"All agreed that every shock seemed to come with a rumbling noise from the westward, beginning with the horizontal and ending with the vorticose (rotary) motion." According to Dolomieu, during the Lisbon earthquake, the shocks were preceded "by a loud subterranean noise like thunder, which was renewed for every shock.... This great shock," he says, referring to one of the great upward shocks, "occurred without the prelude of any slight shocks, without any notice whatever as suddenly as the blowing up of a mine.... Some, however, pretend that a muffled interior noise was heard almost at the same moment." The noises do not generally continue long after the earthquake shocks. In some cases, however, a very loud noise is heard at intervals for a considerable length of time after the principal shock. This was the case at Quito and Ibarra, in which a great noise was heard for from eighteen to twenty minutes after the principal shock. In a similar manner during the earthquake of October, 1746, at Lima, and Callao, South America, peals of underground thunder were heard at Truxillo for fifteen minutes after the principal shock. In such cases it seems probable that the noises were not caused by the same impulses that caused the original shock, but by the forces that caused the subsequent shock. Humboldt relates that in 1784 there were noises heard at Guanajuato, from the 9th to the 12th of February. They were not, however, followed by an earthquake. Humboldt also states that in an earthquake which occurred on the 30th of April, 1812, on the banks of the Orinoco River, in South America, a loud thundering noise was heard, without, however, any shock, but at this time a volcano on the island of St. Vincent, in the Lesser Antilles, although some 632 miles to the northeast, was pouring out streams of lava. Again in the great eruption of Cotopaxi, in 1734, underground noises were heard as if cannon were being fired. These sounds were distinctly heard at as great a distance as Honda on the banks of the Magdalena River. Now, bearing in mind that the crater of Cotopaxi is situated on the high plateau of Quito, in a region full of valleys and fissures, it would seem that for the sounds to have been sent through the 436 miles between the mountains and the valley of the Magdalena River, the waves must, for the greater part, have been transmitted through the solid earth at some considerable distance below the surface. Mallet states that the underground noises which continued for more than a month from the midnight of January 9th, 1784, at Guanajuato, were not followed by any earthquake shocks, that it was if as thunder clouds occupied the space below the surface at that part of the earth and from these clouds there came the slow rolling sounds like short, quick, snaps of thunder.