The mouth is 90 feet from the ground, and its floor is about 95 feet from the summit of the cliff, [Footnote: Martel (A.), Le Réfuge du Roc d'Aucor, Brive, 1895.] which is crowned by the _oppidurn_ of Murcens, the best preserved of all Gaulish strongholds in France, and was held by the English in 1370. The only possible way to obtain access to the interior would be from above, as the plumb-line let down from the summit fell 44 feet wide from the base of the cliff. Accordingly a rope ladder was attached to a tree on the top, and Armand descended furnished with a plumb-line, the end of which was attached to a cord. "Having descended 77 feet, he swung free in the air at the level of the transverse poles. Then he endeavoured to throw the lead-weight beyond one of the poles. He succeeded only after the seventh or eighth attempt, and was well pleased when the weight running over it swung down to our feet, as the position of the poles and the slope of the floor of the fissure did not allow it to rest in the cavern. 'Pull the cord,' shouted Armand. 'What for?' 'You will soon see. Pull'--and speedily the string drew after it one of our stout ropes. 'Now do you understand?' asked Armand. 'I have fastened my rope ladder to the cord that goes over the pole. Four or five of you pull and draw me in towards that pole, and so we shall get the better of the situation. When I have fixed the ladder to the pole you may all mount by the grand stair.'" By good fortune that beam held firm, and first Armand got into the cave and then the others mounted from below. What made the entrance treacherous was that the floor at the orifice sloped rapidly downwards and outwards. When within, it was seen that the posts were still solid and firmly planted in notches cut in the rock on both sides. In line with them were two rows of similar notches for the reception of beams extending inwards for about twenty feet, as though at one time there had been rafters to divide the cave into two storeys, but of such rafters none remained. The back of the cave was occupied by a gleaming white stalagmitic column that certainly from below bore some resemblance to a human figure, but the floor of the cavern was so deep in birds' nests, and droppings of bats, leaves and branches, that it was not possible at the time to explore it. This, however, was done by M. Martel in 1905, but nothing of archaeological interest was found. However, he noticed a sort of ascending chimney that extended too far to be illumined to its extremity by the magnesium wire, and he conjectured that it extended to the surface of the rock above, where was the original entrance, now choked with earth and stone.