The cave is in the form of a tunnel a mile or more in length, with innumerable ramifications, in the lava which has flowed from the Bald Yoekul.  It lies on the edge of the uninhabited waste called the Arnavatns-heidi, in a district described by Captain Forbes as distorted and devilish, a cast-iron sea of lava.  The approach is through an open chasm, 20 to 40 feet in depth, and 50 feet broad, leading to the entrance of the cave, where the height is between 30 and 40 feet, and the breadth rather more than 50.  Henderson found a large quantity of congealed snow at this entrance, and along pool of water resting on a floor of ice, which turned his party back and forced them to seek another entrance, where again they found snow piled up to a considerable height.  Olafsen also mentions collections of snow under the various openings in the lava which forms the roof of the cave.  The latter explorer discovered interesting signs of the early inhabitants of the Surtshellir, as, for instance, the common bedstead, built of stones, 2-1/2 feet high, 36 feet long, and 14 feet broad, with a pathway down the middle, forming the only passage to the inner parts of the cave.  The spaces enclosed by these stones were strewn with black sand, on which rough wool was probably laid by way of mattress.  This could scarcely have been a bedstead in the time of the giants, for a total breadth of 14 feet, deducting for the pathway down the middle, will not give more than 6 feet for the layer of men on either side, unless indeed they lay parallel to the passage, and required a length of 36 feet.  He also found an old wall, built with blocks of lava across one part of the cave, as if for defence, and a large circular heap of the bones of sheep and oxen, presumably the remains of many years of feasting.  Captain Forbes scoffs at these bones, and suggests errant wild ponies as the depositors thereof.

Olafsen had found in his earlier visit that the way was stopped, far in the recesses of the cave, by a lake of water, which filled the tunnel to a depth of 3 feet or more, lying on ice; but in 1753 there was not more than a foot of water, through which they waded without much difficulty.  The air soon became exceedingly cold and thick, and for some hundreds of paces they saw no light of day, till at length they reached a welcome opening in the roof.  Beyond this, the air grew colder and more thick, and the walls were found to be sheeted with ice from roof to floor, or covered with broad and connected icicles.  The ground also was a mass of ice, but an inch or two of fine brown earth lay upon it, which enabled them to keep their footing.  This earth appeared to have been brought down by the water which filtered through the roof.  ‘The most wonderful thing,’ Olafsen remarks, ’that we noticed here, was, that the stalactites of ice were set with regular figures of five and seven sides, joined together, and resembling those seen on the second stomach of ruminating animals.  The condensed cold of the air must have imparted these figures to the ice; they were not external (merely?), but in the ice itself, which otherwise was clear and transparent.’