surrounding the diamond pipes.

Besides Mr. Gardner Williams' works, already cited, the following may be specially referred to:BONNEY. On some rock specimens from Kimberley, South Africa. Geol: Mag: 1897, P. 497.

BUTTGENBACH. Quelques observations sur les champs diamantifères de Kimberley. Ann. Soc. Geol: Belgique. Tome XXXII. 1905. Memoires, p. 3.

ROGERS. Geology of Cape Colony. 1905. Chapter IX.,

P. 331.

HATCH AND CORSTORPHINE. Geology of South Africa. HARGER. The Diamond Pipes and Fissures of South Africa. Trans: Geol: Soc: South Africa, VIII. 1905. p.

II2.

Numerous further references will be found in the works above cited, and in the catalogue of printed books, etc., relating to the Geology of South Africa, by Miss M. Wilman (Trans: S.A.Phil. Soc.) Vol. XV., part 5.

II. GENERAL DESCRIPTION.

It is unnecessary to give any detailed account of the superficial deposits of the area surrounding the mines: it must suffice to say that the soil appears to be thin, and is described as of a red colour; it is probably of a lateritic character, derived from the immediately underlying basic igneous rocks. A layer of what is described as limestone is widely distributed over the surface of the district. This appears to be of the nature of travertine or calc-sinter (the calcareous tufa of some authors). At all the mines except Bultfontein, the surface below the soil is composed of a thick bed of a rock, which is commonly spoken of as basalt, varying in thickness from 50 to 100 feet. This is underlaid by some 200 or 250 feet of black carbonaceous shale, and below this again comes a representative of the well-known Dwyka Conglomerate, which is here very thin. These sedimentary rocks scarcely require petrographical description, and no thin slices of them. have been prepared.

Below the conglomerate in the Kimberley and De Beers Mines we have about 400 feet of a somewhat decomposed igneous rock, commonly known as melaphyre. Then comes another series of sediments; 400 feet of quartzite, followed by 260 feet of shale, according to Williams. As we shall see in a subsequent section, this broad division scarcely holds good on close examination, and, indeed, the distinction between shale and quartzite is here purely arbitrary; both rocks have very similar mineralogical composition, and the differences depend chiefly on the relative sizes of the constituent particles.

At a depth of about 1400 feet in both Mines there begins the great series of acid igneous rocks, the quartz-porphyries of Williams. As we shall see later, this series is by no means uniform, but includes several different petrographical types. At Kimberley this series

extends down to 2470 feet, from which depth comes another specimen of a sediment of somewhat peculiar character. At 2500 feet we reach granite, and the greatest depth represented by the specimens in our possession is 2520 feet. In the De Beers Mine, on the other hand, granite is reached at 1920 feet, so that it appears that the upper surface of the granite is here very uneven, rising 600 feet in about a mile. The presence of a sedimentary rock immediately above the granite suggests that this is a buried land surface, and extreme unevenness is a common character of land surfaces composed of denuded archæan or igneous complexes, e.g., the Lewisian gneiss, and the gneissose area of Brazil, in the neighbourhood of Rio de Janeiro. The other mines, Bultfontein, Wesselton, and Dutoitspan, do not reach a greater depth than 750 feet, so that they throw no light on this question.

In the latter group of Mines the upper part of the series shows some variation in detail, but the general succession is very similar. The differences chiefly occur in the rocks above the melaphyre; in the diagram given by Williams, quartzite is shown both at Bultfontein and Dutoitspan at a much higher horizon than elsewhere. We shall return to this question later on.

Since the Kimberley Mine is the one which has up to the present reached the greatest depth, it may for our present purpose be taken as typical, and will be described first, and in considerable detail. Some of the rock types, however, show better development elsewhere; in such cases description will be deferred till the mine in question is treated of.

Since the ground surface around all the mines is at practically the same height above sea level, varying only a few feet on either side of the contour of 3990 feet, it will be convenient in all cases to use the surface as a datum line, and to speak always of depths below this generalised surface. Since the depths given appear to be approximate, correct to only 10 feet or so, no appreciable error will be introduced, and the descriptions will be more intelligible than if heights above sea-level were used.

III.-PETROGRAPHICAL DESCRIPTIONS.

I. KIMBERLEY MINE.

The present surface of the ground at the Kimberley Mine consists of material which is collectively described as debris, having a total thickness of 30 feet, overlying 5 feet of red soil, which is probably of a lateritic nature. Below this we come to the usual basalt, which is here 50 feet thick, rather below the average. The specimens of basalt from this locality (201) (a) are much decomposed, while excellently fresh examples of what is undoubtedly the same rock come from other mines, and notably from Wesselton, so that consideration

(a) Numbers in brackets refer throughout to the original numbering of the specimens as supplied by the Company.

of this rock type may profitably be deferred for the present. Underlying the basalt are 250 feet of black, carbonaceous shale (202), which is certainly of Karroo age, and probably represents the Upper Dwyka Shale (b) of Cape Colony. Below it comes a thin representative of the well-known fossil boulder-bed, the Dwyka Conglomerate (203), which has here dwindled to a thickness of only 10 feet. These rocks need no further reference here.

Below the conglomerate we come to the next great development of igneous rocks, which is about 400 feet thick. This is the rock referred to by Gardner Williams and others as melaphyre (204, 205, 205a). It is an amygdaloidal, non-porphyritic rock, which is considerably decomposed, so much so that determination of the ferromagnesian minerals is difficult. The rock varies a good deal in texture and structure, since it occurs in very thick masses, and the inner parts appear to be somewhat coarser than the margin. The structure, apart from the amygdaloids, is essentially that of a coarsegrained volcanic rock, and it is apparently not holocrystalline, although decomposition has proceeded so far that it is difficult or impossible to determine the original nature of the interstitial matter. The dominant minerals are a plagioclase feldspar and green chloritic pseudomorphs, representing some member of the ferromagnesian group. The feldspar occurs in idiomorphic, somewhat elongated prisms. The majority show twinning on the albite law, and often Carlsbad twinning also. The extinction angle, measured on the albite twinlamellae, rarely exceeds 5°, so that the feldspar may be regarded as oligoclase. In a few cases a somewhat higher angle indicates andesine, a few sections show Carlsbad twinning only, and these prisms are rather shorter and stouter than the others, so that a little orthoclase may be present. This point is somewhat doubtful. The ferromagnesian minerals are unfortunately very much decomposed, being chiefly represented by green chloritic pseudomorphs. Many of these pseudomorphs enclose, however, a few fibres of a pale or colourless actinolitic hornblende, of the character usual in this mineral when of secondary origin, from the uralitisation of augite, and a few crystals of comparatively unaltered augite may also be seen. It it clear, therefore, that the original mineral was augite, which has undergone the usual cycle of changes, being first converted by uralitisation to actinolitic hornblende, and this in its turn to chlorite.

Some specimens of the rock show a certain amount of interstitial quartz, and in places it is clear that the last substance to crystallise was a eutectic of quartz and feldspar, as is so common in the more acid members of the dolerite group.

There is a good deal of iron ore in small scattered grains, including both magnetite (or ilmenite) and pyrites, and apatite is very abundant in minute needles. There are also a certain number of idiomorphic crystals of a pleochroic yellowish green epidote, which may be an original constituent, but is more probably secondary. As

(b) Rogers. Geology of Cape Colony, p. 147.

before stated, the original character of the small amount of interstitial matter is indeterminable, but probably little if any glass was originally present.

Amygdaloidal vesicles seem to be more common near the margin than in the centre of the mass. They are filled with concentric layers of chlorite, calcite, quartz, and chalcedonic silica, in indeterminable order.

This is apparently the rock which was determined by Stelzner as "olivine-diabase," but I have been unable to find any olivine or any mineral which can be supposed to represent olivine in the specimens at my disposal. One or two of these also show a fair amount of quartz, and it is unlikely that olivine would occur in such a rock. Large masses of sub-basic and basic rock of this type often show a good deal of variation in silica percentage, and a small increase in this respect is sufficient to prevent the formation of orthosilicates of the olivine group. The absence of the mineral in a few slices only is therefore inconclusive. It must be confessed, however, that the general characters of the rock suggest a distinctly more acid type than that to which continental petrologists are in the habit of applying the terms olivine-diabase and melaphyre. The latter, indeed, is defined by Rosenbusch as more or less equivalent to his labradorite-porphyrite, which is an essentially basic rock. However, the material in our possession is insufficient to fully decide this question. It can only be stated provisionally, and subject to modification, that the rock appears to be an intermediate lava, with affinities to the hypabyssal quartzdolerites. However, it is probably more acid than the majority of these, and cannot strictly be referred to any accepted rock type. So far as the feldspathic constituent is concerned, the rock shows some affinity to the mugearites of Skye described by Harker (*), which are essentially oligoclase-rocks, with subordinate orthoclase. However, in these Scotch rocks augite is quite subordinate to olivine and iron ores, whereas in the Kimberley rock olivine is perhaps absent, and, at any rate, not abundant, while augite is the dominant ferromagnesian mineral.

Below the "melaphyre," at a depth of 740 feet begins another series of sedimentary rocks, which has a total thickness of some 660 feet in the Kimberley Mine, and over 700 feet at De Beers. The section given by Gardner Williams shows 400 feet of quartzite above, with 260 feet of shale below, but our specimens from various depths indicate an alternation of these two types of sediment, as shown in the table (fig. A.), (206b., 206k.. 206g., 207, 207b.).

A specimen of the quartzite from a depth of 1000 feet (206b.) consisted originally of well-rounded grains of quartz and feldspar, together with compound grains or small pebbles of various finetextured, siliceous rocks. The feldspar is mostly orthoclase and microcline, and this and the rock fragments together make up at a rough estimate 10 per cent. of the whole. There are also a few crystals of a deep reddish yellow, isotropic mineral, which is probably

(*) Tertiary Igneous Rocks of Skye. Mem. Geol. Survey, 1904, p. 264.

perowskite. The rest of the rock consists of clear quartz, with the usual fluid pores.

The cement is entirely of clear quartz, and is in crystalline continuity with the quartz of the grains, so that grain and cement make up one individual, which extinguishes as such. The result is a

mosaic of irregular-shaped quartz crystals, in which the outlines of the original grains are clearly to be distinguished. The rock is thus a very fine example of a typical quartzite, and it bears a very strong resemblance to stiperstones quartzite, † the basement bed of the Ordovician in West Shropshire. From the appearance of the grains it is evident that they have been derived from an area of metamorphic crystalline rocks, probably gneisses and granulites.

The specimen from a depth of 1200 feet (206k) is of quite different character. It is composed of smaller and much more angular fragments, which consist almost exclusively of quartz, embedded in a large amount of interstitial material, chiefly finelydivided mica, which is probably derived from the decay of feldspar under the action of weathering agents. The general character of the rock indicates somewhat different conditions to the last : it is probably derived from a granitic or gneissose complex under such conditions that the feldspar was decomposed and formed a fine argillaceous sediment, in which the quartz grains were embedded.

The specimen from, at, or near the 1300 feet level (206g.) is on the whole very similar in character to the last, but it includes rather more feldspar, together with some white and brown mica; the chips of quartz are still more angular, and suggest extremely rapid deposition, so that they were buried by the succeeding layers of sediment before any considerable amount of attrition had taken place. When examined under a high power the interstitial matter appears to be almost exclusively micaceous.

So far as regards their petrographical character, these rocks were originally feldspathic sandstones or arkoses which have been cemented into hard, lustrous quartzites of varying degrees of fineness. In some cases a rapid alternation of coarse and fine layers produces a well-marked lamination, so that they are often spoken of as shales, but the so-called shales are of essentially the same composition as the quartzites.

From the small amount of evidence at our disposal, it appears that this sedimentary series has been formed by the rapid denudation of a series of crystalline rocks, probably of a gneissose character. As we shall see later, the rocks which have been reached at the deepest levels of the shafts at the Kimberley and De Beers Mines are quite able to supply the kind of material here seen, and its source is probably to be found in the rocks of this series. It is perhaps dangerous to draw any conclusions as to the conditions under which they are formed, but what evidence there is points to deposition at a rapid rate in the immediate neighbourhood of a region of great denudation.

+ Harker, Petrology, p. 229, Fig. 2.