We descended to the terraces flanking the Balasun river, and struck west along jungle-paths to a loosely-timbered flat. A sudden descent of 150 feet landed us on a second terrace. Further on, a third dip of about twenty feet (in some places obliterated) flanks the bed of the Balasun; the river itself being split into many channels at this season. The west bank, which is forty feet high, is of stratified sand and gravel, with vast slightly-worn blocks of gneiss: from the top of this we proceeded south-west for three miles to some Mechi villages, the inhabitants of which flocked to meet us, bringing milk and refreshments. The Lohar-ghur, or "iron hill," lies in a dense dry forest. Its plain-ward flanks are very steep, and covered with scattered weather-worn masses of ochreous and black iron-stone, many of which are several yards long: it fractures with faint metallic lustre, and is very earthy in parts: it does not affect the compass. There are no pebbles of iron-stone, nor water-worn rocks of any kind found with it. The sandstones, close by, cropped out in thick beds (dip north 70 degrees): they are very soft, and beds of laminated clay, and of a slaty rock, are intercalated with them; also an excessively tough conglomerate, formed of an indurated blue or grey paste, with nodules of harder clay. There are no traces of metal in the rock, and the lumps of ore are wholly superficial. Below Punkabaree the Baisarbatti stream cuts through banks of gravel overlying the sandstone (dip north 65 degrees). The sandstone is gritty and micaceous, intercalated with beds of indurated shale and clay; in which I found the shaft (apparently) of a bone; there were also beds of the same clay conglomerate which I had seen at Lohar-ghur, and thin seams of brown lignite; with a rhomboidal cleavage. In the bed of the stream were carbonaceous shales, with obscure impressions of fern leaves, of _Trizygia,_ and _Vertebraria_: both fossils characteristic of the Burdwan coal-fields (see Chapter I), but too imperfect to justify any conclusion as to the relation between these formations.* [These traces of fossils are not sufficient to identify the formation with that of the sewalik hills of North-west India; but its contents, together with its strike, dip, and position relatively to the mountains, and its mineralogical character, incline me to suppose it may be similar. Its appearance in such small quantities in Sikkim (where it rises but a few hundred feet above the level of the sea, whereas in Kumaon it reaches 4000 feet), may be attributed to the greater amount of wearing which it must have undergone; the plains from which it rises being 1000 feet lower than those of Kumaon, and the sea having consequently retired later, exposing the Sikkim sandstone to the effects of denudation for a much longer period. Hitherto no traces of this rock, or of any belonging to a similar geological epoch, have been found in the