and frequent on those vast savannahs or prairies,* which are scarcely eight or ten toises above the level of the ocean? The earthquakes of Cumana are connected with those of the West India Islands; and it has even been suspected that they have some connection with the volcanic phenomena of the Cordilleras of the Andes. On the 4th of February, 1797, the soil of the province of Quito suffered such a destructive commotion, that near 40,000 natives perished. At the same period the inhabitants of the eastern Antilles were alarmed by shocks, which continued during eight months, when the volcano of Guadaloupe threw out pumicestones, ashes, and gusts of sulphureous vapours. The eruption of the 27th of September, during which very long-continued subterranean noises were heard, was followed on the 14th of December by the great earthquake of Cumana. Another volcano of the West India Islands, that of St. Vincent, affords an example of these extraordinary connections. This volcano had not emitted flames since 1718, when they burst forth anew in 1812. The total ruin of the city of Caracas preceded this explosion thirty-five days, and violent oscillations of the ground were felt both in the islands and on the coasts of Terra Firma. It has long been remarked that the effects of great earthquakes extend much farther than the phenomena arising from burning volcanoes. In studying the physical revolutions of Italy, in carefully examining the series of the eruptions of Vesuvius and Etna, we can scarcely recognise, notwithstanding the proximity of these mountains, any traces of a simultaneous action. It is on the contrary beyond a doubt, that at the period of the last and preceding destruction of Lisbon,† the sea was violently agitated even as far as the *The Llanos of Cumana, of New Barcelona, of Calabozo, of Apure, and of Meta. †The 1st of November, 1755, and 31st of March, 1761. During the first of these earthquakes, the sea inundated, in Europe, the coasts of Sweden, England, and Spain; in America, the islands of Antigua, Barbadoes, and Martinique. At Barbadoes, where the ordinary tides rise only from twenty-four to twenty-eight inches, the water rose twenty feet in Carlisle Bay. It became at the same time as black as ink; being, without doubt, mixed with the petroleum, or asphaltum, which abounds at the bottom of the sea, as well on the coasts of the gulf of Cariaco, as New World, for instance, at the island of Barbadoes, more than twelve hundred leagues distant from the coasts of Portugal. Several facts tend to prove that the causes which produce earthquakes have a near connection with those which act in volcanic eruptions. The connection of these causes was known to the ancients, and it excited fresh attention at the period of the discovery of America. The discovery of the New World not only offered new productions to the curiosity of man, it also extended the then existing stock of knowledge respecting physical geography, the varieties of the human species, and the migrations of nations. It is impossible to read the narratives of early Spanish travellers, especially that of the Jesuit Acosta, without perceiving the influence which the aspect of a great continent, the study of extraordinary appearances of nature, and intercourse with men of different races, must have exercised near the island of Trinidad. In the West Indies, and in several lakes of Switzerland, this extraordinary motion of the waters was observed six hours after the first shock that was felt at Lisbon.-Phil. Trans., vol. xlix, pp. 403, 410, 544, 668; ibid. vol. liii, p. 424. At Cadiz a mountain of water sixty feet high was seen eight miles distant at sea. This mass threw itself impetuously on the coasts, and beat down a great number of houses; like the wave eighty-four feet high, which on the 9th of June, 1586, at the time of the great earthquake of Lima, covered the port of Callao.-Acosta, Hist. Natural de las Indias, ed. de 1591, p. 123. In North America, on Lake Ontario, violent agitations of the water were observed from the month of October 1755. These phenomena are proofs of subterraneous communications at enormous distances. On comparing the periods of the great catastrophes of Lima and Guatimala, which generally succeed each other at long intervals, it has sometimes been thought, that the effect of an action slowly propagating along the Cordilleras, sometimes from north to south, at other times from south to north, may be perceived. Cosmo Bueno, Descripcion del Peru, ed. de Lima, p. 67. Four of these remarkable catastrophes, with their dates, may be here enumerated. MEXICO. (Lat. 13° 32′ north.) 30th of November, 1577. 12th of February, 1689. 27th of September, 1717. PERU. When the shocks are not simultaneous, or do not follow each other at short intervals, great doubts may be entertained with respect to the supposed communication of the movement. on the progress of knowledge in Europe. The germ of a great number of physical truths is found in the works of the sixteenth century; and that germ would have fructified, had it not been crushed by fanaticism and superstition. We learned, at Pasto, that the column of black and thick smoke, which, in 1797, issued for several months from the volcano near that shore, disappeared at the very hour, when, sixty *leagues to the south, the towns of Riobamba, Hambato, and Tacunga were destroyod by an enormous shock. In the interior of a burning crater, near those hillocks formed by ejections of scoriæ and ashes, the motion of the ground is felt several seconds before each partial eruption takes place. We observed this phenomenon at Vesuvius in 1805, while the mountain threw out incandescent scoriæ; we were witnesses of it in 1802, on the brink of the immense crater of Pichincha, from which, nevertheless, at that time, clouds of sulphureous acid vapours only issued. Everything in earthquakes seems to indicate the action of elastic fluids seeking an outlet to diffuse themselves in the atmosphere. Often, on the coasts of the Pacific, the action is almost instantaneously communicated from Chile to the gulf of Guayaquil, a distance of six hundred leagues; and, what is very remarkable, the shocks appear to be the stronger in proportion as the country is distant from burning volcanoes. The granitic mountains of Calabria, covered with very recent breccias, the calcareous chain of the Apennines, the country of Pignerol, the coasts of Portugal and Greece, those of Peru and Terra Firma, afford striking proofs of this fact. The globe, it may be said, is agitated with the greater force, in proportion as the surface has a smaller number of funnels communicating with the caverns of the interior. At Naples and at Messina, at the foot of Cotopaxi and of Tunguragua, earthquakes are dreaded only when vapours and flames do not issue from the craters. In the kingdom of Quito, the great catastrophe of Riobamba led several well-informed persons to think that that country would be less frequently disturbed, if the subterranean fire should break the porphyritic dome of Chimborazo; and if that colossal mountain should become a burning volcano. At all times analogous facts have led to the same hypotheses. The Greeks, who, like ourselves, attributed the oscillations of the ground to the tension of elastic fluids, cited in favour of their opinion, the total cessation of the shocks at the island of Euboea, by the opening of a crevice in the Lelantine plain.* The phenomena of volcanoes, and those of earthquakes, have been considered of late as the effects of voltaic electricity, developed by a particular disposition of heterogeneous strata. It cannot be denied, that often, when violent shocks succeed each other within the space of a few hours, the electricity of the air sensibly increases at the instant the ground is most agitated; but to explain this phenomenon, it is unnecessary to recur to an hypothesis, which is in direct contradiction to everything hitherto observed respecting the structure of our planet, and the disposition of its strata. CHAPTER V. Peninsula of Araya.-Salt-marshes.-Ruins of the Castle of Santiago. THE first weeks of our abode at Cumana were employed in testing our instruments, in herborizing in the neighbouring plains, and in examining the traces of the earthquake of the 14th of December, 1797. Overpowered at once by a great number of objects, we were somewhat embarrassed how to lay down a regular plan of study and observation. Whilst every surrounding object was fitted to inspire in us the most lively interest, our physical and astronomical instruments in their turns excited strongly the curiosity of the inhabitants. We had numerous visitors; and in our desire to satisfy persons who appeared so happy to see the spots of the moon through Dollond's telescope, the absorption of two gases in a eudiometrical tube, or the effects of galvanism on the motions of a frog, we were obliged to answer questions often obscure, and to repeat for whole hours the same experiments. These scenes were renewed for the space of five years, whenever we took up our abode in a place where it was understood * "The shocks ceased only when a crevice, which ejected a river of fiery mud, opened in the plain of Lelantum, near Chalcis."--Strabo. 174 GEOGRAPHICAL ERRORS. that we were in possession of microscopes, telescopes, and electrical apparatus. I could not begin a regular course of astronomical observations before the 28th of July, though it was highly important for me to know the longitude given by Berthoud's time-keeper; but it happened, that in a country where the sky is constantly clear and serene, no stars appeared for several nights. The whole series of the observations I made in 1799 and 1800 give for their results, that the latitude of the great square at Cumana is 10° 27′ 52′′, and its longitude 66° 30′ 2′′. This longitude is founded on the difference of time, on lunar distances, on the eclipse of the sun (on the 28th of October, 1799), and on ten immersions of Jupiter's satellites, compared with observations made in Europe. The oldest chart we have of the continent, that of Don Diego Ribeiro, geographer to the emperor Charles the Fifth, places Cumana in latitude 9° 30'; which differs fifty-eight minutes from the real latitude, and half a degree from that marked by Jefferies in his American Pilot, published in 1794. During three centuries the whole of the coast of Terra Firma has been laid down too far to the south: this has been owing to the current near the island of Trinidad, which sets toward the north, and mariners are led by their dead-reckoning to think themselves farther south than they really are. On the 17th of August a halo round the moon fixed the attention of the inhabitants of Cumana, who considered it as the presage of some violent earthquake; for, according to popular notions, all extraordinary phenomena are immediately connected with each other. Coloured circles around the moon are much more rare in northern countries, than in Provence, Italy, and Spain. They are seen particularly (and this fact is singular enough) when the sky is clear, and the weather seems to be most fair and settled. Under the torrid zone beautiful prismatic colours appear almost every night, and even at the time of the greatest droughts; often in the space of a few minutes they disappear several times, because, doubtless, the superior currents change the state of the floating vapours, by which the light is refracted. I sometimes even observed, between the fifteenth degree of latitude and the equator, small halos |