The top of the upper sandy beds is not seen, but the thickness given is probably well within the mark. The levels of the granite bedrock also vary considerably, and the beds themselves tend to assume a lenticular shape, so that it is impossible to give more than a rough idea of their proportionate development.

The UPPER SANDS, which have been entirely removed by denudation from some of the ridges, and from all the lower ground, are sometimes clayey and stained red by iron oxides, and sometimes fine and white. In a shaft at one spot on the slope of a ridge an actual thickness of 16 feet was passed through before reaching the underlying gravel. It may be stated, however, that so heavy an overburden is met with on few parts of the diamondiferous area.

The GRAVEL itself is composed of beautifully-rounded pebbles in a matrix of sandy clay, sometimes ferruginous. There are some concretionary masses of iron, cemented sandstone, and the gravel is converted in places into a hard conglomerate by infiltrated iron oxides or more rarely by silica. The pebbles are mostly of quartz, frequently rock crystal, but they also include jaspery banded-ironstone, chert, agate, hard sandstone or quartzite, and occasional large and small pieces of silicified wood, as well as fragments of granite and chloritic schist. Large boulders are comparatively rare. The silicified wood, though distributed about in all sorts of positions, may possibly have formed in situ, the granite and schist last mentioned are the only other constituents of the deposit that are not well rounded. The presence of the agate, as already mentioned, shews the deposit to be newer than the lavas of the Forest Sandstone series. In one shaft, 25 feet of gravel had been passed through at the time of my visit, without any indications of approaching the base, and more recent work has shewn that my estimate of 40 feet is probably a moderate one for the maximum thickness.

The LOWER SANDY BEDS have obviously derived most of their materials from the underlying granite. They are micaceous throughout, the upper and lower parts South of the Railway being clayey with a bed of clean sand in between. The base shews fragments of decomposed granite, and it is difficult to fix the point where the bed rock really begins.

It is clear that in these sands and gravels we are not dealing with the insignificant accumulations of the present-day river system. This is no less evident from the character and distribution of the deposits than from their position on the crest of what is now the main watershed of the country. Their extent is quite in keeping with a lacustrine origin; on the other hand, they correspond closely with the alluvial deposits of rivers which have eroded their valleys practically to the lowest possible level, and have for long been chiefly occupied in widening them and spreading the materials furnished by the process evenly over their flood plains. The Somabula beds may therefore be set down, provisionally, as due to the action of an important Tertiary River or river system, probably a feeder of the

great lake which must once have filled the adjacent portion of the Zambezi basin, and draining an area chiefly occupied by granite and the Archaean banded ironstone.

The gravels of the Somabula are interesting from their unique lithological character, as far as Rhodesia is concerned, but their chief interest naturally arises from their being the source of various gem stones, particularly the diamond. The diamonds themselves are peculiar, as almost invariably of a green shade in the rough: this is, however, entirely lost in cutting. They occur in very good crystals, principally octahedra, spinel twins of two octahedra, twinned tetrahedra, twinned hexatetrahedra, dodecahedra, etc. Etched triangles are characteristic of the tetrahedral faces. Worn stones are almost entirely absent. The mineralogical associates of the precious stone are not precisely similar to those of the Kimberley diggings or of the more recent Transvaal discoveries, but they nevertheless present a general resemblance to those of the localities named. Garnets are often common, but are not of the blood-red Kimberley variety. Ilmenite is uncommon. Both magnetite magnetite and haematite, of which grains are numerous, are evidently derived from the Banded Ironstone, while the source of the zircon and of the mica (muscovite) is equally clearly the granite: these minerals have no necessary connection with the original matrix of the diamond. This is also probably the case with the beryls which occur. The typical minerals are (besides the diamond) enstatite staurolite, chrysoberyl, kyanite, and sapphire. Enstatite is the commonest of the minerals popularly grouped together under the name of olivine at Kimberley, but that found here is a brownish variety. Staurolite is an abundant constituent of the sorted material from the puddling machines. Some of the grains are fairly clear, and might almost be taken for garnets on account of their red-brown colour. Chrysoberyl is quite abundant for so rare a stone. The prevailing variety is yellow, but the opalescent (precious) catseye," and the form known as alexandrite," which is green by day and red by candlelight, also occurs. This mineral would seem to be usually a product of contact metamorphism: at the same time, although it is not strictly analogous in a chemical sense, its similarity of composition and isomorphism with olivine are to be noted. As a gem, the ordinary variety is actually called " chrysolite," one of the names properly applied to olivine. Kyanite and sapphire are unequivocal contact minerals, so are rutile and tourmaline, whose occurrence may also be noted. The presence of the first-named is interesting, owing to its softness, which makes its survival rather remarkable. Of the sapphires, both the blue and colourless varieties occur, while true rubies and oriental amethysts are also found, though they are distinctly rare, even for such scarce stones. Another stone whose occurrence may be noted is the so-called "Somabula Blue"; this is harder and heavier than common beryl, and is biaxial with a wide ophiaxial It angle, cleavage flakes shewing the bisectrix normal to them. seems unquestionably, therefore, a variety of topaz: when cut it is one of the most beautiful gems imaginable.

With regard to the origin of the gems, the mere richness of the deposit is sufficient to indicate a near source for the diamond, although many of the constituents of the gravel itself have obviously travelled far. Despite theories to the contrary, it seems evident from the evidence obtained in New South Wales, as well as in this country, that what is commonly called "blue ground" is in all cases the original source of the diamond, and the occurrence of enstatite, as well as the presence of garnets, points to the same origin in the case of the Somabula field. I am aware that Professor Gregory, without making an inspection of the ground, has pronounced the opinion. that the diamond comes from pegmatite veins, but such an idea is so completely at variance with the local conditions, and with all that we know of diamond occurrences, that it scarcely merits discussion. I have little doubt that it will not be long before the pipe which produced the diamonds is discovered, and that it will present, apart from slight local peculiarities, all the usual features of the South African mines already known.

STONE AGE OF SOUTH AFRICA.

By J. P. JOHNSON.

Most of the material described in the following notes was obtained during a recent journey, as the guest of Professor R. B. Young, through Griqualand West. It is of the highest interest, throwing considerable light on the long. obscure succession in South Africa, and on the vexed question of the origin of the Eoliths, as well as introducing some types of implements which have not been previously described from the sub-continent.

PRIMITIVE GROUP FROM LEIJFONTEIN, HERBERT.

In several places on this farm there are patches of gravel lying at the foot of the dolomite escarpment. This gravel consists of subangular fragments of chert and jasper, and is probably derived from patches of very ancient drift that formerly existed on the top of the escarpment.

The chert comes from the dolomite, and is the grey translucent variety of cryptocrystalline silica usually met with in that formation. The opaque jasper is brown inside, but externally has changed to a yellowish-brown, and acquired a high glaze or polish.

While the chert may have been supplied entirely locally, the jasper, on the other hand, has travelled a long way, the nearest source being the Asbestos Hills, some thirty miles to the west.

Mixed with the gravel are quantities of much-worn and highlyglazed jasper Eoliths. A few of these are a little more advanced than the true Eoliths, being made from artificially-produced flakes, but they are a very small minority. Otherwise the group is in every way identical with the typical assemblage met with in the early plateau drifts of southern Britain.

Although attention was drawn to the hacked or rudely chipped stones, 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 them as the work of man, others are equally emphatic in denying their artificial characThe specimens from Leijfontein throw considerable light on this matter, and their testimony, in my opinion, is only capable of one interpretation, namely, that they really are primitive man's first attempts to trim pieces of stone to a useful shape.

*

The Leijfontein Eoliths and flake-Eoliths may be sub-divided in the same way as Prestwich divided the typical Eoliths, that is, into two sub-groups, (1) those in which the pieces of stone have been subjected to little modification, and (2) those in which they have been chipped into definite shape.

It would be difficult to recognise the artificial character of the implements of the first sub-group if found alone. Their great abundance, and the haphazard appearance of the chipping immediately suggests that they have been shaped by the blind forces of Both circumstances have been brought forward as evidence

nature.

against their artificial character. Nevertheless Paleolithic and Neolithic implements are sometimes met with in equal quantity, while, if the Eoliths are, as is claimed, man's first artéfacts, one would expect them to be barely distinguishable from Nature's work. Their association with others, in which the trimming, though of the same rude kind, is arranged in definite patterns, is the sole ground upon which they can be accepted.

Even the better-defined implements of the second sub-group are of so primitive a kind that their artificial character is still the subject of controversy. Yet, apart from the inferior quality of the trimming, and the fact that most are fashioned out of naturally broken fragments of stone, they are identical with the commoner accepted flake-tools of the Paleolithic and Neolithic periods.

Two series of the more differentiated Eoliths and flake-Eoliths, and a set of Neolithic implements of the best quality, for comparison, are represented by the accompanying illustrations (Plates 1, 2, and 3).

Plate 1 shews a series of straight concave and convex-edged scrapers. A., B. and C. are true Eoliths, while D., E. and F. are flake-Eoliths. A. and D. are good examples of the concave scrapers. It will be noticed that there is quite a wide difference in the quality of the workmanship of these two. There is a still greater difference between the better of these and the Neolithic example D. I have South African Palæolithic specimens which, in point of workmanship, 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 recurring varieties of scraper, usually designated by the really descriptive adjectives, circular, rectangular, long and broad. All of these can be matched by Palæolithic and Neolithic examples, while one is still to be counted among the domestic appliances of certain savage peoples. Compare the circular scraper with the Neolithic specimen B. Here, again, I can produce South African Palæolithic specimens, intermediate as regards quality of workmanship. This evolution in delicacy of finish is carried a stage further in some beautiful little examples which I have recently obtained at Riverton, in association with minute scrapers like those from Taaibosch Spruit. They are about one-half the diameter of the Neolithic example.

Plate 2 shews an extremely interesting series of implements. They are very typical of the Eolithic stage of culture, being rarely met with in more advanced assemblages. They are probably all scrapers. A. and B. are double-edged scrapers. It will be noticed that the chipping of the one edge is in the reverse direction to that of the other. C., D., E. and F. are very similar implements, but both edges are chipped on the same side. They are an eloquent testimony to the artificial character of the Eoliths. It is incredible that a long, tapering point like that of F. could be hacked out by blind agencies.

I have above alluded to Prestwich's classification of the Eoliths. His essay in "Controverted Questions of Geology" (1896) is still