snowy mountains of Santa Marta, and the table land of Santa Fe de Bogota, we will proceed to consider their action on the country East of the capital. The commotions were very violent beyond Caurimare, in the valley of Capaya, where they extended as far as the meridian of Cape Codera: but it is extremely remarkable, that they were very feeble on the coasts of NuevaBarcelona, Cumana, and Paria; though these coasts are the continuation of the shore of La Guayra, and formerly known to have been often agitated by subterraneous commotions. Admitting, that the destruction of the four towns of. Caraccas, La Guayra, San Felipe, and Merida, may be attributed to a volcanic focus placed under or near the island of St. Vincent, it may be conceived, that the motion might have been propagated from North-East to North-West * in a line passing through the islands of Los Hermanos, near Blanquilla, without touching the coasts of Araya, Cumana, and Nueva Barcelona. This propagation of the shock might even have taken place, without the intermediate points at the surface of the Globe, the Hermanos Islands for instance, having felt any commotion. This phenomenon is frequently remarked at Peru and Mexico, in earthquakes which have

* Nearly in a line directed South, 64° West..

followed during ages a determinate direction. The inhabitants of the Andes say with simplicity, speaking of an intermediary ground, which is not affected by the general motion, “that it forms a bridge" (que hace puente): as if they meant to indicate by this expression, that the undulations are propagated at an immense depth un er an inert rock.

Fifteen or eighteen hours after the great catastrophe, the ground remained tranquil. The night, as we have already observed, was fine. and calm; and the commotions did not recommence till after the 27th. They were then attended with a very loud and long continued subterranean noise (bramido). The inhabitants of Caraccas wandered into the country; but the villages and farms having suffered as much as the town, they could find no shelter till they were beyond the mountains of Los Teques, in the valleys of Aragua, and in the Llanos or Savannahs. No less than fifteen oscillations were often felt in one day. On the 5th of April there was almost as violent an earthquake, as that which overthrew the capital. During several hours the ground was in a state of perpetual undulation. Large masses of earth fell in the mountains; and enormous rocks were detached from the Silla of Caraccas. It was even asserted, and this opinion prevails still in the country, that the two domes of the Silla sunk

fifty or sixty toises; but this assertion is founded on no measurement whatever. I am informed, that in the province of Quito also, the people, at every period of great commotions, imagine, that the volcano of Tunguragua is diminished in height. It has been affirmed, in many descriptions published of the destruction of Caraccas, "that the mountain of the Silla is an extinguished volcano; that a great quantity of volcanic substances are found on the road from La Guayra to Caraccas*; that the rocks do not present any regular stratification; and that every thing bears the stamp of the action of fire." It is even added, "that, twelve years before the great catastrophe, Mr. Bonpland and myself, from our physical and mineralogical researches, had considered the Silla as a very dangerous neighbour to the city, because that mountain contained a great quantity of sulphur, and that the commotions must come from the North-East." It is seldom that natural philosophers have to justify themselves for an accomplished prediction; but I think it my duty to combat ideas,

See the account given by Mr. Drouet of Guadaloupe, translated in the Trans. of New York, vol. i, p, 308. The author, in giving to the Silla nine hundred toises of absolute height, has confounded the height of the mountain, in my measurement, above the level of the sea, with its height above the valley of Caraccas, which makes a difference of four hundred and sixty toises.

that are too easily adopted on the local causes of earthquakes.

In all those places where the soil has been incessantly agitated for whole months, as at Jamaica in 1693*, Lisbon in 1755, Cumana in 1766, and Piedmont in 1808, a volcano is expected to open. People forget, that it is far from the surface of the Earth we must seek the focus or centre of action; that, according to undeniable evidence, the undulations are propagated almost at the same instant across seas of an immense depth, at a distance of a thousand leagues; and that the greatest commotions take place not at the foot of active volcanoes, but in chains of mountains composed of the most heterogeneous rocks. We have given in the preceding book a geognostical description of the country round Caraccas; we there find gneiss, and mica-slates, containing beds of primitive limestone. The strata are scarcely more fractured or irregularly inclined than near Freyberg in Saxony, or wherever mountains of primitive formation rise abruptly to great heights. I there found neither basaltes nor dolerite, nor even trachytes or trapporphyries; nor in general any trace of an extinguished volcano, unless we choose to consider the diabases or primitive gruenstein, contained in gneiss, as masses of lava, which have filled up

*Phil. Trans. for 1694, p. 99.

fissures.

These diabases are the same as those of Bohemia, Saxony, and Franconia *; and whatever opinion is entertained on the ancient causes of the oxidation of the Globe at its surface, all those primitive mountains, which contain mixtures of hornblende and feldspar, either in veins, or in balls with concentric layers, will not, I suppose, be called volcanic formations. Mont Blanc and Mont d'Or will not be ranged in the same class. The partizans of a universal volcanism, or of the ingenious Huttonian theory, themselves make a distinction between the lavas, which were melted under the simple pressure of the atmosphere at the surface of the Globe, and those layers formed by fire beneath the immense weight of the ocean and superincumbent rocks. They would not confound Auvergne and the granitic valley of Caraccas under the same denomination, that of a country of extinct volcanoes.

I never could have uttered the opinion, that the Silla and the Cerro de Avila, mountains of gneiss and mica-slate, were a dangerous vicinage for the capital, because they contained a great deal of pyrites in subordinate beds of primitive limestone. But I remember having said, during my stay at Caraccas, that the eastern extremity

* These gruensteins are found in Bohemia, near Pilsen, in granite; in Saxony, in the mica-slates of Scheenberg; in Franconia, between Steeben and Lauenstein, in transition slates.