as it is with carbonate of lime, is not easily broken up with the rude tools at the command of the villagers. Whatever the cause I have only found two rude and doubtful fragments which bore evidences of human workmanship brought up in this way. They have since become mixed with the rest of the collection.

The first axe-head picked up was, strange to say, one of the most perfect found: a rapid search was rewarded by the discovery of a pile of specimens weighing over a hundred pounds, and as I was only accompanied by a single attendant I was obliged to make a selection of these, and leave the remainder behind. The next day I pitched my tent on the spot and began my inquiries.

From the large number of implements, and from various other considerations, I concluded that the spot where they were found had been the seat of a manufactory and that the implements had not been drift-borne from over extensive areas. Thus, the whole of the gravel stratum is not equally prolific of implements; indeed they are rare elsewhere. The spalls (i.e., chips) struck in the manufacture of these implements, and the huge primary flakes from which they were manufactured are found here; and I consider that the bulk of my specimens (say 95 per cent.) are unfinished implements.

The implements show signs of rolling, and weathering, and occasionally bear deposits of carbonate of lime. They are very unequally worn, some having the edges sharp, others being much worn and rounded. When broken across on purpose, they show that the material has altered in colour to the depth of a tenth of an inch and often more. The amount of wear and weathering on the celts is the same as that exhibited by fragments of similar rocks in the shingle.

No trace of fossil animal remains was found in the immediate vicinity.

The celts were found in situ, both in exposed sections of the gravel and in sinking pits, where the superincumbent alluvium is from two to three feet thick.

The amount of concretionary deposit on celts naturally weathered out is less than on those won by digging.

All the rocks which occur in the Talchir boulder bed are represented in the collection.

No polished implements occurred mingled with the roughly chipped; nor any implements formed of feldspathic rocks, or of jade. Stone hammers occur in the proportion of about 3 per cent. Flakes are found, but they are very coarse, and possibly doubtful.

About 12 feet of alluvium occurs at various points, but on carefully examining it no implements were found. There are

no indications of celts or rude flakes in the Talchir boulder bed itself. In two or three cases there are chips on the broad ends of the lanceolate specimens which seem to have been caused by use, but as a rule the broad end is unfinished and often bears a piece of the crust of the original pebble. The pointed end, on the contrary, is nearly always finished.

It will now be necessary to give a description of the composition and nature of the gravel.

The gravel stratum varies from two and a-half feet in thickness to one foot in parts. This in the Hinoutee locality is composed of boulders, pebbles, subangular fragments, cubical fragments, masses of limestone, &c. The boulders vary from 18 inches in diameter to tiny pebbles an inch in diameter. The whole is loosely cemented into a mass by carbonate of lime. In places, as opposite Amaharee, the cementing matrix is exceedingly hard and difficult to dig into. Here the superincumbent alluvium is from twelve to fourteen feet thick, and the gravel stratum projects some ten feet into the river's bed in a bold promontory, having so far resisted the erosion of the river, and offering an exceptionally fine field for observation. The gravel here, as elsewhere, rests directly upon the Talchir boulder bed, the lower strata of the gravel actually touching it. The rocks which occur in the gravel are almost identical with those in the Talchir boulder bed, and I find I have noted them as parti-coloured jaspers, jasper-conglomerate boulders, pink gneiss, hornblendic gneiss, porphyritic gneiss, tourmaline granite, lumps of epidote and epidotic granite, pegmatite, vein quartz, quartzites of all colours, cherts, and even graphitic schist.

I cannot identify the quartzite with any existing upper Vendhian' quartzite beds with which I am acquainted in the country between Urgoorh Ghat and Burdhee. The first implement found in situ was a hundred yards lower down than the projecting bed. Here a magnificent section of the drift gravel is exposed for the distance of a quarter of a mile along the east bank, covered with alluvium from 10 to 14 feet thick.

The specimen, an unfinished hache, lay with a portion of the worked point projecting, firmly cemented in the hard mass. Its position was slightly below the middle of the mass, and it required to be chiseled out with a cold chisel and hammer. It is uniformly covered with a fine deposit of carbonate of lime, except on the projecting portion.

The following section will give some idea of the relations of the gravel bed, Talchir, and superincumbent alluvium. The

1 The lower Vendhians seem everywhere to give way and disappear with far greater rapidity than the upper Vendhians. This is very noticeable in the Banda district.

Talchir beds are of very uneven thickness, and the dip rolling. For those who are not acquainted with Indian geology the following brief sketch of this characteristic formation is appended.

SECTION ON RIGHT BANK OF THE BALLIAH NADI, OPPOSITE HINOUTEE. A. Alluvium. B. Gravel, containing the implements. C. Talchir boulderbeds. D. Red sandstone. E. Green sandstone.

The Talchirs form the base of the Gondwana series and rest on metamorphic gneissic rocks: their thickness has been estimated at from five to 900 feet, as a rule, and in the area described, notably, "they form thin, irregular beds, filling up hollows in the metamorphic rocks which latter are often exposed through the Talchirs by denudation" (Griesbach, Mem. Geolog. Surv. Ind., Vol. xiv, p. 14, " Ramkola and Tatapani Coalfield"). The porphyritic gneiss of Pipra is the rock most commonly thus exposed.

The Talchir rocks consist of silty greenish or blackish shales, splitting into angular pieces (being jointed in three directions), or of tolerably compact green and red feldspathic sandstones, occasionally slightly gritty. The terms mudstones and needle shales admirably describe the appearance of the former. The boulder bed is usually green or black silty shale. In this indurated matrix occur pebbles and boulders of all sizes from an oval pebble one quarter of an inch in length to blocks 15 feet in diameter.

The Talchir boulder bed is now generally admitted to be of glacial origin, and is attributed to the close of the paleozoic epoch. It need hardly be said that no single fragment which bore the slightest resemblance to even the rudest implement

has yet been found in the boulder bed, though I have searched it in vain for many miles.

The Talchir boulder bed has been supposed to be of the same age as a very similar formation at the base of the coal-bearing rocks in South Africa. These rocks are described by Mr. Gooch in his paper on the stone age of South Africa ("Journ. Anthrop. Inst.," 1881, page 167), as "fine highly laminated shale with boulders included." It would appear from his geological diagram, that the quaternary alluvium and gravels which have yielded palæolithic implements in such abundance cap this boulder formation at more than one point, but I have not clearly made this out from the letterpress, and may be mistaken. As noted by Mr. Worthington Smith in the discussion that followed the reading of Mr. Gooch's paper, the palæolithic specimens of celts very closely resemble those from Madras, and I may add, the Singrauli gravels.

This brings me to Messrs. Foote and King's discovery of implements in the laterite of the North Arcot District, Madras.

Mr. Foote's discovery was made in 1865, and the results published in the "Madras Journal of Literature and Science," for October, 1866.

Most of his specimens were found in broken-up shingle, the débris of a laterite conglomerate composed of quartzite pebbles; but some appear to have been found embedded in solid laterite itself; this appears, likewise, to have contained pebbles. The laterite conglomerate either rested on metamorphic gneissic rocks, or on rocks which belong to the Upper Gondwana system, the Sri Permatur shales. These shales are of possibly similar age to the Talchir sandstones, and the thickness, composition, and deposit of the laterite gravel is very similar to the Singrauli gravel, substituting lime as the cementing matrix in the place of laterite.

No laterite is found near Hinoutee, but it caps the Pats of Sirgoojah 30 miles south, and even occurs north of the Sone River, near Sookerit, 21 miles south of Chunar, on the Ganges.

I personally compared my specimens with such of Mr. Foote's as were exhibited in the Calcutta Exhibition of 1884, and the specimens are so very similar, that it would hardly be possible to separate them were they mixed together. Every type figured by him is represented in the collection made.

He supposes that the laterite conglomerates and sands were deposited at the bottom of a shallow sea studded with mountainous islands, between which flowed strong and rapid currents, and that the implements were either dropped by accident from rafts or boats, or accumulated by the upsetting of these craft. He divides his implements into three classes :

Class I. Implements with one blunt or truncated end; II. Implements with a cutting edge all round; III. Flakes.

Mr. William King, in an appendix to the above paper, was of opinion that certain of the sites were the seats of manufacture, and with this opinion I agree.

It still remains to account for the extensive spread of the gravel bed described by me over so large an area, and for the fact that many of the celts show traces of grinding and rounding of edges. It must, however, be remembered that the alluvium is very thin and that it is quite possible that if the existing brooks and streams flowed over the bare Talchir rocks and were proportionally larger, enormous quantities of shingle would rapidly form, from the weathering out of the Talchir pebbles. It is easy to understand how some of the implements would be submitted to greater rolling and grinding than others. The variation in this respect, as will seen from the specimens, is very considerable.

The arguments in favour of the site, Hinoutee, having been the seat of a manufactory are so strong as to outweigh any other consideration. The arguments in favour of the site having been a manufactory are:

1st. The presence of the raw material which is identical with that of which the paleoliths are made.

2nd. The presence of recognizable stone hammers in the proportion of 3 per cent.

3rd. The presence of spalls, chips, and flakes.

4th. The fact that specimens in all stages of manufacture occur, and that the great majority are obviously unfinished products.

Neolithic manufactories quite as extensive have been observed by me near Kalnegar, Kalyanpur, &c., and are strewn with chert and agate splinters, used-up stone hammers and broken and unfinished implements.

My conclusion is that the implements lie where they were made, subsequent to their manufacture; and that some 20 feet of alluvium thinly scattered with pebbles from one to two inches diameter was deposited over them by aqueous causes, including possible glacial action.

DISCUSSION.

Mr. C. H. READ observed that the implements found by Mr. Cockburn in Mirzapore formed a very interesting series, although he did not think there was among them any new Indian type. They strongly resemble, as the author observed, those found by Mr. Foote, and appear to be made of the same kind of stone. The great similarity that exists between the implements of the Drift gravels, whether in India or Europe, is a very curious point, and