THE PREHISTORIC PER IN SOUTH AFRICA LIBRARY UNIVERSITY OF CALIFORNIA SAN OICOO THE PRE-HISTORIC PERIOD IN SOUTH AFRICA. O U 02 COCQ g s, gs PM PH * o O PS Z ^ pq 3 Pel _ -. ^5 Ed W o S O a a o s z j ^? O y u a z o 05 O fe " Q a ^ ^ a THE PRE-HISTORIC PERIOD IN SOUTH AFRICA BY J. P. JOHNSON Member of Council of Geological Society of South Africa, Author of "Geological and Archaeological Notes on Orangia," " The Ore Deposits of South Africa,' 1 " The Stone Implements of South Africa," tCr. WITH ILLUSTRATIONS LONGMANS, GREEN AND CO. 39 PATERNOSTER ROW, LONDON NEW YORK, BOMBAY, AND CALCUTTA 1910 [All rights reserved.^ PREFACE. THE object of this volume is to give a brief summary of our knowledge of the pre-historic period of South Africa. With the exception of the part treating of the ruins north of the Limpopo, the book is based almost entirely on my own researches. Much of the data relating to the petroglyphs and rock-paintings, however, were obtained as member of the Com- mission (consisting of Professor R. B. Young, of Johannesburg, Mr. T. N. Leslie, of Vereeniging, and myself) appointed by the Government to report on these relics of primitive art. While much detail remains to be unearthed, the main charac- teristics of the pre-historic cultures of South Africa may now be said to be definitely determined. P.O. Box 6231 JOHANNESBURG. July, 1910. CONTENTS PAGE CHAPTER I. INTRODUCTION 1 CHAPTER II. THE EOLITHS OF LEIJFONTEIN 16 CHAPTER III. ACHEULIC IMPLEMENTS 22 CHAPTER IV. ACHEULIC AND SOLUTRIC GROUPS ... ... 84 CHAPTER V. SOLUTRIC SITES 40 CHAPTER VI. PETROGLYPHS AND ROCK-PAINTINGS ... ... 48 CHAPTER VII. THE PRE-HISTORIC BANTU 60 APPENDIX 77 INDEX TO PLACE-NAMES 86 LIST OF PLATES. PLATE /.ADVANCED ACHEULIC IMPLEMENTS. (a) DOUBLE POINTED AMYGDALITH FROM BARKLY, FORESHADOWING THK SOLUTRIC TYPE. (b) AXE-HEAD FROM VEREENIGING, BESEMBLING THE NEOLITHIC FORM. SCALE f. PLATE II. SOLUTRIC PETEOGLYPHS. (a) ELAND PECKED ON ROCK, VEREENIGING. (6) ELAND SCRAPED ON ROCK, BIESJESFONTEIN. PLATE ///.ZIMBABWE RUINS. (a) EXTERIOR OF MAIN BUILDING, SHEWING CHEVRON PATTERN. (b) INTERIOR OF MAIN BUILDING, SHEWING CONE WITH REMAINS OF WALLED PLATFORM IN FRONT. PLATE IV. ZIMBABWE RUINS. (a) SOAPSTONE POST WITH CARVED BIRD AND CROCODILE. (b) MONOLITHS NEAR WEST ENTRANCE OF MAIN BUILDING. PLATE F. PLAN OF MAIN RUIN AT ZIMBABWE (BASED ON MR. FRANKLIN WHITE'S MORE DETAILED PLAN IN JOURNAL ANTHRO- POLOGICAL INSTITUTED XXXV., 1905). PLATE VI. PLAN OF NANATALI (AFTER MR. POPHAM), SHEWING CEMENT PLVTFORMS. THE PRE-HISTOKIC PERIOD IN SOUTH AFRICA. CHAPTER I. INTBODUCTION. GEOLOGICAL and archaeological research has established a definite sequence in the primitive cultures of the Old World, and since this is here used as a basis for the classification of the South African remains, it is desirable to briefly describe the sequence and the nature of the data upon which it is based. Of the geological data requisite for the determination of the relative age of different early archaeological remains , that afforded by river terraces is the most important. A river in the course of its excavating career frequently leaves strips of drift behind on the sides of the valley it scoops out. Usually it has alternating periods of excavation and deposition, that is, after excavating the valley to a certain depth it proceeds to deposit gravel or other sediment in it for a time, after which it begins to excavate again, and so on. Normally the new channel does not occupy the whole width of the valley bottom , so that a strip or terrace of drift is left behind on one or both sides. Wherever there is a succession of such terraces it follows that the highest is the oldest. Further, owing to the protec- tion from denudation afforded by gravel to the underlying rocks, it often happens that the ancient river terrace eventually comes to occupy the top of a ridge, and is finally reduced to a mere capping to one or more hills. Patches of river drift occupying the tops of hills are, necessarily, very ancient. Contemporary objects, such as stone implements and bones of animals, found in a hill-gravel, will, therefore, be more ancient than those found in the valley-drifts of the same drainage area ; and those from the high-level terraces will be older than those from the low r -level terraces. The value of the evidence afforded by river terraces is well illustrated by the accompanying section (Fig. 1) by my friends, Messrs. Hinton and Kennard, through the Thames Valley, Britain, in which the broader divisions of the culture sequence above referred to are plainly indicated. It is obvious that the gravel on the top of the hill is older than the terraces of drift on the sides of the valley, and that the latter are in turn more ancient than the alluvium bordering the modern river. EOLITHIC GRAVELS. The patch of gravel on Swanscombe Hill is one of a number of similarly situated deposits scattered over the extreme south of Britain. They are generally considered to be the remnant of the drift deposited by rivers which were in existence long before the drainage system of the country had assumed its present form. Their exact geological age is uncertain, for, owing to their porous nature and great antiquity, any animal remains that may have been enclosed in them, and by which they might have been dated, have, as a rule, since perished. There is one exception, however, for in a patch at Dewlish remains of an extinct species of elephant the Elephas meridionalis of Nesti have been found in some quantity. This occurrence, as will be seen later, though not fixing the exact minor geological epoch, is of great significance. Now, these deposits have yielded some rudely chipped frag- ments of flint in other words, Eoliths which are claimed by many to represent man's very earliest attempts to trim a piece of stone to a shape suitable for use as an implement. The first and most important discovery of these Eoliths was made by Mr. Benjamin Harrison, in the gravel on the Kent plateau. These implement- like fragments consist of naturally broken pieces of flint with secondary chipping along the edges, and resemble the small scraping tools of later periods. While the writer is inclined to the opinion that these are imple- w ^^ CO ^ r& 0> IH M ' 111 fc I I to historic times. Of the implements found in the Neo- lithic alluvium, flakes and scrapers constitute the vast majority ; the former are always neat and small and seldom attain the size of the average Acheulic flake, while minute examples with three or more facets and a well-developed bulb of concussion, are not uncommon, which shows that the art of producing flakes had now reached its highest. The other implements are still more different. The amygdaliths of Acheulic times are replaced by thin, symmetrical, and skilfully chipped javelin-heads, which are often neatly and FIG. 3. NEOLITHIC CONCAVE AND CIRCULAR SCRAPERS FROM SUTTON, BRITAIN. (Actual size.) uniformly notched on either side to facilitate the hafting ; by beauti- fully finished daggers, not unlike the javelin-heads, but usually with a distinct handle worked at the end of the flat blade ; and by axe- heads with a straight or slightly curved ground edge. The last mentioned are usually more or less polished all over. While evidence of the knowledge of the bow appears for the first time in the shape of often exquisitely finished arrow-heads. SOLUTRIC SITES. There is yet another group of stone implements to be mentioned , and that is the Solutric. They are found on the sites of ancient settlements, in caves and underneath rock-shelters. The animal remains found with them are similar to those of the Acheulic river deposits, hence many regard the Acheulic and Solutric implements as contemporary. The Acheulic stage of culture is represented solely by the stone implements upon which it is based. While the Acheulic implements are nearly all amygdaliths, the Solutric implements are nearly all small scrapers ; they are thus probably complementary to one another, which circumstance also suggests that they may be con- temporary. It must be remembered, however, that the charac- teristic large Acheulic implements are met with in quantity only in ancient river gravels where, though, no doubt, either near or actually at the site of manufacture, small scrapers would not occur even if they were in general use at that stage ; whereas the Solutric imple- ments are found mostly on sites of habitation w-here small scrapers would naturally predominate. Acheulic amygdaliths have, indeed, been met with on Solutric sites, but the evidence has always indicated them to be of earlier date. The fact that many of the Solutric sites yield contemporary amygdaliths that are far more advanced both in form and finish than any known Acheulic amygdaliths, makes it improbable that the two are of the same age. Solutric sites are widely distributed over West Europe, and many of them have been inhabited at different epochs. The Solutric stage itself was of very long duration in that area, and many archaiologists claim to recognise three .successive phases of it, namely, the Aurignacien, Solutreen, and Magdalenien, each charac- terised by a distinct assemblage of implements, though others see in these assemblages only a different facies of contemporary work. While Eolithic, Acheulic, and Solutric forms are well repre- sented in South Africa , no true Neolithic implements have yet been found in the sub-continent. ACHEULIC AND SOLUTRIC IMPLEMENTS IN SOUTH AFRICA. In South Africa Acheulic amygdaliths are sometimes met with in quantity in association with terraces situated at some height above the present rivers, but it is uncertain whether they were con- temporary with the formation of these terraces or whether they were merely manufactured on top of the terraces at a subsequent 10 period. Such terraces afford suitable material for the making of amygdaliths, and would, no doubt, be resorted to for that purpose. The Acheulic amygdaliths are, however, known to occur in terraces situated at but a slight elevation above the present rivers and in old talus deposits. Isolated surface specimens are not uncommon, and having readily attracted the attention of the curio collector, are now known from many parts reaching from the Zambesi to the South Coast, but, unfortunately, the conditions under which they occurred and which might have afforded some evidence as to whether they were washed out of a deposit or actually dropped on the present surface, have seldom been recorded. Associated with the amygda- liths and possessing the characteristic Acheulic style and quality of workmanship are true axe-heads, such as are rare in the Acheulic assemblage of Europe, though well represented in that of India. Solutric sites are widely distributed over South Africa, but the often abundant implements obtained from them are mainly small scrapers, and only one, that at Tweelingsfontein, has yielded an ainygdalith. Solutric amygdaliths have been met with in quantity only at one locality, namely, the Cape Town flats, where they are clearly surface implements. These occurrences are unquestionably newer than the river gravels and taluses that have yielded Acheulic amygdaliths, and when covered over with deposits these last are merely rain-washes, or blown sand, of comparatively recent date. The South African evidence is thus also in favour of the view that the Solutric stage is more recent than the Acheulic. SOLUTRIC ART. Both in West Europe and South Africa, the Solutric stage was characterised by a great development of the artistic faculty. Its most interesting manifestations remain to us in the form of engrav- ings and paintings found on the walls of caves and on the back of rock-shelters, and between those of the two areas there are remark- able resemblances, though there are naturally many differences also. This work, though primitive , by no means represents man's first attempts. One of its most striking features is the bold .and unfalter- ing delineation of the objects depicted. But whether it had its birth 11 in some earlier phase of culture or whether it originated in the Solu- tric stage is uncertain. Of the rude beginnings from which this comparatively advanced art must have grown there are no traces. No doubt they were of a perishable kind, such as drawings made on the ground with the finger or with a pointed stick , and perhaps only the work of the specially gifted was, for various reasons, done in the durable form on rock. From what is knpwn of existing primitive peoples, who are, so to speak, survivals from the prehistoric period, it is inferred that rnbst of these petroglyphs and rock-paintings are of an utili- tarian character connected with the primitive belief in sympathetic FIG. 5. VIEW OF VICTORIA FALLS. magic, and are not the expression of a delight in art, while others are of the nature of pictorial records. They, nevertheless, indicate a considerable development of the aesthetic taste. So far as at present known the Solutric petroglyphs and rock- paintings of West Europe are mainly confined to a comparatively small portion of that area, namely, the Dordogne and the Pyrenees. These all occur in caves, and often in the inner recesses where the light of day never penetrates, and thus differ from those of South Africa, which are always situated in the open. Prof. Breuil, w r ho has investigated most of these occurrences, says that while they all belong to the one stage of culture, they repre- 12 sent every moment of its existence, which was of long duration, in that area. In those caves that have been occupied at successive epochs one can often establish by a careful examination the relative age of the different designs. Prof. Breuil, as a result of a compari- son of the different occurrences, states that they exhibit a distinct sequence of development that affected the whole area . The engravings and paintings had a parallel evolution, but whereas at the beginning they were separate, at the end the former, apart from numerous light scratchings, were mainly auxiliary to the latter. The objects illustrated are nearly all animals, but geometric figures or symbols also occur. Both engravings and paintings started as silhouettes, crudely executed in outline, but at the time of amal- gamation the engraved outline had attained much merit, and the painted body, which was always in red or black monochrome, was shaded. At this stage a decadent type of painting, in which shading was discarded and the outline merely filled in with a uniform tint, was introduced; it does not seem to have stayed long, but it has a very considerable interest in that it is the dominant type in South Africa. Ultimately fine shaded polychrome paintings came into vogue. Recently rock-paintings in the open have been discovered in the lower basin of the Ebro. The Solutric petroglyphs and rock-paintings of South Africa are distributed over the whole area. The former are mostly found on boulder-like outcrops of rock, either among kopjes or in the open veld, while the latter are chiefly met with on the back of rock- shelters. The objects illustrated are mostly animals and men which, except in comparatively rare instances, are shown in silhouette only. The petroglyphs are usually larger than the paintings, and, as a rule, represent disconnected units only, whereas the paintings fre- quently depict connected objects such as the participators in a hunt or fight. Many of the more advanced paintings constitute intel- ligible records of customs and myths. Geometric figures or symbols are abundant. I have already noted some differences between the Solutric cultures of West Europe and South Africa. There are, however, two of outstanding importance that I have not yet mentioned. One- 13 FIG. 6. ACHEULIC AMYGDALITH FROM LUCKHOFF. (Actual size.) 14 is that the Solutric implements of West Europe, though presenting a large number of identical forms ^ are very much bigger than those of South Africa. The other is that the petroglyphs of West Europe are lacking in a type, namely, the pecking, that is abundant in South (and North) Africa. From which one may infer that the Solutric peoples of Africa did not come from Europe, nor vice-versa, but that the Solutric peoples of Europe and Africa were separate migrations from the East. 15 FIG. 7. ACHEULIC AMYGDALITH FROM LUCKHOFF. (Actual size.) CHAPTER II. THE EOLITHS OF LEIJFONTEIN. IN several places on the farm Leijfontein, which is situated below the Campbell Rand, near Campbell village, there are patches of gravel lying at the foot of the escarpment. This gravel consists mainly of subangular fragments of jasper a material that has travelled a long way, the nearest source being the Asbestos Hills, some fifty kilometres to the west. While it apparently can only have been brought to its present position by water, it has no evident connection with any existing river, and is, therefore, probably of very great antiquity. The jasper has changed externally from its original dull brown colour to a yellowish-brown, and acquired a high glaze or polish. Mixed with the gravel are quantities of much worn and highly glazed Eoliths. A few of these differ from the true Eoliths in being shaped from artificially produced splinters, or flakes, but they are a small minority. Otherwise the group is in every way identical with the typical assemblage met with in the hill drifts of Southern Britain. Although attention w r as drawn by Prof. Prestwich to the hacked or rudely chipped stones discovered by Mr. Benjamin Harrison, which are now termed Eoliths, as far back as 1889, their origin whether artificial or natural is still the subject of controversy. While some authorities unreservedly accept the trimming as the work of man, others are equally emphatic in denying its artificial character. The specimens from Leijfontein throw considerable light on this matter, and incline me to the opinion that they are in truth 17 18 primitive man's first attempts to trim pieces of stone to a useful shape. The Leijfontein Eoliths and Flake-Eoliths may be subdivided in the same way as Prof. Prestwich divided the typical Eoliths, that is, into two sub-groups : (1) Those in which the pieces of stone have been subjected to very little modification ; and (2) , those in which they have been chipped into definite shapes. It would be difficult to see any artificial character in the chipped pieces of stone of the first sub-group if found alone. Their great abundance and the haphazard appearance of the chipping immedi- ately suggests that they have been shaped by the blind forces of nature. Both circumstances have been brought forward as evidence against their possibly artificial character. Nevertheless, Acheulic, Solutric, and Neolithic implements are often met with in equal quantity, while, if the Eoliths are, as is claimed, man's first arte- facts, one w T ould expect them to be barely distinguishable from nature's work. Their association with others in which the trim- ming, though of the same rude kind, is arranged in definite patterns is the sole ground upon w r hich they can be accepted. Even the more implement-like examples of the second sub- group are of so primitive a kind that their possibly artificial character is still disputed by many. Yet, apart from the necessarily inferior quality of the trimming, and the fact that most are fashioned out of naturally broken fragments of stone, they are largely identical with the commoner accepted flake-tools of the Neolithic period. Two series of the more differentiated Eoliths and Flake-Eoliths are represented by the accompanying illustrations (Figs. 2 and 4). Figure 2 shows a series of straight, concave, and convex-edged forms resembling the scrapers of the Neolithic period. A, B, and C are true Eoliths, while D, E, and F are Flake-Eoliths. A and D are good examples of the concave form. It will be noticed that there is quite a wide difference in the quality of the chipping in these two. There is a much greater difference between the better of these and the Neolithic example A (Fig. 3), but I have South African Acheulic and Solutric specimens which fill the gap. There is no essential difference between the disputed Eolithic examples and the accepted Neolithic ones. B, C, E, and F are four commonly D Fia. 9. ACHEULIC FLAKE-TOOLS FROM VEREENIGING, BEZUI- DENHOUT VALLEY AND TAAIBOSCH SPRUIT. (Actual size.) 20 recurring forms, usually designated by the descriptive adjectives, cir- cular, rectangular, long and broad. All of these can be matched by Neolithic examples, while one is still to be counted among the domestic appliances of certain savage peoples. Compare the circu- lar scraper with the Neolithic specimen B (Fig. 3). Here, again, I can produce South African Acheulic specimens intermediate as regards quality of chipping. This progression in delicacy of trim- ming is carried a stage further in some beautiful little examples found among the Riverton series, described in a later chapter ; they are about one half the diameter of the Neolithic example. Figure 4 shows some typical Eolithic forms that are seldom met with among Neolithic implements. In A and B it will be noticed that the chipping of the one edge is in the reverse direction to that of the other, while in C, D, E, and F both edges are chipped on the same side. They are eloquent advocates of the artificial character of the Eoliths, for it seems improbable that long tapering points like that of F could be hacked out by blind agencies. 21 FIG. 10. SOLUTRIC AMYGDALITH FROM THE EMBABAAN VALLEY. (Actual size.) 22 CHAPTER III. ACHEULIC IMPLEMENTS. ONE may safely express the opinion that Acheulic implements are distributed throughout the length and breadth of South Africa. I have found them, usually in association with old terraces, in the valleys of the Zambesi, the Elands-Rustenburg, the Magalakwin, the Selati, the Olifants, the Komati, the Vaal, the Caledon, the Orange*, and the Zwartkops, at Algoa Bay, while Dr. Peringuey has noted their widespread distribution in the old taluses of the mountain ranges to the north-west, north, and north-east of Cape Town. Here, however, I purpose only to describe those discoveries that throw some light on the problem of their age. At this point it is perhaps desirable to explain the terms Chelleen, Acheuleen, and Mousterien. In my introductory remarks I have referred to the amygdaliths with the broader end left un- worked and to the type with the edge continued all round, also to the large flakes trimmed on one face and edge only. Now, in Europe occurrences are known in which each of these forms is pre- sent to the exclusion of the others, though they are more frequently found together, and it is to these distinct assemblages that the above terms have been applied. It is considered by many that the Chelleen flourished at the beginning of the Acheulic stage, and that the Mousterien predominated at the end, but this theory lacks proof. The terms, however, are convenient to indicate assemblages in which one or the other forms greatly preponderates, but are mis- leading when applied to isolated specimens, since it is obvious that * For details of these occurrences see The Stone Implements of South Africa. 23 an unfinished Acheuleen amygdalith is of Chelleen form, while con- temporary Mousterien fonns are nearly always present at Acheuleen occurrences. I have already referred to the axe-heads or tranchets. They exhibit a wide range of variation, but three conspicuous varieties may be mentioned now : (1) A form consisting of a broad rect- angular blade like that of the ordinary modern axe used for tree- felling, (2) a type with incurved sides, and (3) a form with a narrow blade and rounded top, similar to the common and well-known Neolithic type. FIG. 11. SOLUTRIC SCRAPERS OF LYD IAN-STONE FROM THE JUNCTION OF THE RIET AND MODDER RIVERS. (Actual size.) VICTORIA FALLS. A good example of the possibilities that may be suggested by the geological conditions under which implements occur, and of the difficulty of obtaining decisive geological evidence of their age, is afforded by the occurrence at the Victoria Falls. The implements include Acheulic amygdaliths, but are mostly small flakes with an irregular edge-trimming that may be due to natural causes or may be the result of use. Mr. Feilden pointed out that these implements were associated with gravels bordering the river not only above the Falls where they are at but a slight elevation above the river, but also below the Falls where they overlook the canyon in \vhich the river rushes more than a hundred metres below ; and concluded that the pebbles and imple- 24 ments were deposited there by the Zambesi when it flowed over these surfaces prior to the excavation of the chasm. If this con- clusion were correct it would indicate an unparalleled antiquity for the Acheulic period. Mr. Codrington*, on the other hand, after an extended examina- tion of the neighbourhood, concludes that the gravels are not rem- nants of river terraces formed prior to the excavation of the gorge, but merely detritus brought down by spruits during the rains, and that the implements were fashioned from stones found on the old valley floor after the gorge was carved out. This explanation is cer- tainly the more feasible , for while rejecting an unparalleled antiquity for the class of implement it does not preclude a reasonable antiquity for the examples found. BARELY. At Barkly, on the Vaal Elver, between the bridge and the village, there is a terrace in which numerous pits have been sunk by diggers in search of diamonds. It lies at the foot of a ridge of hills, hence the talus element is predominant, though the river gravel is in evidence throughout. The deposit consists chiefly of small boulders and large irregular stones, in which, however, every edge and corner has been well rounded. The most remarkable feature of the bed is the extraordinary abundance in it of Acheulic amygdaliths and the associated characteristic large flakes ; these frequently occur on the top of the heaps of excavated stones, and, hence, come from the bottom of the pits, a depth of about ten metres. Now, out of all the countless numbers of implements exposed in the heaps of stones, all but a very few are equally rolled, being almost reduced to pebbles. The rolled implements present a Chelleen facies, but the few sharp implements, on the other hand, are much more advanced both in form and finish, resembling the assemblage described below from Vereeniging. I have come to the conclusion that there are two distinct series ; the one probably older than and the other contemporary with the deposit. Prof. Beck has recorded the discovery of a molar of the mastodon in the corresponding terrace on the other side of the river. * Quarterly Journal of the Geological Society of London, LXV., 390-407 (190J). 25 VEREENIGING. A large number of Acheulic implements have been found by my friend Mr. Leslie at a site of manufacture situated on the bank of the Vaal Elver, west of the village of Vereeniging. There the river has cut a channel deep in the solid rock, and on top of the cliff thus formed is an old terrace consisting of gravel and small boulders overlain by loam. This for several hundreds of metres has been furrowed and spread out by the rain. There unfinished amygdaliths occur in great abundance, and the flakes produced in their manufacture are to be found by the thousand, while here and there completed specimens were met with. Examples of both extremes of form of the finished amygdaliths are to be noted among the Vereeniging specimens, but, as usual, the great majority are of intermediate shape. An interesting feature, FIG. 12. SOLUTRIC SCRAPERS OF CHERT FROM THE TAAIBOSCH SPRUIT. (Actual size.) noticed also at Barkly, is the distinct foreshadowing of the typical Solutric form by some aberrant members of this group. The axe- head group is well represented. One example found by myself is so neatly and symmetrically shaped that, but for the evidence of the others, it would certainly be taken for a Neolithic axe-head, the fact that the edge was obtained by chipping instead of by grinding being obscured by the slight amount of weathering it has undergone. The class consisting of a big flake worked along one face and edge only, is also abundantly represented. I obtained in addition two or three smaller flake-tools, including the quartzite scraper shown in Fig. 9(D). The quartzite seems to have been of too coarse a grain, as a rule, for satisfactory working, as nearly all the failures and very few finished implements are in that material, the majority of the 26 good specimens being of a green aphanite. Many of these last are as sharp and fresh as on the day they were made, and the obliteration of the sharpness of the facets in the others is entirely due to weather- ing following exposure. One or two unfinished examples of chert were also found. It is uncertain whether the implements are contemporary with or more recent than the gravel bed, but they are, undoubtedly, older than the overlying loam. THE KRUGERSDOKP AND BEZUIDENHOUT VALLEYS. The slopes of the Witwaters Rand are clothed with a deposit of red, sandy loam and stones, resulting from the disintegration of the rocks of which they are composed. This talus yields quite sharp and fresh-looking Acheulic amygdaliths. The Krugersdorp and Bezuidenhout Valleys are situated in the Witwaters Eand, and their slopes are covered with this deposit. My friend Mr. Jobling has a worn but well-shaped amygdalith from the talus on the side of the Krugersdorp Valley, while I have a quite sharp but much stained example from that on the side of the Bezuidenhout Valley. The valley bottoms are occupied by fluviatile deposits, and the present streams alternately flow over and cut deep into them. The sections thus exposed show coarse sediments, consisting of small boulders of quartzite and subangular pieces of quartz, overlain by dark-coloured loam, and in places by peat. The older deposit is probably contemporary with, and the newer deposit is probably more recent than, the implement-bearing talus. From the stratum of coarse detritus I have obtained a number of Acheulic amygdaliths and a quantity of characteristic large flakes of quartzite, chert, quartz, and green aphanite. Most of the amygdaliths were found in the beds of the streams, but are clearly derived from the stratum of coarse detritus. They are all more or less waterworn. From the Krugersdorp Valley an amygdalith of white quartz is noteworthy, as also is an exceptionally thin and symmetrical quartzite example. More interesting still is a fine example of the broad variety of axe-head in quartzite from the same locality. From 27 the Bezuidenhout Valley a very neat, but much waterworn, example in blue chert found in position in the deposit at a depth of about two metres is also worthy of special mention. The rest of the amygda- liths are all of quartzite. A number of flake-tools were also obtained. Two of these are shown in Fig. 9. A is a skilfully-produced long and narrow chert flake, which may have been employed as a spear- head ; while B is a neat concave scraper of quartzite. These implements must once have been scattered over the slopes of the Witwaters Rand, and, subsequently, washed down together with other coarse debris into the bottom of the valleys during a period of heavy rainfall. Afterwards, more tranquil conditions pre- vailed, during which the stream channels became silted up ; these being in turn succeeded by the erosive activity of the present epoch. FIG. 13. "PIGMY IMPLEMENTS" FROM THE SOLUTRIC SITE AT RIVERTON. (Actual size.} THE STEYNSDORP AND EMBABAAN VALLEYS. The slopes ajid upper end of the Steynsdorp Valley arc clothed with a talus consisting of large sub-angular pieces of the different local rocks embedded in a reddish-brown loam. In the middle of the valley this loam has been removed, and the concentrated stones have been rolled and worn by a former stream. This gravelly detritus is now buried under a later accumulation of loam. The present stream has cut its way down through the loam, and in places through the underlying detritus also. Acheulio arnygdaliths and axe-heads occur both in the talus and in the old stream deposit. In the former they are often quite sharp and fresh-looking, but in the latter they are always waterworn. The talus on the higher portions of the slopes is being slowly but continuously washed to a lower level by the rain, so that here 28 and there an Acheulic amygdalith is exposed to view, looking as though it had been dropped on the present surface. The present surface is strewn with small chert flakes of more recent date, among which I obtained a fine scraper of a form that is very characteristic of the Solutric group. I have also a spheroidal perforated stone ball from there that was given me by my friend, Mr. Atkin, who found it associated (accidentally?) with a Bantu interment; a fine unpolished urn, with incised geometric decoration, the bottom of which had been knocked out, was also found with the skeleton. Acheulic amygdaliths occur under similar conditions in the Embabaan Valley. Specimens were recorded* from there in 1898 by Prof. Eupert Jones, who described thirteen examples found by Mr. Kyan. I also have a number of Acheulic amygdaliths from the Embabaan Valley that were given me by my friend Mr. Nash. Together with them w r ere a number of chert flakes trimmed to a point ; most of these are worked on one side only, but two or three of them have been worked on both, including the fine specimen shown in Fig. 10. There can be little doubt, I think, that these are more recent than the Acheulic amygdaliths, their resemblance to certain of the Solutric forms of Europe being unmistakeable. ROBINSON. The country between Krugersdorp and Eandfontein is dotted with pans , most of which , unlike those of Western Orangia , but like those of the Lake Chrissie district, hold water all the year round. One of these, situated north of Eobinson, has been drained, and a pit has been sunk in it for the supply of brickearth. The section exposed shows bluish-black carbonaceous soil pass- ing downwards, at a depth of about 2' 5 metres, into tenacious yel- lowish-brown loam. At 1'5 metres from the surface there is a well- defined line of stones, and at 2'5 metres there is another but less persistent layer. Isolated stones occur to the depth at present reached, namely, 4 metres. The pit is situated in the flat bottom of the pan, but towards the side. * Journal Anthropological Institute, XXVIII., 48-54 (1898). 29 FIG. 14. SCRAPERS FROM SOLUTRIC SITES NEAR BOSHOF. (Actual size.) 30 A larger number of Acheulic amygdaliths and axe-heads, in different stages of manufacture, have been obtained from this pit, but the exact horizon from which they mostly come is uncertain. They were not noticed till the pit had nearly reached its present depth, which would seem to indicate that they chiefly came from - low down in the section, but one of the foremen, on the other hand, thinks that they mostly came from the one-and-a-half metre hori- zon. I myself saw w r orked stones in situ at both the 1'5 metres and 2' 5 metres level. The condition of the specimens shows that none have come from the surface. The implements are mostly of quartzite of both fine and coarse texture several are of chert, while three are of white quartz. There can be no doubt that they are close to the site of manu- facture. Most of them are quite sharp and fresh, though a few show signs of wear, while one is much worn. Two unusual features are the abundance of round hammer- stones and the rarity of the characteristic large flakes. These last, however, which are usually met with in abundance near sites of manufacture, were probably produced in the preliminary reduction in size of small boulders prior to their conversion into amygdaliths ; in this case the implements were fashioned out of large pebbles that did not require reducing in size. It seems certain that they originally rested on the sloping sides of the pan, where they were made, and, subsequently, drifted down on to the bottom where they were buried. The succession of geological conditions indicated is in accordance with that which obtained in the Krugersdorp and Bezuidenhuit Valleys. LlUCKHOFF. There is an Acheulic site close by the village of Luckhoff of special interest. The implements occur on the solid rock underlying the sheet of reddish wind-borne sand which covers the surrounding flat country. They were exposed in the gullies and on the slopes leading down to the dam, where the covering of sand had been removed by the rain, and were associated with numerous flakes and chips, which 31 FIG. 15. SCRAPERS FROM SOLUTRIC SITES NEAR BOSHOF. (Actual size.) 32 were probably produced in their manufacture. They include both amygdaliths and axe-heads. The amygdaliths were by far the most abundant and are much worn and deeply weathered, the facets being nearly obliterated ; they are of lydian-stone , which has changed externally from its original black colour to a reddish-brown. This lydian-stone much resembles flint in its appearance and fracture when fresh, but is softer and weathers much more rapidly. The axe-heads, of which there are only seven, are of the local fine-grained gabbrodiorite , and are also much weathered and worn; they are larger than the amygdaliths, which accounts for the material used, because, while less easy to work owing to its toughness and uneven fracture, it is readily obtained in larger pieces than is the lydian-stone. The great interest of this occurrence lies in the fact that these Acheulic implements occur in exactly the same position as the Solutric sites near Boshof , that is, they are both mingled with debris accumulated in the former case naturally, and in the latter arti- ficially on the solid rock underlying the sheet of reddish wind-borne sand. Yet although manufactured of the same lydian-stone the implements of the one group are much worn and deeply weathered, whereas those of the other group are as sharp as on the day they were made, and have barely suffered a change of tint. 34 CHAPTER IV. ACHEULIC AND SOLUTRIC GROUPS. IN this chapter I shall describe some occurrences in which Acheulic and Solutric implements are associated, and in which their relative age is clearly shown. RlET AND MODDER RlVERS. The junction of the Riet and Modder Rivers has been rendered classical by Rickard's account* of his discovery of Acheulic imple- ments there. Like all the large spruits in central South Africa, these are bordered by a narrow strip of alluvium overlying a stratum of coarse detritus. ' The implements from the Junction were found in the bed of the river immediately below the point where the rivers become con- fluent, lying either on the bare rock or in small hollows containing a little coarse gravel ; I collected upwards of eighty specimens in a few hours, but had to abandon the majority of them on account of the difficulty and cost of transport." He devotes two plates to them. Plate T. shows two typical amygdaliths. Plate II. shows a fine axe-head drawn to actual size. f I myself obtained quite a number of both types there , but they were all very much waterworn, being almost reduced to pebbles. I have no doubt that they come from the gravelly stratum at the base of the alluvium. This was east of the bridge. West of the bridge, and some little distance north of the river, I found a great quantity of quite sharp and fresh scrapers of lydian- * Cambridge Ant. Society, V. (1880). 35 FIG. 17. SCRAPERS FROM SOLUTRIC SITES NEAR BOSHOF. (Actual size.) 36 stone, mixed with flakes and cores. They had been exposed to view by the removal of a thin covering of surface soil. They exhibit a remarkably wide range of variation, only the commoner forms being shown in Fig. 11. Interesting are extremely elongate forms and a variety trimmed at both ends. The assemblage bears a close resemblance to the Solutric scrapers of Europe. These implements are unquestionably newer than the alluvium. Together with them I found three or four chert scrapers, a multiple-grooved cylindrical piece of sandstone, a hemispherical stone with a hole bored to a depth of about one-and-a-half centi- metres from the flat side, numerous ostrich-eggshell fragments, a bead made of same, and the half of a glass bead. This last, prob- ably, has no connection with the other objects. The discovery by Mr. Bain of the remains of an extinct buffalo in the alluvium of the Modder River is interesting in this connec- tion. TAAIBOSCH SPRUIT. The Taaibosch Spruit is a tributary of the Vaal, and flows into that river south-west of Vereeniging. Mrs. Hutt had previously obtained Acheulic implements there, and drew my attention to the fact. My more important finds comprise a group of Acheulic imple- ments from beneath, and a group of Solutric implements from above, the alluvium. From the bed of coarse detritus, at the base of the alluvium, I have obtained, in addition to a quantity of the characteristic large flakes, a number of Acheulic amygdaliths. They are fashioned out of a green aphanite, and are all w r aterworn, some being reduced to the condition of a pebble. I also obtained specimens of the group of large flakes worked on one side and edge only. No axe-heads were found. In places on top of the overlying alluvium I came across quanti- ties of very small flakes of chert, jasper and agate, and here and there among them examples that had been trimmed into minute scrapers. Some of them are shown in Fig. 12. It is difficult to imagine exactly what could have come within the scope of these remarkably small tools. The fineness of the secondary trimming is really wonderful. They are of the same class as, and contemporary with, those from the junction of the Kiet and Modder Rivers. Though the difference in facies is so very marked it is entirely due to the differ- ence in the materials of which they are made. At some localities the only available material was the lydian-stone, and at others the three varieties of flint. Now, the lydian-stone occurs in compara- tively large pieces and permits of the manufacture of compara- tively large scrapers while at the same time affording scope for the wide range of variation exhibited. The chert, jasper and agate, on the other hand, occur only as small pebbles, and, therefore, only FIG. 18. SCRAPERS FROM SOLUTRIC SITES NEAR BOSHOF. (Actual size.) small scrapers could be made from them, while variation was cor- respondingly restricted. PKIESKA. The Orange River, on the north bank opposite the village of Prieska, is bounded by a terrace of sub-angular jasper gravel. This gravel is cemented into a hard conglomerate by sand and lime. It is overlaid by sandstone , consisting of quartz grains similarly bound together by calcareous matter. I saw many much worn characteris- tic large Acheulic flakes, as well as an amygdalith, in situ, in the gravel, but was unable to extract them owing to its hardness. I, however, obtained one very nice, though worn, specimen that had only just been freed by atmospheric disintegration of the matrix. 38 There are many similar jasper amygdaliths, as well as some of quartzite, in the bed of the river, that are evidently derived from this deposit, and of which I brought away some examples. Among sand-dunes overlying this terrace I obtained conclusive evidence of the presence of the Solutric group in the shape of a characteristically small jasper core, coloured chert flake, and grey FIG. 19. TWO SIMILAR IMPLEMENTS OF UNKNOWN USE, THE LARGER ONE FROM THE SITE AT RIETPAN AND THE SMALLER FROM A ROCK-SHELTER AT KRANZFONTEIN. (Actual size.) chert scraper, like those from the Taaibosch spruit. I also found a hemispherical stone like that from the junction of the Eiet and Modder Rivers, but with the hole barely started. Those who believe that the Acheulic and Solutric implements were made by one and the same people will point out that the above 39 data only proves that some Acheulic implements are older than some Solutric ones, and that there may very well have been, on the one hand, Solutric scrapers contemporary with the early Acheulic amygdaliths, and, on the other hand, Acheulic amygdaliths contem- porary with the late Solutric scrapers. They will also explain the frequent occurrence of Acheulic implements in ancient deposits and the restriction of the Solutric implements to quite recent sites, by the theory that of the earliest implements only the large amygdaliths have survived denudation, and that only the more recent sites have escaped. The persistent separateness of the Acheulic and Solutric groups, however, is all against this view. Besides, the Acheulo-Solutric stage of culture is a pure assump- tion. It may eventually prove to be real, but the evidence at present available is totally opposed to such a possibility. This evidence con- sists of the hard facts, (1) that the Acheulic implements frequently occur in ancient deposits, (2) that the Solutric sites are all quite recent, and (3) that no recent Acheulic implements are yet known. This last being negative evidence is admittedly a weak point ; on the collection of further data in regard to it depends the final solu- tion of the problem. 40 CHAPTER V. SOLUTBIC SITES. ONE may safely assume that Solutric sites are scattered throughout the length and breadth of South Africa. The wide distribution of the known sites, which stretch from Bulawayo to Cradock, fully justifies the assumption. The following is a brief account of the more important ones. KlVEETON. Biverton Island is famous for the outlines of animals and the curious symbols that are pecked on its polished rock-surfaces. The large representation of an eland mentioned by Mr. Stow is still in existence, though sadly damaged. Above the alluvium and gravel on the south side of the river is a thin covering of constantly shifting sand. In places where this had been blown away I came across, in great abundance, a most interesting assemblage of implements. Apart from hammer and grind-stones, a perforated stone ball, and grooved cylindrical pieces of sandstone, they may be divided into three series : (1) Scrapers of lydian-stone and green aphanite , resembling in a general w r ay those from the junction of the Beit and Modder Bivers ; (2) minute chert scrapers like those from the Taaibosch Spruit ; and (3) pigmy chert implements of remarkably delicate workmanship, mostly of peculiar form and unknown use. A series of these last are shown in the accompanying illustra- tion (Fig. 13). They comprise six distinct types, including the long- . 41 lived crescent*, besides little borers that may very well have been employed in the manufacture of the ostrich eggshell beads. The smallest crescent measures only 9 millimetres in length. Among the Riet and Modder group the short variety of the wedge-shaped scraper is well represented. Many of the scrapers of the Taaibosch group are much smaller than those from the type locality, though beautifully finished. Equally small circular scrapers made from the half of a pebble occurred in great quantity. Besides these I obtained fragments of pottery and a number of ostrich eggshell beads, as well as spherical and cylindrical glass FIG. 20. INCISED FRAGMENTS OF OSTRICH EGG-SHELL AND MINUTE SCRAPERS FROM SITES NEAR BOSHOF. (Actual size.) beads. These last, though much discoloured by long exposure to the weather, probably have no connection with the implements, there being all sorts of other modern debris associated with them in places. At the present time there are a number of Bantu living in huts on part of the old site. An idea of the extent of the industry that once flourished on this Solutric site may be gathered from the statement that, from the comparatively small area then exposed to sight, I obtained over thirteen hundred flakes, cores, unfinished and rejected implements, as well as seven hundred beautifully finished scrapers. * Which in Europe survived through the Neolithic and succeeding Eneolithic periods right into the Bronze age. 42 ORANGIA.* The country between the Vaal and the Orange Elvers is prob- ably very prolific in Solutric sites. I have already described the occurrence at the junction of the Eiet and Modder Eivers and that at the Taaibosch Spruit, also the site at Eiverton, where the two facies of implements are combined. I have described elsewhere a number of sites near Boshof , also one at Petrusburg and another at Eietpan, between Boshof and Bloemfontein, while recently my friend Mr. Taylor has discovered a site a little distance north of Bloemfontein. The implements from these sites are nearly all scrapers of lydian-stone like those from the junction of the Eiet and Modder Eivers. It is impossible to adequately describe and illustrate them. So multitudinous are the varieties, so remarkable the identity of specimen after specimen from different sites, so similar is the general assemblage and yet so distinct the local facies, that one requires to see the specimens in order to obtain a true idea of them. It is note- worthy that they include typical Aurignacien forms. At some sites very thick scrapers are prominent, while at others thin scrapers pre- dominate. In addition to the scrapers these sites have yielded some "pigmy implements," hammer and grind-stones, small tabular pieces of aphanite with a groove worked in the one side, multiple- grooved cylindrical pieces of sandstone, portion of a stone ring and perforated stone balls. Two large perforated stone balls from Vboruitzicht and a small one from Meerlandsvlei, given me by Mr. Jones, well illustrate the diversity in size of these objects. The largest specimen measures 10 centimetres along the perforation and weighs about 2 kilogrammes, while the smallest is only 4 centimetres in diameter with a weight of about O'l of a kilogramme. The latter is not com- pleted, the hole which has been started from both sides not being finished. At Devilliersrust I obtained a perforated stone that differs in many respects from the ordinary. Whereas these are either * The sites in the country between the Vaal and the Orange rivers are described in detail in my Geological and Archceological Notes on Orangia, and are only referred to in a general way here. 43 spherical or spheroidal in shape, this one is discoidal. The hole has been bored from both sides, and instead of being of uniform size, gradually decreases in diameter till at the middle it has only half the diameter it possesses at the two ends. Further, the striae in the hole are annular, not longitudinal, showing that the motion of the implement when in use was rotary instead of reciprocal. At one of the sites I found a neat little lanceolate amygdalith of typical Solutric form, worked on both sides. In addition to the implements these sites have yielded beads made of ostrich eggshell, bone pins (portions of arrows), and incised fragments of ostrich eggshells that were probably used as water- receptacles. Fragments of a plain hand-made pottery are abundant, FIG. 21. CONCAVE SCRAPERS FROM SOLUTRIC SITES ON THE WITWATERS RAND. (Actual size.} while I obtained one decorated piece, the decoration, which is not complete, consisting of four rows of cord pattern. I found a few similar scrapers of lydian-stone associated with the petroglyphs at Koffyfontein , Biesjesfontein and Baviaanskranz, and at the last-mentioned place numerous very small chert, agate and jasper flakes, one or two of which had been trimmed into scrapers, in addition. I have described elsewhere the occurrence of minute scrapers of chert, agate and jasper, like those of the Taaibosch Spruit, in rock- shelters, containing paintings at Ladybrand, Modderpoort, and Ficksburg, and near Fouriesburg. To these I can now add Grootvlei, situated between Ficksburg and Fouriesburg. At Mod- derpoort I also obtained a bone pin (portion of arrow) like those from the sites near Boshof (Fig. 27). 44 WITWATERS KAND. I have a number of Solutric scrapers from three widely separ- ated localities on the Witwaters Rand. The implements from the farms Vlakfontein and Elandsfontein and the conditions under which they occur, are so similar that they may be conveniently described together. The majority of these implements are minute scrapers similar to those from Taaibosch Spruit. They are, however, manufactured from different materials. Most of them are of white quartz, and bear eloquent testimony to the skill of the makers, for it is one oi the most refractory stones for the purpose. Many are of chert and some of a green aphanite. A series of these are shown in Fig. 22. It will be noticed that although in every case the bevelled edge has been so carefully and symmetrically worked, no attempt has been made to trim the reverse end to any definite shape ; from which one may infer that they were either bound or cemented to a handle. Larger and smaller specimens, as well as concave scrapers (Fig. 21), occur. Associated with the scrapers are quantities of green aphanite flakes of more ordinary size ; from both localities I have specimens of the cores from which they were struck. I think they were made purely for use as flake-tools. At both places the implements were found on the surface among IOW T kopjes bordering spruits. The implements from the farm Waterval are a similar assem- blage to the above. They are made of the same materials, and are associated with similar green aphanite flakes. Together with them I found numerous, and mostly quite sharp, flakes of quartzite, together with two rough discs of the same material. I have also in my collection a portion of a perforated stone found by Mr. Hewitt, on Signal Hill; it appears to have been of the dis- coidal type. OTHER SITES IN THE OPEN. At a spot about three kilometres south of the village of Cradock, a large number of lydian-stone scrapers have been found by Mr. Cottell ; they are a very representative series and comprise most 45 of the common varieties. With them he found, and kindly gave to me, the two arrow-heads shown in Fig. 28. On Blinkklip, near Wolmaransstad , I found numbers of very small chert and agate flakes and three minute scrapers associated with petroglyphs. On the outskirts of the village of Barberton there is a painting, depicting men shooting at an antelope, placed in a hollow in a rock. In the surface soil around this spot are numerous characteristic small chert flakes, and the typical minute scrapers certainly occur there though I did not find any. From the same neighbourhood FIG. 22. SCRAPERS OF APHANITE, CHERT, AND QUARTZ FROM SOLUTRIC SITES ON THE WITWATERS RAND. (Actual size.) Mr. Hulley gave me two perforated stone balls ; they are spherical and spheroidal in form and weigh 5'7 kilogrammes and 4 kilo- grammes respectively, but differ from those from near the village of Boshof , in that the holes are biconical. South-east of Bulawayo are some kopjes where I came across quantities of very small flakes of coloured chert, some of which had been trimmed into minute scrapers, also a very fine example (Fig. 23) of the crescent form of " pigmy implement." On the right bank of the Hart River, at its junction with the Vaal, I obtained a large number of minute scrapers, mostly of coloured chert, some very nice " pigmy implements," several 46 ostrich eggshell beads, a circular stone with flat sides and edge, a cylindrical glass bead and some pieces of pottery. Especially interesting is the very neat little borer shown in Fig. 24 ; it will be noticed that the trimming of the one side is in the reverse direction to that of the other, so that when revolved both edges come into play. At Blauwbank, situated on the bank of the Orange, I obtained a nice series of minute chert scrapers, as well as some neat " pigmy implements" from among sand-dunes; two of these (Fig. 24) are remarkable for their size and thickness. Together with them I found fragments of pottery, two small glass beads and numerous ostrich eggshell beads. These last occu-rred in every stage of manu- A B FIG. 23. VARIETIES OF THE CRESCENT FORM OF "PIGMY IMPLE- MENT" FROM (A) RIVERTON, (B) BULAWAYO, AND (C) HART RIVER. (Actual size.) facture, from a piece ^ of shell with the hole just started, to the finished article ; a series illustrating the process of making is shown in Fig. 24. ASBESTOS HILLS. The structure and constitution of the Asbestos Hills are very favourable to the formation of rock-shelters ; they probably occur throughout the whole range. Those examined by me are widely separated, one series being situated on Kranzfontein, and the other near Griquatown. Quite sharp and fresh spalls of the local jasper were abundant in all the rock-shelters investigated. The majority of the imple- ments are scrapers (Fig. 25) bearing a general resemblance to those previously described in this book, but the other small flake-tools (Fig. 26) characteristic of the Solutric stage of culture, including 47 the "pigmy implements," are well represented. Ostrich eggshell beads and fragments of pottery also occur. I have more fully described and illustrated the implements from these rock-shelters elsewhere.* OTHER ROCK-SHELTER SITES. In a ravine situated either on Doornkloof, or Kleinbuffelspruit, near the village of Carolina, there is a rock-shelter containing a number of paintings (Fig. 43). Immediately opposite is another, but smaller, shelter in which there are no paintings. In front of this I obtained several of the characteristic minute quartz, chert, and aphanite scrapers. They were found at a spot where the rain had washed away the surface soil. Near Rhodes' Grave, in the Matopo Hills, there is a rock-shelter with paintings from which Mr. Balfour showed me (in 1905) some characteristic scrapers. * The Stone Implements of South Africa. 48 CHAPTER VI. PETKOGLYPHS AND KOCK-PAINTINGS. SOLUTEIC petroglyphs and rock-paintings are distributed throughout the length and breadth of South Africa. It would seem as if local conditions determined which were made. Smooth rocks, even though exposed to rain and wind, afforded suitable surfaces for petroglyphs, but only the backs of rock- shelters that were protected from the weather w r ere suitable for paintings. The petroglyphs are mostly peckings which range from crude outlines like those of Vereeniging to veritable bas-reliefs like those from near Klerksdorp. Lately the writer has discovered engravings associated wdth certain occurrences of peckings, and no doubt further research will show them to be more widespread. The most northerly petroglyphs that I know of are those at the Bumbusi ruins, in the valley of the Zambesi. Though almost cer- tainly Solutric they differ from all others that have yet been found in that they consist of peckings of the footprints or " spoor " of game. As very little has been published about the petroglyphs of South Africa I will give a brief description of some typical occurrences. VEKEENIGING. The most primitive series of petroglyphs that I have yet seen are those in the neighbourhood of Vereeniging. They were dis- covered by Mr. Leslie, under whose guidance I subsequently visited them. There are several separate groups which, while apparently PLATE II. SOLUTRIC PETROGLYPHS. (A) ELAND PECKED ON ROCK, VEREENIGING. (Photo Inj MR. LESLIE.) (B) ELAND SCRAPED ON ROCK. BIESJESFONTEIN. 49 all of the same age, range in subject from the one in which animal figures only are represented, to the one in which geometric designs occur almost exclusively. The former group is situated close by the occurrence of Acheulic implements previously described, but is clearly of more recent date. It consists of a large number of outlines of animals pecked on a small outcrop of fine-grained sandstone. The animals represented are all \vild species, such as the eland, giraffe, rhinoceros, and elephant. The latter group is situated some distance north-west of the village. It consists of numerous geometric figures, which, how- FIG. 24. OSTRICH EGG-SHELL BEADS AND TOOLS USED IN THEIR MANUFACTURE. (Actual size.) ever, do not seem to possess any symbolical character like many of those at other localities, and a few outlines of animals, pecked on another outcrop of the fine-grained sandstone. The pecking is very irregular, more so, in fact, than in any others that I have seen, though the general effect produced is quite good. All the groups appear to be of the same age, the petroglyphs being weathered to the same colour as the rest of the rock-surface and being of very uniform workmanship. Characteristic Solutric scrapers of chert are of common occur- 50 rence on the surface in the neighbourhood of Vereeniging, and Mr. Leslie has two or three of the perforated stone balls from there also. WOLMARANSSTAD. A very fine series of petroglyphs, consisting of figures of animals pecked on smooth rock-surfaces, occur in the neighbourhood of this village. They are situated on the farms Blinkklip, Rooipoort, and Leuwf ontein , the best series being on the last-mentioned. At all these places they seem, from their varying degrees of preservation, to range over a considerable period of time. Some are nearly obliterated by the process of disintegration that all rocks FIG. 25. SOLUTRIC SCRAPERS FROM ROCK-SHELTERS IN THE ASBESTOS HILLS. (Actual size.) are slowly but surely undergoing at the surface, while others seem almost as fresh as on the day they were made. Though the evidence of relative antiquity afforded by the degree of disintegration is not always trustworthy, owing to its effects being very unequal even in an apparently homogeneous rock, there is undoubtedly a great gap in time between those peckings that have been completely weathered to the same colour as the rest of the rock-surface and those that still retain most of their original freshness. The figures are pecked both in full and in outline, the style and quality of the pecking exhibiting wide variations. I have endeavoured to indicate these features in the accompanying copies. Many of the more recent examples are battered rather than pecked 51 on the rock. All the latest petroglyphs are of a distinctly inferior workmanship. The figures of animals all represent wild species, and especially the large locally extinct kinds. At Blinkklip there are also one or two obscurely battered figures of men. In addition there are a couple of geometric figures, the one, at Blinkklip, consisting of three concentric circles with a dot in the centre, the other, at Eooipoort, consisting of several concentric circles. In making the copies that are reproduced herewith, accuracy of outline was obtained by carefully tracing the boundaries of the pecked areas on transparent paper, while the shape, size and dense- ness of the pecks was imitated as closely as possible, their dis- tribution, however, being less regular in the originals. FIG. 26. SMALL SOLUTRIC FLAKE-TOOLS FROM ROCK-SHELTERS IN THE ASBESTOS HILLS. (Actual size.) . The pecking is usually very shallow, but in some of the outline figures it forms a comparatively deep groove. In some instances a grooved outline is combined with a shallow pecked body. The more recent and inferior peckings are probably all Bantu imitations of the earlier ones. Together with the peckings at Leuwfontein, I came across the engraving of a zebra shown in Fig. 32. It probably represents the initial stage in the production of a pecking, and may be compared with the pecked zebra shown in Fig. 33. It has weathered to the same colour as the rest of the rock-surface. At Blinkklip, among the boulder-like outcrops of rock, on which the figures of animals are pecked, I came across, in spots where the surface soil had been partially washed away by the rain, numbers 52 of very small chert and agate flakes and three minute scrapers of Solutric type. BlESJESFONTEIN. The most numerous series of peckings that I have yet seen are at Biesjesfontein, some thirty kilometres south-west of the village of Koffyfontein. They are situated on the kopje adjoining the home- stead. As at Vereeniging and Wolmaransstad only wild animals are illustrated, but more species are represented, the eland, however, which was evidently a favourite, predominating here as there. Men are also represented, while geometric figures are abundant. FIG. 27. MIDDLE PORTION OF BUSHMAN ARROW AND SIMILAR TRUNCATED BONE PIN FROM A ROCK-SHELTER AT MODDER- POORT. (Actual size.) Besides these there are several representations in full of the eland that have been scraped* instead of pecked on the rock. Most interesting, however, are two engravings of which there are probably more examples. Tracings of them are reproduced in Fig. 33. The one is a large engraving in outline of a species of hippotragus, the other is a small engraving in detail of a quagga. The former, like the zebra of Wolmaransstad, is probably the initial stage in the production of a pecking. The latter, on the other hand, was clearly never intended to be pecked over. The quagga is specially interesting, as it is one of those recently abundant animals that have become totally extinct through the agency of the European. * In my Geological and Archceological Rotes on Orangia I have used the word scratched, but the above is more correct. These petroglyphs are all in a more or less similar state of pre- servation, the worked portion of the rock being mostly weathered to nearly the same colour as the original surface. It is difficult to decide whether the small range of variation in this respect is due to the unequal resistance to disintegration of the rock or to difference in age. In addition there are three light engravings of animals, repre- senting eland, gemsbok and gnu, done in the same style as the quagga, but of markedly inferior workmanship, that are so fresh in appearance that they might very well have been made only yester- day. They are probably the work of Bantu. The farmer remem- bers them being there when he acquired the farm 47 years ago ; this FIG. 28. SOLUTRIC ARROW-HEADS FROM CRADOCK. shows that the process of weathering on the class of rock (gab- brodiorite) must be a very slow one, and that the earlier petroglyphs may be very old indeed. Among the boulder-like outcrops of rock on which the figures are shown, I came across, in places where the soil had been washed away, some characteristic lydian-stone flakes and scrapers of the Solutric group. BLOEMFONTEIN. In the immediate neighbourhood of Bloemfontein a number of petroglyphs have been discovered by my friend Mr. Taylor, and I have recently had the opportunity of examining them under his guidance. They comprise a group of peckings and engravings on the northern, outskirts of the town, and a few isolated examples. 54 Of the group of petroglyphs, the oldest peckings would appear to be some figures in outline of the eland that are already largely obliterated, the most recent are some figures in full of men that have quite a fresh appearance. On one rock an old pecking is parti- ally covered by a new one both represent an antelope of uncertain species. The engravings are poor, and apart from numerous lines and incomplete figures of animals include only two Solutric examples. These represent, respectively, an eland (Fig. 34) and an animal that might be a quagga or a giraffe, or even an okapi (Fig. 35). A group of horses (Fig. 36) more lightly engraved, and evidently the work of a Bantu, is interesting. Of the isolated petroglyphs, an almost obliterated pecking, in outline, of an eland and a group of engravings are noteworthy. The latter include a number of more or less parallel lines of apparently no significance , one or two of which accidentally take the form of bows , and two arrow-like objects (Fig. 37), all of which are weathered to the same colour as .the rock-surface ; on the same rock is shown an eland and a gnu in characteristic Solutric style, but engraved over and added to by some European, who has also added a bow and three arrows in modern style. BLAUWBOSCHDRIFT. The very interesting series of peckings on the above farm were discovered by Prof. Young and myself in 1905. At that place the artists had selected a naturally polished surface on w y hich to peck their pictures of man and beast. A reproduction of a photograph of the actual slab of stone on w r hich they occur, and which has a special interest to the student of geology , as well as to the archaeolo- gist, has been published elsewiiere.* The figures are confined to the upper half of the slab and number nearly two hundred. A striking feature is the entire absence of representations of the large locally extinct mammals. Domestic forms predominate, the only wild species being the small ones that are still living in the neighbourhood, while men on horses are conspicuous. Geometric figures are also absent. The three humped oxen (Fig. 38) and the three men on horses * Transactions of the Geological Society of South Africa, IX., PL XII. (1906). 55 '' ' '' FIG. 29. TWO RHINOCEROSES FINELY AND COARSELY PECKED IN FULL AND HEAD OF RHINOCEROS PECKED IN OUTLINE, . WOLMARANSSTAD. (Scale J.) 56 (Fig. 39), in both of which groups the figures are shown in the rela- tive positions they occupy on the rock , are very characteristic of the series. The highly conventional rendering of the latter is note- worthy. The representations of an ox and hyaena, shown in Fig. 40, are quite good, the latter, especially, being very faithfully done con- sidering the limitations of this kind of work. Other peckings (Fig. 38) seem to be purely fanciful, but may be merely poor work. The figure (Fig. 41) of a plough is interesting. These peckings have weathered to the same colour as the rest of the rock-surface. On the same slab there are one or two figures that are evidently much more recent. In them the peckings are comparatively fresh and show up well against the dark background. They are larger than the originals, of which they are imitations, and are markedly inferior in execution. None of the peckings, however, can be ancient. Neither the plough nor the men on horses are com- patible with any antiquity. I originally thought these peckings to be Solutric works, but I am now satisfied that they, like the most recent and inferior examples at other occurrences, are the work of Bantu. KOCK-PAINTINGS. The recent publication by Miss Tongue in colour of a large series of rock-paintings from different parts of South Africa renders it unnecessary to describe them in as much detail as the petroglyphs. The petroglyphic art , despite the high degree of excellence that it attained, laboured under obvious limitations unknown to paint- ing. Hence it is that whereas the petroglyphs seldom represent other than disconnected objects, the rock-paintings frequently depict connected objects such as the participators in a hunt or fight. The vast majority of the rock-paintings are merely silhouettes in which the outline is filled in with a uniform tint. They are mostly red, but are sometimes black. The figures of animals usually show real artistic merit, but those of man are always grotesque. Yet there is something life-like about them. In the representation of a Bantu wedding dance (Fig. 44), for instance, the very twist in 57 v?*' V * ** m viil *\*i^ %fe 5^ $$ 'if * tt*/ H'/*'// IT ..':." Q < H -/. - Q W H O q w - o CO O B CO o d 58 the women's bodies is shown, and the actions they are going through plainly indicated. Along the eastern border of Orangia and in the country south of the Orange Eiver polychrome paintings occur in addition. The eland, for instance, which is actually one of the dowdiest of all the v \o vA 'I// ff'fff \v Sr-r' ' ' ,11 i 8 v\ /4' M 1 sX v\ I" ly\ ;"! \i\ :/i; Ait, 'i.y * \\ i In v,A 'i 1 - vw v \ *Vr ij ii'i i ^ ! if '' i 'i 1 1/ w til ! 8 I l\lnl I i V.\j '.I.'/ ii'i M W ? ; S i / | FIG. 31. BANTU IMITATION OF SOLUTRIC PECKING, WOLMARANSSTAD. (Scale .) game , but was a great favourite of the Solutric people , no doubt on account of the quantity of meat it afforded, is frequently depicted in two or more colours ; the neck and under part of the body is usually shown in white, and the upper part of the body in golden- yellow, red or dark brown. Some of the better polychrome paint- ings show distinct, though incipient, shading. 59 It is noticeable, both among petroglyphs and rock-paintings, that while the ordinary representations of animals do not show any termination to the legs, the best examples show the extremities in considerable detail. Many of the better, not necessarily polychrome, paintings depict dances, hunting-scenes, and fights, while they not infrequently represent myths. The art of the Solutric people of South Africa attained its maximum in the painting, published by Miss Tongue, of the ante- lopes at a pool ; in this not only are the antelopes well shaded but even the reflections of their legs in the water is depicted. As in the case of the petroglyphs, crude Bantu imitations are present at some of the occurrences. 60 CHAPTER VII. THE PRE-HISTOKIC BANTU. SCATTERED throughout the now sparsely inhabited bush country of north-eastern South Africa are abundant proofs of the former wide- spread population of the country by the Bantu or Kafir peoples. Evidence of their agricultural activities is met with in every direction. Even in dense mimosa one comes across extensive areas split up into innumerable plots by lines of stones. A typical example is afforded by the Steynsdorp Valley, which I have already mentioned in another connection. This valley is dotted from end to end with remains of old kraals in the form of low circular enclosures of un- cemented and undressed stones. In and around these, large pebbles, worn down on both sides to a flat disc by rubbing, abound, as also do the polished slabs of granite with which they were used, and the pounding stones and hollowed out blocks of hard rock that served the purpose of pestles and mortars. Also associated with these are quantities of flat pieces of soft schist, in which, though otherwise untrimmed , conical holes have been chiselled ; what useful purpose they could have served I am at a loss to conceive. The evidence of the agricultural activities of its former inhabitants occurs in the form of piles of stones that have been sorted out of the soil ; on the hill-sides these have been arranged in parallel rows in order to serve the additional purpose of holding back the soil in terraces. Equally widespread are the remains of their mining and smelt- ing operations. Old workings on iron and copper ores are scattered throughout the whole area, and associated with them are old slag heaps and fragments of clay furnaces. Small boulders of quartz and 61 62 other hard stone, that have been employed in crushing the ore, are frequent at these workings, and close by one can usually find the corresponding hollows in the rock. Very often, as at Palabora, the Solutric perforated stone balls were used for the same purpose. Again, clumsy large wheel-shaped perforated stones of Bantu manu- facture are often associated with the other remains, but their use is uncertain. From some old workings on iron ore situated near Eustenburg I have a small boulder of quartz, in the four sides and one end of which deep indentations have been worn through its having been used to strike an iron rod ; with it were found the iron rod, rusted to the core, and some polished red and black pottery with characteristic geometric decoration. Most of the old workings are situated on small pockets of ore that would not pay to extract nowadays, but some are very exten- sive. The Messina copper belt, which is situated in the west angle between the Zand Biver and the south bank of the Limpopo is about thirty kilometres in length , and is riddled from end to end with old workings which reach a depth of over twenty metres. Many million kilogrammes of copper must have been extracted from them. Even tin was mined and smelted. At the Rooi Berg, some sixty kilometres west of the village of Warmbaths, there are old workings on tin ore covering an area of thirty hectares. Their dis- covery was preceded by the finding of some hemispherical ingots of the metal , weighing about a kilogramme each , in an old kraal some twenty-five kilometres distant. They attain a depth of ten metres, but are represented at the surface by shallow depressions of not more than three metres deep, having been filled up with drifting sand. Hammerstones and picks consisting of wooden handles through the knobbed ends of which iron gads have been inserted, are found in them. At two places large slag heaps, broken tuyeres, and pieces of smelting pots have been discovered. The occurrence of the charac- teristic low circular enclosures of uncemented and undressed stones indicates the former presence of old kraals within the mining area. But the mining and smelting industry attained its maximum in the highly metalliferous country between the Limpopo and the Zambesi, where, in addition to iron and copper, and possibly tin 63 FIG. 33. REPRESENTATIONS OF ZEBRA, HIPPOTRAGUS, AND QUAGGA, PECKED AND ENGRAVED ON ROCK, BIESJESFONTEIN. (Scale J.) 64 (for masses of that metal have been found in the ruins there) , gold was extracted on an extensive scale. It is in this area that the finest ruins occur ; the circular enclosures above mentioned are probably all cattle pens, but these are mostly forts and fortified kraals. Messrs. Hall and Neale have given in their Ancient Ruins of Rhodesia a list of more than 120 localities in which ruins of various sizes are reported to exist. They range from small ruins of a con- struction no better than the ordinary cattle enclosure, to extensive FIG. 34. ENGRAVING OF ELAND, BLOEMFONTEIN. (Scale J.) ruins of altogether superior construction from those with walls built of undressed and irregularly placed stones, to those w y ith walls made of neatly coursed blocks of trimmed stone. The size of some of the latter, such as Khami, Dhlo-Dhlo, Chum, M'Popoti, Mundie, and Zimbabwe indicates that they were important centres, the last, however, being by far the greatest ; according to Mr. Franklin White they controlled an area of nearly 300 kilometres long by more than 150 kilometres wide. Of the smaller and ruder ruins the Inyanga fort affords a good PLATE ///.ZIMBABWE RUINS. (A) EXTEKIOR OF MAIN BUILDING, SHEWING CHEVRON PATTERN. (B) INTERIOR OF MAIN BUILDING, SHEWING CONE WITH REMAINS OF WALLED PLATFORM IN FRONT. (Photos liy MR. FRAXKMX WHITE.) B PLATE IV. ZIMBABWE RUINS. (A) SOAPSTONE POST WITH CARVED BlRD AND CROCODILE. (B) MONOLITHS NEAR WEST ENTRANCE OF MAIN BUILDING. (Photux by MR. FIUNKLIN WHITE.) 65 example. It is one of a number of similar buildings in the neigh- bourhood, and has been described in detail by Mr. Hall, who has also published a plan of it. The fort occupies a strong strategetic position on the highest point of a prominent ridge, and commands an extensive view of hill and dale dotted with the remains of stone walls and with ruined dwellings. In construction it differs little from the ordinary cattle enclosure being built of undressed stones irregularly placed one on top of the other, while in plan it is merely a number of similar enclosures joined together. The walls, however, are modified to suit the pur- poses of defence. FIG. 35. ENGRAVING OF QUAGGA ( ?), BLOEMFONTEIN. (Scale .) The exterior walls average 2' 5 metres in height, and, as in the other forts, are made up of two parts of about equal thickness, namely, an outer wall and a banquette running round the inside, which in this case is about a metre lower. The entrances are just the height for a man to pass through in a stooping posture and are roofed with lintels of single slabs resting upon the sides. The exterior walls vary much in width from point to point a characteristic feature of all the ruins and are riddled with loop- holes. Within the fortifications are situated the foundations and walls of some score of circular huts. 66 There is abundant evidence of contemporary agricultural activity in the district, including an extensive system of irrigation ditches. But, as remarked by Dr. Randall-Maciver, while the farmers industriously tilled the land, they felt themselves so little secure against the raids of other tribes that even their dwellings were built with a view to defence, while on the heights above were built forts which would serve as places of refuge when the home- steads were no longer tenable. In all fundamental respects the Inyanga fort is the prototype on a small scale and in inferior construction, of the more imposing or FIG. 36. BANTU ENGRAVING OF HORSES, BLOEMFONTEIN. (Scale .) Zimbabwe type of ruins. One has only to compare Mr. Hall's plan of the Inyanga fort with Mr. Franklin White's plans of Dhlo-Dhlo, Xanatali, and even the grandest of the ruins, namely, the main building at Zimbabwe, to realise this fact. The superiority of the latter lie almost entirely in their size and construction. In external appearance, indeed, there is a marked difference between the first and second types. The low walls, built of undressed and irregularly placed stones of the former, are superseded in the latter by high walls built of neatly coursed blocks of trimmed stone. 67 Granite is the prevalent rock in the country over which the Zimbabwe type of ruins are scattered ; sometimes it weathers into kopjes resembling piles of big boulders, and at others it forms great bosses. The ruins are usually situated on these bosses and in com- manding positions. Their outer walls, approximately, follow the contours of the summit, hence their irregularly elliptical shape, and are constructed of roughly rectangular blocks of granite laid in even courses. FIG. 37. ENGRAVINGS OF ARROW-LIKE OBJECTS AT BLOEMFON- TEIN, AND PAINTING OF SIMILAR OBJECT FROM THE NIAUX CAVE. (See Messrs. Cartlwtilac and Breuil, L'Anthropologie, XIX., 39, 1908.) The blocks, it must be explained, are not hewn out of solid granite. As first pointed out by Mr. Mennell, the granite scales off the bosses in large sheets up to 30 centimetres thick, and the blocks, which probably average 25 centimetres in length by 10 centimetres in height, were obtained by merely splitting up these sheets. The best walls are solid throughout, but many are merely faced with stone, the space between the faces being filled up with rubble. No cementing material was placed between the blocks. A note- worthy feature of all the walls is that they unintentionally vary in thickness from point to point. When a large boulder was encountered during the construction 68 FIG. 38. BANTU PECKING OF TWO FANCIFUL FIGURES AND OF HUMPED OXEN, BLAUWBOSCHDRIFT. (Scale J.) - "' FIG. 39. BANTU PECKING OF MEN ON HORSES, BLAUWBOSCHDRIFT. (Scale .) FIG. 40. BANTU PECKING OF HYAENA AND OX, BLAUWBOSCHDRIFT. (Scale .) 69 of a. wall, the former was neither removed nor the latter deflected, but the wall was built across the boulder. The walls are ornamented w r ith one or more rows of geometric design. By leaving out alternate blocks in successive courses, and by replacing some of the courses by inclined slabs of different stone, the chessboard, cord, herring-bone, and chevron patterns were pro- duced. More rarely the dentelle pattern, obtained by arranging the blocks with one corner pointing outwards, was employed. Courses of rock of a different colour were also frequently inserted. Untrimmed elongated slabs, or monoliths of various rocks, were placed upright on the walls of some of the ruins. At Zimbabwe pillars of soapstone, carved at the top to represent perched birds of prey, occurred in addition. Probably the most remarkable structure connected with these ruins is the celebrated masonry cone in the main building at Zim- FIG. 41. BANTU PECKING OF PLOUGH, BLAUWBOSCHDRIFT. (Scale J.) babwe. It is rather more than 10 metres high and measures 5'5 metres in diameter at the base, and is built solid in the same way as the walls. What it signifies is an unsolved problem. By its side are the remains of a smaller circular erection, but it seems uncertain whether it was another cone or not. Circular platforms of cement, reached by means of steps of the fame material, are a prominent feature of the interior of these ruins. In many cases they support hut walls of similar material and almost certainly contemporaneous,* though some regard them as of more recent date than the ruins proper. The same kind of cement often lines the more roughly built interior walls! * Though the writer cannot claim to speak about these ruins with the authority of an original investigator, he feels bound to remark that he was much impressed by the apparent contemporaneity of the remains of the cement-walled huts at the Khami ruins. 70 MAIN ENTRANCE /O 2O PLATE V. PLAN OF MAIN RUIN AT ZIMBABWE (BASED ON MR. FRANKLIN WHITE'S MORE DETAILED PLAN IN JOURNAL ANTHRO- POLOGICAL INSTITUTE, XXXV., 1905). M = MONOLITHS. X = PROBABLE SITES OF CEMENT PLATFORMS ACCORDING TO DR. BAND ALL- MAGI VER. XM = PLAT- FORM WITH MONOLITHS. The portion of the enclosing wall decorated with Monoliths and Chevron pattern is indicated by dotted line. 71 The defensive character of these ruins is everywhere apparent apart from the enclosing walls. At Zimbabwe, for instance, 'the "hill" (kopje) ruins have protecting walls at every point where ascent is possible, and the path leading up to them winds between boulders and along faces of rock, supplemented, where necessary, by walls, in such a manner as to be impassable to an enemy. I have already pointed out that in plan there is no essential difference between the Inyanga and the Zimbabwe types of ruin. The difference is almost entirely in size and construction. The fore- going description of the main features of the latter class of ruin will FIG. 43. PAINTING OF MAN DRESSED UP AS BIRD AND OF TWO ANTELOPES FROM A ROCK-SHELTER NEAR CAROLINA. (Scale J.) show how great this difference is. But between the one and the other quality of construction every gradation is to be met with. In fact the numerous ruins present a complete picture of the evolution of the Zimbabwe type of building.* There can be no doubt that all the ruins are the work of a Bantu people that attained a more advanced stage of culture than any of the * In this connection it is interesting to note that in 1907 I came across a little ruin perched on the top of a kopje at the extreme north-eastern end of the Murchi- son Range. It is oval in plan, measures about 8x5 metres, and is built of square slabs of schist. In neatness of construction it resembles, but is smaller than and lacks the ornamentation of, the better ruins north of the Limpopo From it a good view can be obtained of the Palabora kopjes, where there are extensive old workings in iron and copper ore, and where the peculiar marali were formerly manufactured. 72 73 74 surviving tribes. The ornaments, weapons, and pottery found associated with the ruins are characteristically Bantu. Apart from glass and porcelain beads comparatively few imported articles are met with. BB 5 /o FIG. 46. PLAN OF INYANGA FORT. (After Mr. Hall.) Zimbabwe was the fortified kraal of the head chief. Inside its walls were situated the huts of his wives and children and of his head men and their families. The other similar buildings were the residences of the lesser chiefs. Around or near these imposing structures were grouped the huts of the common people. 10 20 30 METRES. PLATE VI. PLAN OP NANATALl (AFTER MR. POPHAM), SHEWING CEMENT PLATFORMS. FROM THE PROCEEDINGS OF THE RHODESIA SCIENTIFIC ASSOCIATION. 75 The connection of these people with the mining industry is shown by the crucibles and ingot moulds, and by the beads, bangles, and other ornamental objects of gold found in the ruins. It is probable that their chiefs traded gold with the Arabs on the east coast, and it is not impossible that Arabs helped in the mining operations.* It is not difficult to guess what happened to the people who erected these buildings. That an ever present danger threatened them is shown by the defensive character of the buildings. No doubt, then, as within historic times, the country was periodically over-run by war-like tribes who made a living by. killing and robbing FIG. 47. TWO MARALI FROM PALABORA. (After Dr. Haddon.} Supposed copper currency. The length of the straight portion is about 49 centimetres in each. the more peaceful and industrious tribes. It was to a series of these raids that the builders of the ruins succumbed. Bantu culture, then, received a setback from which it has never recovered. ****** It is manifestly impossible in a work of this character to give any other than a very brief and general account of the interesting ruins above dealt with, and, indeed, the wealth of data recorded in the several books and numerous papers specially devoted to them renders this unnecessary. Since Mr. Bent's investigations pub- lished in his Ruined Cities of M ashonaland , a vast amount of facts have been accumulated, mainly by Ehodesians, among whom * See Appendix. 7G Mr. Franklin White and Mr. Hall must be specially mentioned the former mainly on account of the splendid plans he has prepared of various of the better ruins, and the latter chiefly because of the quantity of information he has collected regarding Zimbabwe while Dr. Eandall-Maciver, in his Mediaeval Rhodesia, has made us acquainted with that remarkable and extensive complex of pit-dwell- ings and forts which he terms the Niekerk ruins, and, what is more important, has given us the true interpretation of the accumulated archa3ological data. 77 APPENDIX. BROADLY speaking, South Africa may be divided into the western desert region, the north-eastern bush country, and the central high veld, which encroaches on the east and south. In addition, there are the mountain ranges of the extreme east and south, and the narrow low-lying coastal belt. The sites described in this book are mainly situated on the high veld and in the bush country. Those who desire to gain an idea of the past and present environment of the sites on the high veld should read Chapters I. and V. of my Geological and Archaological Notes on Orangia. The pre-historic bush country was probably no different to the present, and the following account from my diary of an expedition through a characteristic portion of it will present to the mind a very good picture of this type of country. ****** [From Pietersburg, through Chunies Poort to Onverwacht, in the Lulu Mountains, thence up the Ohrigstad Valley and across the Drakens Berg to Leydsdorp and back to Pietersburg ; 9-25 Sept. , '08. Party : Messrs. Nash and Hambidge and self with Kafir driver. Outfit : Wagonette and six mules.] Pietersburg is the northern terminus of the Central South African railway system. The immediately surrounding country is typical high veld like that of Orangia, and consists of undulating grass plains, broken by an occasional kopje. The geology, however, is different, the plains being mostly underlain by granite and the kopjes mainly composed of schists. It is sparsely dotted with mimosa. About twenty-five kilometres south-east of Pietersburg is a range of mountains trending roughly north-east and south-west, separating the high from the low veld. Our immediate objective was a pass through them known as Chunies Poort. We left Pieters- burg at mid-day and arrived at our first camping place, the entrance to the pass, where there is a store, towards night. Starting early the next morning we entered the pass. Chunies Poort begins as a breach in the precipitous quartzite escarpment and continues for a considerable distance as a picturesque gorge, thence developing into a narrow valley hemmed in on either side by rugged mountains. Its bottom is shared by the wagon track termed a road, and the Chunies River, which, though a raging torrent during the rainy months, is at this time of the year (the end 6f the dry season) merely a boulder strewn channel. Its only tenants are the baboon, leopard, and other wild animals. The gradient is all the way markedly downwards. There are some old Kafir workings in copper-ore in the dolomite formation in the neighbourhood. The pass is well clothed with bushes and small trees, and towards the further end the peculiar euphorbia and other more tropical forms appear. After three hours' trek we reached a store by M'Phatleles loca- tion, well out of the pass, and outspanned for a couple of hours for breakfast. We were then in the low veld a tract of undulating country covered with thorn-bushes and diversified by parallel ranges of kopjes. Thence we left the Chunies River, which turns abruptly eastward, and continued due south to the Olifants Eiver, where we again outspanned. Taking advantage of a glorious moon-lit night we made a further trek and camped opposite Adriaan's Kop, a prominent landmark interesting to the geologist as being a siliceous segregation in granite a little distance up the river. The wide, tree-bordered channel of the Olifants at this point is now only occupied by a series of crocodile-infested pools linked together by a feeble meandering stream. The temperature up to now had been fairly high, but the next day it was perfectly torrid, so we decided to give our animals a rest and spent the time fossicking around. 79 In the night we trekked by moonlight back down the Olifants, and the following day, when it was cool and windy, continued in the same direction until opposite the Lulu Mountains, where we crossed the river and camped. The next day we proceeded along the foot of the mountains as far as the " farm " Twickenham. During the morning it was windy and overcast, but the weather improved in the afternoon. The Lulu Mountains are the most conspicuous of a series of parallel ranges that constitute a tract of country of singular scenic beauty. They are composed of gabbrodiorite , a rock that, like granite, frequently weathers into fantastic piles that are specially the habitat of the equally bizarre euphorbia. Here and there one comes across a Kafir kraal, the huts of which are picturesquely perched among the great boulder-like outcrops. At Twickenham we were hospitably received by Mr. Winter, junior, who is acting as collector of native hut rents for one of the big land companies owning some of the tracts of unutilizable wilder- ness, termed farms, in the neighbourhood, and w r ho has a prettily situated little house perched on the lower slopes of the range and commanding a fine view of the valley between this and the next. About three-quarters of an hour's drive further on is a charac- teristic pile-like kopje known to the Kafirs as Swali, where there is a natural recess or rock-shelter, on the back of which are a number of Bushman paintings. They depict various antelopes, a large elephant, and men and women, in silhouette, some of the animals being very well represented. They are all in one colour red and have suffered much from the action of the atmosphere , most of them being now very indistinct. A large slab of the rock with paintings on has fallen away. On the hill opposite (Mosego) a former chief of the local Kafirs, named Sekwati (the one who was captured by Gen. Wolseley and taken to Pretoria) , is buried in the cattle enclosure of his kraal , and sacrifices of black goats or oxen are still made there when rain or other favours are wanted. The next day we trekked on to Onverwacht, which we reached towards night, stopping on the way to see the paintings at Swali 80 kopje and making one outspan only, this being at a pool by a trader's store, about an hour and a half from our starting point. We passed several Kafir kraals on the way, the females of which turned out to celebrate the occasion. One group of intombis, encouraged by a greeting in their own language, followed us for a long distance, and we caused considerable merriment by chaffing them. After much persuasion, for they have an inherent fear of all such contrivances as cameras, we induced them to submit to be photographed , though only by consenting to include ourselves in the group. At Onverwacht we were hospitably entertained by old Mr. Winter, who is the father of our host of the night before and a gentleman of considerable education. One can imagine our surprise at coming across in these wilds a man conversant with the dead languages and possessing an extensive library of philosophical works ! He came to the Lulu Mountains twenty-eight years ago as a missionary. Then the country teemed with elephants and big game of every description ; now the largest animal met with , apart from an occasional stray impala or koodoo, is the rooi-rhebok. The lions have disappeared, but leopards are still plentiful. As a result of his intimate knowledge of the local Kafirs and their language , old Mr. Winter has become a storehouse of their traditions and beliefs, many of which are of great interest. One tradition particularly attracted my attention. There are many old workings on iron and copper ores in the Lulu and neigh- bouring mountains, while the slags produced in smelting them are to be found on almost every " farm " in the vicinity. The tradition is that there was once a race of miners who were not Kafirs, in this part of the country, and the last of whom were killed by a former chief, Tulare. The Sekwati already referred to was the son of Tulare and father of Sekukune. The miners had guns which were buried by Tulare in a cave. Sekukune had the guns exhumed and converted into hoes. This last operation was performed by Seku- kune's blacksmith, who was personally known to old Mr. Winter, and who described the guns to him. They differed from all the other guns with which he was acquainted in having long curved stocks. This seems to point to the miners being Arabs. 81 Tulare had his kraal on the Steelpoort River, and was noted as a kind, just, and powerful ruler. To be addressed as "Son of Tulare ' ' is the greatest compliment that can be received by a Kafir of this neighbourhood. Among the numerous interesting beliefs related to me by old Mr. Winter the primitive idea that the enviable characteristics of a dead man can be acquired by eating his flesh is noteworthy. This was done as recently as 1901 during the faction fights that accom- panied the British-Boer war. Old Mr. Winter seemed quite contented with his isolated exist- ence, the presence of a large family, comprising wife, three more sons and a couple of daughters, being, no doubt, a compensating factor. All the sons are veritable giants, but with constitution weakened by the dread malaria. The next morning, after making tracings of some Bushman paintings (Fig. 42) that occur in a rock-shelter near by on the mountain side, we proceeded on our journey in an easterly direction, leaving the Lulu Mountains at our back, and, after an excessively hot mid-day, during which we outspanned, and an overcast and cool afternoon, arrived at the Steelpoort Eiver, where we camped at night. Break of day revealed a pleasing and unexpected scene. Wind- ing through the brown bush veld, like a great serpent, and following the course of the river, as far as the eye could see, was a long belt of green corn. Here and there in it nestled a homestead. Very rest- ful was the sight of this strip of civilization traversing the wilder- ness. A similar scene was encountered at the Spekboom Eiver, which we reached at mid-day. It was again excessively hot. After a couple of hours outspan we continued our journey. First through a narrow and stone- strewn, but well- wooded, valley, then over a big hill-range with bare undulating surface, and, finally, down into flat bush country again. Darkness then overtook us, but we struggled on. After what seemed to us an interminable time we entered a little kloof in a randje, the country beyond which was lit up by the glare of a big grass fire, itself as yet concealed from us. An exciting dash through the further end of the kloof, where the flames, which were 82 sweeping along the slopes on the other side of the randje, nearly caught us, another quarter of an hour's trek through long tangled grass, and we had reached the Ohrigstad River, where we camped for the night. In the morning we discovered that we had camped immediately opposite the site of Ohrigstad. This village was founded by the Boers in the early days but was abandoned after a long struggle with the dread malaria. Most of its former inhabitants l;e in the grave- yard. Of recent years malaria has been very much less prevalent, and as in the case of the Steelpoort and Spekboom Rivers, farmers have established themselves all along the river, which is now y bor- dered by a pleasing belt of green corn. Close by was a homestead from which we obtained some bread. Learning that there was a farrier across the river we inspanned after breakfast and went in search of him, our mules having already dropped no less than five shoes. We found him engaged in building a new house. He was a Canadian who had married a Boer girl and settled down there. Luckily he still had his tools. So after the usual cup of coffee we started to work, all lending a hand. The smithy was an open-air one sheltered by the spreading branches of a big marula, and consisted of a small prospecting forge and anvil. A heap of scrap iron supplied the material for their shoes. The task was completed by mid-day. In the latter part of the afternoon we proceeded on our way up the Ohrigstad Valley. This portion of the journey was one of the most delightful that it has ever been my lot to undertake. The dusty w r agon-track , sheltered from the blazing sun by overhanging trees, skirted on the one side the bush-covered slopes of the valley and on the other the waving corn of the broad alluvial flats that separated it from the tree-fringed river. HOW T restful to the eye was the fresh green of that stretch of cultivated land after the monotonous brown of the parched wilderness. Here and there the road crossed the river where the latter zig-zagged over the flats. At the first crossing beyond the village site we found the exit from the drift a narrow cutting through the river bank blocked by a big tree which, w r e learnt, the farmer had placed there with the aid of a team of oxen. Our views, however, as to the desirability 83 of returning to the drift we had already crossed in the morning, not coinciding, we chopped the tree into three pieces and removed them with the help of the mules. We continued our journey till dark and started again at day- break, then outspanned for breakfast at a spruit. Ahead were the Drakens Berg. These we had to cross, but the narrow breach through which the river rushes affords no dry passage. So we con- tinued along the old track which here leaves the valley on the right and crosses the mountains in a north-westerly direction. A steep winding ascent, a long succession of ups and downs and what seemed to us an almost perpendicular descent, and we were once more on the banks of the Olifants, though some hundred and fifty kilometres away from the point where we had first encountered it. I shall never forget that descent. The last ridge terminates in an abrupt escarpment, consisting of steep granite slopes capped by quartzite precipices and towering above the gently undulating bush country beyond. Our track led through a break in the cliffs to the top of the slopes and thence along a narrow kloof down them. It is a wonder to me that our wagonette was not smashed to pieces in its career down this dry stone-strewn boulder-blocked water course. A score of times it nearly capsized and a score of times the mules were all but a huddled heap under it. The bottom of the pass brought us to the river at the point where it breaks through the mountains. We camped on the river bank, just within the gorge, under one of those gigantic trees that are only met with along the big rivers, and amidst some of the finest scenery that I have ever seen. Close by w r ere some fantastic baobabs and a Kafir kraal. After a bathe in a rock-pool (it is not safe to go in the main stream on account of the crocodiles) and a good meal, by which time it was dark, we indulged in a well-earnt rest, watching the fire-flies and listening to the running water until sleep closed our eyes. The next morning we set out to find a drift, but had not gone far before the disselboom was broken in crossing a small spruit ; this we repaired. A little further on we reached a drift and crossed the river. We outspanned towards night at a spruit, overtaking a Kafir 84 with his wife and daughter carrying his belongings, who were making for Leydsdorp. Here we had to make big fires round the camp owing to the frequent presence of lions, which are very fond of mules, in the neighbourhood. We left the spruit after breakfast and at mid-day outspanned at a tributary of the Macoutsie. Here we came across a four and a half metre python, which one of us shot, lying in wait for game at the bottom of a pool. Thence we trekked again and outspanned for the night at another tributary of the Macoutsie, where we secured a photograph of an interesting basic aphanite dyke traversing the" granite. In the night a strong wind arose followed by rain. The next morning we pushed on again. The sky was overcast and the weather cold. Yesterday the temperature attained 36 C. in the shade, to-day it is 11 C. We arrived at Leydsdorp towards night. On the way we saw a couple of hippotragi. The country traversed from the Selati Eiver to Leydsdorp is beginning to get green. It was cold all day with fine rain. This is my second visit to Leydsdorp. It is a fever-stricken spot surrounded by kopjes, and comprises a single street, built on one side only, the government offices, which occupy a building apart from the others, and a few scattered houses, mostly unoccupied. Its present white population , which we found congregated at the inn bar, numbers six souls in all. With the surrounding country bathed in mist (a very unusual spectacle) it looks very different to when I saw it last. Only two of the mines in the neighbourhood the Blue Jacket and the Sutherland are now being worked, and only on a small scale. The next morning we continued our journey, outspanning by. the store on the Thabina Eiver (3| hours distant) for breakfast and reaching the Letsetele Elver, where we camped, towards night. All day it was overcast, with a cold wind and some drizzle. Instead of, as before, taking the direct road from Thabina store to Pietersburg, it was decided to take the much longer, but easier, road that goes by the Tzaneen estate and thus avoid the there steep ascent from the Groot-Letaba Valley. We found this road good but heavy owing to the wet. 85 Next morning we started off again and outspanned for break- fast just across the Groot-Letaba by the store opposite the tobacco factory of the Tzaneen estate. In spite of the extremely unhealthy climate and the distance from railways and markets, there are quite a number of farmers in this part of the valley. Cotton, tobacco, fibre, coffee, tea, and rice all thrive there, but have not yet been cultivated on a commercial scale. After inspecting the estate we proceeded till night, when we reached another store where we camped. The same peculiar weather continued throughout the day, but the delightful scenery through which we passed amply compensated. This section of the country is very rugged and well-wooded, and under the present abnormal meteorological conditions, with its dank vegetation, clammy atmosphere and mist-draped hills, reminded me forcibly of Grampian scenery. But for the paw-paws, bananas, palms, and other tropical types here and there, the resemblance would have been complete. The next day we went on to the Koodoos Eiver, where we out- spanned, then over to and up the Middle-Letaba Valley and thence up a long sinuous path to the high veld again. The high veld here is not quite typical, not being nearly so high, for instance, as at the point where the other road reaches it, and hence one finds it dotted with a mixture of thorn-bushes and euphorbias. The weather had improved considerably, but the sky was still overcast. We camped for the night at Kelly's store, on the farm Netrecht. In the morning we were up early, made a three hours' trek and outspanned opposite a Boer homestead. We were now on typical high veld again with only occasional patches of sparse bush. The sky was again clear and the country bathed in its accustomed sunshine. Another three hours' trek and we were back at Pieters- burg. 86 LIST OF FIGURES. FIG. PAGE. 1. Section through Terraces of the Thames Valley, Britain ... 3 2. Eoliths from Leijfontein ... 5 3. Neolithic Circular and Concave Scrapers, from Button, Britain 7 4. Eoliths from Leijfontein ... 9 5. View of Victoria Falls ... 11 6. Acheulic Amygdalith, from Luckhoff 13 7. Acheulic Amygdalith, from Luckhoff 15 8. Small, much worn, Acheulic Amygdalith, from beneath the Alluvium of the Taai- bosch Spruit ... ... 17 9. Acheulic Flake-Tools, from Vereeniging, Bezuidenhout Valley and Taaibosch Spruit 19 10. Solutric Amygdalith, from the Embabaan Valley ... ... 21 11. Solutric Scrapers of Lydian- Stone, from the Junction of the Eiet and Modder Rivers 23 12. Solutric Scrapers of Chert, from above the Alluvium of the Taaibosch Spruit ... 25 13. " Pigmy Implements," from the Solutric Site at River- ton 27 14 18. Scrapers from Solutric Sites near Boshof , 29, 31, 33, 35, 37 19. Two similar implements of un- known use, the one from the Site at Rietpan, and the other from a Rock- Shelter on Kranzfontein 38 FIG. PAGE. 20. Incised Fragments of Ostrich Eggshell and Minute Scrapers, from Solutric Sites near Boshof ... ... 41 21. Concave Scrapers, from Solu- tric Sites on the Wit- waters Rand 43 22. Scrapers of Aphanite, Chert, and Quartz, from Solutric Sites on the Witwaters Rand ... 45 23. Varieties of the Crescent Form of " Pigmy Implements," from Riverton, Bulawayo, and Hart River ... ... 46 24. Ostrich Egg-shell Beads and Tools used in their manu- facture 49 25. Solutric Scrapers, from Rock- Shelters in the Asbestos Hills 50 26. Small Solutric Flake-Tools, from Rock-Shelters in the Asbestos Hills ... ... 51 27. Middle Portion of Bushman Arrow, and similar Trun- cated Bone Pin, from a Kock-Shelter, at Mod- derpoort ... 52 28. Solutric Arrow-Heads, from Cradock 53 29. Two Rhinoceroses, finely and coarsely pecked in full, and Head of Rhinoceros pecked in outline, Wolmaransstad 55 30. Eland, shewn by means of elongated pecks, Wolmar- ansstad 57 87 FIG. PAGE. 31. Bantu Imitation of Solutric Pecking, Wolmaransstad ... 58 32. Engraving of Zebra, Wal- maransstad ... 61 33. Representations of Zebra, Hippotragus and Quagga, pecked and engraved on Rock, Biesjesfontein ... 63 34. Engraving of Eland, Bloem. fontein 64 35. Engraving of Quagga ( ?), Bloem fontein 65 36. Bantu Engraving of Horses, Bloem fontein 66 37. Engravings of Arrow-like Ob- jects at Bloemfontem, and Painting of Similar Object from the Niaux Cave ... 67 38. Bantu Peckings of Humped Oxen, and Two Fanciful ( ?) Figures, Blauwboschdrift 68 | FIG. PAGE. 39. Bantu Pecking of Men on Horses, Blauwboschdrift ... 68 40. Bantu Pecking of Ox and Hyaena, Blauwboschdrift... 68 41. Bantu Pecking of Plough, Blauwboschdrift 69 42. Portion of Rock-Painting in the Lulu Mountains ... 70 43. Painting of Man dressed up as Bird, and of Two Ante- lopes, from a Rock- Shelter, near Carolina 71 44. Painting of Bantu Wedding Dance 72 45. Two Paintings, representing Custom and Myth 73 46. Plan of Inyanga Fort 74 47. Two Marali, from Palabora ... 75 .88 INDEX OF LOCALITIES. (frames of Farms in Italics.) Adriaan's Kop Asbestos Hills PAGE. 78 ... 16,46 Khami Ruins Kleinbuffelsspru it ... Klerksdorp PAGE. 64,69 47 48 Bauiaanskranz Barkly 43 24 Koffy fontein 43 Bezuidenhout Valley Biesjes fontein Blauwboschdrift Blauwbank ... 26 ... 43,52 54 46 Komati River Koodoos River Kranzfontein Krugersdorp Valley 22 85 46 26 Blinkklip 45 50 Ladybrand ... 43 Bloemfontein Boshof 53 32 L/eij fontein ... 11,12,16 Bulawayo 45 Letsetele River 84 Bumbusi Ruins Caledon River Campbell 48 22 16 Leuw fontein Leydsdorp Limpopo River Luckhoff 50 77,84,84 62,71 30 Campbell Rand Capetown 16 10, 22 Lulu Mountains 77,79,81 Carolina 47 Macoutsie River 84 Chuna Ruins Chunie's Poort Cradock 64 77, 78, 78 44 Magalakwin River Matopo Hills Meerlandsi-lei 22 47 42 Devilliersrust Dhlo-Dhlo Ruins Doornkloof ... 42 ... '64,66 47 Middle-Letaba Valley Modder River Modderpoort 85 34 43 Drakens Berg Elands fontein Elands-Rustenburg River Embabaan Valley Ficksburg ... 77,83 44 22 27 43 Mosego .Hill M'Phatleles Location M'Popoti Ruins Mundie Ruins Murchison Range . . . 79 78 64 i.e., 64 71 Fouriesburg 43 Nanatali Ruins 85 Griquatown Groot-Letaba Valley Grootvlei 46 84 43 Niekerk Ruins Ohrigstad Valley ... 76 77,82 Hart River 45 Olifants River 22,78 77 79 Inyanga Ruins 64, 66, 71 Orange River ... 22,37,42,58 89 Palabora Kopjes Petrusburg PAGE. 62,71 42 Taaibosch Spruit ... Thabina River PAGE. 36 84 Pietersburg ... 77,78 Twickenham 79 37 Tweeling&fontein 10 Rhodes' Grave Riet River 47 34 Vaal River Vereeniging 22,25,45 25,48 Rietpan 42 Victoria Falls 11,23 Vlakfontein 44 Riverton Island Robinson 28 Vooruitzicht 42 Rooi Berg 62 Warmbaths 62 Rooipoort 50 Water val 44 Selati River . . 22,84 Witwaters Rand 27,44 Signal Hill 44 Wolmaransstad 50,52 Spekboom River Steelpoort River . . . Steynsdorp Valley Swali Kopje 81 81 27,60 79 Zambesi River Zand River Zimbabwe Ruins Zwartkops River ... 10,11,22,48,62 62 ... 64,66,69,71 22; BY THE SAME AUTHOR. Geological & Archaeological Notes on Orangia. WITH ILLUSTRATIONS. PRICE 10 -. "A valuable contribution to the geological and archaeological literature of South Africa. " Knowledge. "An important contribution to the economic geology of South Africa." South African Mining Journal. " Contains probably the best geological description of the diamond mines of South Africa that has yet been written." - Mining World, Chicago. "An indispensable guide to the Orange Colony." London Mining Journal. LONGMANS, GEEEN & CO., 39, PATERNOSTER Eow, LONDON ; NEW YORK, BOMBAY, AND CALCUTTA. BY THE SAME AUTHOR. THE ORE DEPOSITS OF SOUTH AFRICA. Part I. BASE METALS. WITH DIAGRAMS. PRICE 5 - NET. " This volume, the author states in his preface, is intended to meet a demand among those technically connected with the mining industry, for a co-ordinated and condensed account of the base-metal ore-deposits at present known in South Africa. It is also intended as a guide to the prospector. The author claims that only an elementary knowledge of geology and some mining experience are necessary in order to understand it, and that, with these qualifications, it will mateiially assist the prospector in his search for metalliferous mineral occurrences. . . . This volume constitutes a work of reference that no one interested in the mining industry can afford to be without. It is at once a concise and co-ordinated description of the different known occurrences, a complete record of output, and a guide to the scattered literature of the base metals of South Africa. "- South African Mining Journal. Part II. THE WITWATERSRAND AND P1LGRIMSREST OOLDFIELDS AND SIMILAR OCCURRENCES. WITH Two LARGE GEOLOGICAL MAPS AND DIAGRAMS. Price 5/- net. CROSBY LOCKWOOD & SON, 7, STATIONERS' HALL COURT, E.G., & 12lA, VICTORIA STREET, S.W., LONDON. THE STONE IMPLEMENTS OF SOUTH AFRICA. SECOND EDITION, REVISED AND ENLARGED. WITH ILLUSTRATIONS. PRICE 1O -. "Our knowledge of the prehistoric period of South Africa, like that of other parts of the world, is mainly derived from the almost imperishable stone implements which were then used, and which remain when much else has gone. In this volume the author gives a co-ordinated account of his different discoveries in South Africa. The new edition describes a large number of fresh finds, and includes many new illustrations. One of the most striking features is the very wide geographic distribution of Mr. Johnson's finds, which has afforded a sound basis for classification. Four distinct periods are now recog- nised the Eolithic, Strepyic, Acheulic, and Solutric. Very interesting are the repro- ductions of aboriginal paintings, peckings and engravings, which were contemporary with the implements of the newest period." South African Mining Journal. LONGMANS, GREEN & CO., 89, PATERNOSTER Row, LONDON ; NEW YORK, BOMBAY, AND CALCUTTA. Printed by EDWIN SEAKS & Co., 4, Dorset Buildings, Salisbury Square, London, E.C. 120298 University of California SOUTHERN REGIONAL LIBRARY FACILITY 405 Hilgard Avenue, Los Angeles, CA 90024-1388 Return this material to the library from which it was borrowed. UC SOUTHERN REGIONAL LIBRARY FACiLiTy A 000 756 783 7 DA 570 D? B7 British Museum. Antiquities. Dept. of British and Mediaeval Hod Hill ... 1962- (Card 2) 1. Hod Hill, England. 2. Dorset, England - Antiquities. 3. Excavations (Archaeology) Eng- land. I. Brails ford, J W. II. Burden, Henry. III. Richmond, Sir Ian Archibald 1902-