and the under face or cleavage plane is quite unworked. The flake is apparently unassociated, and little is known of its history, but it is most certainly a Levallois flake. Where this stands in relation to the South African cultures it is impossible to say. In Europe the Levallois flake is being more and more regarded as a late Acheulean implement, and as more typical of this than of the Mousterian. Whether the culture represented by this flake should fall into our Earlier Stone Age group or the Middle Stone Age it is quite impossible to say. Dr. van Hoepen links it with the Pniel material, but it is diflicult to find proof for his reasoning, and further material with a complete history will be necessary before anything definite can be said.

The Middle Stone Age.

Originally when working on the prehistoric period in South Africa it was found necessary and convenient to divide the entire time-field into two main periods, known as the Earlier Stone Age and the Later Stone Age respectively. As a result of further research it has become more and more evident that these terms do not comprehensively cover our prehistoric period. In the first instance these two terms were regarded as representing the South African equivalents of the Mediterranean Lower Palaeolithic period in the first group, and all later industries and cultures in the other. The Earlier Stone Age thus originally comprised the Stellenbosch, the Fauresmith, and the Victoria West Industries, the internal time-relationship between these three being entirely unknown. The Later Stone Age was regarded as embracing the Still Bay Industry (originally included under the term “Eastern Culture”) and two interrelated industries, the Smithfield and the Wilton. Within the Earlier Stone Age were thus grouped all coup-de-poing industries, while in the Later Stone Age appeared a variety of flake industries, though the division was not made on any such arbitrary basis.

With the accumulation of material directly resulting from the more intensive study of the archaeological field, it was forced upon our notice that we were dealing, in South Africa, with a series, not of two, but of three main invasions, either of a migratory or of a purely cultural type. It was therefore found necessary to redivide the Later Stone Age and to form a third group. The cleavage is perfectly natural. In the Later Stone Age were originally included two groups. The one consisted of the old “Eastern Culture,” comprising the present Still Bay Industry and a variety of more or less allied material. These industries cannot be regarded as in any way typically Neo-anthropic; they can now be shown to be of distant date, and the Still Bay has proved to be associable with a physical type peculiar to itself. The other group consisted of the Smithfield-Wilton complex, composed of two main industries, typically Neo-anthropic in character and in technique, both associable with a physically modern type, both to be regarded as extending well into modern times, and to both of which can be attributed artistic feeling. In short we can now definitely assign this Smithield-Wilton group to the physical race known rather vaguely as the “San” or Bush-Hottentot peoples, while on the other hand we can assign the Still Bay Industry to the physical type described by Dart as the “Boskopoid Race.”57

Over and above the obvious and even basic differences between the modes of technique employed by the makers of the Still Bay implements and those preferred by the Smithfield-Wilton groups, additional material which has subsequently accumulated has had to be accounted for. From time to time sites have been found revealing implements and facies bearing a distinct family resemblance to the Still Bay material, but by no means includable under that industrial term. As an instance of this, the writer discovered the Glen Grey Falls site in December 1926, and it was immediately apparent that this site showed strong affinities to the Still Bay material, but could not fairly be included as belonging to the same industry. The matter was left in abeyance for the time, and Mr. M.C. Burkitt was invited to visit the site in August 1927. He immediately pointed out the necessity for a complete revision of the group which had previously been called the “Eastern Culture,” and as a result the term Still Bay Industry is now to be regarded as applicable only to an industry which is truly represented by the name-site, and which is, so far as our present knowledge goes, strictly confined to our southernmost littoral. The term “Eastern Culture” has long been discarded as unscientiiic and misleading; its place is taken by the term Middle Stone Age.

This revision of the terminology applied to South African prehistory brings out more clearly the fact that we are dealing with at least three main invasions. If we were to fall back on the terminology employed round the Mediterranean Basin we might term these (a) Lower Palaeolithic, (b) Middle Palaeolithic or Mousterian, and (c) Neo-anthropic. These terms have unluckily become more and more wide in their connotations, and have come to be applied to concepts and to groups of ideas whose associations are only true in their entirety in Europe. Racial, cultural, and temporal concepts have merged through continued associations, so that it is now possible to speak in a loose way of Lower Palaeolithic Times, Mousterian Man, and Neo-anthropic Industries. This bastardisation of European terminology has rendered it relatively useless in this or in any other extra-European country, but in spite of the imperfections of this medium, we can still with perfect clarity draw a broad and purely cultural parallel between these three main European periods and the three main periods of South African prehistory, although closer and more detailed collation is at present unsafe.

As has been already pointed out, the Fauresmith Industry of the Earlier Stone Age would seem to mark the first appearance of “Middle Palaeolithic” elements in the subcontinent. We are here faced for the first time with the appearance of a conventionalised flake technique in addition to the more normal coup-de-poing forms so typical of the Earlier Stone Age. The flake type represented shows a series of flake-implements, frequently with faceted butts, and showing a tendency to convergent rather than parallel flaking on the outer face of the flake. This faceting of the butt in a certain proportion of the flakes may well be accounted for by presuming that the flake technique here employed was the result of an evolution from a core technique and the obvious effect of using flakes or spalls struck from the coup-de-poing as primarily unconventionalised tools. The presence of convergent flaking might similarly be accounted for, as this would appear to be entirely due to the angle at which the blows removing the flakes were struck. The sudden appearance of conventionalised flake tools is not so easily explicable. The elements described above may more safely be regarded as intrusive, as they are to become typical of the Middle Stone Age industries. We may, however, infer one of two things, either that the Middle Stone Age is a direct evolution from the Fauresrnith Industry: or that certain elements which were appearing in the south of Africa mixed with a basic industry already present to produce the Fauresmith, and that these elements increased in intensity and strength, through the further infiltration of purer stock, to lay the foundations of the Middle Stone Age. The similarity between the Middle Stone Age Industries as a whole and the Mousterian of North Africa, etc., has already been pointed out by Mr. Burkitt on his recent visit. Very strong affinities are seen to exist, most notably the faceting of the butts of the flake tools, and the tendency to convergent flaking. Mr. Reid Moir has pointed out58 that this latter tendency is typical of the Mousterian, so much so that it must be taken into account in differentiating between that phase and the two chronologically adjacent periods. The Rev. Neville Jones has long recognised a “Mousterian” element in Rhodesia.59 It therefore seems more than probable that we owe both the flake implements in the Fauresmith Industry and the basis of the Middle Stone Age to a “Mousterian” influence or infiltration, not necessarily from Europe, but certainly from the north.

The elements which are peculiar to the Middle Stone Age and to which reference has been made above are worthy of further study, and may best be dealt with here. We have already seen that the Earlier Stone Age was typified by a core technique, that is, by the shaping of a core to the maker’s needs. This technique has already been sufficiently described in the Stellenbosch paper, and nothing more need be said of it here. We will find that the Later Stone Age is typified by a pure flake technique, consisting of the preparation of a core, and the subsequent removal of a flake which is finally trimmed for use. The preparation of the core may be divided into two stages, the making of a flat percussion or striking platform, and the removal of a series of longitudinal and parallel trimming flakes struck from about the edge of the striking platform, and running down one face of the stone to flHute it at right angles to the platform. The final flake struck off is thus composed on its inner face of the positive cleavage-scar, while the outer face is made up of a series of parallel flutings or negative cleavage-scars: all of these scars show their origins to have been at the striking platform.

The makers of the Middle Stone Age implements employed a technique in some ways intermediate between the Earlier and Later Stone Age methods. The artefacts are all of a flake type, and no true core-tools are discernible. Two basic differences are noticeable between the technique here employed and that to be used in the Later Stone Age. In the Middle Stone Age the striking platform is not flat, but is distinctly faceted. The trimming flakes are not parallel (as they are in the Later Stone Age), but tend to be convergent; as a result of this preparation the final flake removed is eminently suitable for use as a point, and, indeed, the typical implement throughout the Middle Stone Age Industries is the worked point in a variety of forms.

This difference between the parallel and convergent flaking is due to the angle of incidence of the blow removing the flakes; if the blow is struck at right angles to the percussion platform the flakes removed will be parallel, whereas if the blow is struck at a “wide angle” the flakes will tend to converge.60 As a result the main body of the flake lies at right angles to the fragment of flat striking platform left on the spall in the Later Stone Age flakes, and the shape will tend to be that of a rectangle. In the Middle Stone Age, on the other hand, the flakes tend to be triangular, while the strongly faceted butt or fragment of the striking platform is distinctly visible from the cleavage face, and therefore lies at an obtuse angle to the cleavage face, and at an acute angle to the outer face which bears the signs of convergent flaking (see text-fig. 1).

In this preliminary statement it is intended to give the reader a fairly clear conception of the general types of implement which are here grouped together. It reveals the purely technical basis of the differentiation, while chronological and other grounds for the division will be dealt with under the headings of the various individual sites and industrial groups, and will be finally summed up in the conclusion.

Our first difficulty in the study of the Middle Stone Age lies in the fact that we still have insufficient material, while much of the material which we have at our disposal has been sent in without data. As a result we are at present unable to divide this main period into industries with a sufficient degree of certainty. It is, therefore, necessary to utilise two terms, the word Industry being employed where a group is certain and definable; but in cases where uncertainty may still exist either in our knowledge of the exact relationship of the group to the other groups and industries, or where insufficient data have been collected to form a basis of a definite classification, the term Variation will be employed. This does not, of course, preclude our raising a “Variation” to “Industry” status should evidence of a sufficiently definite type accumulate to justify such action.

This difficulty is immediately encountered in any attempt to deal with the following four groups, which I have termed:

• The Glen Grey Falls Industry.
• The Pietersburg Variation.
• The Still Bay Industry.
• The Howieson’s Poort Variation.

If these four groups alone comprised the Middle Stone Age our task would be relatively simple, as they form what at first sight would appear to be a direct evolutionary series, linking up with the Fauresmith Industry in the first place, and finally evolving into the Later Stone Age Industries. Unhappily, a number of other variations occur which must be accounted for; perhaps these are sports and offshoots thrown off from the main evolution suggested by these four. Such a view would be easy, save for the fact that other factors have to be taken into consideration. First amongst these is that we do not know that this hypothetical evolutionary series necessarily implies a time-sequence. Investigation along these lines would prove useful, and would most certainly be suggestive. The “dominant” would appear to be the Mousterian influence, appearing with the Fauresmith and ousting the palaeo-anthropic coup-de-poing, or at least reducing it in size to the lance-head. Final contact with Neo-anthropic industries would seem to have forced the appearance first of the Still Bay Industry, and eventually the more obviously mixed HoWieson’s Poort Variation.

We have already seen that the Fauresmith Industry marks the appearance of a probable invasion, and may thus have to be regarded as a mixed industry. The similarity between this and the Glen Grey Industry is evident, though most, if not all, of the Earlier Stone Age characteristics have gone. The Pietersburg Variation is directly comparable with the Glen Grey Industry, but there is a very considerable fining of technique and a greater symmetry of implement. From this it seems but a step to the Still Bay Industry, in which a technique and a beauty of form very comparable to the Lower Solutrean of Europe make their appearance. Mr. Burkitt suspects that he can see the beginnings of a Neo-anthropic invasion in the forms of the Still Bay, while in the Howieson’s Poort Variation this influence becomes more certain and dominating, so that this last is to be regarded as a mixed group, Still Bay and Neo-anthropic elements appearing side by side and in the same deposits, thus throwing the road open for the Later Stone Age. The Howieson’s Poort Variation thus forms as strong a link between the Middle and Later Stone Ages as the Fauresmith did between the Earlier and Middle.

Those variations which fall outside this series will have to be dealt with separately when these four units have been disposed of.

The Glen Grey Industry.

Some fourteen miles east-north-east from Queenstown is Driver’s Drift, a ford and bridge by which the Queenstown-Lady Frere Road crosses the White Kei River. Some three miles up this river, which here flows south, an interesting formation is met with. The river falls suddenly into a gorge, perhaps 150 feet in depth; the sides are precipitous and remain so for about a mile as they follow the river; below this the dolerite gives way to alluvium and shales, and the gorge automatically widens out into down-country.61 At this opening of the gorge a dolerite-capped kopje guards it on its eastern side; on the talus forming the foot of this hill and reaching as far down as the high-water mark of the White Kei is an implement site of surprising richness if we consider the area which it covers. The material seems to lie on, and immediately beneath, the present surface, and shows no signs of having been rolled by water-action; both this and the compactness of the site point to comparative modernity, archaeologically speaking.

On a farm, Rockwoods, in the great Bongolo Basin, some four miles from Queenstown and a mile north of the Queenstown-Lady Frere Road, a kloof cuts down from the high mountain lands towards the homestead which stands in the basin below. The eastern krantz or cliff-face has fallen somewhat, and a rough cart-road has been built up the kloof on the talus. The slight cutting necessary to build the road reveals an extensive stone implement site in the talus and above the fertile land which borders the stream below. A short search reveals the similarity between this material and that found at Glen Grey Falls, and shows that here it forms part of the mountain talus directly upon which lies Later Stone Age material.

Other sites in the Queenstown district point to the presence of this industry over the whole of this area, but the evidence from these sites is not yet suzficiently definite to merit quotation.

Dr. Lebzelter of Austria, when visiting South Africa, discovered a quantity of very similar material at Keilands Mission Station in the Stutterheim district, some fifty miles south-east of Queenstown. Here a donga or erosion gully cuts through the yellow ochreous clay in the fields below the mission station, and reveals a site about a foot below the present surface of the Kei River alluvium. At a little distance from this site Dr. Lebzelter found Smithfield “B” material lying directly upon the surface of the alluvium. The site is neither sufficiently extensive nor typical to give much conclusive evidence, but the material is definite enough to point towards Glen Grey affinities.

Some years ago an interesting collection came into the hands of the South African Museum; these implements constituted the life work of Mr. Alfred Brown, a recluse living at Aliwal North in the north-eastern province of the Cape. Mr. Brown had spent a very considerable part of his life forming a private museum in his home. He kept copious notes, conscientiously recorded on a “double entry” system, one in diary form, the other under a variety of headings, archaeology, palaeontology, mineralogy, etc. The stone implements themselves were kept in separate drawers housed in a series of boxes: each box had a number, while each tray or drawer was similarly numbered. At his death, for purposes of packing, it was thought fit to combine the contents of the several boxes together, with the result that the material has become inextricably mixed, and almost the whole of his life’s work has been lost, as his entries refer almost entirely to the box and tray numbers. A little has been retrieved, and as a result a certain amount is known. If the individual implements had been marked by him, or if more care had been taken after his death to keep the various implements in their proper boxes, a most extensive survey of a single archaeological area would have been opened to us.

The implements consist almost entirely_ of material showing very strong affinities to the Glen Grey facies, and of Wilton types. Only in one instance can any sequence between these two industries be reconstructed from his notes and implements. In a yellow clay deposit west of Kriedfontein sluit and north of a small dyke through which flowed an evil-smelling stream, Mr. Brown discovered a hoard of Wilton implements. These seem to have been the material from a workshop site, but the maker, apparently to conceal all evidence of his presence, had collected the flakes into a heap and had placed large stones on either side, with a hammer stone on top. Subsequent denudation had removed the soil which had apparently been placed over this heap, and disclosed the oblong hammer-stone. Some 200 flakes were found embedded in the clay which had washed into the hiding place; these were of Wilton type and perfectly clean and unpatinated. About this site a number of other flakes and implements had also been revealed by the removal of the overlying soil, showing that the cache had been erected over or buried into an old site. Mr. Brown collected some sixteen flakes and implements from a radius of a few yards. These last have mostly been displaced in his collection, but three are known with certainty: these consist of lance-head types worked over the whole of the outer face, and are similar to Glen Grey elements.

Similar implements, all assignable to the Glen Grey Industry, were found by him on sites at Middleplaats, Melkspruit, the racecourse, and on the boundary of Grassridge farm, all in the Aliwal North district.

During May 1928 Mr. H.P. Thomasset of Weenen, Natal, sent me a considerable number of implements, some of which were returned to him and some of which he kindly gave to the Cape Town University collection. These he discovered along the banks of the Bushman’s River on which Weenen stands. The site appears to be rich, and the implements are very like the Glen Grey material and that discovered by Dr. Lebzelter in much of this area. Points and roughly made lance-heads abound and form the bulk of the finds, a core-scraper and a double concave-scraper also appear. The workmanship is not very advanced, but this again parallels the Glen Grey material.

Implements showing similar facies are represented at the Grahamstown Museum from the Tugela basin, from Ravenshill Tarkastad district, from Gowie’s Kloof overlooking Grahamstown, and from the gravels exposed in making a playing-field at St. Aidan’s College, Grahamstown.

Dr. Lebzelter’s discoveries in the neighbourhood of Mangeni, some sixty miles south of Vryheid show strong similarities. He has named three groups, the Isikwenenian, the Inxobongoan, and the Ingeleduan respectively. All show affinities to the Glen Grey material, though the Inxobongoan may be more closely allied to the Pietersburg Variation.62

Facies.

Judging from the material collected from the Glen Grey Falls site and housed at the South African Museum (S.A.M. 4552 and 4690), this industry shows primarily a flake technique, though there are signs of its having had an origin in a core industry. The most symmetrical and carefully made implements of the series fall into a class midway between the small neat coups-de-poing of the Fauresmith Industry and the beautifully made lance-heads of the Still Bay. The workmanship is not good, tending to be coarse. The more finished types are rare, and only three or four good specimens appear from the Falls site. These consist of small amygdaloid implements up to 75 mm in length, about 40 mm in width towards the butt, and fairly thick. They are made on longitudinal flakes, the bulb of percussion sometimes appears at the hinder end, but has often been removed by flaking; the remainder of this cleavage face is untouched. The outer face is strongly rounded and is worked all over: where the striking platform is present it is faceted.

A second type of implement is the point; this is wide and flat, some 60 mm long, by 40 mm across at the base. The shape is roughly that of an isosceles triangle, the butt is again faceted, and working is confined to retouching along the edges only. These points are typically Mousterian.

Among other implements are high-backed points, side-scrapers, and neatly made discoidal artefacts, or fabricators bounded by a wavy edge. Small cores verging upon core-scrapers, and crude, unevenly made end-scrapers complete the industry (Plates XV and XVI).

Throughout this industry consistent faceting of the butt is apparent, and a core shows this to have been done previous to the removal of the flake. The primary longitudinal trimming flakes are struck from the percussion platform, and tend to be convergent.

Distribution.

It will be seen from this survey that this industry appears to be confined to the Natal-Eastern Province strip (the Eastern Highway) so far as our present knowledge goes, and from north to south the known distribution is this:

• Mangeni (Lebzelter).
• Tugela Basin (Albany Museum, Grahamstown).
• Bushman’s River, Weenen (Thomasset, and University of Cape Town).
• Aliwal North (Brown Collection, and South African Museum).
• Ravenshill, Tarkastad (Albany Museum).
• Glen Grey Falls (South African Museum).
• Rockwoods, Bongolo, Queenstown (South African Museum).
• Bower’s Drift, Queenstown (South African Museum).
• Keilands, Stutterheim (Lebzelter).
• Gowie’s Kloof, Grahamstown (Albany Museum).
• St. Aidan’s College, Grahamstown (Albany Museum).

The Pietersburg Variation.

The Glen Grey Industry stands out very distinctly from the Still Bay, but an illuminating link seems to be present in the Pietersburg Variation. While on the one hand this may be regarded as merely a reined Glen Grey, yet on the other hand it shades into Still Bay forms, even equalling them in symmetry and beauty in a few instances.

The technique evidenced in the Pietersburg Variation is far finer than that shown in the Glen Grey Industry. The lance-heads are well formed and show at least two distinct types. One is wide and almond-shaped, lenticular in section with a rounded butt; the shape is clear-cut and well defined; working is common on the under face. A second type is narrow, and about the same length, 60-75 mm, as the first type, but if cut across would show a section approximating either to a semicircle or to a right-angled isosceles triangle, the diameter or hypotenuse in these respective sections being formed by the original cleavage face. This second type is often markedly keeled along its length. A third type, less common but more advanced, is the long, leaf-shaped lance-head, very similar to the usual Still Bay form. A good example of this type forms the cover-design of J.P. Johnson’s books. It should thus be obvious that the two main criteria differentiating between the Glen Grey Industry and the Pietersburg Variation are the better workmanship and the greater variety of forms evident in the latter.

Sites

Mr. E.G. Paterson of Grace Dieu submitted a number of implements to the Albany Museum, Grahamstown. These he had discovered in dongas cutting into the surface of the soil. The implements do not seem to appear on the surface of the open veld. Mr. Paterson tentatively associates pottery, but this seems to be of a modern Bantu type. The implements are far superior to, though similar to, the Glen Grey material. The outer face of the lance-heads is consistently worked all over, and in some instances the same treatment is evident on the cleavage face. The length varies from 50-65 mm, the shape being that of a wide almond. In the better finished specimens the butt is perfectly rounded, while in the unfinished specimens it is faceted. The angle between the striking platform and the cleavage face is about 115°, and the butt is thus plainly visible from this face. The material used is a surface quartzite (Plates XVII and XVIII).

It has been found necessary to use this as a name-site in spite of the fact that the finest specimens and the greatest variety occur in Swaziland, as the best known site, at Ezulweni, shows the presence of coups-de-poing and modern Bantu metal ornaments in close association with the lance-head forms.

The finds from this site, at Ezulweni in Swaziland, have already been mentioned in speaking of the Earlier Stone Age material associated. The association of the various artefacts was recognised by Dr. Péringuey as interesting if it was to be regarded as reliable. We have here what would seem to be the association of implements of three distinct periods: Earlier, Middle Stone Age, and modern Bantu. The association of the first and last is obviously untenable, and the further question immediately arises as to how far we can accept the association of coups-de-poing with well-made lance-heads.

During August 1921, Dr. Péringuey received some iron bangles, some coups-de-poing, and a number of lance-heads of Pietersburg type from the tin-bearing gravels of Ezulweni. These objects were found in apparent association, when excavation was being done with a powerful Monitor jet at a depth of 25 feet below the present surface, and lying upon bed-rock. The implements were found within 6 feet of the rings, and no implements were noted as occurring above this level. Dr. Péringuey pointed out that the iron could not possibly have lasted until the present time, if its age is to be gauged by the association with Earlier Stone Age elements, and it seems more than likely that we have here an association of three industrial groups, brought about by gravitation to rock-bottom by means of dongas. The coups-de-poing are very typical of the Stellenbosch Industry, and are made of indurated shale and of granites from the Swaziland Series. The lance-heads are typical of the Pietersburg Variation, and are of indurated shale.

These lance-heads are worthy of further study: they are of two types, in each case worked over the whole of both faces, while the butts are carefully rounded. The one type is a long narrow variety rising to a keel down the outer face; the other type is wide and flat. Various points are also present in this group; they are triangular in shape, and very symmetrical. One specimen is of a quartzitic sandstone, and is very similar to the specimens to be noted later from Mossel Bay and Knysna. Together with these was submitted a large (50 mm. × 15 mm.) crescent or lunate of light green chert. This recalls the rare crescents to be found in association with Still Bay material at the Cape and elsewhere (Plate XIX).

Various sites appear in the Transvaal, all showing similarities to the Pietersburg material. From Barberton (S.A.M. 3016) are lance-heads of the same two types, the one triangular in section and long and narrow, the other flat and wide. From Klerksdorp come two very well made examples of this narrow lance-head; one is remarkably well-made, and is identical with the specimen to be mentioned later from Victoria West. The outer face is worked all over, while the bulb of percussion is carefully worked away on the cleavage face; the butt is again rounded (S.A.M. 1236). Rustenberg, Transvaal, shows similar specimens (S.A.M. 789) so far as shape is concerned, but the workmanship is coarse.

From Genesa and Morokwen, 65 and 120 miles respectively northwest of Vryburg, come implements of similar type. A well-shaped butt from a lance-head certainly shows very close affinities with Pietersburg.

Apart from these sites which seem to form a wall across the Transvaal, there are a few outlying sites which are of interest from the point of view of distribution. From Dordrecht, in the Eastern Province, the Rev. T.W. Green has presented the South African Museum with a number of points and well-made lance-heads (S.A.M. 1585), the latter being worked over the whole of the outer face, and also at the butt end of the cleavage face. The butt is rounded, and the section semicircular.

A certain number of the implements collected by Mr. Alfred Brown in the Aliwal North district show a tendency to be more like the implements of this Pietersburg Variation than the Glen Grey Industry. Unluckily our lack of knowledge makes it impossible to use this material, or to draw any conclusions from the evidence he has collected.

The Albany Museum, Grahamstown, shows excellent specimens from Dr. Atherstone, sent in by him from Kleinemond, Bathurst district. These are all of the almond, and of the long narrow keeled types, worked over the whole of the outer face and at the bulbar end of the cleavage face, the butt being rounded in each instance. A single specimen of the long narrow type is also represented from Cossakspost, Middelburg district (A.M.G. 1619), and is worked on the outer face only.

All these sites seem to show that this variation had a strong hold in the Eastern Province, and in the Transvaal. Evidence of a distribution in the central portion of the Union is available from Victoria West. In 1912 the Rev. J.G. Perold submitted a number of small implements from the golf course at Victoria West; the series was sent in as though from a single site, but they show every sign of having belonged to two different periods, if not two different sites. The earlier of the two shows very distinct signs of strong patination on the black indurated shale used, while the more recent specimens are entirely unpatinated and sharp. This later material consists of typical Smithfield end-scrapers and the like, and is fairly definitely assignable to Smithfield “A.” Of the patinated implements one consists of a large point, steeply and badly worked along the two edges, forming a rough equilateral triangle; with this is associable a well-made lance-head, broken at the tip, but otherwise a well-finished specimen. It is a narrow type, triangular in cross-section, the outer face being worked over the whole surface to a central keel, while the bulb of percussion is worked away on the cleavage face, the remainder of this face being unworked.

Distribution.

Again, treating the distribution from north to south, so far as is at present known it is as follows:

• Pietersburg (Albany Museum, Grahamstown).
• Rustenburg (South African Museum).
• Klerksdorp (South African Museum).
• Swaziland (South African Museum).
• Balfour (South African Museum).
• Barberton (South African Museum).
• Dordrecht (South African Museum).
• Middelburg (Albany Museum, Grahamstown).
• Aberdeen (South African Museum).
• Kleinemond (Albany Museum, Grahamstown).
• Morokwen (South African Museum).
• Genesa (South African Museum).
• Victoria West (South African Museum).

The Still Bay Industry.

It has already been pointed out that the Pietersburg Variation seems to form a step from the Glen Grey Industry towards the Still Bay, and in fact implements assignable to this first group which almost exactly parallel the finest Still Bay material are in the possession of the Transvaal Museum, the South African Museum, and Mr. C. Van Riet Lowe, and are all from the Transvaal-Swaziland area.

The typical implement of the Still Bay Industry is again the lance-head, worked evenly and neatly over both faces. The most usual shape is the laurel leaf, with a semicircular, or a wide-angled pointed butt. The implement is thin, never exceeding a centimetre in thickness, even in the largest known specimen (120 mm. × 47 mm.), though this thickness seldom falls below 5 mm. in the smallest specimens (38 mm. × 2 mm.). An allied, but less common type is the willow-leaf, or medium lance-head; it is long and leaf-shaped, and is similarly worked over the whole of both faces. The section is lenticular, and the semicircular- and triangular-sectioned types are both missing, and as a result, the central keel so typical of the Pietersburg Variation is absent. With these types appear points, which may be divided roughly into two classes: the normal triangular point, apparently common to all Middle Stone Age industries; and an oak-leaf type, so called from the wide and rounded scalloping along the edges. The secondary flakes removed from the edges of this latter type are spaced out and the result is a strongly scalloped edge, the waves being too widely separated to allow of the implement being termed “serrated.” These oak-leaf types are in many instances curved, either to left or to right, thus presenting an asymmetrical face. The point proper is symmetrical, straight, and not scalloped, and is similarly worked on the outer face only, trimming being confined to the edges, leaving the primary longitudinal trimming flakes visible on the outer face (Plate XX). The butts in both these types, and indeed in all implements of this group where working is not complete, are formed by strongly and evenly faceted fragments of the percussion platform.

In 1866 Sir Langham Dale, who had been greatly interested in the final results of the Boucher de Perthes controversy in Europe, discovered a large flake of chert, trimmed longitudinally, with a strongly faceted butt, some four miles along the Maitland road, and on the Table Bay side (west) of that road. As a result of further research, he and his sons may be said to have laid the foundation of South African archaeology. Considerable interest was at once shown by the local savants of the day, and from time to time articles appeared in the Cape Monthly Magazine (New Series) and in the publication of the Philosophical Society. Notable among these papers are Dr. Dale’s articles in the Cape Monthly for 1870, and the interesting paper by Mr. E.J.  Dunn in the Transactions of the Philosophical Society for 1880. To a large extent these papers centred about the Cape Peninsula, and as a result the history of the Still Bay Industry has been a long one.

The Dale Collection.

The greater part of Sir Langham Dale’s collection has been presented, by both himself and his sons, to the South African Museum.

The most interesting and typical implement represented is again the lance-head; the shape is normally leaf-like, and the two aspects consist of an outer face, trimmed all over to give a slightly curved section, and an under face, similarly trimmed, but flatter. In some instances working is almost entirely confined to the outer face, the removal of the bulb of percussion by lateral flaking, or a thinning towards the point sufficing for the under face. In the finest specimens the whole of each face is worked evenly. With this beautiful technique the even-grained surface-quartzite normally used produced a delicate implement.

Wide and medium “leaves” predominate, the narrow, heavily keeled type common in the Pietersburg Variation being absent. The butt is normally rounded, but in some instances the implement is butted by a point less sharp than that at the forward end. The size of the implements in the Dale Collection varies to a very considerable degree, specimens ranging between 75 mm. × 34 mm. × 10 mm., and 38 mm. × 23 mm. × 5 mm.; the workmanship on the smaller implements is clumsier and less delicate than that on the larger. This seems to have been due to the size of the implement and the resultant diliculty of dealing with it.

A few rough side-scrapers also make an appearance, and the Dale Collection contains a single well-made lunate or crescentic scraper, almost identical with the specimen referred to above from Swaziland. This is from a Fish Hoek site, and it is larger (30 mm.) than the Wilton type of crescent, which will be mentioned in a later paper. It will be seen that Colonel Hardy also associates the lunate with his Still Bay material (Plate XX).

The Hardy Collection.

The sites worked by Colonel Hardy are of extreme interest. The Cape Peninsula consists of a mass of mountains joined to (or separated from) the mainland of Africa by a flat, sand-covered plain, seldom rising more than 70 or 80 feet above sea-level. The northernmost mountain-mass of the Peninsula reaches its highest point in Table Mountain, which backs Cape Town to the south. The whole forms a single group which is cut off from the remainder of the Peninsula by a long valley, 3 miles from side to side, which cuts through the mountain masses, opening on to False Bay at Fish Hoek and on to the Atlantic Ocean at Noord Hoek some 5 miles away. Along the centre of this valley for the greater part of its length runs a hill which is covered by wind-blown sand at its eastern extremity, but rises free from sand at the centre of the valley. Toward the top of this hill, and in the centre of the valley, is a cave overlooking the southern branch of the valley, with a magnificent view of both oceans; it is known as Skildegat, and will be referred to later.

At the foot of the sandy portion of this hill, and on the southern side, Colonel Hardy discovered a site in the shifting sands, and has constantly revisited it over a number of years, collecting material as the caprice of the prevalent winds allowed, The material used at this site is the fine-grained surface quartzite, which is preferred throughout the Still Bay Industry, having a clean, even, dependable fracture, making it eminently suitable for the purpose to which it was put. A few specimens occur in Table Mountain Sandstone, a granular quartzitic sandstone with a dependable, though coarse cleavage. This latter material weathers faster than the surface-quartzite, and is more easily pitted and rounded by the wind-blown sands.

As at Sir Langham Dale’s Maitland site the lance-head type predominates, but the variation is here even greater, the length varying from 100 mm. to 40 mm. Here also the better finished specimens are worked over the whole of both faces, the original cleavage face tending to be slightly flatter than the outer face, giving a lenticular section. The shapes are similar to those appearing from Maitland, wide and medium leaves predominating. The forward point is usually formed by an angle of 40° to 50°, and the sides curve back to a butt which is either semicircular, or pointed with an angle of about 90°. In one or two instances diamond shapes occur, both ends being evenly pointed. The rounded or pointed butt coincides with the original position of the bulb of percussion, and in a few instances the cleavage face is left intact or the two ends alone are worked; the object seems to have been to give the implement a good “stream-line,” and it is difficult to class these partly worked implements as rejects, as the workmanship is often extremely fine. One or two instances of the “oak-leaf” make their appearance, both straight and curved types, but the workmanship is cruder than that found on the true lance-head.

At the opposite end of the valley, some four miles away, stands Noord Hoek. The valley has here passed the central hill, and opens out into a wide flat, mostly sand-covered, and running perhaps a mile inland. Towards the coast this flat is covered by shifting sand-hills, but further inland this gives way to a hard turfy earth, sometimes covered with a limy deposit. It was directly upon this hard earth that Colonel Hardy discovered his Still Bay material, often covered by shifting sands. The material here is exactly similar to that found at the opposite end of the valley, lance-heads, oak-leaves, etc., being discoverable at both sites.

Covered by the shifting sand-dunes, to the seaward side of an inland lagoon or pan are to be found a number of Kitchen Middens containing pottery and the usual associations of the local Later Stone Age sites. This site would appear to be entirely separate from the Still Bay Industry site, but the material from here has very naturally been thinly scattered over the area occupied by the Middle Stone Age material. It would appear from the positions of the two sites that the sand on the seaward side of the pan had accumulated since the Still Bay peoples left their site, and that the later arrivals had lived on the shores also, but on a later beach, built up of sand, blown and drifted from the west.

Colonel Hardy has also done a considerable amount of research work at a site between Milnerton and Maitland, some five miles across the Cape Flats from the town, on what is possibly a continuation of Sir Langham Dale’s site. It is situated on a shale floor, and as a result of the presence of iron the implements have taken a deeper patination than the sand-covered Noord Hoek and Fish Hoek specimens. The material used is the same surface-quartzite, and the types represented are similar to those found by Sir Langham Dale and Colonel Hardy at other Cape Peninsula sites save for the fact that, like Sir Langham Dale’s specimens, they tend to be smaller than those found in the Fish Hoek valley. The specimens in the Hardy Collection from here lie between 79 mm. × 28 mm. and 48 mm. × 17 mm. They have usually been worked over both faces, but in a few instances only the outer face has been touched.

Usually associated with these Still Bay types, Colonel Hardy has found a number of “anvil-hammer-stones,” flattened spheres, usually river pebbles or beach material, which have been used about their equators as hammer-stones, the perimeter being thus chipped and scarred by the action of hammering. The centre of each face, too, often shows signs of pitting, as though these stones had also been used as anvils in direct or indirect rest percussion. In course of time and as a result of usage, flakes have been removed from these hammer-stones, and Colonel Hardy has collected an interesting series, some specimens showing a few flakes struck off, others showing a little of the original surface of the stone only, a series of flake-scars, all showing an origin about the equator, taking up the whole of both faces. This forms the “discoidal artefact” or fabricator, which seems to be common to all our South African Industries.63

Colonel Hardy has discovered a number of large crescents at Fish Hoek. These vary in size from 80 mm. to 25 mm. in length, and are therefore much larger than the crescents or lunates of Wilton type which lie between 20 mm. and 8 mm. These large crescents are not worked along the chord, or straighter edge, though some show possible signs of usage along this part. The curved edge is worked with a steep “backing” technique, directly across the depth of the implement. One specimen shows working on the under (cleavage) face very similar to that employed in the manufacture of the lance-head: this fact may point to a weak link between this and the Howieson’s Poort material.

Skildegat Cave.

We have already had reason to mention the presence of this cave which is situated on the southern face of the hill dividing the Fish Hoek-Noord Hoek valley, and locks southward across a stream running through the flat country below.

Once the hill is ascended the cave is most easily approachable along a natural pathway passing along the foot of the rock krantz or cliff in which the cave is situated. The cave is large, measuring about 100 feet across the mouth, by 50 feet deep, and rising to a domed roof varying from 6 to 30 feet in height above the present surface of the deposit. This deposit is known to reach a depth of over 10 feet, but is so filled with roof debris fallen from the ceiling of the cave that work below this level is dangerous.

In 1925 the writer [A.J.H. Goodwin], at the instigation of Mr. J.C. van der Poll, dug a trench across the cave from front to back; further work proved inconvenient and it was abandoned, only the upper midden refuse of Wilton type having been disturbed. A short time after this Mr. Peers and his son visited the cave and commenced investigation, discovering a number of skeletons in the deposit towards the back of the cave. They very kindly invited me to visit the site from time to time and to take notes. Excavation is still being done in a methodical and careful way, and although a few preliminary reports of a popular type have appeared, no scientific paper has yet made its appearance on the subject of this cave.

In the upper deposits Mr. van der Poll had discovered a fragment of a bored stone and other evidences which proved that the cave had been inhabited by a people bearing a Wilton culture, and following a “strandloper” (beach-dweller) mode of subsistence. Mr. Peers later discovered three complete skeletons of the type usually associated with Wilton material, and the fact that these had been laid upon beds of bushes and branches which had not appreciably decayed, together with the presence of a fragment of twisted iron lying under one skeleton, go to prove that the cave had been inhabited by these folk until the European occupation of the Cape. It is of interest here to note that Van Riebeek in his journal mentions that this valley was inhabited during his governorship, and in fact seems to have been the home, or at least the refuge, of Herry or Harry, his amusingly dishonest interpreter.

While the makers of the Wilton implements are most certainly to be associated with cave-paintings, yet this cave shows no signs whatsoever of cave-painting having been practised. The name Skildegat (painted hole) is more than suggestive, but save for a few fragments of red ochre in the cave deposits of Wilton type, no signs of painting appear. It is, of course, possible that the sea air attacks the material used as paint, and this might account for the complete absence of cave-art.

Well below the Wilton material, and not buried from those deposits, occurs a single, almost complete skeleton of an entirely different type. Physically the skeleton is markedly heavier in every bone than the dainty infantile skeletons of the “San” race associated with the Wilton material. Eyebrow ridges are here present, the nose is wide, and the whole make-up is rugged. Little more need be said here, save to state that this skeleton appears to fall within the group described by Professor Dart as of “Boskopoid” type.

A full description of the implements before the publication of a monograph on this cave would be premature, and it is only necessary to state here that the implements are of Still Bay type, showing a tendency to affinities with implements to be described later from Howieson’s Poort.

The importance of this cave cannot be overestimated. It proves the sequence of Still Bay and Wilton, it associates physical types with each, and it proves that the later types continued until a comparatively modern date (seventeenth century).

Pringle Bay.

Pringle Bay is a small inlet, bounded on the east by Cape Hangklip, and lying on the other side of False Bay from Somerset Strand. The land bounding the bay to the north rises slowly, and merges into the talus of Cape Hangklip on the east, and into the other hills bounding the bay to the west. It was well up on the slopes of the western hillside that I [A.J.H. Goodwin] discovered a large reedy loess deposit, which had formed below a small but persistent spring. The blown sand had been caught up in the reeds and their roots until a cushion of black loess some 5 feet thick had been formed. Above this layer the loess had changed, and the superficial deposit of about 5 feet in thickness is composed almost entirely of blown sand, with a minimum of roots and vegetable content. This 5-foot sandy layer has been denuded, and is only left in peculiar tectiform lumps scattered about on the exposed surface of the darker loess. Before the accumulation of this final sandy layer there had been two human occupations of this site. The irst shows very fine specimens of the Stellenbosch Industry, made in the white Table Mountain Sandstone prevalent about the site. The later occupation consists of Still Bay material made in a dark surface quartzite, while only a few flakes of the Table Mountain Sandstone make their appearance (text-fig. 2). It is obvious that the comparatively damp climate, which is apparently marked by the presence of the reedy loess, gave way at some time to a very much dryer and more windy climate, during which the sandy loess accumulated, and it was previous to this change of climate that both these occupations took place. This point will be referred to later in dealing with the climate of this period as a whole.

Still Bay.

Some three or four miles west of the Kaffir’s Kuil River-mouth, south of Riversdale, the coast rises steeply to form a hill overlooking the sea. This hill stands out conspicuously from the surrounding sand-covered hills, owing to the presence of a scar of red sandy earth which faces the sea. Over and on the crest of this hill is a wide area covered with a limy deposit, much the same as that present at the Noord Hoek site, but here it overlies the red sandy soil. Embedded deeply in this lime are to be found coups-de-poing of normal Stellenbosch type; they are of quartzitic sandstone and seldom well made. Bones of a large buck or buffalo have been found in association with these implements by Mr. Heese in the same limy deposit. Sporadically over the surface are to be found pottery, implements, and midden-refuse of the type usually associable with Wilton types and the coastal sites which occur on this coast, and from this point they run from the top of this hill, down a slight valley, across this and up the opposite slope. Here the surface is again covered by the lirny deposit which has washed away in places to expose the red earthy subsoil. It is this end of the series of man-made deposits that constitutes the Still Bay name-site.

It will he seen that Mr. Heese has found three industries in close proximity to one another: Stellenbosch material on the first hill, Wilton material stretching across from here to the second hill, and Still Bay material on the latter. It is unluckily impossible with our present knowledge to date the three industries respective to one another from evidence available at this site, but it is worthy of note that the Wilton site does not in any way coincide with the Still Bay site, but is situated lower down the slope, and only touches the Still Bay site at the latter’s south-eastern end, the remainder being free from Wilton material towards the top of the hill and across it. The position is thus exactly similar to that presented at Noord Hoek and at Sir Langham Dale’s Maitland site, in each case the nuclei of the sites are definitely separate, though their outer edges tend to overlap.

The actual material from the Still Bay sites is very much the same as that discoverable at Noord Hoek and elsewhere. If anything, the workmanship is a little finer than that found at the Cape Peninsula, but the forms remain very little changed. The material here, also, is the beautiful surface quartzite preferred by these folk, but here and there flakes and refuse from rough attempts to utilise quartzitic sandstone also occur.

Over the crest of the rise, to the immediate westward of this site, a considerable quantity of sand has accumulated. From its position, relative to a few flakes discoverable beneath the sand, it would appear that this accumulation is subsequent to the appearance of the Still Bay workers; similarly the aggregation of sand on the western slopes points to a period of strong westerly winds. Elsewhere in this neighbourhood Wilton material overlies the driven sand.

Other Sites.

Sites are known at Stellenbosch, and one or two possible areas occur towards Port Elizabeth, but apart from these, very little further is known of the distribution of this industry. It would appear then from our present knowledge that the Still Bay Industry proper is confined to the southernmost coast of Africa, from the Cape Peninsula to the region of Port Elizabeth.

Mr. M.C. Burkitt, on his recent visit to South Africa, pointed out the possibility that the Still Bay technique and industry would appear to be the outcome of a mixture of Middle Stone Age and Neo-anthropic influences. This seems more than probable in the light of the Howieson’s Poort finds, which show a mixture of Still Bay and Neo-anthropic types very clearly, and we can presume that the racial types represented by the two industries were not mutually abhorrent, but sufficiently alike physically to allow of mixture of some sort. From our knowledge of modern races we may presume that such a mixture implied the presence of two types of implement makers (i.e. the men), and not merely concubinage resulting from wars, as in this latter case the two types of implement could not both survive.

The Howieson’s Poort Variation.

Some three miles south-west of Grahamstown, and situated high up in a krantz or cliff facing the Howieson’s Poort Hotel, across a wide valley is a cave well-hidden by trees and looking eastward. Mr. J. Hewitt and the Rev. P. Stapleton, S.J., first reported their finds from this site in 1925, and work has been done in this cave from time to time since that date. In this paper64 they refer to a series of peculiar flakes from the Kasouga River-mouth, paralleling specimens from Bell, Peddie district, and from this Howieson’s Poort cave. The distinctive feature of these particular flakes, which only form a part of the cave finds, is the appearance of secondary trimming on the under face of the flake. Work of this type is very uncommon, and is usually to be associated with Middle Stone Age material. Later Stone Age implements only show working on the outer face, the cleavage face remaining intact. In a later paper65 these two writers enter into a much fuller account of the finds, while further details will appear in another paper in the S.A. Journal of Science for this year (1928).

The rock shelter is peculiar in the complete absence of cave-paintings, and for the fact that the type of refuse marks it as being more in the nature of a cave-workshop than a home-site. No animal remains appear, nor are there any of the shells, pottery, or beads usually associable with a Later Stone Age home-site. The deposit is small but rich. The surface accumulation is perhaps a foot thick, and consists of vegetable matter, guano, wind-blown sand, etc. Below this is the inhabited layer which consists of a black deposit about eight inches to a foot thick, which lies directly on a pure sandy layer.

The cave is of extreme interest, showing as it does a mixed industry consisting of Still Bay elements combined with Neo-anthropic elements, the latter of a type not known in a pure state elsewhere in the Union.

The Still Bay implements consist of the usual lance-heads worked over the whole of both faces. The shape is much the same as that found in the normal Still Bay Industry sites, but the size is generally small, lying in the neighbourhood of 50 mm × 30 mm, some being as small as 40 mm × 18 mm (Plate XXI, ll-14).

The Neo-anthropic elements are peculiar and typical. Crescents appear, but they are far larger than the Wilton types, ranging as they do from 50 mm × 15 mm to 32 mm × 14 mm, with a thickness of 4 mm towards the thicker back. In the Wilton crescent the curved edge is usually trimmed vertically, the Howieson’s Poort finds are bevelled, while the working is in some instances confined to the two ends of this curved edge, the chord or straighter edge of the crescent being unworked. The large size of these crescents immediately recalls the Still Bay crescents, which though scarce are apparently widespread and form a definite part of that group. The material used at Howieson’s Poort is again the surface quartzite so beloved by this group of people. The Howieson’s Poort crescents seem to be a variation of the transverse arrow-heads typical of North Africa and the Iberian peninsula, and a single specimen of this type actually occurs here (Plate XXI, 15).

With these crescents can be associated pointed blades, having a characteristic trimmed oblique edge towards the apex. The point is sometimes obliquely recurved. A series is represented, and is illustrated in the papers above referred to, which reminds one of an oblique-burin technique. A flake has been detached from the apical end, and the facet thus formed runs obliquely from the tip to a point about one-third of the way down the implement, on the opposite edge. A number of “rod-scrapers,” rod-like pieces of worked flake, ranging from 58 mm to 26 mm in length are represented. They are a constant type, and recall the “redirecting flakes” appearing sporadically in the Middle Stone Age. Scrapers of various types also appear together with end-scrapers, horse-shoe scrapers, hollow-scrapers (spokeshaves), and so on, but the small thumbnail scraper so typical of the Wilton Industry is completely absent. These scrapers shade into the flake-knives worked on the under face, which in this aspect resemble the lance-heads of Still Bay type.

One extremely interesting group described by these writers under the term “gravers” represents the true burin of Europe. The burin technique employed consists of the working of an edge obliquely or directly across the end of a flake, and the subsequent removal of a single fluting flake from the prepared end, the scar running down one edge of the flake. This leaves a chisel- or gouge-like edge at the intersection of the trimmed end and the flake scar. This scar is best known as the burin facet. Some seventeen burins are illustrated in Mr. Hewitt’s papers, and these were the first true burins to be recognised in South Africa. One or two examples have appeared from Skildegat Cave, but they are still uncommon.

The lance-heads associated with these elements are often worked over the whole of the outer face, and the faceting of the butt is also common. In a few instances work appears also on the under (cleavage) face towards the butt end only; the normal type, worked over the whole of both faces, is, of course, fully represented. No specimens appear to reach the degree of finish apparent at Still Bay.

Mr. Hewitt draws attention to the similarities existing between these Howieson’s Poort finds and material discovered by him at the Kasouga River-mouth, and again to the similarities present between these two and the Still Bay Industry. The further comparison between these finds and the Cofimvaba specimens does not appear to be tenable.

In a note to the second paper, Mr. M.C. Burkitt adds:

The Howieson’s Poort finds as a whole are exceedingly interesting and quite unlike those of the well-known Wilton Industry. They are undoubtedly to be correlated in culture, though probably very much later in date, with the Upper Palaeolithic. They are, however, elements of Middle Palaeolithic culture still surviving here, and these have undergone modification, the result being a sort of Still Bay type of tool. That Middle Palaeolithic man has left his mark as far south as the Grahamstown district can be abundantly proved. Can we consider the true Still Bay types as a hybrid of Mousterian and Neo-anthropic cultures? There is analogy for a similar phenomenon in North Africa.

It is of interest to note that the uppermost layer (layer B) at Montagu cave shows slight affinities to the Howieson’s Poort material, though it is also strongly allied to the Wilton types, to which group it was originally relegated.

Other Variations.

As has been noted before, these four groups appear to show a sequence from the Fauresmith, leading through the Middle Stone Age to the Neo-anthropic Industries. That this sequence is not purely evolutionary in the strict sense of the word, but rather to be regarded as a progressive hybridisation, is more and more clear as we study the northern forms which have come down to us, sometimes unchanged, but as often metamorphosed in a striking manner.

The Stellenbosch appears to be a pure industry; the Fauresmith is apparently a hybridisation of this with a possible Mousterian influence. The Fauresmith seems to have branched into two industries, or more correctly two industries show affinities to the Fauresmith, but not to each other. The first of these, with which we have already dealt, is the Glen Grey; the second will be dealt with later by Mr. Lowe, and is the Smithfield “A.” The Glen Grey Industry is not markedly different, save in fineness of workmanship, from the Pietersburg Variation. A further advance marking, as Mr. Burkitt points out, possible Neo-anthropic influence is the Still Bay. The final step is the Howieson’s Poort Variation.

So far so good, but is this all? Unluckily it is not. We have a number of groups which are either local variations, that is, in the nature of deteriorations or sports cut off in space from contact with advancing cultures; or else variations forced upon the workers by the presence of differing material. The discovery of the true relationships is going to prove difficult.

The Mossel Bay Variation.

It will be easiest to deal with a group of which we can presume the origin before attempting to speak of the disconnected variations scattered throughout the country.

For a considerable time stone implements from the great cave at Cape St. Blaize, and from other sites in the Mossel Bay district, have called forth a deal of speculation. Implements from these sites at the South African Museum show a large number of unworked points and flakes of Table Mountain Sandstone. All are trimmed by parallel or convergent fluting flakes; in all cases the butt, where visible, is faceted. There seems to be a minimum of worked points, though one similar to a lance-head type is represented in the Transvaal Museum collection.

The cave at Cape St. Blaize has been spoilt from an archaeological point of view by the surface having been levelled, and a seat overlooking the sea having been added. Some two years ago I [A.J.H. Goodwin] visited the cave, but was unable to add any tangible evidence towards the elucidation of the mystery of the true position of this variation. A number of Wilton implements of surface quartzite were discovered, and these appeared to be mixed up very considerably with the Mossel Bay types. This led me to the conclusion that the Mossel Bay implements formed a local variation of the Wilton. I feel, however, that further search at this cave might reveal a sequence of industries which should be of immense value, and which would entirely contradict my original conclusions.

A recent visit to Knysna has thrown a considerable new light on the Mossel Bay types there found. For some years Mr. William Brown has taken a very considerable interest in the local archaeology, and he very kindly showed me a number of sites, mainly consisting of Mossel Bay implements. Across the mouth of the Knysna River from the village known as The Heads, stands the Western Head, a high hill facing the sea and overlooking the river. The approach to the top of this hill is made from above a peculiar natural bridge, formed of two double arches, perhaps a hundred feet high, in series with two caves, which do not cut right through the rock. These two caves have been inhabited at some time by a people with a strandloper type of subsistence, the deposits having been covered with wind-blown sea sand deposited by a south wind.

From here the ascent is steep until the topmost point of the headland is reached. The top is then found to be a large sand-covered flat. For over a mile the surface soil is composed of hard sandy loess partially cemented, but cut into by dongas or erosion gullies. The uppermost 4 feet is then seen to consist of red loess, while below that a considerable amount of vegetable matter produces a harder black loess. The upper more sandy layer seems sterile of implements, but directly this is denuded large implement-covered areas appear. The implements and flakes are all of Table Mountain sandstone of a coarse type. The erosion of the patches where implements are found has released the sand, and great sandhills move over the surface from year to year revealing and concealing parts of the site.

All the implements are of one general shape; they may be described thus: Longitudinally trimmed flakes, trimmed by the removal of two, or at the most three, convergent flakes. The third flake removed is usually the central one, and is nearly always “stepped” at about one-third of the distance along the spall. This is so usual that it would appear that it was intentionally done to give a “stop” on one face in order to make hafting more easy. The flakes generally form points with a general shape much that of an acute-angled isosceles triangle. A number of parallel-sided rectangular flakes, perhaps 50 mm to 75 mm in length, point to this form being also a desired type. Secondary trimming seems uncommon, but is present in some specimens. Some few oak-leaf types appear. In all instances the butt of the implement or flake is faceted and gently rounded (Plate XXII).

Mr. Brown found, and presented to the South African Museum, an excellent point of agate, or some similar flint-like material, showing a typical Mousterian shape and technique. One edge is trimmed very heavily, the other but slightly, a discrepancy due in all probability to the thickness of one edge of the flake originally removed. The butt is similarly faceted (text-fig. 3).

The quartzitic sandstone used at this Knysna site seems to be the only material available in any quantity; a few specimens of chert, cloudy quartz, agate, etc., do occur, but they are very uncommon. The quartzitic sandstone is brittle and coarse, and as a result fine work is impossible, and a technique similar to that found on the Still Bay material is quite out of the question. In the agate specimen mentioned above the workmanship is immediately finer, and is very like the Mousterian of Europe. The oak-leaf types point to a connection between this industry and the Still Bay, and it is very possible that an almost complete lack of fine-grained material created the Mossel Bay Variation directly from the Still Bay or some very similar industry. The lack of secondary working is largely due in all probability to the fact that the coarse-grained sandstone used produces a natural, fine, saw-like edge. A further point in favour of the view given above is that at all Still Bay sites there are a certain number of flakes showing distinct similarities to the Mossel Bay material, and a number of flakes of Table Mountain sandstone are always present at these sites.

The sites at which this variation appears are Cape St. Blaize (Mossel Bay), Gouritz River, both the Eastern and the Western Head at Knysna, and one or two other neighbouring sites.

In a railway cutting on the forest narrow-gauge railway, Mr. C. van Riet Lowe found a skeleton of “San” (Later Stone Age) type in 1922. In 1928, Mr. FitzSimons66 of the Port Elizabeth Museum started excavations at this site, and recovered a number of skeletons with rough implements of the type usually to be associated with the coastal midden peoples. Some little time later Mr. William Brown discovered a deposit of Mossel Bay types at a depth of 22 feet, directly below the skeletal remains which were near the surface. This proves a sequence of time between the two industries, and the fact that the interjacent deposit consisted of sand will be referred to again in the paper.

The Alexandersfontein Variation.

Some six miles south of Kimberley a wide undrained valley has formed, and is known as the Alexandersfontein Pan. During the rainy season this forms the centre of a large drainage area, and as a result the high lands surrounding the pan have been inhabited by a great variety of people of different cultures over a long period of time. Lying on the higher lands as they do, the implements left by these peoples have been open to the action of the elements, and as a result material of various types has washed down towards the pan below. In this way a large number of types and implements of different ages have become almost inextricably mixed, and relative age can only be judged by the uniform patination within each group of implements.

Messrs. Swan & Power of Kimberley have expended a great deal of time and energy in collecting material from here for the McGregor Museum, Kimberley, and as a result of their work we find that there is (1) a considerable quantity of relatively unpatinated Smithfield material, and (2) a large number of other implements much more heavily patinated, and, therefore, to be regarded as of earlier date. These last implements are of great interest. They are all of a flake type, showing longitudinal convergent flaking and a faceted butt. The outer face is seldom worked with any certainty, and though what may be regarded as rough secondary chipping does appear, this might be due to wear. The under (cleavage) face shows in a very considerable number of cases the presence of flaking of a type usually associable with pressure (wide, shallow flake-scars) and very definitely of human origin.

What relationship exists between this type and the other “Mousterian” or Middle Stone Age elements in the Union it is impossible to say with any certainty. Exactly similar material was found by me [A.J.H. Goodwin] scattered over the Middelburg (Cape) Golf Course, but this, too, shows a lack of definition and a degree of wear making it difficult to say whether the slightly worked flakes discovered are the desired implements, or whether the makers had more specialised implements in view. The extreme patination on the indurated shale, and the relatively slight patination on the Smithfield implements of the same material, make it more than likely that the Alexandersfontein Variation is considerably earlier in age than the Smithfield “B” types found at Alexandersfontein.

The Hagenstad Variation.

Major Collins67 describes a site at Hagenstad, some thirty miles north of Bloemfontein. He here discovered a series of stone implements under two distinct layers of peat, each about 8-10 feet thick. These artefacts seem to have occurred in association with bones of extinct animals, burnt fragments of wood, and wooden pins. All the implements but two were lost; these two are described as a spear-head and a knife. It seems to have been at this site that Captain R.G. Helme discovered a number of shale implements of interest. He describes the site as being situated at the radio-active springs of Hagenstad, Brandfort district. He discovered implements at a depth of 20 feet below ground-level, and in association with the teeth and bones of an extinct buffalo, Bubalus bainii. The implements consisted of triangular points, worked on flakes of indurated shale, and step-flaked along the edges of the outer face. They were 2 or 3 inches in length and half as broad.68

Mr. Lowe visited this site some time later, and was shown a number of flakes with certain similarities, embedded in the hard crystalline floor of Floris mineral baths,69 and found there by Mr. Venter (S.A.M. 4671). The deposit here looks as though it might produce some interesting fossils of human date, as in the small volume of material submitted by Mr. Lowe were found hippo teeth and a large variety of other bones firmly embedded in the mineral matter, together with human-made flakes. The most interesting implement he has sent in from this site is a long point, dagger-like, some 254 mm in length and about 64 mm across the greatest width. Together with this were found a number of small spheres, about 76 mm in diameter, similar to types which have been sent in from time to time from various museums.70 A number of smaller flakes very similar to Captain Helme’s material are also represented. These show facetted butts in a few instances only. The working on the long point is simple, and is well done, flakes having been removed with considerable evenness from the two edges for the entire length and on the outer face only.

The McGregor Museum, Kimberley, shows some similar points which may prove to belong to this variation. Mr. Wohlfahrt has submitted six specimens from a depth of 12 feet in gravels in the Vaal River at Windsorton, a similar specimen found with these is present in the South African Museum. These were not associated with coups-de-poing, which were discovered at a still greater depth in gravels directly below these points. Mr. Skill has presented a point of exactly similar type to the McGregor Museum from Droogveld in the Barkly West area. The Port Elizabeth Museum also contains a very similar point from the Sidney Estates, Vaal River (P.E.M. 718). This specimen is some 6 inches long. Whether all these types represent a single Middle Stone Age Industry, or whether there is merely an affinity between these elements, it is difficult to say. The triangular flakes found by Captain Helme are very similar to points represented in the Albany Museum, Grahamstown, from the Peddie coast and from Roberts’ Vlei, Grahamstown. The workmanship on these implements is good, but the retouch is confined entirely to the edges of the implements and to the outer face.

The Sawmills Variation, etc.

Implements from Sawmills, described by the Rev. Neville Jones in his Stone Ages of Rhodesia, seem to fall into the Middle Stone Age; in fact they show certain affinities to the Aliwal North material described above.

A further similarity to the Sawmills implements is visible in the material represented in the South African Museum (S.A.M. 2998) from Butterworth in the Transkei. The implements from here are of a rich brown chert or jasper, and are delicately worked, the material allowing of very fine workmanship. One implement of a black glass-like stone is neatly worked with a technique very similar to that found at Sawmills. The other elements resemble Aliwal North material more. The workmanship at Sawmills is well advanced, and Mr. Jones says on p. 53:

Mr. Stevenson has recently found a point, neatly worked on both faces, which nearly approaches the quality of the Solutrean “laurel-leaf” point, and exhibits the skill to which these people attained.

Later he observes (p. 113):

Some of the more advanced examples from Sawmills and the caves of the Matoppo Hills make a near approach to these Cape Flats points, but they fall a little short in the degree of finish they exhibit.

This may prove a very useful link between the two areas, but it seems a little too early to group these various types under a single industrial or variational term.

Conclusions.

It would thus appear that we have in the Middle Stone Age a number of more or less allied groups. Some of these fall excellently into a presumptive evolutionary series, while others fall well out of this series and appear to consist of local variational sports. There is a certain similarity running through the entire Middle Stone Age, notably the continued presence of convergent flaking, the use of points, and the faceted butt. They would all appear, therefore, to have had a common origin, presumably to the north, and in all probability in that culture to which we may safely refer as the “Mousterian.” Of some of these groups we know sufficient to assemble the material into definite industries; in other cases considerable research of a wider type is needed before we can with any fairness judge of what should constitute an industry and what a variation. Material from a single site, if it forms a wide and truly comprehensive collection, may be regarded in some instances as sufficient to found a new industrial or variational group. Lack of knowledge is not the only reason why certain groups have been termed variations. The Pietersburg Variation has been so called because we do not yet know where it merges into the Glen Grey Industry and into the Still Bay Industry respectively. The Howieson’s Poort Variation has been so called as it is definitely a mixed industry showing an apparent mixture of peoples, a mixture which can be analysed, and is not an intimate combination as the Still Bay might appear to be.

Climate.

Let us retrace our steps and look at the various circumstances governing our finds in the extreme south between Cape Town and Port Elizabeth. In the Fish Hoek Valley the implements of Still Bay type occur on what would seem to be a turfy ground formed during a moist and fertile climatic period. Over a part of this, and concealing some of the finds, has accumulated a layer of wind-blown sand. This sand has been blown and drifted to this part of the coast by a west wind. We noted that midden refuse, Wilton implements, pottery, etc., of the Later Stone Age type lay directly upon this sandy deposit, but nearer the present coast, and I deduced from this that the accumulation of sand from the west had shifted the shoreline further out at this point by perhaps half a mile between the two occupations. At the Pringle Bay site I found Still Bay and Stellenbosch types of implements lying directly upon a turfy loess, the product of a moist but windy climate; this loess contains a maximum of vegetable matter, the remainder of the deposit being made up of wind-blown sea sand. Above this stratum and directly above the implements, the deposit changes to a pure sandy loess containing no vegetable matter. At Still Bay the position is much that occurring at Noord Hoek, though here instead of a beach, sand has accumulated on the western slopes of the neighbouring hills, similarly deposited by a west wind. At Knysna the Mossel Bay implements lie on a like deposit composed of sand and vegetable matter, and have had a 3- or 4-foot sandy layer accumulated above them, even at a height of some 400 feet above sea-level, by a west wind. A similar sand deposit overlies the Mossel Bay material found by Brown at the site described by Lowe and by FitzSimons near Knysna.

What has happened at Howieson’s Poort? Here we seem to have an exactly opposite state of affairs; the Howieson’s Poort Variation is lying directly upon a sandy layer, and is in turn covered by a humus layer.

What do these facts imply? The prevalent winds in this southern-most part of Africa should be westerly, but owing to the uplift of air produced by the heated surface of the Kalahari desert this wind is diverted, and becomes a south wind when it strikes the Cape. Can we then presume that the climate of the Middle Stone Age period in the Kalahari was cooler and perhaps more fertile than it is at present? We have the apparent fact of west winds depositing sand at the Cape, while on the other hand we have the hippo teeth at Hagenstad, and evidences of a more amenable climate at Taungs.

If the sand-depositing west winds are all of a single period, and if the Howieson’s Poort sand deposit was accumulated by this means (which is unlikely), we have a rough means of dating the Middle Stone Age relative to the sandy period, and the entire Middle Stone Age save for the Howieson’s Poort Variation would seem to have preceded this climatic change.

Can we correlate this sand-accumulating period with the European glacial periods? After the end of the Bühl Stadium in Europe, the storm centre, which had hitherto formed a reasonable climate in the Sahara, shifted northwards, leaving a hot desert area behind, and producing a humid climate of winter rains in the Mediterranean basin. Now the climate of Africa west of 28° East is symmetrical about the equator at the present time. Can we presume that this symmetry is normal and balanced, and that the creation of a desert in the Sahara region would imply a simultaneous creation or accentuation of desert conditions in the Kalahari area? If so, we can date the Kalahari desert and the westerly accumulated sand of the southern-most littoral as of post-Bühl date (say 7000-6000 B.C.). If it were possible to prove such a correlation, we would be in a position to lay the entire foundations of chronology and recent geology in South Africa.

Sequence.

One point must not be overlooked. We have been using the term Middle Stone Age to cover the group here dealt With. Is this justifiable? I [A.J.H. Goodwin] think it is. At Windsorton the material is datable as being intermediate between the Earlier Stone Age deposits in the deeper gravels, and the Later Stone Age material of the surface. At Rockwoods, Queenstown, the Glen Grey material lies in the mountain talus; Later Stone Age material lies above it. At Skildegat cave the Wilton types are shown to be later than the Still Bay types. At Knysna Mr. Brown has shown that a similar state of affairs exists, “strandlooper” skeletons lying at a depth of 8 feet, and Mossel Bay types at a depth of 20 feet. At Still Bay we find Stellenbosch types of implement embedded in the hard red earth, lying upon which is Still Bay material. At Middledrift the writer [A.J.H. Goodwin] found Middle Stone Age material lying upon gravels containing water-worn implements of Stellenbosch type.

This and other evidence mentioned in this paper justifies us in regarding the Middle Stone Age as lying between the Earlier and Later Stone Ages, and the term used would appear to be an adequate label for the whole.

The Later Stone Age.

The Earlier Stone Age comprises, basically, a variety of core industries; small flake tools appear only sporadically and with little consistency, and flake cores are almost entirely absent. The Stellenbosch Industry shows little if any advance on the Lower Palaeolithic of North Africa and Europe; the Victoria West a variation without a necessary advance. In the Fauresmith Industry, however, we see the flake coming into its own. Here attempts at longitudinal flaking and the appearance of faceted butts imply either a lithi-cultural evolution or a cultural admixture.

With the Middle Stone Age we get our first view of pure flake industries. Lance-head and point types predominate, faceted butts are common, and longitudinal flaking comes into its own. As the highest point of an evolutionary series we get the Still Bay Industry, shading into the Howieson’s Poort group. With the former we can now associate a definite physical type, but apart from our knowledge of the stone implements and the physical remains we know little of this phase, and indeed hardly anything of the arts, crafts, and means of subsistence, in spite of the fact that this latter appears to have been largely based on animal foods.

With the dawn of the Later Stone Age we arrive at a wider and better known field, we are in a position to give not only the lists of implements made and used, but also the foods, clothing, industries, arts, and physical characteristics of its people. We recognise a Neo-anthropic group, and, definitely belonging to it, those “Bushmen” our immediate forbears so sanctimoniously annihilated in a fit of righteous indignation that lasted well over a century. The term “Bushman,” however, has always been loose and unscientific, and the only definition that is now possible is “a number of peoples enjoying a cattle-thieving and a hunting mode of life, living in the more arid parts of the Union, and consisting of members of broken-down Bantu clans, Hottentot groups, San tribes with a slight admixture of European blood.” Along the coastal regions the term Strandlooper has been applied to any person of no fixed abode, with a predilection for shell-fish and a habit of leaving the shells about.

In both cases we must regard the terms as unscientihc and for our purpose useless. Physically we have a group of people, named by various writers the “San” tribes, of fairly definite physical type at present inhabiting parts of the Kalahari. This term might, at the present state of our knowledge, be best used to imply the physical type predominant in all the Later Stone Age peoples.

The Smithfield and Wilton Industries.

Some years ago Dr. Kannemeyer from Smithfield, a friend of Mr. Alfred Brown of palaeontological fame from Aliwal North, took considerable interest in archaeology. He corresponded for some years with Dr. Péringuey, and at his death almost his entire collection came to the South African Museum (S.A.M. 1830), of which Dr. Péringuey was then Director. Unluckily the collection, though almost entirely of the group we have called Smithfield, has confused with it individual specimens of Middle Stone Age type.

Mr. Alfred Brown also took a keen interest in archaeology, and one point, perhaps psychological, is of interest with regard to these two collectors. Dr. Kannemeyer concentrated almost entirely on Smithfield material, while Brown, his confrere, appeared to disregard — or perhaps failed to recognise — Smithfield types, and to have set out to discover only Middle Stone Age material. The fact that Middle Stone Age implements do occur in and near Smithfield, and that Lowe has discovered a number of Smithfield sites in the immediate vicinity of Aliwal North, show that both types appear at and near both towns.

For some years the writers of these papers have been working on the Later Stone Age, and the results here set forth have been arrived at by a co-operative method of field and museum work, an interchange of ideas, and a sifting of theories only possible between field and museum workers. The museum worker needs a quick confirmation or denial of his theoretical ideas, while the field worker needs the great accumulation of material at the disposal of the museum worker for his comparative knowledge. Work must necessarily be shared. Many other workers have done a very considerable amount on the stone implements of South Africa, but they have always been either pure field workers or simply museum men, and the results of their labours are often not only misleading but, more often than not, of little real use. The field worker cannot know what to seek, or fails to appreciate the true significance of his finds, and the museum worker cannot check his theories, especially the more negative ones, unless their researches are complementary. It may very truthfully be said that on the contents of each parcel received depends the validity of the museum worker’s every theory.

In 1925 the term “Smithfield” was first used to describe the industry investigated by Dr. Kannemeyer at Smithfield, and best represented in Péringuey’s book by the photographs of Cottell’s Cradock specimens.

The term “Wilton” was first made use of at the same time, and was taken from a cave on a farm of that name near Grahamstown. Mr. J. Hewitt excavated the site for the Albany Museum and defined the industry there represented.

At first sight it would appear that the Smithfield Industry represents an offshoot from the Neo-anthropic Capsio-Aurignacian cultures of the Mediterranean basin — most probably from the Lower Capsian peoples — while the Wilton, with its marked affinities, appears to be more closely related to the Upper Capsian group. Culturally, a strongly marked similarity is present between these three industries, but very much more research will be necessary before we can discover the true relationship between the Smithfield, Wilton, and Capsian groups. Unluckily one road is not yet open to us, for nothing is known of the physical characteristics of the people who carried the Capsian industries across North Africa.

It is, however, almost certain that the Smithfield Industry is an evolved and localised form of a Neo-anthropic group belonging to South Africa, and perhaps even confined in its evolution the Upper Orange and Vaal Rivers catchment area. Several workers — notably Neville Jones, Lowe, van Hoepen, Burkitt, and Goodwin — have at different times sought for Smithfield material in Southern Rhodesia, but so far no traces of this industry have been found north of the Limpopo, or, for that matter, beyond the confines of the Union.

Wilton material, on the other hand, is found from the Cape to the Zambesi; and to those who constantly look north for origins, this rather suggests the possibility of this industry having been the parent and the Smithfield a later variant. That this, however, is not wholly the case can only be proved after a closer examination of the industry than can be afforded in this introduction.

It is sufficient here to say that it has been found necessary to divide the Smithfield into three phases, presently designated “A,” “B” and “C.” Of these, Smithfield “A” has every appearance of being the earliest, is most probably largely indigenous, and therefore least like the Wilton and Capsian; Smithfield “B” is a variation of “A” toward the Wilton, while “C” is very like the Wilton and Upper Capsian; the whole indicating strongly convergent evolution on the part of the Smithfield Industry, while the Wilton remains fairly constant.

From the evidence presently available it seems as though we should adopt as our working hypothesis the assumption that the bearers of the Wilton Industry represent an offshoot from the Capsian Group of North Africa, and that during the southward migration they moved in successive waves that probably diverged into two main channels — one down the central zone where the Smithfield variations are so well developed, and the other down the east coast where the industry remained fairly constant. A possible third channel is down the western desert route.

So far as the Smithfield Industry is concerned, it is only those Neo-anthropic waves that came down the central zone that interest us. Here the earliest arrivals came into contact with two great moulding influences: (1) The presence of Middle Stone Age folk still inhabiting the great area drained by the Upper Orange and Vaal Rivers, and (2) the presence of a new, desirable, and eminently suitable material. In this area, liberally focussed on the Orange Free State, this new material, lydianite or indurated shale, provided a vast amount of stone with excellent fracturing qualities, and, in addition, the Middle Stone Age folk seem also to have given the invaders certain new ideas as to stone implements. It is to this first contact between two entirely different cultures that we must perhaps look for the beginnings of the Smithfield Industry, just as we must look to later waves and later contacts for the changes in the convergent evolution above referred to. Thus we see that although the Smithfield may represent an offshoot from the purer Neo-anthropic Wilton, it is indirect, and it is not to the latter alone that we must look for possible beginnings or “parentage.”

Most typical of these Later Stone Age Industries — Smithfield and Wilton — is the flat striking platform, the line longitudinal parallel flaking, and the even, steep secondary trimming.

The Smithfield Industry in the Orange Free State.

In his chapter on chronology in the Handbook of Aboriginal American Antiquities, Holmes says:71

This statement so completely and effectively describes and reflects the state of affairs in South Africa that the words might, with absolute application, be substituted for America wherever the latter occurs.

The confusion that arose out of the adoption or adaptation of European classifications and nomenclature to suit local needs was such that in 1925 the writer [C. Van Riet Lowe] first strongly urged the necessity for a complete break-away from the European school.72 Meanwhile Mr. A.J.H. Goodwin, senior lecturer in Social Anthropology at the University of Cape Town, was also working in this direction, and the ultimate break-away is undoubtedly due to his larger influence.

At the annual session of the South African Association for the Advancement of Science, held in Pretoria in July 1926, the founding of this new system of describing local prehistoric periods was accepted by all local prehistorians, and it was decided definitely to abandon the direct use and application of European terminology.

The following table gives a broad outline of the scheme, the Palaearctic counterparts being quoted for comparison:

South Africa			European and North African Counterparts
Age	Industry
Earlier Stone Age	Stellenbosch		Chellean
	Victoria West		?
	Fauresmith		Acheuleo-Mousterian
Middle Stone Age	A variety of industries with Mousterian affinities. Incompletely studied.		Mousterian
Later Stone Age	Smithfield	“A”	Capsio-Aurignacian
		“B”
		“C”	Upper Capsian
	Wilton

From this table it is seen that one of the periods of the Later Stone Age — Upper Palaeolithic in character — was named the “Smithfield.” Mr. Goodwin suggested the name, not only on account of the fact that at that time the best assemblage of artefacts of this period was from Smithfield, but also to honour and commemorate the pioneer work of the late Dr. Kannemeyer in that area.

The term implied and embraced a lithicultural group that was distinctly recognisable and seemingly well defined at the time, but one that, it was later found, called for closer scrutiny.

But for the recognition of certain Capsian features,73 it was impossible to establish either European or North African parallels, for among the implements that belonged definitely to the Smithfield were types that were fairly, if not wholly, typical of the various periods that constitute the Palaearctic Upper and Epipalaeolithic Ages.

Historical.

Working in the Riet River Valley (see accompanying map, Plate XXXVII) toward the close of 1926, the writer [C. van Riet Lowe] first discovered at De Kiel Oost No. 101, ten miles north of Koffiefontein on the Koffiefontein-Jacobsdal main road, not only the possibility of, but also the necessity for, dividing the Smithfield Industry into two clearly deined lithi-cultural groups, and suggested a Lower and an Upper phase,74 the time sequence of the industrial groups being indicated by a marked variation in incrustation and patina between the different and differing assemblages.

The following extracts from a letter to Goodwin, dated 16^th October 1926, are therefore interesting:

This was the beginning of the breaking-up of the Industry and a sharp look-out was kept for occurrences of concavo-convex and circular scrapers. It was intended to make a special study of associations. By good fortune, the discovery of other and entirely separate Lower Smithfield sites soon followed, notably at Lockshoek, No. 191, (fig. A), and at Blaauwheuwel, No. 425, district Fauresmith. These sites made it immediately apparent that there was a difference, and showed also that where the Lower Smithfield stations were always in the open and apparently confined to the south-western portion of the Orange Free State, the Upper included both open and cave sites that existed over the entire Upper Orange and Vaal Rivers catchment areas, i.e. including the whole province shown on the accompanying map and extending considerably beyond its borders. In the course of his field-work, the writer [C. van Riet Lowe] gradually came to recognise such variations between the great majority of Upper Smithfield cave and open sites, that the necessity for grouping the former under a different though distinctly relevant sub-head arose. Obviously the cave-Smithfield belonged to the same culture, but such typological differences and differentiations existed, that it became essential to establish another sub-group.

Meanwhile, however, owing to the complete absence of stratification between the so-called Lower and Upper phases, and after discussion with Mr. Neville Jones, it was decided to name these Smithfield “A” and “B” respectively.

Throughout this period of more or less intensive fieldwork, I was in constant communication with Mr. A.J.H. Goodwin, and it is worthy of special note that toward the close of 1927, both Goodwin and I found it necessary, simultaneously and independently, to subdivide the Smithield into at least three groups. In a letter to Goodwin, dated 21st October 1927, I wrote:

Our decision, after a great deal of correspondence and fruitful interchange of notes and ideas, to avoid the use of the terms Lower, Middle, and Upper as applied to these now distinct lithicultural groups, was, and still is, due to an entire absence of stratification and the consequent impossibility of dennitely establishing sequences or time horizons.

It would, however, seem that the earliest manufacturers of Smithfield implements appeared in Southern Africa a very considerable while ago and here evolved — in a variety of ways: (1) either independently to produce a sequence of types, or (2) by infiltration and a series of contacts with the manufacturers of the more Neo-anthropic Wilton types — three distinct and partially contemporaneous groups.

That the three groups — “A,” “B,” and “C” — are most intimately interrelated and that two of these — “B” and “C” — are closely related to the Wilton is abundantly clear, and although the surface and general indications lead one to visualise a straightforward evolutionary process — “A” to “B” to “C” — it is as yet quite impossible to say not only whether or how this took place, but also what part the Wilton played.

It is, however, my considered opinion that the earliest Neo-anthropic arrivals found certain areas of the Riet River Valley still occupied by men of an earlier period, and that the first contact gave rise to a “fusion” that marks the beginnings of the Smithfield Industry. Certain aberrant, though often improved, types of implements appear in “A” that also appear in the most highly developed Fauresmith assemblages, and it would seem as though there occurred a partial overlapping of two otherwise entirely separate and distinct industrial groups and periods. It is, however, much more probable that, as a result of the first Neo-anthropic wave, “A” is an indirect offshoot from and an improvement on a not yet fully appreciated Middle Stone Age that either evolved from or, by infiltration, came into contact with and borrowed from the Earlier. The Middle Stone Age has, for lack of adequate field data, been imperfectly studied, and the scanty nature of occurrences in the Orange Free State gives rise to complexes that are extremely difficult to unravel.

It may be taken, however, that the Smithield Industry as a whole does form a distinct and composite, though in itself somewhat complex group. The rise of “A,” which heralds the dawn of the Industry, is obscure, but does appear to represent a contact and partial fusion between the earliest Neo-anthropic wave and the indigenous Middle Stone Age folk. All we can say with certainty is that the heavier points and scrapers (end and side) that occasionally occur in the best Fauresmith and Middle Stone Age assemblages reappear in an improved form on unadulterated “A” sites; that certain other Smithfield types, as will be shown later, reveal also Middle Stone Age or Mousterian-type influences, and that rare and/apparently “foreign” Wilton specimens are also occasionally found. This may be due to partial contemporaneity, overlapping and borrowing one from the other, or it may indicate a lithicultural interrelationship that resulted from, so far as “A” is concerned, contact between the earliest Neo-anthropic wave or waves and Middle Stone Age folk.

Until such time as we find complete stratification, the Middle Stone Age must remain imperfectly understood, and until it is understood, the problems that now confront us must remain. On the other hand, we have definite stratification to prove that Smithfield “B,” which is less like the Middle Stone Age and much more like the purer Neo-anthropic Wilton, followed long after both the Middle and Earlier Stone Ages, and, despite the absence of stratification within the Smithfield proper, the greater age of “A,” when compared with “B” and “C,” is clearly manifested inthe weathered and incrusted surfaces.

As one would naturally expect, there is, apart from occasional overlapping and indications of partial contemporaneity, a distinct shading off of “A” to “B,” just as it is equally obvious that, despite recurrent overlapping, there is a shading off of “B” to “C.” Unadulterated “A,” “B,” and “C” sites are known, but this overlapping of “A” and “B” and of “B” and “C” — not, so far as we know, of “A” and “C” — is frequent, and any absolute demarcation is a matter of extreme difficulty.

As it is impossible not only to describe, but also to appreciate the subdivisions of the Industry until all the implements and occurrences have been described, we shall first turn to the necessary descriptions.

Description of Implements.

Duckbill End-scraper.

By far the commonest and most typical implement, and one that appears throughout the three groups, is a small rectangular end-scraper known from its shape as a “duckbill.” This invariably is a slender flake from the outer face of which have been struck two to three long fluting flakes, these latter having been removed before the whole was struck from the core. The striking platform is untouched and therefore flat. In the normal or typical specimen, the end remote from the platform and bulb is trimmed by the removal of a number of steep, small flakes struck from the underside or flake-surface until the end, being secondarily trimmed, acquired the shape and appearance of a bevelled arc (Plate XXV, fig. 1).

From factory-site debris it is quite clear that the manufacture of a normal duckbill comprised:

1. preliminary shaping of the core by the removal of light fluting flakes from the edge selected;
2. the removal of the main flake by a blow struck immediately behind these fluting flakes (Plate XXIV); and
3. the trimming of the flake so removed, a flake of necessity slender if the workmanship was accurate (Plate XXIV).

The trimming of the core by the preliminary removal of fluting flakes was a straightforward process that does not call for description.

The removal of the main of primary flake was achieved by means of a detaching-hammer, apparently by direct rest percussion, i.e. the core was held at rest on the ground in one hand as shown in Plate XXIII, and the detaching-hammer in the other. By constant use the latter gradually became a faceted spheroid or polyhedral stone, averaging about 3 inches in diameter (Plate XXII, fig. 1). Cores and detaching-hammers are common and, in passing, it may be mentioned that these detaching-hammers are indistinguishable from those of the other and older industries. Flakes suitable for secondary trimming were then selected from those so struck and the process of further (secondary) trimming commenced.

The trimming of the flake. — To achieve the finer and more accurate workmanship required in secondary trimming, it is essential that the fabricator have an angular working edge, the angle to be as near 90° as possible. Detaching-hammers are too irregular and round-edged, and the manufacturer therefore discarded both these and the core for the trimming-stone and anvil. The latter was any large fixed or movable stone with a flat upper surface bounded on one side by a right- or acute-angled edge, and the trimming-stone (Plate XXIII, fig. 2) a heavy flake struck specially for the purpose, the working edges of both being — after use — considerably bruised and step-flaked.

The flake to be trimmed was held on the anvil with its flake surface uppermost, the end to be trimmed projecting slightly over the right-angled edge of the anvil below (Plate XXIV). By this indirect rest-percussion method, and under short, sharp, and precise blows from the trimming-stone, the implement rapidly acquired a bevelled edge across its end and emerged as a small rectangular chisel-like tool perhaps an inch or two long and half an inch or an inch wide, the length in all cases being greater than the width (Plate XXV, fig. 1).

Variations of this simple type occur. Occasionally the sides as well as the end remote from the bulb are trimmed (Plate XXV, fig. 2), or the fluting flakes from the back are replaced by a keel (Plate XXV, fig. 3). In very rare cases, both ends are trimmed and we have a double-ended scraper (Plate XXV, fig. 4).

Apart from the double-ended scrapers, which are known to occur in “A” and “B” only, these duckbills appear in all three phases, the only differences being in size and workmanship. In “A” they tend to be large, though normal types and sizes do appear; in “B” the size is entirely normal and the retouch surer than in “A”; in “C” the implements not only tend to be still smaller, but incline rather to a thumbnail shape, so much so, as a matter of fact, that thumbnail shapes far outnumber duckbills.

Circular Scrapers.

A sub-variety of the duckbill is the circular scraper. This was manufactured in an exactly similar fashion, but the secondary trimming is more elaborate, and the whole, instead of being angular in plan, is circular (Plate XXVI, fig. 1). The worked edge forms a continuous circular arc beginning at one side of the striking platform and ending at the other. The platform is always left.

Large circular scrapers that fluctuate about the size of crowns, 60 mm in diameter, are abundant in “A,” but do not appear either in “B” or “C.” I have only found one circular scraper on a “B” site, Aliwal North ●75; it is about the size of a shilling, some 25 mm in diameter.

Concavo-convex Scrapers.

The concavo-convex scraper, sometimes referred to as an axe-edged implement or chopper-scraper, is a particularly interesting tool inasmuch as it shows an unusual mastery of stone technique.

It consists of a flake trimmed on the outer face by a single blow and then removed from the core by a second blow struck immediately behind the first, the resulting flake thus having a negative bulb on one face and a positive on the other. It is therefore bounded by two flake surfaces, and in shape and appearance is concavo-convex (Plate XXVI, fig. 2).

Across the width of the flake opposite the striking-platform, and invariably on a slight curve, the implement is trimmed to a bevelled edge, so that, in plan, it is roughly isoscelene or forms the segment of a circle with the two bulbs, the negative and the positive, both at or near the centre of the circle. The two confining or outer radii are normally unworked, but the base or are is always trimmed to a bevel.

Two types appear: In one the radii to the outer corners and mid-point of the trimmed arc are all roughly equal, and the shape approaches that of an equilateral triangle (Plate XXVI, fig. 2d). In the other, the line from the centre (bulbs) to the mid-point of the arc is about one-third of the true radii from the bulbs to the outer edges or ends of the arc, and the finished shape is that of an obtuse-angled isosceles triangle (Plate XXVI, fig. 2c).

The concavo-convex scraper appears in “A” only, and is entirely and definitely absent from both “B” and “C.”

Rare occurrences of this type are, however, known in the best Orange Free State Fauresmith assemblages, and it is this contributory factor that leads one to believe that the earliest Smithfield arrivals found men of an earlier period still in occupation of the Riet River Valley. It would seem that the morphology of the implement is too definitely characteristic for its occurrence on two entirely different lithicultural horizons to have been quite fortuitous.

Trimmed Points.

Normally, these are simple flakes struck from a core and then trimmed along the two bounding lateral edges to the shape of a slender acute-angled isosceles triangle; the process of manufacture being similar to that resorted to for duckbills, i.e. the secondary trimming is all on and from one side, and the flake surface and butt are untouched (Plate XXVII, fig. 1).

The workmanship and symmetry of these are such that their use as side-scrapers only is suggested. In instances, however, the secondary trimming is so symmetrical, and has been so carefully done, that it would seem as though working points were aimed at (Plate XXVII, fig. 2). Lastly we come across true Still Bay types, i.e. points worked on both sides in Solutrean fashion (Plate XXVII, Hg. 3).

From ●81, the Wepener “B” factory site, is the lanceolate type (a) carefully worked all over one face, but only slightly worked on the other, while from ●80, the Mook “B” factory site, is the fragment (b) carefully worked with fluting flakes over both faces. Another site that has yielded a similar point is ●55, Hagenstad. These are the only known cases, and it must therefore be emphasised that the occurrence is exceedingly rare. Also it cannot yet be said that these Solutrean-type points definitely belong to the Smithfield Industry. The side-scraper types and those points worked only on one face, however, do belong, but are commoner in “A” and “B” than in “C.”

From the Smithfield “A” factory site at De Kiel Oost No. 101 are two trimmed points of particular interest and importance. The first (Plate XXVIII, fig. 1) is broken, but has on one surface artificially incised or engraved lines. From the nature of the lines it would seem that they once extended beyond the surface here shown and that they were cut when the flake was still fresh, for microscopic examination shows no difference in the weathering in and out of the lines. This occurrence is of particular interest because, as will be explained later, it seems that the art of rock engraving in the area under consideration was practised by the manufacturers of “A” implements.

The second (Plate XXVIII, fig. 2), also from the De Kiel Oost “A” site, is an ordinary trimmed point with a keel and heavy, beaked point from the end of which, and down the length of the tool, a long fluting flake has been struck by a blow at the point. The point is therefore rather like the prow of a ship — the flake surface being the deck — sharp-nosed, steeply “backed,” and keeled. Have we not perhaps here the tool used for rock engraving? Some sort of artefact must have been used, and, apart from this specimen, we have nothing that is even suggestive of the ordinary graver or burin of Palaearctic regions. Although its exact purpose is naturally obscure, and despite the fact that orthodox gravers have been found in Southern Africa,75 the suitability of this point as a graver is undoubted, and I therefore venture to figure it as a graving tool. The notes under Smithfield “A” that follow later are, apropos of this, illuminating.

Serrated Scrapers.

This is a variety of circular scraper that appears to belong rather to “A” than to “B,” and one in which we again see a skill in, and unusual mastery of, stone technique. The type is rare, and the great majority of those so far found are from “A” sites; only a few specimens have been recovered from “B” assemblages; none from “C.” Indications certainly lead one to believe that, essentially, it belonged rather to “A” than to “B.”

A typical specimen is shown in Plate XXIX, fig. 1. The whole is worked on a flake, and, although the flake surface is entirely untouched, the obverse shows an intensity of secondary trimming that is exceedingly rare in this Industry. We have (i) the general shaping of the primary flake by the removal of flakes struck off the back, most probably by direct free-hand percussion; (ii) the preparation of the bevelled edge by indirect rest percussion as in the case of an ordinary duckbill, i.e. with a trimming-stone and anvil; and (iii) the removal of small fluting flakes at regular intervals round the working edge, flakes removed, it would seem, by pressure or, at least, indirect free-hand percussion, i.e. the removal of the small, regularly pitched fluting flakes from the implement held in one hand, by a punch held between the fingers of the same hand and striking the punch with a hammer-stone held in the other. The resulting edge is rather like that of the sickle flints described by Miss Oaton Thompson in The Neolithic Industry of the Northern Fayum Desert.76 The regularity of the “wave-lengths” or pitch of the serrations is most striking, and there can be no doubt that some specially precise method of flaking must have been resorted to.

Occasionally these scrapers are simple, flat flakes with no appreciable secondary trimming beyond that required for the serrations.

Notched Scrapers or Spokeshaves.

These are flakes on which the secondary trimming is confined to a shallow notch or notches (Plate XXIX, fig. 2). No occurrence is known in “A,” only a few have been found in “B,” while in “C” they are common. Occasionally the notch appears in the side of another implement, such as a duckbill.

Stone Borers.

Invariably these are heavily though neatly trimmed points 5 to 7 inches long and triangular in cross-section (Plate XXX, figs. 1 and 2). An average specimen would be about 14 cm long and 3-5 cm across the widest section of the butt, whence it tapers to a point at the end. These tools were used for pecking and abrading the holes through bored stones. The initial pecking process damages the point and the abrading wears down the angular and sharp edges to such an extent that the point ultimately becomes wholly circular in cross-section.

Occasionally naturally pointed stones were also used as abraders (Plate XXX, figs. 3 and 4). In such cases there is little or no secondary trimming — just a pointed stone worn circular in cross-section and conical over the length used for abrading.

These tools occur throughout the Industry, though the best and most elaborately worked are undoubtedly from Smithield “B” sites. The four specimens shown on Plate XXX are from the type “B” station ●27, Avalon, Riet River Valley, O.F.S.

Bored Stones.

These artefacts occur in all three phases, but it is to “B” that we must look for the best, for it was undoubtedly during this phase that intense specialisation in their manufacture took place.77 Those associated with “A” are large and coarse, and no specimen with a polished exterior is known from this phase. Also, as in “C,” they are markedly less common than in “B.”

The commonest type is the spheroid with hour-glass perforation (Plate XXXI, fig. 1). Apart from subvarieties, there are two main types: the digging stone and the ring.

The Digging Stone.

This includes four conventionalised types:

1. the spheroid (artificially rounded);
2. the ovoid (either artificially or naturally rounded);
3. the pyriform (artificially rounded); and
4. the pseudo-rectangular (artificially shaped).

In the same material the spheroid varies considerably in size and weight. Specimens ranging from 8 kg and about 23 cm in diameter to 35 g and 3·8 cm in diameter are known. The perforation is usually hour-glass, but cylindrical holes are known. Of the hundreds of bored stones I have found and examined, only one has a complete hole bored through from one side only. The general practice was to start boring at both ends at the same time, the hole being always medially constricted when fresh, though after much use the constriction becomes less marked and the tendency is toward a cylinder. In the specimen bored from one side only the diameter of the hole at one end is 30&mbsp;mm and at the other 6 mm, and it is entirely conical. The over-all diameter of the stone is 7·5 cm; the material, soapstone. The softness of the stone probably influenced the maker to work through from one side only.

But for their shapes, the ovoid and pyriform types (Plate XXXI, figs. 2 and 3) are similar to the spheroids. Naturally rounded stones were seldom used, the vast mass having been artificially shaped throughout.

The pseudo-rectangular (Plate XXXI, fig. 4) appear to have been ordinary spheroids worn cube-shaped by use as grinders. There are either two or four grinding surfaces, usually four, not so far as I know either one or three. At the centre of the grinding surface we often find the small concavity so characteristic of the grindstones of this period. (See notes under Grindstones.)

It is generally presumed that these bored or digging stones were used as make-weights for digging sticks, and, although we know that this was largely the practice among the aboriginal San folk, we cannot, with the freest rein to our imagination, assign all those known to this category. A 35 g specimen would not add appreciably to the weight of any stick that might be used for digging.

In cave paintings, definitely associated with “C,” bored stones appear as club-heads as often as they appear as make-weights for digging sticks, so that at least we have two uses to which they were put (Plate XXXII). My own opinion is that, in addition to utilitarian needs, they also had certain ceremonial significances.

The Ring.

This is a circlet of stone that resembles an armlet or anklet (Plate XXXIII, fig. 1). The size does not vary much. It is usually about 10 cm in external, and 7 cm in internal, diameter. It is a beautifully worked and highly polished artefact oftenest made of slightly indurated shale. Specimens of pure shale are also known. In cross-section, the “body” of the ring is an acute-angled isosceles triangle with a rounded base — concave outwards.

The process of manufacture comprised (i) the selection of a suitable slab of shale, either indurated or pure, the rounding off of the outer edge by flaking, and the beginning of the perforation by pecking and chipping prior to (ii) abrading and (iii) polishing. When about one-third completed, the artefact resembles a bladed disc.

These rings, always rare — it is believed that only eleven “finished” specimens have so far been recovered — are known to occur in the “B” and “C” phases only. Interesting is the fact that exactly similar stone rings are used as arm-rings by the Tuareg.78

Grooved Stones.

These are natural boulders or pebbles, either fixed or movable, in which hollow-ground grooves of varying sizes and shapes occur. It would appear as though there are four types.

The first may be classed as a tool-sharpener or whetstone. Either fixed or movable, it has on its surface grooves in various directions and of varying sizes. These grooves are often striated and were obviously worn by having had points rubbed in and along them, the points, most probably, being of bone.

The second is a bead-stone (Plate XXXIV, fig. 1). This is a small water-worn pebble of sandstone, usually about 10 cm long, with one or more perfectly U-shaped grooves that measure about 5 mm wide and 3 to 5 mm deep. The groove is straight and fits the characteristic ostrich egg-shell bead of average size. The regular choice of sandstone, invariably friable, is significant.

It is probable that this stone was also used as an arrow-straightener, for by heating it and passing a green stick or reed up and down the groove, the tendency would be for the shaft to dry out in parts and so enable the artificer to straighten out any irregularities.

These deductions on usage are based on actual observations made among the surviving San tribes. Such artefacts as these are still in use, and the occurrence of identical specimens on Smithield factory-sites and settlements is rather indicative of similar usage in prehistoric days. Or perhaps it would be more correct and more descriptive to say “in the past,” for the prehistory of Southern Africa, anomalous as it may sound, is certainly not all prehistoric, and it is my considered opinion that both the “B” and “C” phases of the Smithfield Industry were extensively practised in this country long after European invasion and settlement. It is because of this that one feels entitled to make deductions such as the above.

The third type is known as a poison-stone. Here the groove is broad and shallow-perhaps 12 mm across and up to 5 mm deep. Again from observations among recent and surviving San folk, it would appear that the poison used on arrow-tips was poured into or smeared in this groove and the arrows to be poisoned worked up and down or round and about in the poison in the groove.

In the fourth type, the groove or grooves are again perfectly U-shaped and similar in size to those in the bead-stone, but in longitudinal section the groove is an arc, and the direction of the groove often oblique to the axis of the stone (Plate XXXIV, fig. 2). This obliquity of the groove, and the fact that, longitudinally, it is in itself not straight but curved, suggests that the artefact was held in one hand and worked or pulled up and down an insecure object such as a raw-hide thong — much as a cobbler often uses his wax.

One is inclined to regard this tool, which is always of soft, friable sandstone or shale, as a brayer used for cleaning, stretching, and shaping leather or gut thongs — perhaps bow-strings — the leather, of course, being raw-hide and wet.

Grooved stones occur throughout the series, and the poison and braying stones, so far as we know, in “B” and “C” only — though I see no reason why they should not appear in “A” also.

Grindstones.

Upper and lower grindstones are found throughout the Smithfield series, and a striking feature of both the lower and upper stones is that they are invariably made either of dolerite or of fine-grained tough sandstone. Where partially indurated shale has been used, the grindstones are always small, rather like palettes. The upper stone is usually a ball of dolerite, the characteristic weathering of which is spheroidal, or a river pebble, one or more faces of which have been rubbed down by use to a flat or very slightly convex surface. Typical of the upper-stones is the occurrence of a small, pitted hole or concavity about 1 cm across and 3 or 4 mm deep at or near the centre of the grinding surface.

Mr. Goodwin tells me that this pit is still made in the grindstones of surviving San folk, and says that it serves to catch up and turn over the tsama and other wild seeds they grind. The perfectly smooth grinder will not “catch” on the seeds otherwise, as the material being ground forms a smooth paste on the lower stone which resists grinding.

These upper stones are of a size which implies that they were held in one hand, and many were used as pounders as well as grinders, for where on one face, usually at one end, we have a smoothly rubbed and polished surface, we have on another, usually at the opposite end, a roughly abraded and pitted area that indicates the fact that the stone was used as a pounder or pestle too.

The lower grindstone is usually a flat slab perhaps 30 cm square, and has a shallow groove some 5 to 8 cm wide and up to 2 cm deep, the groove running almost the full length of the face.

Artefacts with Ground Edges.

Apart from the artefacts referred to above where pecking, abrading, and polishing appear, the Smithfield Industry has no parallel with the orthodox Neolithic of Palaearctic regions. The makers of Smithfield implements never, so far as we know, attempted the typical Neolithic axe or celt, and, with exceedingly rare exceptions, none of their implements show that they ever attempted to make a true cutting or chopping edge by polishing. One of the exceptions referred to is a thin slab of sandstone one edge of which has undoubtedly been ground to a cutting edge (Plate XXXIII, fig. 2). This is from the type “B” station ●27, Avalon.

Mr. Heese has figured ground and polished fragments from Britstown, Cape,79 not unlike the above, and these, with a few other similar finds, constitute our nearest approach to cutting implements, and are the only types that it has been possible so far to associate with the Smithfield Industry.

Bone Tools and Ornaments.

Tools.

It has not yet been discovered whether bone was definitely used in “A” or not, but its use is certainly indicated by the presence of grooved stones of type (a), the tool-sharpener. In “B,” however, unconventionalised bone points do occur, and that figured in Plate XXXV, fig. 1, is from ●81, the Wepener “B” station. In “C” they are common (Plate XXXV, fig. 2). These points are merely unconventionalised awl-like tools, first scraped and then rubbed to a point at one end, the other being left untouched, small straight bones of birds and animals being used.

In the Cape Province, however, Goodwin has been able to associate conventionalised bone arrow-tips with Smithfield remains. These conventionalised types are: (a) a point that consists of a bone rubbed to a pencil shape, about 10 mm thick and 80 mm in length. The one end is pointed, the other squared off to a platform at the butt, the platform being about 4 mm in diameter. The shape is thus very much that of an elongated torpedo, with one end cut oil square; (b) the other type is thinner, about 6 mm thick and 80 mm long, pointed at both ends and usually more shapely than the other. The use of similar points by the surviving San folk in South West Africa has frequently been described.

Ornaments.

Plate XXXV, fig. 3, shows a small artificially pierced fragment of bone, apparently used as an ornament. This is from ●103, Ventershoek “C” site. The perforation is hour-glass.

Interesting in this connection is the fact that Neville Jones figures an exactly similar specimen from a Rhodesian Wilton site.80

Ostrich Egg-shell.

Utilitarian.

Extensive remains of ostrich egg-shells on all Smithfield settlements indicate their use throughout the Industry, but the only utilitarian need that the shell proper is known to have served is that of a receptacle for storing and carrying water. One end of the shell was opened and the contents emptied out. The opening was then trimmed to a neat circular shape and the sides near the hole bored for the “strings” from which the egg was suspended. These receptacles were often ornamented by having ladders or a series of chevrons incised into the shell round the periphery near the mouth. All drilled holes are conical — except after considerable use.

Ornamental.

Pendants.

Fragments of shell were cut into crude ellipses and each ellipse perforated by a small hole near one of the foci. Occasionally the periphery was scalloped (Plate XXXV, fig. 7).

Beads.

Egg-shell beads are exceedingly common in “B” and “C.” The average bead is a small ring of shell with an external diameter that seldom exceeds 8 mm and averages about 5 mm (Plate XXXV, fig. 8). The hole through the centre is large and, when fresh, always hour-glass or conical. The beads were strung on gut or sinew and the “chain” often of appreciable length. From the neck of a skeleton (San) exhumed on a “B” site — ●29, Slagtkraal — I collected sufficient beads to complete a 4-foot chain.

Where it is definitely known that whole-egg water-carriers or receptacles, pendants, and beads, particularly beads, were extensively made and used during the “B” and “C” phases, we cannot be so sure that the same may be said of “A,” despite the fact that unadulterated “A” sites have yielded quantities of shell fragments. In this connection, however, it must always be borne in mind that the openness of the vast majority of sites has not been conducive to the preservation of bone and shell tools and ornaments, and in the comparative age of “A” may lie the explanation of the meagreness of the occurrence of these artefacts.

Shell, Bone, and Pottery Borers.

The implements used for piercing and boring holes in ostrich egg-shell beads, bone pendants, and pottery are actually miniature or pygmy stone-borers, on which, it may legitimately be said, the secondary trimming is barely macroscopic (Plate XXXV, fig. 9). These ostrich egg-shell borers, as they are called, are usually triangular in cross-section, taper to a point, average about 2-5 cm in length and about 4 mm in greatest diameter at the butt. It is likely, however, that tapered flakes of suitable size and shape and with suitably sharp edges were also used.

Pottery.

Sherds are found on all Smithfield settlements and factory sites, but these are so uncommon on “A” sites that it is not yet felt possible definitely to associate the art of the potter with the maker of “A” implements. In “B” and “C,” however, pottery makes a definite appearance. The pot and sherds illustrated on Plate XXXVI are from “B” factory sites — the former from ●42, Eagle’s Nest, the latter from ●27, Avalon — the type “B” station.

The holes in the fragments are conical, tapering inwards, and were obviously bored after the pot had been baked. The Eagle s Nest pot is U-bottomed, about 11 cm high, and stands in unstable equilibrium. It is the only known whole pot from the high-veld open sites.

All the pottery contains grit obtained from ground rock, was well baked but not glazed. The openings are wide, the rims always slightly everted, and the general contours and size often strongly reminiscent of those of an ostrich egg. When any design appears it is always of simple type — impressed or incised. No colouring matter was used.

In association with Smithfield remains in the Cape Province comes a type of pot of particular interest, and I am indebted to Mr. Goodwin for the following description:

In size the pot is generally large, standing perhaps 40 cm high. The neck is wide, with a slightly everted rim. Below the neck the pot bellies out to an egg-like shape, the point of the egg forming the roundly pointed base. Immediately below the neck are two lugs, mammiform in shape and situated at extreme opposite sides. These lugs are pierced from side to side, parallel with the pot wall and horizontal, by a hole about 5 mm across. In a few instances grass rope has been found passed through these lugs as though for suspension. Decoration, incised or impressed on the clay before baking, appears.

Excellent illustrations of these pots appear in Péringuey’s The Stone Age of South Africa.81 Lugs with grass ropes passed through them are to be seen in the McGregor Museum, Kimberley.

Wood.

A single wood point (Plate XXXV, fig. 4), charred at one end and suggestive of a fire-stick, is all we have. It is from the Ventershoek “C” assemblage.

Glass.

The use of modern glass for implements is a subject that has given rise to a great deal of controversy, and, although many scores of so-called glass implements are to be seen in various local museums, it would appear as though very few of these are acceptable. Among the many thousands of implements, collected from over two hundred factory sites and settlements, all of which were unknown and unworked when first visited, I have only two glass implements and one of these is doubtful. It is, however, patent that the other, definitely associated with “B,” has been artificially trimmed. This implement (Plate XXXV, fig. 5) is a thumbnail scraper from ●71, Bethulie. It is from an open site situated on the right bank of the Orange River, some three miles from the nearest habitation, and there is nothing about the locality either to attract or invite attention. I mention this because it is no picnic spot, and few civilised men, I am sure, have passed this way.

The site yielded a typical assortment of “B” types (see Site List), and among the scrapers was this one of glass, and no other glass was found at or near the site. It is a typical thumbnail scraper, rather scratched, and apparently pitted by wind-blown sand.

The occurrence, to my mind, is by no means surprising; rather is it surprising thatimore genuine glass implements have not been found. There can be little doubt that both the “B” and “C” variations were practised by the aboriginal San, and glass was introduced into this country long ere these aboriginals disappeared from the area under review. A parallel is to be found in Australia, and we should expect to find glass tools indubitably associated with the remains of primitive man.

Crescentic or Lunate Scrapers.

Among the mass of Smithfield implements collected, only two crescentic scrapers have so far been found. Strangely enough, these are both from the De Kiel Oost “A” site. Plate XXXV, fig. 11b, shows the specimen in my private collection. The other was found by Burkitt and is in his Cambridge collection. The occurrence is provocative and strange, and one tends to regard it as of extraneous origin — as though the maker of Wilton implements dropped his passport en passant.82 The nearest known Wilton site is near Kimberley, about sixty miles north-west of De Kiel Oost.

A few crescents were found in the ●48, Dieplaagte — ●54, Meerlandsvlei area, over seventy miles north of De Kiel Oost.83 These also appear in Smithfield assemblages, but the data at our disposal is insufficient to justify any expression of opinion.

An essential feature of a crescentic scraper is that it be isoscelene in lateral cross-section as shown in Section X-X, fig. 116, Plate XXXV.

Skeletal Material.

But for a single exception, all the human remains so far recovered from Smithfield settlements — “B” and “C” only — have been pure “Bush” or San.

When a skeleton is exhumed or found at or near any site it is, of course, impossible to say whether the remains belonged to it or not, but so many widely separated sites have yielded pure San remains that it appears reasonable to assume that these folk were the makers and users of “B” and “C” types.

But for the exception referred to above, indicative of ceremonial inhumation of an impure “Bushman,”84 all the skeletons so far found suggest rather a haphazard shoving away of the body into any convenient crevice or cavity either in rock or in the ground and there covering it slightly with earth.

The Smithfield “A” Industry.

Remains of “A” are fairly extensive, but few factory sites are known and these are all in the open, along or in the immediate vicinity of the Riet River Valley in the south-western corner of the Orange Free State. The appended map shows the known extent of the industry, and, although Smithfield remains are found beyond the borders of the area shown, this area has been specially selected because (i) it has been more elaborately and carefully studied than any other, (ii) we have an unusual mass of field observations and data on which to build our assumptions, (iii) the Industry is more freely and fully developed here, and (iv) the entire series — “A,” “B,” and “C” — appears in the field.

The total number of factory sites recorded is 104, and of these only five are unadulterated Smithfield “A.” The best site is ●37, on the farm Lockshoek No. 191, district Fauresmith, illustrated on the field sketch opposite (fig. A). The implements and artefacts collected here represent a typical factory-site assemblage, and include:

1. Concavo-convex scrapers.
2. Large circular scrapers.
3. Duckbill end-scrapers.
4. Side-scrapers.
5. Trimmed points.
6. Stone borers.
7. Bored stones.
8. Grooved stones.
9. Grindstones.
10. Founders and grinders.
11. Fabricators:
    1. Cores.
    2. Detaching-hammers.
    3. Trimming-stones.
    4. Anvils.

So far as the Smithfield Industry is concerned, the concavo-convex and large circular scrapers are known to occur in “A” only. Notched scrapers do not appear, and other elements that make a definite appearance in the industry, but that have not yet been found to belong to “A,” are thumbnail scrapers, ostrich egg-shell beads and their borers — though the presence of grooved bead-stones is indicative of bead manufacture. The association of pottery has also not yet been satisfactorily established, despite the fact that sherds have been found.

The workmanship generally is cruder than in the other and, in my opinion, later groups, and the implements are larger. The largest duckbill found measures approximately 110 mm × 60 mm × 12 mm, while the average fluctuates about 60 mm × 24 mm × 5 mm, which is not only above the average of the other groups but actually ranks amongst the largest specimens.

Considered as a whole, there can be no doubt that the “A” group has every appearance of having antedated both “B” and “C.” “A” and “B” sites frequently overlap — at De Kiel Oost, for example — and such occurrences tend, to the uninitiated, to be most misleading; but fortunately, in almost every such instance, the “A” implements tend generally to be more deeply patinated and more heavily incrusted than those of “B” — an exceedingly helpful fact. The overlapping of “A” and “C” sites is not known.

It is, however, only when we examine an unadulterated “A” site that the true significance of its separateness and distinct morphology are appreciated. Such a site is at Lockshoek, where a full range of “A” implements was found and where there was no admixture of implements characteristic, in the order of importance of occurrence and technique, of the “B” and “C” phases.

It was during this phase that I believe rock engraving was practised, for, apart from the almost regular contiguity of “A” sites and rock engravings, we have in typical “A” assemblages, rock fragments and one implement that have artificially incised or engraved lines on their surfaces. The implement has already been described (Plate XXVIII, fig. 1). The artificially incised lines on rock fragments, a specimen of which is shown on Plate XXVIII, fig. 3, comprise little figures of simple geometric design, mainly ladders or chevrons. The engraved fragment here illustrated was found on the De Kiel Oost “A” site and is in Burkitt’s Cambridge collection.85

The subject of rock engravings constitutes a study in itself, and it is not possible to discuss the matter in detail here. It is sufficient to note that these engravings usually depict animals. One finds on a single site: elephant, giraife, rhinoceros, eland, a variety of antelope — large and small — baboons, jakhal, meerkat, human beings, snakes, simple geometric figures, etc. The technique varies considerably, and it would seem that there are four distinct styles:

1. Profile only; in long, thin scratched lines. Deep patina.
2. Profile only; in short, stubby lines. Superimposed on (a). Less deep patina.
3. Profile only, but broad and pecked. No lines. Light patina.
4. Pecked profile with body either wholly or partially pecked in. Eye usually shown. Light patina.

Each object is depicted as a separate entity. Grouping or composition does not appear.

One of the finest examples of rock engraving known is reproduced as a double-page illustration in the Illustrated London News, vol. 173, No. 4656, of 14^th July 1928. The discoverer’s attribution of great age to this work of art, however, finds no support among local prehistorians.

For an analysis of rock engravings, I would refer my readers to Schapera’s interesting statement on certain stylistic affinities with the Capsian.86 Schapera deals with both engravings and paintings, and, quite apart from any consideration of associated or contiguous artefacts, concludes that the engravings belong to a more primitive and earlier period than the paintings — an illuminating and important conclusion.

One point of extreme interest in connection with the “A” phase is the fact that in the assemblage of material collected on the Fauresmith Industry type-station on the farm Brakfontein, No. 231, district Fauresmith, O.F.S., appear a few unusually well-developed end- and concavo-convex scrapers. Had these implements not been found so deifinitely associated with Fauresmith material, it would, despite the unusually deep incrustation, have been the most natural thing to allocate them to Smithield “A.” It is just such an occurrence, singular as it is, that inclines me to the belief that the earliest Smithfield arrivals actually came into contact with, if they did not here evolve from, a Palaeo-anthropic type.

It is also illuminating to note that the Fauresmith Industry, advanced Acheulean to Mousterian (La Micoque) in type, appears to be an offshoot from and an improvement on the Stellenbosch, 87 Chellean-type, and that in the hey-dey of its practice we find distinct Mousterian influences that recur in the Middle Stone Age and in Smithfield “A.” The heavy circular scraper with serrated edge, for example, is here suggestive of a Mousterian influence, for occurrences of other implements with serrated edges have, in South Africa, been found only in assemblages that reflect Mousterian affinities.

It is difficult otherwise to explain why this phase not only differs so from, but preceded the “B” and “C” with their more neo- anthropic types, and why the latest or “C” phase shows such a leaning towards the typically Neo-anthropic Wilton Industry.

I find it difficult to escape the suspicion that the earliest Smithfield arrivals actually came into contact with, if they did not evolve from or were influenced by, a strong but earlier local element that left its mark in the forms that survive and appear in the “A” phase only, and that these forms were discarded when the later Smithfield peoples, with their more fully developed neo-anthropic characteristics, put in an appearance.

The Smithfield “B” Industry.

This group is confined almost entirely to open sites and not only covers the entire province shown, but extends considerably beyond its borders. The best and fullest development and most extensive remains are undoubtedly to be found in the Orange, Caledon, Riet, Kaffir, and Modder River valleys — particularly the Riet. The accompanying map (Plate XXXVII) shows 92 factory sites, 35 of which are in the Riet River Valley. All of these are in the open.

The type station is ●27, Avalon, a sketch map of which is shown opposite (fig. B). An idea of the richness of this site may be gleaned from the fact that, when it was first discovered, I was able in the course of an afternoon’s work to co-ordinate and collect over fifteen hundred exhibition specimens from a section of the station that covered less than an acre. Later I added considerably to this already fine and representative assemblage, and was able to give many hundreds of specimens to various local museums and universities. It is from here that the best assortment of grindstones, grooved stones, pottery, bored stones, and stone borers comes, and, although many sites have yielded similar assemblages, none has produced such a quantity of “finished” specimens. There can be little doubt that in this area there once thrived a dense population or that it supported settlements over a long period of time.

The complete list of “B” types includes the following:

1. Duckbill end-scrapers.
2. Thumbnail scrapers.
3. Side-scrapers.
4. Notched scrapers.
5. Trimmed points.
6. Stone borers.
7. Bored stones.
8. Stone rings.
9. Grooved stones.
10. Grindstones.
11. Pounders and grinders.
12. Stone palettes.
13. Ostrich egg-shell borers.
14. Ostrich egg-shell beads and pendants.
15. Incised ostrich egg-shells.
16. Bone points.
17. Pottery.
18. Glass implements.
19. Implements with ground edges.
20. Fabricators
    1. Cores.
    2. Detaching-hammers.
    3. Trimming-stones.
    4. Anvils.

This list makes it immediately apparent that (i) concavo-convex and large circular scrapers are missing, and (ii) that thumbnail and notched scrapers, ostrich egg-shell beads and borers, pottery, bone, glass, and the art of grinding edges definitely appear for the first time.

The implements tend generally to be smaller and more wieldy, and the retouch surer than in “A.” By far the commonest implement is the duckbill end-scraper. Its numerical increase and preponderance over the other types have now multiplied in an almost ludicrous manner. I have collected as many as seventeen without moving my feet, and well remember Neville Jones’s amazement when he collected fourteen in a similar fashion.

The largest duckbills seldom exceed the average of “A,” i.e. 60 mm × 24 mm × 5 mm, and the average of “B” would measure about 30 mm. × 15 mm by 4 mm.

The entire assemblage is definitely neo-anthropic in character, and, excepting those implements that include polishing, it may rightly be said that the characteristics reflected are paralleled in the Mediterranean and European areas by Capsio-Aurignacian types, and that the general indications suggest an evolution from Smithfield “A.”

It is during this phase that the manufacture of bored stones develops into such a marked feature, and it may legitimately be claimed for “B” that it was an age of both serious and intense specialisation in these and other implements that bespeak the presence of a nomadic hunter of the great open veld. We have all the tools required to cut and shape sticks, clubs, bows and arrows, to sharpen wood and bone points, to clean and dress skins or scrape the meat off bones, to bray raw-hide thongs, to dig and grind wild roots and edible foodstuffs that it was possible to procure in the plains so frequented and favoured by these people.

The “B” phase also shows an occasional contiguity to both rock engravings and cave paintings when, in the former case, typical “A” implements are absent, and in the latter, typical “C” implements in the order of importance of occurrence are also absent, and it would seem as though the industry is part-associable with both these artistic developments.

While it is impossible to gauge the full age of the industry, it is definitely known that its practice persisted until long after the arrival of Europeans. The use of glass alone proves this, and of further interest in this connection is the fact that, toward the close of 1927, I actually came into contact with one of the very few survivors of the now, so far as the Union is concerned, almost extinct San folk.

On the farm Schaapplaats, on the left bank of the Vet River, immediately opposite ●64, Il Paradiso, about sixteen miles due west of Theunissen, district Winburg, O.F.S., was the aged “Bushman” (the popular term is used) whose photograph is here reproduced (fig. D). He was a child when his father was captured and “tamed,” and was brought up in service on the farm where his father had lived and hunted before him, and where I actually found him — civilised but slim. In addition to Afrikaans, in which tongue we conversed, he spoke an original San dialect and told of his father’s use of stone for the manufacture of scrapers for paring down and shaping wooden clubs, bows and arrows, for cleaning skins, preparing karosses and taking the meat off bones, and although the material to hand where we met was intractable — coarse sandstone and dolerite — he actually struck flakes and demonstrated the secondary trimming of end- and side-scrapers. He also explained that sharp chips of stone were used as arrow-tips, but knew of no conventionalised shape. The chips were fixed by means of a natural gum-cement and gut. He was eager to take me out on the veld to show me the plants that furnished their poison and medicine, but, most unfortunately, I had not sufficient time at my disposal to make an investigation.

But for his chin, which is insufficiently receding and small, all his features are pure “Bush”:

Height	4 feet 81 inches.
Frame	Dwarish.
Face	Triangular and fox-like. Forehead almost vertical.
Eyes	Small and sunk.
Nose	Broad, flat, no bridge.
Jaws	Prognathus.
Ears	No lobes.
Hair	Peppercorn. Twisted and wiry.
Colour	Yellowish brown.
Skin	Loose.
Feet and hands	Diminutive.
Limbs	Slender.
Back	Hollow.
Chest	Well developed.

I was able to help Miss Bleek, our noted student of Bushman languages, to visit him, and in a recent letter (10^th June 1928) she says:

I got down enough words to place the language. The difference between this speech and the Colonial Bushman’s is distinct. It is much nearer that of the Bushman of Griqualand West and Gordonia — how much I can’t tell until I am at home and can compare.

She is satisiied that he is a “Bushman,” and we may expect to hear more of him later. To me, he is a link with Smithfield “B.”

The distribution of the industry extends considerably beyond the borders of the Orange Free State, but the area shown contains the best and most fully developed remains.

The list of factory sites known, as well as the materials used, is given in the Site List.

One aspect of “B” that cannot be ignored here is discussed in a statement made two years ago, and that is the probable relationship between the Smithfield and Wilton Industries.88 That these industries are closely interrelated is beyond question, but when this earlier statement was made we had not yet recognised the great territorial extent of the Smithiield, neither did we appreciate the time covered by, nor did we know the many types included in, the lndustry. The possibility of the existence of the groups or subdivisions now described was also not known. “A” and “C” had not yet been recognised, and the Smithfield Industry referred to described “B” particularly but included also certain cave or “C” types. Now the combination of these latter groups, with no apparent relationship to any industry other than the Wilton, seemed at the time to represent a deterioration of this purer and more typically neo-anthropic prototype, but the subsequent discovery of “A,” and the recognition of this obviously earlier and probably true prototypic phase, throws so much fresh light on the problem that a revision of this earlier statement is necessary.

Where now I do not accept the idea that the “B” phase represents a degraded Wilton, I must draw attention to the great probability of Wilton influences, either pure or impure, making themselves felt on the slowly evolving Smithfield groups, and point out that actually we may be dealing with a number of neo-anthropic invasions, and that both the “B” and “C” phases may largely owe their differentiations, not only from the earlier Smithfield, but also from each other, to successive waves of more typically neo-anthropic influences, i.e. to successive waves of Wilton peoples coming down from the north. The occasional occurrences of crescentic scrapers and of points worked on both sides — Solutrean-type — so characteristic of the Rhodesian Wilton89 in typical Smithfield assemblages rather supports the suggestion.

Whether the incoming and migratory Wilton remained pure or not, it is impossible to say, but the probability is that certain changes did take place. That the Wilton does vary we know, for, whereas bored stones are exceedingly rare in Rhodesia and have not yet been definitely associated there with the Wilton, they do occur in the Wilton in the Cape, and the significance of this, to my mind, is that the Cape Wilton represents the re-emergence of migratory movements after the passage through and contact with the zone of Smithfield influence.

The Smithfield “C” Industry.

This is essentially the cave phase of the Smithfield Industry and, as such, includes certain new and characteristic features that appear only in the other groups, when they do appear, in positions of considerably less prominence. The most marked differences are: (i) the implements generally have diminished in size and the workmanship is finer; (ii) that, whereas in “B” the duckbill end-scraper is the most common, in “C” it is almost entirely replaced by the thumbnail (Plate XXXV, fig. 6); (iii) notched scrapers, pottery, and bone points are common; (iv) cave paintings are now a regularly associated feature; and (v) the concavo-convex and large circular scrapers of “A,” as well as the rock engravings of this earlier phase, are now definitely absent.

The best station is ●103, Ventershoek, on the Basutoland border, near Wepener, O.F.S., shown on the field sketch overleaf (fig. C). The site is in a cave about a quarter of a mile off the Wepener-Mafeteng main road, and most of the implements found were in the debris from and talus immediately below the cave. The walls are richly decorated with paintings-typical “Bushman paintings” — and here, too, is one of the finest panels I have ever seen. On an area of about 9 square feet we have depicted a scene that is best described as “The Raid.” The work is in polychrome, black, browns, reds, yellows, and white, and shows a party of “Bushmen” (red ochre) chasing off a head of cattle (piebald-in all the colours mentioned above), while another body of “Bushmen” (red ochre), armed with bows and arrows, is fighting a rearguard action against a horde of pursuing Bantu (black, but dressed in skins and ornaments in various colours). Arrows (white) literally rain on the oncoming Bantu, and the entire effect is a masterpiece of realism and action.90

A complete assemblage of implements and artefacts of “C” comprises:

1. Thumbnail scrapers.
2. Duckbill end-scrapers.
3. Side-scrapers.
4. Notched scrapers.
5. Trimmed points.
6. Solutrean-type arrow-heads (?)
7. Stone borers.
8. Bored stones.
9. Grindstones.
10. Pounders and grinders.
11. Grooved stones.
12. Ostrich egg-shell beads.
13. Ostrich egg-shell borers.
14. Pottery.
15. Bone points.
16. Painting materials.
17. Fabricators
    1. Cores.
    2. Detaching-hammers.
    3. Trimming-stones.
    4. Anvils.

The trimmed points, side and notched scrapers are identical with those of “B,” the latter occurring much more frequently.

Bored stones and their borers become less common, but the polishing of the former is of a higher quality. Pottery and bone become much more common and occur now as regular and important features. Quantities of ochreous painting materials are found, and we are able definitely to say that the cave paintings were done by the makers of “C” implements.

The tanged arrow-heads shown on Plate XXXV, fig. 10, are from the vicinity of ●97, Modderpoort, on the right bank of the Caledon River, near Ladybrand, O.F.S., and, although I have made exhaustive searches in and near the vicinity of these finds, I have found only Smithfieldian remains, and I it does not seem impossible, though admittedly the idea is incautious, that these are Smithneld. They are the only three known, and I record the occurrence as one unusually problematical and interesting. The case, however, may be a parallel with the De Kiel Oost crescents.

It would appear as though the practice of this phase was confined entirely to the mountainous areas of the north-eastern Cape, the eastern Free State, Basutoland, and the western borders of Natal, with possible extensions along the Drakensberg to the north and south of the areas described. The best development undoubtedly took place in the Caledon River Valley, i.e. along the border of the Orange Free State and Basutoland. The list of known factory sites and materials used is given in the Addendum.

The question of materials to hand is important and a study of the data in the Site List illuminating. When we review the entire Smithfield range, we immediately notice the decrease in sizes of implements: “A” to “B” to “C.” The ubiquitous duckbill end-scraper of “A,” for example, fluctuates from 110 mm to 30 mm in length; in “B” from 40 mm to 20 mm, and in “C” from 30 mm to 10 mm. The workmanship becomes finer and, on the whole, the retouch surer. It is probably this decrease in size that gives rise in “C” to the characteristic thumbnail scraper, for this actually is a diminutive duckbill with a tendency to fan out over the trimmed end. Another influence, apart from the probability of neo-anthropic waves from the north, and an important one on this tendency to decrease, lies in the material to hand. Where in “A” and “B” lydianite was almost exclusively used, in “C” it very seldom appears. The makers of “C” implements used chalcedony, chert, quartz, agate, jasper, and similar flint-like materials — collected in quantity in the rivers and streams that drain the volcanics of the Drakensberg Mountains (see Geological Notes). All this material was obtainable in small sizes only, chiefly as water-worn pebbles, and primitive man was therefore obliged, once he had taken to the mountains where there is almost a complete absence of lydianite or indurated shale, to be content with small flakes.

For reproductions of rock or cave paintings of this phase, I must refer my readers to the remarkably fine reproductions in Miss Helen Tongue’s Bushman Paintings.91 It is also anticipated that the Carnegie Institute will soon publish Stow’s equally fine collection. Miss Bleek is presently busy preparing this for publication.

Stratification.

When the term Smithfield was originally invented, it was intended to imply and describe only that industry here referred to as “B,” but this paper, it is hoped, makes it clear that the results of later and more elaborate fieldwork oblige us to recognise at least three closely interrelated, but distinct, industrial and lithicultural groups. The term Smithfield Industry now includes the subdivisions herein referred to as “A,” “B,” and “C,” these possibly being the Lower, Middle, and Upper phases.

So far as time-sequence within the Smithfield proper is concerned, it has been shown that the only fairly definite sequence is that “A” has every appearance of having preceded both “B” and “C,” whereas it is more than probable that these latter phases were largely, if not wholly, coeval *mdash; “B” in the open and “C” in the caves. It would also seem as though they flourished contemporaneously with the Wilton. The affinities between “B” and “C” and the Wilton are astonishingly marked. This applies most particularly to the cave or “C” Smithfield and the Wilton, for it is almost correct to say that Smithfield “C” is a crescentless Wilton. This typically neo-anthropic Wilton, so strongly reminiscent of the Upper Capsian of North Africa, is immediately recognisable from “A” and largely so from “B,” but it is only just separable from “C.” The vast majority of “C” Smithfield and Wilton sites are under rock-shelters or in caves, and both are definitely associable with cave paintings of Eastern Spanish style, and the fact that they are closely allied is irrefutable, but just how it is presently impossible to say. My personal leanings are toward the “successive wave” theory, i.e. successive waves of typically Neo-anthropic (Wilton) people moving southwards and there influencing the already “settled” communities.

So far as Southern Africa is concerned, Rhodesia, in my opinion, is the real home of the Wilton Industry, despite the fact that equally typical, though occasionally somewhat altered, remains are found at the southern extremity of the continent. It is possible that most, if not all, the Cape sites represent the remains of those migratory waves that succeeded in penetrating the steppes of the high-veld and ultimately “settled” farther south.

En route, however, certain “contacts” were made and the migration was marked by apparently extraneous characteristics and influences left in its trail. Thus “A,” for example, may have been slowly evolving into “B” when one of the earlier “waves” made contact and, in certain areas, directed this evolution. And what more likely area can we find than the Orange Free State with its water-courses bounded in the west by the arid and waterless deserts of the Kalahari and in the east by the great chain of Drakensberg Mountains? It is otherwise difficult to account for the presence of what actually are extraneous elements in the Smithfield — occasional crescentic scrapers and Solutrean-type points of Wilton type — and for the occurrence of bored stones within the Wilton in the Cape.

Another and perhaps later wave made contact with an already changing “B” in and near the more mountainous areas of the Caledon Valley in the eastern Free State and so gave rise to “C.” But it is impossible to theorise on the meagre archaeological evidence at present at our disposal. A great deal has still to be done before our problem is satisfactorily solved. The possibilities of degeneration must also be borne in mind. I emphasise this, because the physical degeneration of the aboriginal San folk, long ere they disappeared from the area under review, is an accepted fact.

A thorough study of primitive art, particularly cave paintings, will throw much light on the subject and, incidentally, perhaps also provide a clue to routes followed.

Although we have no stratification within the Smithfield proper, and cannot therefore establish chronological horizons, we have a mass of evidence to prove that the “B” phase followed long after both the Earlier and Middle Stone Ages. Apart from the great differences in patina and incrustation already referred to, the following are the two most convincing and illuminating instances:

1. At ●56, Floris, O.F.S., we have a “B” surface site overlying a 7 to 9 feet sterile stratum under which occur implementiferous gravels with Middle Stone Age — Mousterian-type — remains. The accompanying sketch (fig. E) shows a cross-section that clearly demonstrates the state of affairs.
2. At ●36, Lockshoek, No. 191, O.F.S., we have a “B” surface site overlying a 7-feet sterile stratum under which occur implementiferous gravels with Earlier Stone Age — Fauresmith or Acheuleo-Mousterian-type — remains as shown on the accompanying diagram and photograph (figs. F and G).

From these and similar occurrences elsewhere, it is definitely known that “B” followed long after (1) the Middle Stone Age, and (2) the Earlier Stone Age.

Despite recurring types, and a certain overlapping, every indication, whether in patina or incrustation, suggests that “B” also postdated “A” and probably antedated “C” in its beginnings.

Summary.

The Smithfield Industry heralds the arrival of Neo-anthropic Man, and with him the dawn of the Later Stone Age in South Africa. The entire assemblages of implements and artefacts are essentially Upper and Epipalaeolithic in character.

The Industry comprises three distinct phases or sub-divisions, “A,” “B,” and “C”; these, despite a certain overlapping and contemporaneity, apparently represent the Lower, Middle, and Upper Periods.

Our present knowledge of the Industry is perhaps best summed up as follows:

Smithfield “A.”

This has every indication of being the earliest phase of the Industry, and betokens the appearance in South Africa of Neo-anthropic Man. It shares certain common features with both the Middle and Earlier Stone Ages, but whether these imply a successive overlapping of these otherwise entirely distinct lithicultural groups, or suggest an evolutionary development, it is as yet impossible to say. We must wait for a more critical and exact study of the Middle Stone Age. Characterised by concavo-convex and large circular scrapers (Plate XXVI), it is associable with rock engravings, and its position on the time scale is the base of those Periods or Industries that follow the Middle and belong to the Later Stone Age. Tentatively we may therefore regard it as the Lower Smithfield.

Smithfield “B.”

This is the best represented and most fully developed phase of the Industry. It is definitely neo-anthropic in character, and, although it includes types reminiscent of the Middle Stone Age, it cannot be said to share any integral or regular features in common either with this or the Earlier Stone Age. Characterised by the duckbill end-scraper, intense specialisation in bored stones and a complete absence of concavo-convex and large circular scrapers, it is associable with both rock engravings and cave paintings, and has every indication of being an offshoot from and advance on “A,” thus constituting the Middle Smithfield. Excepting those implements that have ground edges, or are in any way polished, it is paralleled in the North by Capsio-Aurignacian features, not only lithiculturally, but also artistically. When Goodwin first drew attention to the affinities between the Smithfield and the Capsian,92 the Smithfield he visualised was essentially this “ B ” or Middle phase.

Smithfield “C.”

This is also definitely neo-anthropic, and, apart from certain typological diiferentiations, it varies from “B” only in that it is essentially a cave industry regularly associated with cave-paintings. Characterised by a preponderance of thumbnail scrapers (Plate XXXV, fig. 6), a regular use of bone and notched scrapers, it is, but for the absence of crescentic scrapers and any appreciable presence of “pointes” worked on both sides, duplicated by the Wilton (“Upper Capsian type”) — to which it appears to be as closely related as it is to “B.” Concavo-convex and large circular scrapers are definitely absent, and, although its exact position is obscure, it would seem (1) to have followed long after “A,” and (2) to have been largely, if not wholly, coeval with “B” and Wilton. Indications lead one to conceive it to be the latest offshoot from, and development or culminating phase of, the Smithfield — whatever infiuences brought about its creation — and therefore the Later Smithfield.

Geological and Geographical Notes.

(With special reference to the influence of material to hand.)

A consideration of the map of the Orange Free State immediately reveals the fact that, whereas the western two-thirds, more particularly the western half, is a monotone of undulating plains in the form of high-veld or steppes where rock outcrops are composed entirely of shales — Ecca and Beaufort of Permian and Triassic Age — and doleritic intrusions, the eastern third is mountainous and made up almost exclusively of sandstones of the Stormberg Series of Rhaetic-Jurassic Age.

The altitude of the western two-thirds fluctuates about 4500 feet above sea-level, whereas that of the east rises rapidly until, along the eastern border, it exceeds 10,000 feet above sea-level. lnevitably, therefore, the main drainage is east-west, the rivers becoming more sluggish and tending over the flat western districts to meander.

In the eastern mountainous areas the sandstones are either coarse-grained and intractable, or soft and friable, and always entirely unsuitable for the manufacture of stone implements. The shales of the central and western districts are even more unsuitable when pure, but vast and frequent intrusions of plutonic rock — Karroo Dolerites — occur throughout this latter area, and the shales at and near the contact planes are baked or indurated and form “indurated shale” or lydianite — the extent of the induration or metamorphosis depending on the nearness of the shale to the intrusion. It is this lydianite that is so eminently suitable for the manufacture of stone implements, and primitive man turned to it eagerly and used it in preference to any other material when and wherever he could. Outcrops are particularly extensive and frequent in the south-western districts, and this was probably the most influential factor in man’s choice of a settlement.

Depending on the extent of induration and the consequent impregnation of iron compounds, lydianite varies from a rich silky dark-grey to black when freshly broken; it is a flint-like stone with an even, conchoidal fracture; all flakes have typical bulbs and so even is the texture of the material that “ripples” seldom appear.

On the other hand, the only suitable material available in any quantity in the eastern districts is that which originated in the volcanics of the Drakensberg — the culminating feature of the Stormberg Series. These lavas cap only the highest chains of mountains, but in the loads of streams and rivers that drain the heights are quantities of chert, chalcedony, agate, quartz, and similar flinty materials that are constantly being brought down to the valleys, and it was this supply that constituted practically the only material suitable for man’s needs. I say “practically,” because there are in the Stormberg Series occurrences of shale that have, in places, been indurated by dolerite or other igneous intrusions. Indurated outcrops, however, are infrequent.

While lydianite occurs in bulk in the steppes of the central and western districts, the flinty rocks from the mountainous east are broken up and invariably reduced to water-worn pebbles, so that where it is possible to spall a good, large flake in the west, it is impossible to do so in the east, and in this, very probably, we have a factor that contributed largely to the comparatively pygmy Smithfield “C” types and, incidentally, to the finer technique.

The importance of material to hand cannot be over-emphasised. Instances are known where men of the same lithicultural group produced good implements where the material was good and bad implements where the material was bad. A striking example occurs in our local Stellenbosch Industry. At Pretoria the material is bad and the implements are crude and ill formed, whereas along the Vaal, north of Kimberley, the material is good and the implements, manufactured by people of the same lithicultural horizon, are good and well formed. Further, it cannot be doubted that the material to hand in the Riet River Valley in the south-western Free State gave a considerable impetus to the development of the Fauresmith Industry. And so, too, we must take into account the possible influences of material to hand in the Smithfield, for, where lydianite alone is procurable, the “B” phase is “heavier” than it is where chalcedony, agate, and quartz are also available. The “B” sites along the Orange River, for example ●70 and ●71 near Bethulie, ●74 Goedemoed, ●75 Aliwal North, and ●76 Doctor’s Drift, have very many smaller implements than those along the Riet River Valley ●1 to ●36 inclusive, but the implements and the order of importance of occurrence are the same, and the smaller sizes along the Orange River due entirely to the use of those flinty pebbles brought down from the volcanics of the Drakensberg. The Riet River does not drain volcanic areas, and its valley was populated not only because of a convenient water supply, but rather on account of the great abundance of suitable material in the valley and its immediate surroundings.

This desirable material (lydianite) was probably the greatest attraction, for dense settlements thrived here not only during the Later Stone Age, but also during the Middle and Earlier Stone Ages.

The Weathering of Lydianite.

When exposed to the elements away from the action of a regular water-supply, i.e. when out on the veld, the iron contents are oxidised and the colour changes first to khaki and then through a rich light chocolate to a rusty brown. The discoloration exists in a crust that, in instances, reaches 3 mm in thickness.

When exposed to the action of water or very humid conditions, i.e. under the action of hydration in addition to oxidation, the exterior assumes a lighter hue and there is no appreciable crust. For example, all implements found in springs, or where they have been under water in sluggish streams, are light grey. Under the action of running water the implements inevitably become water-worn, but the light grey exterior remains.

Notes on some Typical Smithfield Industry Factory Sites.

(See Plate XXXVII.)

Twelve sites only have been chosen — four from each phase, and it may be taken that each is essentially characteristic of the phase represented. To give such detailed lists for all the sites known is merely to repeat — with negligible variation — the lists of implements, etc., associated with each particular phase.

To the lists of implements given here — this is the implication of “factory site” as distinct from settlement only — must always be added the usual variety of fabricators: detaching-hammers, trimming-stones, anvils, and debris; cores, flakes, and rejects. All these fabricators and debris occur on the factory sites shown.

Smithfield “A” Sites.

Site ●37. (Type.)

Locality: Farm Lockshoek, No. 192 — Fauresmith, O.F.S.

Situated about 16 miles north-east of Jagersfontein on the Jagersfontein-Bloemfontein main road. A well-known farm easily accessible from Jagersfontein. In gorge above homestead, immediately below dam wall, and on the surface over broken ground adjacent to the main road as shown in fig. A, remains of a fully developed “A” site were found. The collection from here includes:

• Concavo-convex scrapers — Lydianite or indurated shale.
• Circular scrapers — Lydianite or indurated shale.
• Duckbill end-scrapers — Lydianite or indurated shale.
• Side-scrapers — Lydianite or indurated shale.
• Trimmed points — Lydianite or indurated shale.
• Stone borers — Lydianite or indurated shale.
• Bored stones — Sandstone.
• Grindstones — Dolerite or Diabase, Sandstone.
• Pounders and grinders — Dolerite or Diabase.
• Grooved stones — Sandstone.
• Rock engravings in vicinity.

Site ●17.

Locality: Farm Blaauwheuwel, No. 425 — Fauresmith, O.F.S.

Situated on the right bank of the Riet River, immediately opposite Koffiefontein and downstream of the bridge, we have a surface site particularly rich in remains. The list of implements collected includes:

• Concavo-convex scrapers — Lydianite or indurated shale.
• Circular scrapers — Lydianite or indurated shale.
• Serrated circular scrapers — Lydianite or indurated shale.
• Duckbill end-scrapers — Lydianite or indurated shale.
• Trimmed points — Lydianite or indurated shale.
• Engraved stone — Shale.
• Stone borers — Lydianite or indurated shale.
• Bored stones — Sandstone.
• Grindstones — Dolerite or Diabase, Sandstone.
• Pounders and grinders — Dolerite or Diabase, Sandstone.
• Grooved stones — Sandstone.
• Rock engravings in vicinity.

Site ●14.

Locality: Farm De Kiel Oost, No. 101 — Jacobsdal, O.F.S.

Situated on the left bank of the Riet River, immediately downstream of the bridge some 10 miles north-west of Koffiefontein on the Koffiefontein-Jacobsdal main road, are extensive surface remains of both “A” and “B” settlements and factory sites. The entire factory site here referred to covers about six acres, and the “A” site is at the extreme downstream end. Unfortunately there is overlapping, but the “A” implements are distinct from those of “B,” and, even if a lot is mixed, it is possible to sort them into their correct groups by patina and incrustation only. As an experiment, this was actually done by two independent workers.

The “A” factory site extends over an area of broken ground about a quarter of a mile downstream of the bridge, and the implements and artefacts collected here include:

• Concavo-convex scrapers (Lydianite or indurated shale);
• Circular scrapers (Lydianite or indurated shale);
• Serrated scrapers (Lydianite or indurated shale);
• Duckbill end-scrapers (Lydianite or indurated shale);
• Trimmed points — one of which has artificially incised lines on the unworked surface (Plate XXVIII, fig. 1) (Lydianite or indurated shale);
• Stone borers (Lydianite or indurated shale, Shale);
• Bored stones (Sandstone);
• Grooved stones (Sandstone);
• Pounders and grinders (Dolerite or Diabase, Sandstone);
• Grindstones (Dolerite or Diabase, Sandstone);
• One engraved stone (Plate XXVIII, fig. 3) (Shale);
• One Solutrean-type point (pygmy) (Plate XXVI, fig. 30) (Agate);
• Two crescentic scrapers (Plate XXXV, figs. 11a and b) (Lydianite or indurated shale, Agate);
• Fragmentary pottery.
• Fragments of ostrich eggfshells.
• Rock engravings in vicinity.
• Fauresmith coups-de-poing heavily weathered and rolled, re-used by Smithfield man as fabricators; and
• Fauresmith points retrimmed by Smithfield man.

Site●22.

Locality: Farm Brakfontein, No. 231 — Fauresmith, O.F.S.

The farm Brakfontein is 8 miles south of Koffiefontein on the main road to Fauresmith. On the right bank of the spruit that runs past the orchard in front of the homestead, and about 200 yards upstream of the orchard, were found extensive remains of an “A” site. Implements and artefacts collected here include:

• Concavo-convex scrapers (Lydianite or indurated shale);
• Circular scrapers (Lydianite or indurated shale);
• Duckbill end-scrapers (Lydianite or indurated shale);
• Trimmed points (Lydianite or indurated shale);
• Stone borers (Lydianite or indurated shale);
• Bored stones (Sandstone);
• Grooved stones (Sandstone);
• Grindstones (Dolerite or Diabase);
• Pounders and grinders (Dolerite or Diabase);
• Fragmentary pottery;
• Rock engravings in immediate vicinity.

Smithfield “B” Sites.

Site ●27. (Type.)

Locality: Farm Avalon, No. 554 — Fauresmith, O.F.S.

Situated about 25 miles north-east of Jagersfontein on the Jagersfontein-Bloemfontein main road. On the left bank and in the bend of the Riet River about a quarter of a mile upstream of the causeway as shown in fig. B are extensive remains of a Smithfield “B” settlement and factory site. All implements are found on the surface, but shifting sands are apt to cover and uncover these at different times. The list of implements and artefacts collected here includes:

• Duckbill end-scrapers (Lydianite or indurated shale);
• Thumbnail scrapers (Agate, Chalcedony, Jasper, Lydianite or indurated shale);
• Side-scrapers (Lydianite or indurated shale);
• Trimmed points (Lydianite or indurated shale);
• Stone borers (Lydianite or indurated shale);
• Bored stones (Sandstone);
• Grooved stones (Sandstone);
• Grindstones (Dolerite or Diabase, Sandstone);
• Pounders and grinders (Dolerite or Diabase, Sandstone);
• Stone ring and fragments of incomplete rings (Shale);
• Pottery (fragmentary) (Plate XXXVI, figs. B, C, and D);
• Ostrich egg-shell fragments, beads, and pendants;
• Ostrich egg-shell borers (Chert);
• Stone palettes (Shale).

Site ●42.

Locality: Farm Eagle’s Nest, No. 550 — Boshof, O.F.S.

Situated about 16 miles north of Petrusburg on the Petrusburg-Boshof main road and on the left bank of the Modder River immediately downstream of the approach to the bridge, and also partially upstream, are extensive remains of a Smithfield “B” settlement and factory site. All the material is on the surface, and the list of implements, etc., collected here includes:

• Duckbill end-scrapers (Lydianite or indurated shale);
• Thumbnail scrapers (Agate, Quartz, Chalcedony, Lydianite or indurated shale);
• Side-scrapers (Lydianite or indurated shale);
• Trimmed points (Lydianite or indurated shale);
• Stone borers (Lydianite or indurated shale);
• Bored stones (smallest 1.25 oz.) (Sandstone);
• Grooved stones (Sandstone);
• Grindstones (Dolerite or Diabase);
• Pounders and grinders (Dolerite or Diabase, Sandstone);
• Pottery;
• Ostrich egg-shell fragments and beads;
• Skeletal material.

Site ●65.

Locality: Farm Paardenvallei, No. 668 — Winburg, O.F.S.

Situated about 8 miles south of Theunissen, 2 miles downstream of the railway bridge at Vet River Station and on the left bank of the Vet River in the immediate vicinity of the homestead are extensive remains of a “B” settlement and factory site. The material is all on the surface, and the list of implements, etc., collected here includes:

• Duckbill end-scrapers (Lydianite or indurated shale);
• Thumbnail scrapers (Agate, Quartz, Chalcedony, Lydianite or indurated shale);
• Trimmed points (Lydianite or indurated shale);
• Stone borers (Lydianite or indurated shale);
• Bored stones (Sandstone, Soapstone);
• Grooved stones (Sandstone);
• Grindstones (Dolerite or Diabase);
• Pounders and grinders (Dolerite or Diabase);
• Fragmentary pottery;
• Ostrich egg-shell fragments and beads;
• Bone points (found by owner of property) (?).

Site ●80.

Locality: Farm Mook, No. 54 — Wepener, O.F.S.

Situated about a quarter of a mile from Van Staden’s Rust village, 24 miles south of Wepener, and in the immediate vicinity of the homestead, orchards, and kraals are extensive remains of a “B” settlement and factory site. The material is all on the surface, and the list of implements, etc., collected here includes:

• Duckbill end-scrapers (Lydianite or indurated shale, Agate, Chalcedony, Jasper, Chert);
• Thumbnail scrapers (Lydianite or indurated shale, Agate, Chalcedony, Jasper, Chert);
• Trimmed points (Lydianite or indurated shale);
• Notched scrapers (Lydianite or indurated shale);
• Stone borers (Lydianite or indurated shale);
• Bored stones (Sandstone);
• Grooved stones (Sandstone);
• Grindstones (Dolerite or Diabase);
• Pounders and grinders (Dolerite or Diabase, Sandstone);
• Fragmentary pottery
• Fragments of ostrich egg-shells;
• One Still Bay type point (Plate XXVII, fig. 3b) Lydianite or indurated shale.

Smithfield “C” Sites.

Site ●103.

Locality: Farm Ventershoek — Wepener, O.F.S.

Situated about 5 miles from Wepener on the Wepener-Mafeteng main road, half a mile upstream of the bridge over the Ventershoek Spruit, in and near the caves shown in fig. C are excellent remains of a “C” settlement and factory site. The list of implements etc., collected here includes:

• Thumbnail scrapers (Agate, Quartz, Chalcedony, Chert, Lydianite or indurated shale);
• Duckbill end-scrapers (Chert, Lydianite or indurated shale);
• Trimmed points (Lydianite or indurated shale);
• Notched scrapers (Lydianite or indurated shale);
• Stone borers (Lydianite or indurated shale);
• Bored stones (Sandstone);
• Grooved stones (Sandstone);
• Grindstones (Sandstone, Dolerite or Diabase);
• Pounders and grinders (Dolerite or Diabase);
• Bone points;
• Bone pendant;
• Pottery (fragmentary);
• Fragments of ostrich egg-shells;
• Bead borers;
• Fragment of stone ring (Sandstone);
• Fragments of red ochre;
• Cave paintings (polychromes).

Site ●100.

Locality: Clarens Town Lands — Bethlehem, O.F.S.

Under rock shelters and in talus immediately below these in gorge above the village are remains of a “C ”site. The list of implements, etc., collected here includes:

• Thumbnail scrapers (Agate, Quartz, Chalcedony, Chert);
• Duckbill end-scrapers (Chalcedony, Lydianite or indurated shale);
• Trimmed points (Lydianite or indurated shale);
• Notched scrapers (Chalcedony, Lydianite or indurated shale);
• Bored stone fragment (Sandstone);
• Grindstones (Sandstone);
• Pounders and grinders (Sandstone);
• Bone point fragments;
• Fragmentary pottery;
• Cave paintings (monochromes).

Site ●101.

Locality: Farm Schaapplaats — Bethlehem, O.F.S.

Situated about 5 miles south-east of the village of Clarens and in the gorge immediately adjacent to the homestead (de la Harpe) is an unusually spacious and deep cave that has been used as a shelter for sheep for years. The ground is therefore considerably disturbed, but in, and in the immediate vicinity of, the cave is an abundance of “C” remains. The collection from here includes:

• Thumbnail scrapers (Quartz, Chalcedony, Agate, Lydianite or indurated shale);
• Duckbill end-scrapers (Chalcedony, Lydianite or indurated shale);
• Trimmed points (Lydianite or indurated shale);
• Notched scrapers (Chalcedony, Lydianite or indurated shale);
• Grindstones (Sandstone);
• Pounders and grinders (Sandstone);
• Bone points;
• Pottery (fragmentary);
• Ochreous painting materials;
• Cave paintings (polychromes and monochromes).

The paintings here are of an unusually high order. It is worthy of note that bored stones have been found in the neighbourhood.

Site ●102

Locality: Van Reenen Town Lands — Harrismith, O.F.S.

Situated about a mile east of the village of Van Reenen and very near the crest of the Drakensberg Mountains is a deep gorge that contains numerous caves and rock-shelters. Many of these caves were inhabited by “C” Smithfield folk, and from one I was able to collect:

• Thumbnail scrapers (Quartz, Chalcedony);
• Duckbill end-scrapers (Chalcedony, Lydianite or indurated shale);
• Notched scrapers (Lydianite or indurated shale);
• Fragment of grindstone (Sandstone);
• Pounders and grinders (Dolerite or Diabase);
• Fragments of bone points;
• Fragmentary pottery;
• Red ochre.

Bored stones have been found in the vicinity. The walls of the caves contain many fine examples of cave painting, but excursionists to the mountains have, unfortunately, destroyed most of these. Initials of every Dick, Tom, and Harry literally cover the walls of the finest cave and so many of the finest paintings.

Addendum on the Further Distribution of the Smithfield Industry.

Mr. Lowe’s monograph on the Smithfield sites of the Orange Free State has been so thorough that little need be added to his paper; but, although everything points to the Orange-Vaal basin having been the original home of the industry and its variations, yet he has carefully avoided leaving the impression that they are confined to that area, and requested me [A.J.H. Goodwin] to add this complementary note on distribution.

It will have been seen that the evidence available is in favour of the hypothesis that variations “A” and “B” both evolved in this area, and that “C” is largely regardable as a variation produced by two factors — the change of material in the eastern part of the sub-continent, and continual attacks of a new industry appearing from the north, which seem to have had their greatest effect on the eastern edge of the main Smithfield area and near the Basutoland mountain masses. This influence is the Wilton lndustry, which is relatively pure in Rhodesia, and seems to have been the main factor in the evolution of the almost microlithic “C” variation of the Smithfield, and at the same time to have introduced the cave-dwelling habit which, too, would be affected and suggested by the presence in the Stormberg series of “cave sandstones,” and the great mountain folds of Basutoland and the Eastern Cape.

Mr. Lowe’s map, and our knowledge of the distribution of the variations outside the Free State, point to a confined “A” variation, appearing almost entirely in the South-Western Free State, Kimberley, and the various districts immediately south of this area. The “B” variation is far less confined, and seems to cover the whole Free State and a large part of the Cape Province. Smithfield “C” appears to extend towards the south-east in much the area occupied by the Wilton peoples, though this latter group seems to have extended westward along the mountain folds of the south as far as the Cape Peninsula, and even further to the north-west.

Smithfield “A.”

The McGregor Museum, Kimberley, shows a very fine assortment of Smithfield “A” implements, including the typical concavo-convex of this variation, sites being well represented from Zoutpan (Jacobsdal, O.F.S.), Alexandersfontein, Wittepan, Rietpan (all Kimberley district), and from Kimberley itself. From near Modder River station Mr. Power has collected Smithfield “A” implements from a site amazingly rich in Smithfield “B” implements. South of the Free State the tell-tale concavo-convex and its associated types appear from the Victoria West Golf Links, and are here associated with a number of odd scrapers made from the weathered outer faces of blocks of shale; this natural face forming a deeply patinated patch covering the greater part of the outer face of the flake. These implements are in no way conventionalised, but are probably of the Smithield “A” variation, as the utilisation of the outer crust of the shale is a very typical trait, noticeable at many Smithfield “A” sites, but largely lost in the other variations.

Craigie Glen, Wodehouse district, and Culmstock, Middelburg district, both show excellent examples of the concavo-convex. All these latter sites are represented in the South African Museum.

Britstown offers a possible diiliculty — an apparent evolution of a local character from Smithfield “A.” A fine specimen of the concavo-convex is represented at the South African Museum, but it is of a type not usual to Smithfield “A” sites elsewhere. It is very much wider and more shallow than the normal type (perhaps 4 inches by 1 inch). This type is exactly paralleled from East London.

From Britstown, too, come a number of beautifully serrated circular scrapers (in the collection of Mr. Heese at Riversdale), only paralleled by specimens in the collection of Mr. Swan at Kimberley. With the Britstown specimens are associated fine tanged arrow-points, but it is difficult to place these with any certainty with our present knowledge.

It is worth noting here that tanged arrow-heads were already appearing in the later Mousterian of North Africa, and, if the Smithfield “A” variation owes much to the Middle Stone Age, it is possible that the tanged arrow-head forms part of its debt. The peculiar scraper — showing affinities to the concavo-convex — illustrated by Péringuey95 from Matatiele, Griqualand East, possibly belongs to the Smithield Industry; it is fairly typical, but is quite unassociated. This specimen similarly shows the weathered surface of the shale over the outer face.

Smithfield “B.”

Most of the other open-air sites represented in the South African Museum collection and in other of our museums consist of Smithfield “B” sites. This variation is by far the most widespread and the richest industry of the Union.

Of these sites little need be said: all are open-air sites, save one or two cave sites at Britstown, Naauwpoort, etc.; and all show much the same range of implements described by Lowe.

Along the Orange River proper, west of its junction with the Vaal and south-west of Kimberley, the local jaspers have forced a slight variation; square scrapers, etc. appearing among specimens submitted to this museum by Mr. Bryant, but this area has been very insufficiently studied as yet.

A site at the Half-way House to Barkly West from Kimberley has yielded Mr. Power a single tanged arrow-head similar to Mr. Heese’s specimens from Britstown; the site at Half-way House is a “B” site, associated with rock engravings.

The Moonlight Kop specimens from Victoria West show a tendency towards Smithfield “C,” so far as size is concerned, but the workmanship and the types of implements represented seem to class it as of Variation “B.”

Smithfield “C.”

This is typically a cave industry; and while it shades into “B” and is often difficult to separate from that group with any degree of certainty, yet even greater difficulty is encountered when we attempt to differentiate between the “C” group and the Wilton. So much is this so, that Mr. Burkitt96 says:

Mr. van Riet Lowe has lately come to the conclusion that the Smithfield Industry found so often in rock-shelters with paintings can be distinguished as a Smithfield “C” group from the above-mentioned divisions which he calls Smithfield “A” and “B” respectively. Personally I am not convinced that such a Smithfield “C” can be thus separated. I fancy the occurrence of thumbnail scrapers, etc. in some of these industries merely indicates some contact with the Wilton Culture.

In a recent letter [26/08/1928], speaking of this statement by Burkitt, Lowe says:97

The occurrence of thumbnail scrapers, not only in the “C” but also in the “B” phase, does, as pointed out in my paper, indicate some contact with the Wilton; but it must be borne in mind that, apart from the typological differentiations that exist between “B” and “C,” and between “C” and the Wilton, the presence of this thumbnail scraper and its associated microlithic forms, may also very largely be due to material to hand, just as its numerical increase may be due to the exigencies of new environmental conditions. The fact that crescents have not been definitely associated with this “C” or cave-Smithfield, and that bored stones do not appear in the Rhodesian Wilton, are facts sufficiently significant to indicate the necessity for recognising this “C” phase as a variation, not only of the Smithfield but also of the Wilton. The associated arts are also different.

It will be seen, therefore, that Smithfield “C” may be regardable purely as a mixed industry. But it is necessary to name it; and rather than regard it as a branch of the Wilton the term Smithfield “C” is preferred. It is better to use the name Wilton Industry for the purer stock which Mr. Neville Jones already recognises in Rhodesia — an area where the Smithfield does not appear.

It is to be admitted that any scientific attempt to differentiate in certain instances becomes almost farcical, the Smithfield “C” being sometimes only a crescentless Wilton; the absence of the crescent being made up for by a greater number of small end-scrapers, thumbnail scrapers, and horseshoe scrapers, which in themselves are often indistinguishable from the Wilton varieties. Many sites fall absolutely within the Smithfield “C” group, or in the Wilton group.

Many sites have been insufficiently studied or are insufficiently represented to allow of their inclusion under either of these three headings. Further work on the Smithfield should make the three-fold division clearer, or may perhaps necessitate its re-arrangement on a sounder scientific basis.

The Wilton Industry.

The term “Wilton” was first used very tentatively by Mr. J. Hewitt, Director of the Albany Museum, Grahamstown, in July 1921. Some few years later, when attempting to sift out the various groups of implements into definite industries, I [A.J.H. Goodwin] used the descriptive terms Pygmy and Microlithic in referring to this industry. In 1926 when finally submitting my scheme to the Science Association at Pretoria, the term Wilton was again suggested, and accepted.

Our first knowledge of the Wilton Industry comes from the Cape Peninsula, various crescents, thumbnail scrapers, and the like, appearing from a number of kitchen middens and sand-dune sites in this district. Unluckily, Dr. Péringuey often used the term “Cape Flats type” to include both the Still Bay Industry and the Wilton, as they are sometimes found in contiguity at the Cape.

Mr. Hewitt’s careful and meticulous scientific methods have rendered a great service to archaeology, and have given us an interesting insight into the material found at the Wilton cave. This site was originally discovered by Mr. C.W. Wilmot (the discoverer of the Cofimvaba material), while he was postmaster at Qumbu. The farm Wilton stands some 5 miles east of Alicedale, and forms part of the original Hoffman’s Kloof farm. The rock-shelter is some 2 miles from the homestead, and overlooks a narrow kloof. The shelter is long and shallow, backed by a steep wall of overhanging quartzitic sandstone, which is covered with paintings. To quote Mr. Hewitt’s original paper:98

Mr. Hewitt gives a more detailed description of this cave, which may well be quoted here in full.

The floor of the rock-shelter at Wilton is covered very largely by ashes, which in places have a vertical depth of 4 feet, but no layers are traceable therein. This is the unanimous conclusion of the three investigators (Revs. P. Stapleton and Kilroe, and Mr. Hewitt), who devoted five days to its exploration.99 Moreover, with few exceptions, the Wilton implements seem to constitute a homogeneous assembly, despite considerable range in size amongst the scrapers. … Amongst the numerous paintings on the inner wall of the rock-shelter are some spirited representations of antelopes in profile. The technique is quite superior, and a number of the antelopes are in two colours, red and creamy white, but there are no group scenes. A very distinctive feature is the treatment of the human figure, the limbs and body being tremendously elongated. These, which are wholly red, may not belong to the same period as the antelope pictures. … Thus we arrive at a conclusion which has long been anticipated, but not hitherto so well supported by actual data as now detailed, that the short-headed Bushman made the delicate ostrich-shell beads, the pygmy crescents, and the tiny scrapers, and was also the author of rock paintings of superior and characteristic technique.

In this series of quotations I have deliberately left out his references to the possible presence of a second and later industry (apparently Smithfield “C”), as we are not dealing with this industry here. Were he more sure on this point we would be in a position to prove an interesting sequence, i.e. that Smithfield “C,” at any rate in the south-eastern districts, is to be regarded as later than the Wilton; and also to point to the possibility that both pottery and the bored stone belong primarily to the Smithfield Industry and were only introduced into this area at the end of the Wilton period. For further information the reader is referred to Mr. Hewitt’s paper. Against Mr. Hewitt’s suggestion of the presence of two industries it is worth noting that Mr. Burkitt states:100 “only one industry occurs; that is to say, the rock-shelter has only been inhabited by folk belonging to one culture.”

Mr. Hewitt kindly allowed me to see his collections from this site. The implements represented consist mainly of small crescents of two types, thin flakes worked to a curve on one edge, and slightly thicker flakes worked on two edges. The crescent is usually analysed into an arc and a chord, the chord is the straighter of the two sides, the arc curving round to meet this at each end. In the single crescent only the arc is worked, the working being of a steep backing type, consisting of a series of evenly spaced trimming flakes struck or pressed off from the under (cleavage) face. In the double crescent the arc and chord are both worked. The section across the crescent is much the same as the section of a pocket-knife blade, or a scissors blade, the thicker edge being formed by the arc, the thinner by the chord.

Pygmy end-scrapers of a shape and size similar to the Smithfield “C” type occur here also. Whether Mr. Hewitt is right in regarding these as belonging to a later appearance of Smithfield “C” man it is difficult to say. The presence of a bored stone points to this possibility.

Thumbnail scrapers make an appearance at this site, and right through the Wilton Industry. These are very similar to the Smithfield “C” types, and shade into the little circular horseshoe scrapers, and both may best be described as variations of the end-scraper, but both types are worked across the end and along each lateral edge giving a small (1 cm²) square or circular scraper of which only the striking platform remains unworked. With these microlithic implements appear tiny cores, some without secondary working, some very definitely made into core scrapers, tiny trimming flakes having been removed from about the circumference of the striking platform. With these may be mentioned little hammer stones, measuring an inch or less across, which appear to have been used for the purpose of making the secondary trimming on the crescents and scrapers. Small rod-like flakes, steeply worked on either edge, perhaps an inch in length, by a third of an inch across, by half an inch thick, are also found definitely associated with the Wilton implements. These last make a sporadic appearance in the Smithfield “C” variation and in a few Middle Stone Age sites, though in the latter they tend to merge into the peculiar re-directing flHakes of that period.

Rough bone points (“awls”) also appear from the Wilton Cave. These consist of split bones sharpened at one or both ends to a torpedo-like point, by the action of rubbing. Mr. Lowe has mentioned these points in his Smithfield paper. They appear to be in every way identical with the bone arrow-points used by the present-day Bushmen of South-west Africa, the Kalahari and Angola, which may therefore be shortly described. The arrow-head consists of three parts, measuring some 13 cm over all. The foremost part consists of the point, a splinter of bone rounded, and sharpened at the forward end, measuring about 7 cm in length by 4 mm at its greatest thickness. The butt, which is squared off, is bound into a reed collar, and rests directly against the forward point of the link-shaft, which is also bound into the reed collar. The collar usually measures 15 mm in length by 7 mm in diameter. The link-shaft consists of a rounded splinter of bone, pointed at each end, and measuring some 6 cm in length by 1 cm in thickness. The hinder end is inserted into the reed shaft of the arrow, which measures perhaps 40 cm in length. The point is poisoned and the whole fore-shaft can be reversed in the reed arrow to guard the poisoned point when not in use (text-fig. 1). It seems more than likely that the points found at the Wilton shelter and elsewhere are actually the parts of bone arrow-heads.

Ostrich egg-shell beads and their tiny stone borers, showing delicate working across one edge, are also found in the Wilton site. Exactly similar beads occur among the modern Bushmen and the tribes with whom they have come into contact. In the Congo strings of beads of this type occur as currency, and many of the North African Neolithic sites show exactly similar beads made by the same process. (See Addendum.)

Among the ornaments found by Mr. Hewitt may be mentioned a pierced univalve (Nassa) marine shell, apparently bored for suspension and a dassie tusk (Hyrax sp.) similarly bored towards the root.

Having introduced a description of the name-site it is necessary before going further to explain exactly why the term Wilton has been used rather than a term such as “Smithfield D.” At first sight, and judging only from the prehistory of the Union, it seems as though we were in the presence of an evolutionary series, starting with Smithield “A” and evolving to the Wilton.101 Directly we pass out of the Union it becomes immediately obvious that the Wilton of Rhodesia is earlier than that of the Union, and that no Smithfield types appear. Where mixture is present it is with Middle Stone Age types. Mr. Neville Jones illustrates a series of Wilton types from the later deposits at Sawmills.102 The industry here appears mixed, crescents, lance-heads, and circular scrapers appearing side by side. In further illustrations (figs. 24, 25) he depicts material from what would appear to be a mixed site (Bambata), and two pure Wilton sites (Nswatugi and Kalanyoni).

Judging from the purer sites, the Wilton in Southern Rhodesia appears to consist of crescents, thumbnail scrapers, horseshoe scrapers, ostrich egg-shell beads, and their tiny stone borers, bone points, and spatulae, pigment and paintings. The bored stone seems to be entirely lacking at cave sites, though a few have been recovered unassociated from other sites. Rock engravings have been found in a cave near Bumbuzi and are depicted by Jones (fig. 37).103 Further engravings showing animal tracks, whirls, ladders, etc. have been reported and photographed by Mr. W.O. Montague Owen, Kasamba district, Northern Rhodesia, from a site 200 feet above the Munwa stream, 4 miles from its junction with the Luapula River. These and the bored stones seem to be our only indication of the road taken by the original folk from whom the Smithfield Industry finally obtained these two elements.

Within the Union we first encounter the Wilton Industry at Lichtenburg, whence material has reached the McGregor Museum, Kimberley, from Doornlagte farm. On the Vaal River it appears once more at Riverton (M.M.K., 1152) in close conjunction with Smithfield “B” types. Johnson104 describes implements from an island at Riverton. Mr. J. Swan has retrieved a number of Wilton crescents from the Riverton site; these have been presented by him to Professor Balfour of Oxford. Messrs. J. Power and S. Tapscott have obtained similar collections. The site is an open site, situated on the south-eastern bank of the Vaal River; it has been cut up by dongas to a considerable degree, and the material is being drifted down to the river which at this point contains Stellenbosch Industry gravels. Similar Wilton implements are known from Rooipoort in the Kimberley district. The Port Elizabeth Museum also shows Wilton types from Kimberley district (P.E.M., 509). Further down the Orange River, Wilton sites are sporadically known, one site at Abiam, Gordonia, being represented at Kimberley. A few sites are known in the Kalahari (Gamotiep, Aries, etc.) and are represented at the South African Museum and in the Port Elizabeth Museum, while Johnson mentions crescents from the Hart-Vaal River junction. It is evident that the Wilton Industry is well represented in this part of the Union. Many of the sites are in the open, and it has thus proved difficult to get a complete series of implements typical of this area. Crescents, small end-scrapers, and pottery seem typical here as elsewhere. Down the eastern side of the subcontinent the Wilton is not yet well known. The Free State shows sporadic occurrences (see Lowe’s statement on De Kiel Oost crescents in the previous paper), but insufficient is yet known of the eastern portion of this area to enable us to deduce much.

Further north Barberton (A.M.G.) and Carolina (S.A.M.) have yielded implements which might be classed as of Wilton type. South of this new difficulties are encountered, as we enter native territory which has not been studied to any extent. The area south of the native territories is not yet too well known, but Queenstown shows Wilton types at a rock-shelter overlooking Rockwoods farm in the Bongolo basin. The shelter contains a weird series of window and ladder designs in red, yellow, white, and black.105 Associated with these Wilton types I found a European brass pin, the head of which had been clamped on, and which probably dates back a hundred years.

Many other caves in the Eastern Province, especially about Grahamstown, contain pure Wilton material. Another cave in this area excavated by Mr. Hewitt is the Spitzkop Cave, Springvale. Here he discovered Wilton types of implements, pierced fragments of sea-shells, bone points (better made than the Wilton specimens, one exactly similar point is of wood), ostrich egg-shell beads, remains of a firestick, portions of a European clay pipe, and china beads of white banded with red; the last two point to modernity. The pierced sea-shells are relatively common in this district, and worthy of description. They consist of round fragments of nacreous shells about an inch across which have been bordered by a series of cuts about the edge, spaced at about one-tenth of an inch apart, and of similar length. This gives a roughly scalloped edge.106 Together with these he found a section of bone, some 5 inches long, incised all about with cross-hatched lines, and a few “palettes,” flat fragments of slate, which suggest a parallel to the European painter’s palette.107 The Albany Museum shows similar palettes from the Port Alfred kitchen middens (A.M.G. 1717), here associated again with Wilton implements. A similar association occurs at Waterkloof, Stanhope (A.M.G. 1739, 1741), where at two shelters containing paintings, bone points, Wilton implements, ostrich egg-shell beads, and quartz crystals were found with a palette. At Groot Kloof, Stanhope (A.M.G., 1742), the same association exists, ostrich egg-shell beads, a large incised bead, similar to the pierced sea-shell fragments mentioned above, incised pottery on which the incisions have been done before baking, giving an impression of corrugation, fragments of fibre string, very neatly made, and palettes all appearing together. From Kabeljauw’s Cave, Jeffrey’s Bay, a very similar assortment appears, Wilton implements, incised ostrich egg-shell, ostrich egg-shell beads, pierced sea-shell pendants, bone points, a bored stone, and pottery. From a neighbouring midden site Wilton flakes are again associated with pottery.

It will be thus seen that in the south-eastern province of the Cape the normal associations of the Wilton are:

• Crescents.
• Thumbnail scrapers.
• Horseshoe scrapers.
• Small end-scrapers.
• Small cores and hammer~stones.
• Bone points.
• Firesticks.
• Pottery.
• Ostrich egg-shell beads.
• Pendants of nacre, or of ostrich egg-shell.
• “Palettes.”
• String.
• Paintings.
• Paint stones.
• Skeletons of a “San” (Bush-Hottentot) type.

At Burnt Kraal (Grahamstown), a small rough cone, pierced longitudinally and similar to the bowl of a clay pipe, but of Wilton pottery, was found in a similar series of associations. It seems necessary to include also the bored stone, though this is by no means typical of the Wilton in Rhodesia.

Middledrift, in much this region, shows an open-air site, situated on the banks of the river, where Mr. Gladwin and the Messrs. Wilson have found crescents and end-scrapers of surface quartzite, indurated shale, and other materials associated with pottery and ostrich egg-shell beads.

Westward of this area, along the mountain ranges, various Wilton sites occur, but have not been properly explored. Mr. Fitzsimons’ excavations in several caves in the Tzitzikama Mountains are of interest, and several papers, by himself, and by Mr. Gear and Dr. Gordon Lang of Johannesburg, point to the presence of two racial types, an earlier “Boskopoid race” followed by the light-boned San peoples. No sequence of culture has been noted, but from the larger objects retrieved it is obvious that the later folk had a later Stone Age culture, perhaps Wilton or Smithfield “C.” Owing to the fact that these people have taken to a “strandloping” type of subsistence, small implements are in all probability difficult to find. It seems likely that a series much the same as the Skildegat Cave sequence has been overlooked at these sites.

In the shell caves at Knysna filled with midden refuse, I was able to discover Wilton crescents, thumbnail scrapers, a bone point, ostrich egg-shell beads, etc., forming a typical Wilton group. No pottery was represented at all in the caves. The implements were mainly of chert, quartz crystal, etc., and distinct in type from the Mossel Bay Variation implements which are more typical of Knysna. Many flakes of this industry seem to occur in the midden material, pointing either to a mixture of culture or to the Wilton midden makers having used these long flakes for shell opening.

Pottery and bored stones have been found by Mr. William Brown at Duthie’s farm, and also at a site between the Inner and Outer Obelisks (the naval terms for the West Head and the signal hill), about 100 feet above sea-level. They are not associated with implements. Paintings seem to be entirely absent, though Mr. Brown states that Messrs. Bain and Henkel, while excavating a coastal cave on the Western Head, discovered a large bone with an eland engraved upon it. This was lent with other interesting objects to the British Museum, but was stolen on the return voyage.

The Cango Cave at Oudtshoorn opens out of a shelter containing paintings; the debris from the floor shows crescents, thumbnail scrapers, ostrich egg-shell beads, etc. Some of these are housed in the Kimberley Museum. On a visit there, some years ago, I found work to be impossible as the floor deposit had been concealed by a roadway having been built over it. A certain amount of material is still discoverable below this road, however, having apparently been thrown out.

The Cape St. Blaize cave at Mossel Bay is very similar to the Knysna caves, both Wilton and Mossel Bay types apparently occurring together. Whether this is due to later mixing of the deposits by the local authorities, or to an actual mixture of cultures it is impossible to say. The Wilton material also shows affinities to Smithheld “C.” The material used is surface quartzite, the Mossel Bay types being of quartzitic Table Mountain sandstone.

Péringuey mentions and illustrates108 an interesting object, consisting of a chipped core-scraper mounted on a stick with some type of cement. This was discovered by Mr. R.E. Dumbleton at the end of the last century, in a cave situated near the mouth of the Touw River, which passes southward a few miles east of George. Describing the cave109 Mr. Dumbleton states that the cave floor was covered with a thick layer of guano which the owner of the farm wished to utilise, and a skeleton was recovered, but partially destroyed. Mr. Dumbleton commenced investigations towards the middle of the cave; a further skeleton was revealed lying on its left side in a fully flexed position and enclosed in a buck-skin; he continues:

On coming to the head I discovered immediately in front of the face two tortoise shells. … With these there was the lumbar vertebra of a large ruminant, several flint scrapers, and also at peculiar instrument consisting of a piece of flint fixed in gum-cement, in which was inserted a piece of wood about four inches long, serving as a handle. The latter was perfectly rotten and broke off short.

Happily the pieces were retained and the whole is now in the South African Museum. Little description of this object is necessary. It is some 6 inches in length, and consists of a round stick perhaps half an inch thick; over half of the stick is enclosed in a mass of gum-cement, which is out of all proportion to the rest of the implement, the greatest thickness being almost 45 cm. Attached to the stick and set into the gum-cement is a core-scraper of surface quartzite.

One other hafted stone implement from this area is housed in the South African Museum. It consists of a similar mass of cement attaching a flat unworked flake of slate to a stick. This was discovered in a cave near Knysna, and is mentioned in a letter from Mr. Henkel to Péringuey (S.A.M., 11^th October 1912):

I attach a sketch with measurements of an implement found in a cave. Undoubtedly it is of the same kind as figured in Plate xix, 150. I can give you no further particulars at present, but am sending this hurried note to report the find, as apparently the implements are rare. [Plate XLII].

This specimen (S.A.M. 1588) is also housed in the South African Museum, and is of particular interest as it seems to imply that unworked flakes, if of a suitable shape, were used, and actually mounted for use. The type of mounting would not allow of much lateral strain being exerted, and the implements would only stand relatively direct pressure. This mode of hafting is still to be found among modern Bush tribes among certain of whom arrows are tipped in this manner with quartz crystal.110

The next area in which sites of Wilton types are known is the Riversdale district. Some years ago Mr. H. Harger submitted a number of implements from the sand-dune middens on the Still Bay coast to Mr. Hewitt of Grahamstown. Since then Mr. C.H. Heese of Riversdale has collected a very considerable number of implements of this type from the middens west of the Kaffirs Kuil River-mouth. Visiting this spot a few years ago, I found a number of Wilton crescents associated with pottery of the usual type, and with midden refuse. Mr.  Burkitt some time later discovered identical material both here and in a nearby plantation.111

Some few miles inland from here and north of Riversdale Mr. Heese is at present excavating a cave which shows a most perfect series of Wilton implements of quartz crystal, white quartz, surface quartzites, and the usual types of material. The cave is known as the “Cave of Hands” from the silhouettes of hands painted in red and the series of red dots, apparently ochred finger prints, on the walls. These at once link up with the original Wilton site, and various similar Eastern Province sites showing the same peculiarities. Dr. K.H. Barnard retrieved a number of Wilton implements from a similar cave in Garcia’s Pass, Riversdale.

Caves at Rooi Els, some few miles from Gordon’s Bay, east of the Cape Peninsula, show an accumulation of midden refuse, containing a few rough bone and stone implements typical of the midden sites of the south coast, but not definitely associable with any particular industrial group.

In the Cape Peninsula itself pure Wilton sites occur here and there; from Diep River a very fine series of crescents showing a wonderful evenness of workmanship and similarity of form have been submitted to the South African Museum. They were collected about the edges of the vleis or meres typical of this portion of the Cape Flats. Some Wilton implements appear in the Dale Collection from Maitland, while Mr. J.M. Bain has sent in specimens from a variety of sites on the Flats (Plate XLI). Colonel Hardy has found a very considerable number of Wilton implements at various sites in the Peninsula;112 and the association of pottery, crescents, thumbnail scrapers, horse-shoe scrapers, ostrich egg-shell beads and their borers, with now and then a few bone tools, seems certain for this region also.

Mr. Drury of the South African Museum recently excavated a cave at Witsands in the Cape Peninsula at the instigation of Mr. C. van der Poll. In this cave was discovered a large accumulation of midden refuse of the normal coastal type. The bone material at this cave is exceedingly abundant and interesting, and consists of bone points (“arrow-heads”), bone awls, bone tubes, comparable with specimens known from inland sites but without ornamentation, ostrich egg-shell beads, bored egg-shell fragments, a nacre disc, similar to the Eastern Province specimens, a spoon or spatula, and a peculiar bone ornament or thin spatula (Plate XLIII). Pottery and rough flakes of midden types are associated. The bone implements are identical with those appearing in such abundance from the caves at Robberg near Knysna, and described by Péringuey in his work. There seems to have been a large bone industry at these coastal sites directly resulting from the presence of strong-winged sea-gulls, duikers, etc., whose wing bones make admirable implements with the minimum of exertion. The bone tubes often found are of interest; they are sometimes decorated with incised lines, sometimes plain — no paint has been found in them, otherwise we would have an exact parallel with the Upper Palaeolithic paint tubes of Europe.

At Bloembosch, Darling district, Dr. Péringuey found a collection of beautifully made Wilton types, apparently associated with bones of Equus capensis and Bubalus bainii. The site was partly covered by a shifting sand-dune. He associates a similar series, and has illustrated specimens in his work. He states (p. 140):

Mr. J.M. Bain and I discovered small cores, scrapers, and borers; also a few of the crescent-shaped pygmies; perforated ostrich egg-shell beads, but of a somewhat larger size than those of Fig. 144, and not so carefully concentrically rimmed; rough scrapers … large scrapers, but also of the Cape Flats type; a small mortar; nuclei of different types of rock; broken !kwes (i.e. bored stones); a small grooved stone for sharpening awls, etc. Here also we found two brass buttons of the same pattern as those usually found in the Cape Flats middens, as well as a few pieces of Oriental china, derived probably from the wrecks of some Dutch East Indian merchantmen sunk in Saldanha Bay. … Diligent search, three times repeated, failed to reveal any other kind of implement that could lay claim to great antiquity — no boucher or vestige of it was met with, no long knife-scraper showing sign of old age.

The double association of extinct animals of Recent age, and china sherds is not very convincing, but the site is of great interest in itself.

At Montagu Cave, Dr. K.H. Barnard of the South African Museum found a very considerable number of Stellenbosch implements, in three deposits, each separated from the next by a sterile layer. Above a further sterile layer were discovered a number of implements showing strong similarities to the Wilton Industry. Crescents, Wilton cores, blades, some of Châtelperron type, etc., were common. A sandstone fragment also occurred showing signs of having been used to grind down bone points. With these were found a duckbill end-scraper extremely like Smithfield types, and one or two unusual flakes of a blade-like character.

I excavated a cave near the source of the Krom River (Clanwilliam district), which rises in a pass skirting the Sneeuwberg. At this cave an almost identical industry was revealed. Two or three trial pits were dug:

Pit A at the edge of the cave showed a total depth of 18 inches. The top 6 inches contained pottery. The entire depth showed Wilton implements with apparent additions of various flake knives, etc. The excavation reached rock bottom.

Pit B was dug in the centre of the cave. Pottery was completely absent. The top 5 inches was sterile; below this to a depth of 18 inches Wilton types appeared. The excavation cut into an accumulation of roof fragments containing no implements.

Excavation was next carried on in the talus, where exactly the same types of pottery and of implements appeared. This cave contained a few associable paintings, consisting of conventionalised signs, dots, imprints of small hands, etc., which link up with Riversdale and Wilton. The colour is dark red, and the floor deposit similarly contained fragments of dark red ochreous pigment, pointing to the paintings being contemporary with the deposits. Dr. K.H. Barnard and Mr. Primos, who were present, suggested that these paintings were the efforts of modern herd-boys.

Little is known of the Western Province and South-west Africa. Wilton-like flakes appear from Zak River, one or two crescents from the Orange River, and a few Wilton sites are known in the Kalahari. I have also been shown Wilton types, crescents, beads, etc., from Walvis Bay, which do not differ at all from the types common elsewhere.

Normal Associations.

The normal and constant associations composing the Wilton Industry are (Plate XLI):

• Crescents.
• Small end-scrapers.
• Thumbnail scrapers.
• Horseshoe scrapers.
• Pygmy cores.
• Ostrich egg-shell beads.
• Bead borers.
• Bead shapers (grooved stones).
• Pottery, of the eared and pointed-base type.
• Bone points (“awls”) and their sharpeners.
• Paint fragments.
• Paintings.

To these may be added objects which have not been specifically mentioned in the sites given above, but which are almost always present, namely, lower and upper grindstones, which in no way imply a knowledge of agriculture, and of a type identical with those used by the modern Bushmen of the Union and South-west Africa. The lower grindstone consists of a river stone, on one or both faces of which a shallow groove, usually about 2 inches wide by 6-8 inches long, appears. The upper grindstone is a worn fragment of a pebble, the grinding face of which is slightly curved or flat, and has at its centre a slight pitting. The modern Bushmen use exactly similar stones for grinding the seeds of the tsama melon. It seems more than likely that these folk had no knowledge of agriculture, as we find no signs of agricultural implements, no signs of domesticated plants, or of used agricultural lands, etc., while the modern descendants of these folk, who have not really advanced in any way, nor appreciably deteriorated where relatively pure, have no knowledge of agriculture even of a primitive type.

Less constant elements appearing at Wilton sites are:

• Painted grave-stones (Plate XLII).
• Bored stones.
• Bone spatulae (Plate XLIII).
• True awls (Plate XLIII).
• Nacre and egg-shell pendants.
• String of twisted fibre.
• Rope of plaited fibre.
• Grass sieves.
• Basketry.

Painted grave-stones are known from Wilton, Robberg, Tzitzikama, Coldstream, and one or two other sites. The specimen from the Wilton site is painted plain red. Robberg shows animal figures and human forms; Tzitzikama shows a whale, a group of three human figures (Plate XLII), etc. Coldstream shows three female figures and a child.113 Bored stones may be an intrusive element from the Smithfield as Lowe has suggested, but occur fairly commonly, especially at the Cape Peninsula. Spatulae are uncommon. I have found a racquet-shaped flat bone specimen at Noordhoek, and two or three others are known. True awls, bone points with the joint left as a handle (Péringuey, Plate xxiii, 172), occur in large numbers at the Robberg Caves, Knysna, and at Witsands; they occur sporadically in cave and midden sites of the south coast, and are usually made of the bones of sea-birds. Nacre and shell pendants seem to be confined largely to the Eastern Province. String and rope are uncommon owing to their short life, but they are known. Plaited sieves, made by the twined or wrapped method (a straight warp, with twined or wrapped woof), are known, but are only fragmentary. Basketry is really unknown, but fragments of pottery occur upon which twined basketry has been impressed before baking as a design, proving that it was used by these folk.

Types of Sites.

There are in general three types of Wilton sites:

Open-air sites.

These occur normally in parts of the Union where caves are uncommon (dolerite formations, etc.). Nothing need be said of these save that they are usually surface sites, in some instances implying the presence of huts of some sort by the distribution of implements. The unprotected state of the implements results in an almost complete absence of bone, ostrich egg-shell, and often of pottery, though we may presume that they existed. Only stone implements are to be expected at such sites. Burials and paintings are so far unknown.

Cave sites.

A full range of objects may be expected. The drier the cave the greater the chances of an interesting haul of implements and their associated objects. Paintings and skeletal remains are common and important.

Midden sites.

These sites are of interest to the archaeologist because they seem to represent an inland people, who have taken to a coastal type of subsistence. The term “Strandloper” has been badly applied to these folk, and it would be better to use the term as a verb, implying a strandloping type of subsistence. Mr. Burkitt has suggested a new industrial grouping for these people, but this would seem unnecessary. The midden sites often show a full range of Wilton implements, pottery, bone points, and ostrich egg-shell beads, though many of the other elements have entirely disappeared, owing apparently to the presence of moisture. The cave middens do, however, show a full range of the commoner associations. Added to these are two elements not found inland, vast accumulations of fish-shells of modern type; and a number of rough instruments, apparently water-worn boulders, hacked into shape by use (not for use, so far as can be judged) in removing shells and breaking them open for eating purposes. These are unconventionalised and formless, though they often simulate the coup-de-poing and discoidal artefact (fabricator) of the Stellenbosch Industry.

It is, however, necessary to note the changes apparently resulting from this mode of subsistence. Bone awls increase, amorphic stone instruments appear, grave-stones are more common, and appear to have been substitutes for the cave-art of the inland peoples, and in fact show human and animal figures of comparable types. At the same time flesh foods seem to have been largely replaced by fish, and the bulkiness of the shells compared with their contents gives an immediate impression of a vast population or a long continued residence. The populational density most certainly was greater than that of the inland folk, owing to the higher rainfall of the southern and eastern coastal belts of the Union, and to the increased supply of vegetable foods directly resulting from this, together with the abundant shell-fish supplies of the rocky coast. But even with this increase of density, only a comparatively small number of people could have subsisted in a given area.

Mr. Drury, who has carried out a large number of excavations for the South African Museum, suggests that the inland peoples trekked periodically to the coast, and he brings forward remains of sea-shells in inland caves to support his contention. Mr. Hewitt shows similar proof in his caves, where nacre pendants of marine origin make their appearance.

There is, on the other hand, the fact that many of the midden deposits fail to reveal Wilton or Smithfield “C” implements, the only stone objects appearing being the bored stone and the formless unconventionalised stones typical of the midden folk, and occurring throughout the world. Pottery of the usual type is often associated. This is very largely negative evidence. The extreme difficulty encountered in any attempt to excavate a midden without a carefully graded series of at least two sieves makes positive evidence almost impossible to obtain, and I have in several instances discovered Wilton implements in middens previously regarded as sterile, while, on the other hand, I have failed to obtain any small implements at all from other middens. It is thus possible that many middens are refuse heaps from an evolved or deteriorated Wilton, which has discarded the microlithic side of the industry as unnecessary.

One further subject is of extreme importance, but has not hitherto been sufficiently studied: the possibility of the middens either being of different ages, or having been made in some instances by peoples of mixed cultures. The Knysna and Mossel Bay cave middens almost certainly show a mixture of Mossel Bay types with Wilton implements; both series of caves have, however, been so disturbed that it is impossible to say whether an actual mixture or a sequence is here represented. Similarly Mr. Fitzsimons has deduced a mixture of Stellenbosch and Wilton types with a possible further addition of the Mossel Bay variation, from his caves at Tzitzikama, but the similarity often existing between the formless midden instruments of the coastal sites and rejects from the Stellenbosch Industry has already been pointed out. From photographs published by him it seems very probable indeed that the “coup-de-poing” depicted are actually the shell-cleavers normally associated. He also deduces an early heavy boned race followed by a series of San peoples, a conclusion similar to that reached at Skildegat, but shows no sequence in the types of implements discovered, a point of immense importance which seems to have been overlooked. The discovery of stratification and of associations are the first and last duties of the prehistorian.

Stratification.

The fact that our knowledge of Wilton material is largely confined to cave sites, together with the general absence of Earlier or Middle Stone Age implements from caves, makes stratification relative to Wilton implements a difficult subject. The richness of the bone, fibre, wood, and paint to be found in Wilton deposits, together with the common association of modern objects with Wilton types, point to this industry lasting well into the modern period.

The Montagu Cave114 shows distinct stratification, determining the fact that the Wilton material is here later than the Stellenbosch. Similar conclusive evidence is available from Riverton and elsewhere on the Vaal River where Wilton implements occur on the surface above gravels containing Stellenbosch types. Middledrift in the Cape Province shows evidence of a like sequence. Much the same type of proof is forthcoming from Stellenbosch, where the implementiferous gravels are of an age much greater than the surface sites of Wilton type occurring on the Stellenbosch Flats.

Stratification between Middle Stone Age material and the Wilton Industry is similarly difficult to find, and has apparently been overlooked in many instances. At the Skildegat Cave, in the Fish Hoek valley, Still Bay material underlies Wilton debris of a “strandloper” type. Other Cape Peninsula sites showing both Wilton and Still Bay types are in sand-dunes, and stratification is difficult to obtain or to prove, but attention has already been drawn in the Middle Stone Age paper to the possibility at the Noord Hoek site of the Still Bay implements having been made near the seashore, and to the Wilton types having been left subsequently on a “false” or “apron” beach which had by then accumulated to the westward to enclose the Noord Hoek vlei or lagoon.

The sequence of physical types at the various Tzitzikama caves and at similar sites is of extreme importance, as, while it does not give us direct evidence, it re-affirms the sequence of an earlier Boskopoid type, which is proving to be associable with Middle Stone Age material, and a later San type which can be definitely associated with Later Stone Age industries.

If we regard the presumption that Smithfield “C” and the Wilton are contemporaneous as valid, further proof of sequence might be deduced.

It does not seem unnecessary to appeal once again for more careful archaeological excavation with the intention of discovering stratification and association, with considerably less of the body-snatching methods of the ingenuously amateurish grave-robber.

South African Neolithic Elements.

It is an extremely difficult thing either to define or to confine the term Neolithic. The polishing of stone is not in itself a sufficient criterion. In Europe the term has been applied to a number of cultures or civilisations which may or may not have been related to one another. The signs by which their Neolithism (if such a vile word may be coined) is recognisable are the presence of the polished stone celt, of pottery, domesticated animals, and agriculture, to name the four most evident criteria. In the Union we have no evidence of the presence of these four elements in complete association. The Smithfield “B” and “C” Industries and the Wilton all show the presence of pottery. Smithfield “B” most certainly shows polished stone-work, but there is no evidence whatsoever of the polished stone celt in actual association with any of the Later Stone Age Industries. It has already been shown that the presence of grindstones does not necessarily imply a knowledge of agriculture in spite of Mr. Burkitt’s suggestion.116 Of domesticated animals there is no sign; even the dog is missing from cave or midden deposits, though we may presume its presence from gnawed bones of various large buck which appear from time to time.

The Rev. Neville Jones117 objects to the application of the term Neolithic to the Later Stone Age Industries.

The South African Upper Palaeolithic is sometimes spoken of as “Neolithic,” but such a course appears more unsafe and misleading than to apply the cultural equivalent names, since to use this word is at once to suggest an age of polished stone implements, through which, as far as we know, this country has never passed. A few polished stone implements have been found in Rhodesia, but the only one about which I can obtain any information was found in an ancient working at Penhalonga. I am inclined to believe that it is not of very ancient date, and it cannot at least be produced as evidence of the existence of a Neolithic culture.118

On the other hand, Mr. C.H. Heese of Riversdale is keen on the term “Neolithic” being used to include all ground stone-work of human origin.119

It is necessary to admit that if the term Neolithic is to be applied only to the slight and relatively obvious cultural advance which is implied by grinding and polishing implements (vastly increasing the ability to design tools and decreasing the number of rejects as it did), then the term must necessarily be extended to include the Smithfield, in spite of the fact that the normal Smithfield implements are basically of Upper Palaeolithic or Mesolithic origin. If, on the other hand, the term is intended to apply to a phase typified by a particular implement (the celt or polished axe), then the term must be limited in applying it within the Union. Personally I would prefer to limit the term to industries showing at least three of the four elements mentioned, above — the polished celt, pottery, agriculture, and the domestication of animals, together with definite burial, the building of huts, a predilection for art, etc.; but our knowledge of the associations of the few known specimens of the celt which exist in the Union is insufficient to allow of such limitation.

The finds which I wish to describe here under the heading of “Neolithic Elements” are of two types. The first consists of a single, beautifully made “axe,” possibly ornamental, but certainly designed with a distinct cutting edge; the second group consists of three individual celts in various museum collections in the Union.

The first type is described by Hewitt:120

This specimen was found in a rock crevice at Paradise Kloof, Grahamstown, by Master Keith Glass, who presented it to Albany Museum. Within 100 yards of the site of discovery there is a “Bushman” cave with paintings; and from the floor of this cave various pygmy implements have been taken. The implement has some resemblance to a Zulu battle-axe in miniature, being sub-triangular in shape and having a well-defined but short tang. There is a sharp cutting edge of 65·5 cm long, formed by grinding from both surfaces. The implement was evidently made from a flattish slab of stone, originally about one-third of an inch thick, and has deliberately been converted into its present shape by grinding. The surfaces generally and the sides have been well smoothed and no angular ridges have been left anywhere. The short tang seems to have fitted into a haft, and is slightly grooved at the base for binding purposes.

This last statement I disagree with, but the point is not fundamental; the tang appears to be more in the nature of a knob by which the “axe” might be hung by string as an ornament or a “pocket knife” (text-fig. l). The shape and size of this object are reminiscent of Bantu razors, though these are always of metal. I make no apology for divorcing this find from the Paradise Kloof material described previously in the Wilton paper. It is an isolated find, not occurring in the cave, and differing from all Wilton elements in that it has a polished cutting-edge, a conception apparently foreign to the makers of our Later Stone Age material. Mr. Hewitt, perhaps rightly, links this element up with the ground or scratched “palettes” previously described as of Wilton type in this volume.

It is necessary to point out the following facts which differentiate this “axe” from other South African specimens:

1. The entire implement is shaped and is symmetrical.
2. The specimen is tanged.
3. The implement is flat and thin, and could not have been used as a true axe.

Let us turn to the celts of more normal Neolithic type. In contra-distinction to the Paradise Kloof specimens,

1. Only the blade or working edge is ground or worked at all.
2. Specimens are not tanged.
3. The body of the implement is always heavy and robust, the working edge being formed by an angle of perhaps 40°, and is eminently suited to use as an axe.

Within the Union there are three specimens of this latter type, only one of which has been at all adequately described. These specimens are housed in three museums, the South African Museum, the Albany Museum at Grahamstown, and the Stellenbosch University collection, and I am indebted to Professor de Villiers for permission to photograph and describe this last. I believe a further note is to be published on this specimen by Mr. C.H. Heese in the South African Journal of Science. Mr. Lowe has also submitted a note to Man on the implement.

The implement which has been most fully described121 is from the farm Vaal Krantz, a few miles from Spitzkop, near Grahamstown. It was picked up from among a number of stones which had been thrown out in digging a water-furrow, and nothing is known of its history apart from this. Péringuey quotes the analysis of Du Toit:

The axe is manufactured out of a small, weathered lump of dark-greenish quartzite — the material of which can be matched at many localities on the border of the Karroo. The lump of rock shows on one side the surface which has been exposed to the air, i.e. it is smooth, brown, and slightly polished. The under-surface is lighter in colour, rough, and slightly pitted. The lump was subjected to scarcely any trimming and the edge was then ground. The scratches were made in the operation of grinding, and are most decidedly not of glacial origin. The shape and scratches of the lump entirely negative the idea that it was a glaciated pebble from the Dwyka.

This last statement was evoked by Dr. Schonland’s belief that the pebble was glacially striated and hence of Carboniferous age in origin. He did not in any way imply that the worked edge was of Carboniferous age. A further remark on these scratches is made by Hewitt, who states that “there are abrasion marks, clearly indicating that it has been bound up in attachment to a haft.” The natural shape of the original stone has been utilised and only the working edge has been ground. The shape is much that of the front view of a helmet, the straighter edge being worked (see Péringuey’s illustration).

The second specimen, housed in the South African Museum (S.A.M. 3449), is described as having been found “7 miles from the Peddie coast,” and thus comes from the same general region as the Vaal Krantz specimen. The maker has similarly utilised the original shape of a water-worn pebble, and has ground the one end to an edge. The shape is roughly that of a tapering cylinder, the wider end of which has been ground and polished. The extreme edge shows signs of use. One face of the polished portion shows signs of having been peeked previous to the grinding operations. The thinner end of the cylinder is of a size fitted to the hand, though this in no way implies that the specimen was not originally hafted. The total length is 12·5 cm, the width of the cutting edge is 6 cm, the width tapers towards the back, and the greatest thickness of the stone is 3 cm. The material is a quartzitic sandstone similar to the Table Mountain Sandstones, and the particular stone was originally a water-worn pebble (Plate XLIV).

The third specimen, housed at Stellenbosch University, was discovered by Mr. P. Krige of Stellenbosch, near Piquetberg, Western Province. It consists of a celt, oval in outline, one end of which has been sharpened to an axe-like edge. The material is an argillaceous sandstone from the Malmesbury Series, and the size is larger than that of the Vaal Krantz specimen, though the shape is similar (Plate XLV). In this specimen also there are signs of abrasion, the ground portion having been roughly peeked to shape before being finally rubbed down. The striations show that the grinding was done with a lateral motion (i.e. parallel with the edge, not at right angles to it).122

Further comment on these two main types of “neolithic” axe would be rash at this stage, but attention must be drawn to other occurrences of very similar implements from Central Africa. The one consists of a small neolithic type of axe, some 3 inches long, by 1 inch or so in thickness, housed in the South African Museum and found on the Uasin Gishu plateau, 1° N. by 35° E., north-east of Lake Victoria. In this specimen, made of a greenish granite, the edge only is ground, while the hinder end of the cylinder shows signs of pecking, as though it had been used as a hammer-stone. The remainder of the stone is completely unworked. A second specimen is depicted by Stainier.123 It was discovered by Commandant (Christiaens at the confluence of the Wele and Bomokandi. It consists of an elongated drop- or pear-shaped stone, the thicker end of which has been beautifully ground and polished to an edge. About one half of the stone has been so treated, the demarcation being neatly defined by an equatorial line; behind this line the stone shows signs of having originally been a water-worn pebble. Schweinfurth124 writes of similar Neolithic axes from the Congo. Other specimens of like type have been reported from further south, from the Katanga Province, from Battlefields, Southern Rhodesia, a specimen from Northern Rhodesia, etc.

It seems probable from the evidence produced that we have, in Africa generally, and specifically in the Union, elements which can only be adequately described as Neolithic, and to which this term should be confined. The elements consist of edge-polished river-pebbles, chosen for their shape, but not otherwise worked. Pecking or abrading seems to have constituted the preliminary shaping of the chopping-edge, which was then completed by grinding and polishing.

How these elements fitted in with the Later Stone Age Industries it is as yet impossible to say. Whether the idea behind the ground objects of the Smithfield Industry and the palettes of the Wilton was taken over from the bearers of the Neolithic elements by these folk; whether we can associate the ground chopping edge with a living people, an early invasion of Bantu or Hottentot peoples; whether the Neolithic elements constitute a normal part of our prehistory or are merely sports thrown off in the evolution of local industries, are all questions which must remain unanswered until such time as our prehistory is a known science, rather than an inchoate feeling into the past.

Addenda.

A Few Notes on the Archaeology of Sheppard Island.

The text-figures here reproduced demonstrate an occurrence of stratification at Sheppard Island, the complete details of which were collected only after the remaining papers of this volume had gone to press. The discovery, however, is of such moment that it is deemed desirable to include a brief statement as an addendum.

Stimulated originally by Professor Dart’s report on “Mammoth and Man in the Transvaal,”93 I paid a special visit to the island, and as a result of Mr. Sheppard’s invaluable assistance, my own observations in “diggings” untouched until the day of my arrival, and a detailed tacheometric survey, it is now possible briefly to report as follows:

The island, with a superficial area of about 35 acres, is situated in the Vaal River, about 10 river miles upstream of Bloemhof: lat. 27° 40′ S., long. 25° 45′ E. It may, therefore, be described as belonging either to the South-western Transvaal or to the North-western Orange Free State. Confined by the old river bed on the north and the new channel on the south, it is an island in the proper sense only during the summer (rainy) season. In winter, when the run-off is at a minimum, all the water passes down the new channel, and the old, elevated bed is left high and dry-the bed of the new being about 10 feet lower than the lowest point of that of the old. It is extremely fortunate that the opportunity for a visit not only coincided with the opening of the new diamond “diggings” but also that it occurred during one of the driest seasons on record. It would otherwise have been impossible to obtain so much information, for during the pleasanter summer months — October-April — both channels are liable to be flooded and the island is then an island in the literal sense.

Professor Dart was led to believe that the island is a relic of the old valley. This does not appear to me to be the case. Rather does it seem to have been formed by a process of siltation while the river was altering its course — an alteration that is still going on. Mr. Sheppard has lived on the island for the past twenty-one years, and he assures me that its contours have changed appreciably during this time — an assurance that I anticipated. A “slice,” “several feet thick,” has been eroded from the Orange Free State or left bank, while to the southernmost arc of the island, on and near the line of section, has been added a slice that reaches a maximum thickness (on the X-X line) of 10 feet. This growth of the island is clearly indicated by a row of willows that originally grew along the water’s edge and now only appear half-buried in silt, some 10 feet back from the stream.

My impression is that while the heavier and older D gravels (fig. 2) were in process of deposition, more particularly at and near the initial phase of the deposition, the river conformed more or less to the contours of the old river bed — in a channel perhaps somewhat wider than it presently is. On account of the inclination in the underlying rock, the tendency of the river was in a southerly direction, and because of the depression marked “Deep Pool” on the site plan, a barrier was created. This gave rise to a tendency in the stream to erode the southern or left bank and to silt up the original or old channel.

This process of erosion and deposition was extremely slow, so slow in point of time that not only were the lighter and later C gravels piled up before the river had reached its present channel, but the D gravels had changed considerably in nature and composition — appearing over half the present channel as lighter and less compact D₁ gravels. This linking of the D and D₁ gravels is due entirely to the fact that both are diamondiferous, for they differ in all other respects.

In addition to the slope of the underlying bed-rock of shale and its influence on the direction of the stream, the deposition or piling-up of the later and lighter C gravels provided the culminating phase of the obtruding influences and hastened the alteration of the stream. The further left or south the channel moved, the greater the eddies on the right, and these eddies initialled the deposition of the heavier clays that presently constitute the island; the lighter loam and sand being kept in suspension and so carried away with the current.

All the material that appears in the island, a uniform heavy, black clay — known locally as pot-clay or turf — appears also in the confining banks, but in these latter there is a preponderance of lighter, yellowish sandy loam. Thus it would seem that the deposition of the lower, heavier, and more compact D gravels, separated from those above by a distinct line of demarcation, took place during a definable geological phase, and that during the greater part of this phase the main flow of the Vaal was more or less confined to the old river bed, with an underlying, ever-present and ever-developing tendency to scour toward the left.

The Gravel Beds.

The “D” Gravels.

Due largely to undulations and irregularities in the underlying bed-rock, these lower gravels vary in depth from 6 inches to 6 feet. They comprise a compact mass of heavy, water-worn boulders and contain both rolled and unrolled types of Stellenbosch implements lying cheek by jowl with completely mineralised and very slightly damaged remains of mammoth: (i) Archidiskodon transvaalensis, sp. nov. and (ii) Archidiskodon sheppardi, sp. nov. (Dart, op. cit.).

The occurrence of both worn and unworn implements — the wear in some is such that they are barely recognisable, while in others the implements are so fresh that they might well have been made yesterday — indicates occupation during the deposition of the gravels. Whether or not this occupation was shared by the mammoths described, is, as yet, impossible to say, but the teeth show so little evidence of having been rolled and are so jagged, yet complete at the roots, that contemporaneity is strongly indicated. I have little hesitation in venturing the opinion that we are here dealing with the remains of co-existing mammoth and man, and therefore do not agree with Dr. Broom when he says: “… it may be regarded as almost certain that they (the teeth) are very much older than the lowest gravels of the Vaal …”94 The geological formation of the Vaal catchment upstream of Sheppard Island is such that had these teeth come from some older formation, they must have come a considerable distance, and must therefore have been subjected to such knocks and rolling as to suffer considerable damage en route. The teeth, however, are so well preserved that it is extremely difficult to account for their occurrence in these gravels unless they were freshly dropped at or near the scene of the discovery while the gravels were in process of deposition. When Dr. Broom sees these teeth and the surroundings from which they come, I am sure he will find it difficult to adhere to this expression of his opinion.

The possibility of mammoth and man having been contemporaneous in Southern Africa is a vital point in the study of local prehistory, and it is unfortunate that authoritative opinions should be expressed when the available data are incomplete.

The continuity of the D gravels under the island is demonstrated in two shafts. These are indicated on the site plan. In shaft No. 1 the gravel occurs at a depth of 27 feet and rock 2 feet lower down; in No. 2, at a depth of 37 feet and rock 3 feet lower down. Also along the extreme eastern and western shores of the island, the gravels are D — as shown on the site plan.

The “D₁” Gravels.

Where the gravels reappear along the southern shore, we find they are no longer compact and heavy but comprise a mass of light, loose, water-worn pebbles (the term boulders is hardly applicable), sand and clay. The only feature shared with D is that they are diamondiferous, but even in this there is a difference, for whereas from a given volume from D we may obtain ten diamonds, from an equal volume from D₁ we may obtain only one. The deposition as a whole, i.e. the apparently continuous D-D₁ bed, suggests a petering out of original or D elements as we proceed from the old to the new channel, and indicates an appreciable lapse of time between the beginning and the end of the deposit as measured between the old and the new channels. Because of this, and because we seem here to have the tail-end, as it were, of the original deposit, I have named these southernmost new-channel gravels D₁.

Due again to undulations or surface irregularities in the underlying bed-rock, these D₁ gravels vary in depth from a few inches to a few feet. They contain both rolled and unrolled implements of Fauresmith type. No animal remains have been found and no implements of Stellenbosch type.

Every indication leads one to assume that this new channel deposit is the tail-end of the original D bed. The depreciation in diamond content, supported by the occurrence of implements of a later age, suggests most strongly that the deposition of these lighter D₁ gravels constituted the culminating phase of these original deposits. This assumption is strengthened when we examine the next or C deposit.

The “C” Gravels.

Due partly to the inclination of the underlying D gravels and largely, I imagine, to eccentricities of cross-currents and eddies, these overlying C gravels vary in depth from 18 inches to 16 feet. The deposit is made up of light, loose, water-worn pebbles and sand, interspersed with a few boulders, and contains both rolled and unrolled remains of Fauresmith type. The line of demarcation between these upper C gravels and the underlying D deposit is clear. There is no gradual merging of heavy to fine material, and the differences between the two layers are unmistakable. The deposit is identical in appearance and texture with that at D₁, and differs from it only inasmuch as, where in D₁ we find diamonds, we do not find them in C.

The Smithfield Occurences.

On or near the surface of the gravels and not only on both the Free State and Transvaal banks, but also on the island itself, are occurrences of Smithfield types; rare, it is true, but there nevertheless.

In the old river bed I recovered a fragment of a stone ring in shale and on the island a typical trimmed point of side-scraper type in lydianite. Bored stones, duckbills, and thumbnail scrapers have been found on both banks. These remains are Smithfield “B,” and the period of occupation appears to have been entirely recent.

Discussion.

Here, therefore, we have a clear case of Stellenbosch-Fauresmith stratification, the old valley having been occupied at different times by people practising different industries.

There is just one complexity: while the remains in the later and upper D₁ and C gravels respectively are typical of the Fauresmith Industry and include the characteristic small and neatly made flake coups-de-poing, trimmed points with faceted butts and unconventionalised scrapers, yet there occur a few such advanced specimens, largely unrolled, that we may be in the presence of either one or two further probabilities:

1. If the assemblage from the upper deposit is entirely Fauresmith, then we are here in contact not only with an outlier of the Industry, but with the most advanced phase hitherto recorded. This, if it be so, may necessitate a revision of the chronological horizon of the Industry, for the affinities of the assemblage are markedly Middle Stone Age. We have more evidences of “Mousterian influence” here than is the case in those assemblages from the real Fauresmith terrain — the South-western Free State. Or
2. We may be in the presence of a mixture, an overlapping of two distinct yet partially contemporaneous cultures — the Fauresmith, with implements rolled and unrolled, and an early phase of the Middle Stone Age, with implements lightly, but largely unrolled. If this is the case, then the state of preservation of the implements inclines one to look upon the appearance of the more typical “Mousterian influences” or affinities as the more recent of the two. The implements, however, appear throughout the upper C and D₁ deposits, and wider and more detailed investigation will be necessary before any more definite opinion can be expressed. Several artificially rounded stones have also been recovered from these later gravels, and as it has been possible in other parts of the Union (notably at Elandslaagte near Heilbron, and at Floris Mineral Baths in the Free State) to associate these with the Middle Stone Age, the occurrence is significant.

My own leanings are strongly inclined towards this latter probability, i.e. that we have an intermixture and partial overlapping of two distinct industrial groups.

We, therefore, have the sequence: Stellenbosch-Fauresmith-Smithfield “B,” with the possible addition of a phase of the Middle Stone Age between and separating the Fauresmith and Smithfield, thus giving us the most important and illuminating stratification hitherto recorded: Stellenbosch-Fauresmith-Middle Stone Age-Smithfield “B.”

In this brief summary it is possible to review only four further points of interest:

1. The gravels as a whole combine to form the most recent deposits or “terrace” of this type in the Vaal, and the lower D deposit may therefore be regarded as Pleistocene — probably Late. There is no reason to presume geological antiquity; rather is the reverse the case, i.e. that these gravels, despite their implementiferous and fossiliferous content, are relatively modern.

2. In the Stellenbosch assemblage from the older D gravels, coups-de-poing and biseaux occur in even numbers. All the biseau shapes illustrated in text-fig. 2 of Mr. Goodwin’s Stellenbosch paper in this volume appear, and I fail to see how even the sceptic could escape the conviction that both types were designedly made.

   In its essentials, the general assemblage is certainly characteristic of the Stellenbosch Industry, but it shows a general advance over the “Southern Mountain” group. It is much more closely allied to the “Vaal River” group — a group that I incline to regard as the most advanced and probably the culminating phase of the parent or Stellenbosch Industry. We find such a general improvement not only in the implements, but also in the technique, that we may well be in the presence of the latest developments of the Industry and so near the beginning of that transition that led — largely due to the use of a new material, lydianite — to the Fauresmith. The Fauresmith in turn, in this very zone (a zone which we may well term the Lydianite Zone) later evolving to the Middle Stone Age — either as an autochthonous development or due to the iniltration from the north of what we may tentatively term “Mousterian influences,” or perhaps both.

   I am more convinced than ever that in the valley of the Vaal lie hidden the greatest and most important secrets of the prehistory of this country, and that not until it has been more thoroughly explored shall we solve the many problems that presently confront us.

3. While the Stellenbosch implements are all of materials taken from the lower D gravels — quartzites, amygdaloidal lavas, diabasic rocks, etc. — the Fauresmith and Middle Stone Age types are almost exclusively of lydianite — the material so favoured during these industrial developments.
4. While the Stellenbosch implements were originally discovered as incidentals to the occurrence of diamonds, the position over this area of the Vaal is such to-day that the presence of a large coup-de-poing or biseau in a newly discovered gravel is almost as sure an indication of the presence of diamonds as is the “bandom” or “bantam.” In other words, if an experienced digger discovers Stellenbosch-type remains in a gravel, he immediately feels that the chances are a thousand to one that he will also discover diamonds!

Before concluding this brief statement, I must record my great indebtedness to Mr. H. Sheppard, the owner of the island, not only for his invaluable assistance, but also for such information during my all too brief visit that, but for him, my work would have been hopelessly inadequate. As a guide to those hopeful of furthering research in this area, I must also add that my entire Sheppard Island collection, as well as the details of the survey, have been given to the University of the Witwatersrand.

The Manufacture of Ostrich Egg-shell Beads among Modern Bush-folk.

Mr. James Drury of the South African Museum has for the last twenty years been perfecting a process for casting posed figures of the living body. The results of his work are housed in the Museum, and give an interesting survey of the racial types grouped under the term “Bushmen.” He has given me an interesting account of the process of manufacturing ostrich egg-shell beads employed by the Naron tribe,115 near Sandfontein (about 20° E. by 22° S.). The bead-making is done by the women of the horde. A woman takes an ostrich egg-shell which she breaks into large pieces; she then bites these into bits about half an inch across. One fragment is now placed on a hard piece of skin and an iron-pointed wooden drill some 20 inches long is applied to the centre of the fragment. The drill is held between the open palms of the hands, and rotated rapidly to and fro. When the shell is almost pierced, it is reversed and the drilling concluded from the opposite face. A pointed rimer, consisting of a small handle of wood in which is a fragment of iron, is inserted and the hole carefully rounded and finished off. The fragment of shell is roughly trimmed to a disc by placing it on an anvil stone and chipping away the unwanted material with a hammer-stone. The shell now shows a rough symmetry. These fragments are strung on a length of gut which is knotted tightly at each end, compressing the drilled shells into a solid stick, and a piece of hemp is twisted between the shell fragments further consolidating the whole. This stick of shell fragment is now placed on the hard skin, and a stone in which is a groove of the desired size is rubbed along it, so as to grind the beads to a perfectly symmetrical circular shape, all of one size. The beads are now released and cleaned, and are added to the string necklace, etc., which is being made. Mr Drury adds:

The woman broke three out of every four beads while making them, so one can imagine what a tedious process it was. The string she was making was nearly 12 feet long, and the time taken to make this was some three months; it would be sold for two handfuls of tobacco.

If we reckon twelve beads to an inch, this means 1728 beads. The “cash value” seems out of all proportion to the labour, but the only gauge is the value of the tobacco to the beadmaker and the value of the ornamentation to the buyer.

Text-fig. 2 shows fragments of beads, stone bead-borers, and a grooved bead stone, implying that the method now employed is identical with that originally used by our Wilton and Smithfield folks.

Footnotes

1. du Toit, “The Carboniferous Glaciation of S. Africa,” Trans. Geol. Soc., Oct. 1921, p. 188. Also Coleman, “Ice Ages,” London, 1926.
2. See Schwartz, “The Kalahari Project,” Soc. de Geog. Geneve; also ibid., A South African Geography, London, 1921, and in his various other works. For Jubaland and Uganda see Hobley, London Geographical Journal, 1914, p. 44, and L. Harger, Uganda Natural History Society, vol. vi, p. 192. Similar evidence for various parts of Africa can be found in Brooks’ “Climate through the Ages,” London, 1926.
3. Penning, “Gold and Diamonds,” London, 1901.
4. Cf. the Batoka Gorge, Victoria Falls, and the gorge below the Aughrabies Falls, Orange River.
5. L’Anthropologie, vol. xxiii, 1912, p. 151.
6. Péringuey, Annals S.A.M., vol. viii, p. 17.
7. Op. cit., p. 47 et seq.
8. Op. cit., p. 49.
9. But compare Burkitt, South Africa’s Past in Stone and Paint, Cambridge, 1928, p. 59.
10. Op. cit., locality described, p. 65 et seq.
11. Op. cit., p. 51.
12. Goodwin, “The Montagu Cave,” Annals S.A.M., vol. xxiv.
13. U.C.T. Engineering and Scientific Society Journal, 1922, pp. 36-50.
14. Trans. Roy. Soc. S.A., vol. xvi, 1.
15. J.R.A.I., vol. xlv, pp. 81 et seq.
16. Prehistoric Period in South Africa, pp. 25, 26. See also Trans. S. Air. Phil. Soc., 1905, vol. xvi, p. 107.
17. “Stone Ages of South Africa,” Ann. S. Afr. Mus., vol. viii, 1912.
18. Nature, vol. cxx, 10th Dec. 1927, 3032, p. 41.
19. Ibid., vol. cxxi, 3rd Mar. 1928, 3044, p. 324.
20. Trans. Geol. Soc. of S.A., 24, 1922.
21. Johnson, op. cit., pp. 28 and 54.
22. Van Hoepen, Die indeling en relatiewe ouderdom van die Suid-Afrikaanse klipwerktuie, S.A.J.S., vol. xxiii, pp. 793-809, 1926. See also Oor die Pnielse Kultuur, S.A.J.S., vol. xxiv, pp. 566-570, 1927.
23. See Péringuey, op. cit., Plate VIII, figs. 53 and 56. See also fig. 2 in this paper.
24. See also Péringuey, op. cit., p. 59 et seq.
25. E.T. Hodkinson, The Stone Cultures of the Vaal River Diggings, S.A.J.S., 1926, pp. 876-886.
26. Beck, Geol. Mag., 1906, p. 49.
27. Fraas, Fauna aus den Diamanteifen von sudafrika, Zeits. der deutsch G. G., Heft lix, 2, 1907.
28. Goodwin, Archaeology of the Vaal River Gravels, Trans. Roy. Soc. S.A., xvi, 1928.
29. Haughton, On some fossils from the Vaal River, Trans. Geol. Soc. S.A., 1921 vol. xxiv.
30. Du Toit, Annual Report Geol. Com., vol. xi, 1906, p. 171.
31. Penning, Gold and Diamonds, London, 1901.
32. Van Hoepen, S.A.J. S., vol. xxiii, 1926, p. 793.
33. G. Leith, On the Caves, Shell-mounds, and Stone Implements of S.A., J.R.A.I., 1899.
34. Péringuey, op. cit., p. 15.
35. Neville Jones, The Stone Age in Rhodesia, Oxford, 1926.
36. See also Neville Jones, Implement-bearing Deposits at Taungs and Tiger Kloof, J.R.A.I., vol. 1, p. 412.
37. Lebzelter and Bayer, Stone Cultures of the Zululand Highveld and Northern Natal, Ann. Transv. Mus., vol. xii, iii. March, 1928, Pretoria.
38. One is in the collection of Rev. Fr. Gardner, SJ., Gokomere, Fort Victoria, S. Rhodesia.
39. See Van Hoepen, S.A.J. S., 1926, vol. xxiii, p. 793.
40. Mr. Jansen states that the south hill is a dyke, not a true tafelberg. The local dolerite is of two types, coarse and Hue, and the coarser dyke appears here to have forced its way through the liner.
41. Jansen: “A new type of Stone Implement from Victoria West,” S.A.J.S., xxiii, 1926, pp. 818-825.
42. J. Hewitt and Rev. P. Stapleton, S.J., “On some remarkable Stone Implements in the Albany Museum, Grahamstown,” S.A. Journ. Nat. Hist., vol. v, December 1925.
43. See Man, July 1909, pp. 100-103.
44. In view of this last statement it would be as well to state here that we are not now dealing with the long, curved “knives,” etc., which appear to be purely the results of insolation, and not the result of human handiwork; these specimens appear at first sight to be “plunger-flakes,” but actually show no signs of artificial shaping.
45. MS., 19th November 1926.
46. These must not be confused with the peeked stone spheres found elsewhere and of apparently later age. Mr. Heese here refers to a variation of the “discoidal artefact” or fabricator, or perhaps facetted stone ball.
47. van Hoepen, S.A.J.S., 1927, vol. xxiv, p. 566, “Oor die Pnielse Kultuur.”
48. British Museum Handbook to “Stone Age Antiquities,” 1926, p. 33.
49. “Oor die Pnielse Kultuur,” hm. cit., p. 567.
50. Letter, 25th April 1927.
51. The Fauresmith “Coup-de-poing,” S.A.J.S., vol. xxiv, 1927, p. 502.
52. Goodwin, “The Archaeology of the Vaal River Gravels,” Trans. Royal Soc. S. Africa, vol. xvi.
53. Now called the Stellenbosch Industry.
54. Hewitt, S.A.J.S., vol. xxii, 1925, p. 443.
55. “The Stone Age in S.A.,” article in East London Dispatch, 25th March 1927.
56. “Ouderdom van die Suid-Afrikaanse Klipwerktuie,” S.A.J.S., vol. xxiii 1926, pp. 804, 805.
57. “Boskop Remains from the South-East African Coast,” Nature, vol. cxii, No. 2817, and elsewhere.
58. Man, May 1928, No. 58.
59. The Stone Age in Rhodesia, Oxford, 1926, passim.
60. This is visible in modern Australian work. Compare Spencer and Gillen, The Arunta. Macmillan, 1927, chap. xxvii.
61. See Rogers and du Toit, Geology of the Cape Colony, London, 1909, p. 268, for a, section at this fall.
62. Lebzelter, Annals Transvaal Museum, XII, iii, 1928.
63. Compare Neville Jones, op. cit., p. 113.
64. Hewitt and Stapleton, “Some Remarkable Stone Implements, etc.,” S.A. Journal of Natural History, V, 23-38, Dec. 1925.
65. Ibid., “Some Stone Implements from Rock Shelter, etc.,” S.A.J.S., pp. 574-587, 1927.
66. S.A.J.S., vol. XXV, 1928.
67. Journal Royal Anthrop. Inst., xlv, 79, 1915.
68. For associated fauna see R. Broom, “Man Conternporaneous with Extinct Animals in South Africa,” Annals S.A. Mus., vol. xii, p. 13, 1913.
69. For a section see p. 191.
70. Compare Dart, “Round Stone Culture of South Africa,” S.A.J.S., xxii, 427, 1925. For an illustration of the long point see Goodwin, “The Middle Stone Age,” S.A.J.S., xxv, 1928.
71. Bureau of American Ethnology, Smithsonian Institution, Bulletin No. 60, pt. 1, 1919.
72. Lowe, “Stone Implement Workshops in the Orange Free State,” S.A.J.S., vol. Xxii, 1925.
73. Goodwin, “Capsian Affinities,” S.A.J.S., vol. xxii, 1925.
74. Lowe, “Bored Stones,” S.A.J.S., vol. xxiv, 1927.
75. Stapleton and Hewitt, “Stone Implements from Howieson’s Poort,” S.A.J.S., vol. xxiv, 1927.
76. J.R.A.I., 1926, vol. lvi, pl. xxxvi, figs. 9 and 10.
77. Lowe, “Bored Stones,” S.A.J.S., vol. Xxiv, 1927.
78. F. Rennell Rodd, “People of the Veil,” pl. xxxvii, p. 285.
79. Heese, “Ground and Polished Stone Implements from the N.W. Karroo,” S.A.J.S., vol. xxiii, 1926.
80. Neville Jones, “The Stone Age in Rhodesia,” Oxford, fig. 24, p. 73, No. 5.
81. Ann. S. Afr. Mus., vol. viii, 1911, pl. Xxiv. See also Burkitt, South Africa’s Past in Stone and Paint, Cambridge, 1928, fig. Xix.
82. Vide infra, p. 187, “ Stratification.”
83. J. P. Johnson, The Stone Implements of South Africa, 1908 ; Geological and Archaeological Notes on Orangia, 1910.
84. Lowe, “ Modeler River Man,” S.A.J.S., vol. Xxiii, 1926.
85. Burkitt, op. cit., fig. xviii.
86. Schapera, “ Stylistic Affinities,” S.A.J.S., vol. Xxii, 1925. See also Burkitt, op. mit., pp. 145-152.
87. Lowe, “ The Fauresmith Coup-de-Poing,” S.A.J.S., vol. xxiv, 1927. Also Burkitt, op. cit., pp. 167-168.
88. Lowe, “ Wilton and Smithfield Industries,” S.A.J.S., vol. Xxiii, 1926. Compare Burkitt, op. cit., pp. 93-94.
89. Neville Jones, “Stone Age in Rhodesia,” Oxford, 1926. When this book was written, the new nomenclature had not yet been fixed, but the Wilton Industry is described on page 29 in the Classification Table under “South African Stages, (b) : Cave and surface deposits, with implements of generally accepted Bushman origin.” An excellent illustration of a Wilton assemblage is given on page 73, fig. 24. Nos. 32 and 33 are crescentic scrapers; No. ll is a Solutrean-type point ; No. 15 is a duckbill; and Nos. 17, 19, and 20 are thumbnail scrapers.
90. Compare Sollas, “Ancient Hunters,” fig. 233, p. 434. Compare also Péringuey, op. cit., pp. 166-169, pl. xxv.
91. Bushman Paintings, Oxford, 1909.
92. Goodwin, “ Capsian Afiinities,” S.A.J.S., vol. Xxii, 1925.
93. Nature, vol. cxx, No. 3032, 10th Dec. 1927.
94. Nature, vol. cxxi, No. 3044, 3rd Mar. 1928, p. 324.
95. Péringuey, Annals S. A. Museum, vol. viii, Plate XVI, fig. 132.
96. Burkitt, South Africa’s Past in Stone and Paint, Cambridge, 1928, p. 94.
97. Burkitt, Chapter IX, p. 133.
98. “ On Several Implements and Ornaments, etc.,” S.A.J.S., xviii., 1921, 454-467, with Plates ix-xii.
99. This was their first excavation at this site; they have since paid frequent visits, and have done considerably more excavation, but their original conclusion has not been appreciably changed.
100. Burkitt, South Africa’s Past in Stone and Paint, Cambridge, 1928, p. 47.
101. Lowe, “ The Interrelation of the Wilton and Smithfield Industries,” S.A.J.S., xxiii, 1926, p. 869.
102. Neville Jones, Stone Age in Rhodesia, Oxford 1926, fig. 13.
103. See also Guide to Rhodesia, 1924, p. 364.
104. J. P. Johnson, The Prehistoric Period in South Africa, Longmans, 1912.
105. An interesting parallel is visible in Frobenius’ “Das unbekannte Afrika,” Munich 1923, Plate xi.
106. See Hewitt, op. cit., Plate xi; cf. Plate XLIII, 10, in this paper.
107. Compare Péringuey, Plate xxvi, p. 196.
108. Péringuey, op. cit., pp. 149-151 and Plate xix, 150. See Plate XLII in this paper.
109. Cape Town Diocesan College Magazine, 1892.
110. Goodwin, “ Sir Langham Dale’s Collection,” S.A.J.S., xxv, 1928.
111. Burkitt, op. cit., fig. xix.
112. See Goodwin, “The Hardy Collection,” S.A.J.S., Xxiii, 1926, p. 826. Also Péringuey, op. cit., chapter ix.
113. See Burkitt, op. cit., fig. xx ; also Péringuey, op. cit., Plate xxvii, figs. 199, 200, and 201.
114. Goodwin, “The Montagu Cave,” Annals S.A.M., xiriv.
115. See also Miss Bleek, The Naron, University of Cape Town, 1928.
116. Burkitt, South Africa’s Past in Stone and Paint, p. 102.
117. Neville Jones, Stone Age in Rhodesia, p. 22.
118. The Battlefields specimen was not then known to him.
119. C. H. Heese, S.A.J.S., vol. Xxiii, 1926, p. 789.
120. Hewitt, “ Peculiar Elements in the Wilton Culture,” S.A.J.S., vol. xxiii, 1926, pp. 901-904, with pl. Xix.
121. See S. Schonland, Records of Albany Museum, vol. ii, 1907; Péringuey, Ann. S. Afr. Mus., viii, 1912, chap. Xviii and fig. 26; Hewitt, S.A.J.S., vol. Xviii, 1921, pl. X, and ibid., S.A.J.S., vol. Xxiii, 1926, p. 902.
122. A further specimen, pierced for hafting, has been described to me by Mr. Lowe, from Koster.
123. Annales du Musée du Congo, Serie III. L’age de la pierre au Congo, pl. i.
124. “ Note sur des objets en minerai de fer provenant du pays des Monbouttous” Bull. de I’Inst. égyptien, 2^e series, No. 4, 1883.

Bibliography

1. Burkitt, M.C. 1928. South Africa’s Past in Stone and Paint. Cambridge.

General Website Disclaimer | About | Advertise | Contact

Created and Copyright by Robert I. Sadler, 2002-2012

http://southerncape.co.za/recreation/books/the_stone_age_cultures_of_south_africa_1929.php

3 January 2012