it fills up such hollow places as it occasionally meets with. It is owing to this circumstance, that we find several pools of stagnant water below those places in the channel of the river that had been stopped: but no such are to be seen to the north of this place; and, as the vast quantity of water which was continually rushing down from the ice-mountains could not by any means make itself a sufficient outlet through the lava, the nearest hollows and crevices of which, had been already filled up with the impeded waters, it became necessary that some of it should discharge itself farther on, along the side of the rocky hills, and on the east of the lava; where a considerable broo is now, consequently, visible. This stream forced its passage in many places with great violence, especially to the eastward of the farm-house of Eystridalur, and thence continually ran in a direction parallel with that of the lava, stretching toward the sandy districts to the south. On the western side of the lava there was likewise a brook running past the farm of Thvera, and parallel with the lava. This, also, at length, empties itself into the southern sandy dis

trict; yet it is by no means my opinion, that this stream can have taken its rise solely from the Hverfisfliot. Indeed, the color of its waters sufficiently proves a mixture of the mountain, and of the jökul streams; 'the latter discharging themselves into the plain from the jökul stream of the Hverfisfliot, which was choked up by the lava; whereas the other stream proceeded from various small rivulets and springs, which descended on the western sides of the mountains.

Without much reflection we might be inclined to believe that these waters would at length rejoin, and run along the original channel of the river; but it must be remembered, that the channel itself, was originally not deep at this place, but had been subject to many alterations and shiftings, and likewise that, wherever it stretched out into the sandy plain (being divided into different branches) it carried along with it, and heaped up continually, the loose sand. On the 22nd and 24th of July, 1784, when, in company with Mr. Livetzen and some other persons, I travelled over this

district, we could scarcely observe any traces of the bed of the old river, and it is not a little remarkable that the whole of the vast quantity of water, which had here spread itself over the sandy plain, was still smoking in many places. In some parts so great was the heat, that we could scarcely bear to hold our hands in it; a circumstance that was rendered more particularly unpleasant by our being under the necessity of riding at a foot-pace for three long Danish miles through this hot, and, in general, very deep water. Every where, too, we were enveloped in a thick sulphureous fog and haze, that arose from the surface of it. In all probability the confined water will increase still more beneath the lava; particularly when it becomes thoroughly cooled, and nothing is lost by evaporation. A new channel for the Hverfisfliot will consequently be formed; for the water, now held in an unnatural state, from being as it were dammed up, must force itself a passage, either through the lava, or by breaking down its sides; yet I do not apprehend that any inundation will be the consequence, or that any damage will ensue, except indeed to the two

farm-houses, Selialand on the eastern and Thvera on the western side. Till, however, such time as a new route is effected across the lava itself, between Siden and Fliotshverfet, and also between Thvera and Selialand, it will be very troublesome for persons who may have to travel between these two places; as they will be obliged to pursue a long and tedious course round the whole extent of the lava.

Skaptaa and other

waters.

§ XXIII.

Having now arrived at the Skaptaa, and examined, not merely this great stream, but also the subordinate rivers and brooks, which had been stopped in their course or had wholly lost their waters, I found that this river still continued to flow uninterruptedly from its origin as far as a place below Uxatindur and opposite to Hordubreid, where the torrent of fire that had from the northward broken through Ulfarsdal filled up the channel, as has been before described. That the river has not been impeded in its progress before it reached this place, I conclude from the circumstance of its being here

that we first observed stagnant water, which I consider a sure sign. It was not possible to have a view of the river itself, on account of the lava having, above this spot, accumulated to a considerable height; and because the atmosphere was every where filled with a thick smoke, caused by the dreadful heat, which still existed both here and throughout the whole extent of the upper part of the lava.

A little below this place the volcanic matter becomes still narrower, and there, on its western side, as well as upon the lava itself, we remarked several large pools of stagnant water, collected from the rivers Efri and Sydri-Ofæra, which, as well as the Skaptaa, were here choked up. Some streamlets, indeed, had found a passage along the west of the hraun, but nevertheless in this spot the confined waters became larger, and were more connected with each other.

Towards the east no body of water appeared to have made its way, until we came where the lava had filled up the channel of