would, however, be desirable to ascertain once for all with certainty what are the extreme limits between which the velocities vary. It is probable that the more violent commotions by no means always possess the greatest velocity. The measurements, moreover, do not always relate to the same direction which the waves of commotion have followed. Exact mathematical determinations are much wanted, and it is only at a very recent period that a result has been obtained with great exactitude and care from the Rhenish earthquake of the 29th of July, 1846, by Julius Schmidt, assistant at the Observatory of Bonn. In the earthquake just mentioned the velocity of propagation was 14,956 geographical miles in a minute, that is 1466 feet in the second. This velocity certainly exceeds that of the waves of sound in the air; but if the propagation of sound in water is at the rate of 5016 feet, as stated by Colladon and Sturm, and in cast iron tubes 11393 feet, according to Biot, the result found for the earthquake appears very weak. For the earthquake of Lisbon on the 1st of November, 1755, Schmidt (working from less accurate data) found the velocity between the coasts of Portugal and Holstein to be more than five times as great as that observed on the Rhine, on the 29th of July, 1846. Thus, for Lisbon and Glückstadt (a distance of 1348 English miles) the velocity obtained was 89.26 miles in a minute or 7953 feet in a second; which, however, is still 3438 feet less than in cast iron."

27

27 Julius Schmidt, in Nöggerath, Ueber das Erdbeben vom 29 Juli, 1846, s. 28-37. With the velocity stated in the text, the earthquake of Lisbon would have passed round the equatorial circumference of the earth in about 45 hours. Michell (Phil. Transact. vol. i, pt. ii, p. 572) found for the same earthquake of the 1st November, 1755, a velocity of only 50 English miles in a minute, that is, instead of 7956, only 4444 feet in a second. The inexactitude of the older observations and difference in the direction of propagation may conduce to this result. Upon the connexion of Neptune with earthquakes, at which I have glanced in the text (p. 181), a passage of Proclus in the commentary to Plato's Cratylus, throws a remarkable light. "The middle one of the three deities, Poseidon, is the cause of movement in all things, even in the immovable. As the originator of movement he is called 'Evvoriyalos; to him, of those who shared the empire of Saturn, fell the middle lot, the easily moved sea" (Creuzer, Symbolik und Mythologie, Th. iii, 1842, s. 260). As the Atlantis of Solon and the Lyctonia, which, according to my idea, was nearly allied to it, are geological myths, both the lands destroyed by earthquakes are re

Concussions of the earth and sudden eruptions of fire from volcanoes which have been long in repose, whether these merely emit cinders, or, like intermittent springs, pour forth fused, fluid earths in streams of lava, have certainly a single, common causal connexion in the high temperature of the interior of our planet; but one of these phenomena is usually manifested quite independently of the other. Thus, in the chain of the Andes in its linear extension, violent earthquakes shake districts in which unextinguished, often indeed active, volcanoes exist, without the latter being perceptibly excited. During the great catastrophe of Riobamba, the volcanoes of Tungurahua and Cotopaxi, the former in the immediate vicinity, and the latter rather further off, remained perfectly quiet. On the other hand, volcanoes have presented violent and long-continued eruptions, without any earthquake being perceived in their vicinity, either previously or simultaneously. In fact, the most destructive earthquakes recorded in history, and which have passed through many thousand square miles, if we may judge from what is observable at the surface, stand in no connexion with the activity of volcanoes. These have lately been called Plutonic, in opposition to the true Volcanic earthquakes, which are usually limited to smaller districts. In respect of the more general views of vulcanicity, this nomenclature is, however, inadmissible. By far the greater part of the earthquakes upon our planet must be called Plutonic.

That which is capable of exciting earth-shocks, is everywhere under our feet; and the consideration that nearly ths of the earth's surface are covered by the sea (with the exception of some scattered islands) and without any permanent communication between the interior and the atmosphere, that is to say, without active volcanoes, contradicts the erroneous, but widely disseminated belief that all earthquakes are to be ascribed to the eruption of some distant volcano. Earthquakes on continents are certainly propagarded as standing under the dominion of Neptune, and set in opposition to the Saturnian continents. According to Herodotus (lib. ii, c. 43 et 50), Neptune was a Libyan deity, and unknown in Egypt. Upon these circumstances-the disappearance of the Libyan lake Tritonis by earthquake-and the idea of the great rarity of earthquakes in the valley of the Nile, see my Examen Critique de la Géographie, t. i, pp. 171 and 179.

gated along the sea-bottom from the shores, and give rise to the terrible sea-waves, of which such memorable examples were furnished by the earthquakes of Lisbon, Callao de Lima, and Chili. When, on the contrary, the earthquakes start from the sea bottom itself, from the realm of Poseidon, the earth-shaker (σεισίχθων, κινησίχθων), and are not accompanied by upheaval of islands (as in the ephemeral existence of the island of Sabrina or Julia), an unusual rolling and swelling of the waves may still be observed at points where the navigator would feel no shock. The inhabitants of the desert Peruvian coasts have often called my attention to a phenomenon of this kind. Even in the harbour of Callao, and near the opposite island of San Lorenzo, I have seen wave upon wave suddenly rising up in the course of a few hours to more than 10 or 15 feet, in perfectly still nights, and in this otherwise so thoroughly peaceful part of the South Sea. That such a phenomenon might have been the consequence of storm which had raged far off upon the open sea, was by no means to be supposed in these latitudes.

To commence from those commotions which are limited to the smallest space, and evidently owe their origin to the activity of a volcano, I may mention in the first place how when sitting at night in the crater of Vesuvius at the foot of a small cone of eruption with my chronometer in my hand, (this was after the great earthquake of Naples on the 26th of July, 1805, and the eruption of lava which took place seventeen days subsequently), I felt a concussion of the soil of the crater very regularly every 20 or 25 seconds, immediately before each eruption of red hot cinders. The cinders, thrown up to a height of 50-60 feet fell back partly into the orifice of eruption, whilst a part of them covered the walls of the cone. The regularity of such a phenomenon renders its observation free from danger. The constantly repeated small earthquake was quite imperceptible beyond the crater,- -even in the Atrio del Cavallo and in the Hermitage del Salvatore. The periodicity of the concussion shows that it was dependent upon a determinate degree of tension which the vapours must attain, to enable them to break through the fused mass in the interior of the cone of cinders. In the case just described no concussions were felt on the declivity of the ashy cone of Vesuvius, and in an

exactly analogous but far grander phenomenon, on the ashcone of the volcano of Sangai, which rises to a height of 17,006 feet to the south-east of the city of Quito, no trembling of the earth28 was felt by a very distinguished observer, M. Wisse, when (in December, 1849,) he approached within a thousand feet of the summit and crater, although no less than 267 explosions (eruptions of cinders) were counted in an hour.

For

A second, and infinitely more important kind of earthquake, is the very frequent one which usually accompanies or precedes great eruptions of volcanoes, whether the volcanoes, like ours in Europe, pour forth streams of lava; or like Cotopaxi, Pichincha, and Tunguragua of the Andes only throw out calcined masses, ashes and vapours. earthquakes of this kind the volcanoes are especially to be regarded as safety valves, as indicated even by Strabo's expression concerning the fissure pouring out lava near Lelante in Euboea. The earthquakes cease, when the great eruption has taken place.

Most widely distributed, however, are the ravages of the waves of commotion which pass sometimes through completely non-trachytic, non-volcanic countries and sometimes through

28 The explosions of the Sangai, or Volcan de Macas, took place on an average every 13".4, see Wisse, Comptes rendus de l'Acad. des Sciences, tome xxxvi, 1853, p. 720. As an example of commotions confined within the narrowest limits, I might also have cited the report of Count Larderel upon the lagoons in Tuscany. The vapours containing boron or boracic acid give notice of their existence and of their approaching eruption at fissures by shaking the surrounding rocks (Larderel, Sur les établissements industriels de la production d'acide boracique en Toscane, 1852, p. 15).

29 I am glad that I am able to cite an important authority in confirmation of the views that I have endeavoured to develope in the text. "In the Andes the oscillation of the soil, due to a volcanic eruption, is, so to speak, local, whilst an earthquake, which, at all events in appearance, is not connected with any volcanic eruption, is propagated to incredible distances. In this case it has been remarked that the shocks followed in preference the direction of the chains of mountains, and were principally felt in Alpine districts. The frequency of the movements in the soil of the Andes, and the little coincidence observed between these movements and volcanic eruptions, must necessarily lead us to suppose that in most cases they are occasioned by a cause independent of volcanoes" (Boussingault, Annales de Chimie et de Physique, t. lviii, 1835, p. 83).

trachytic, volcanic regions, without exerting any influence upon the neighbouring volcanoes. This is a third group of phenomena, and is that which most convincingly indicates the existence of a general cause, lying in the thermic nature of the interior of our planet. To this third group also belongs the phenomenon, sometimes, though rarely, met with in non-volcanic lands, but little disturbed by earthquakes, of a trembling of the soil, within the most narrow limits, continued uninterruptedly for months together, so as to give rise to apprehensions of an elevation and formation of an active volcano. This was the case in the Piedmontese valleys of Pelis and Clusson, as well as in the vicinity of Pignerol in April and May, 1805, and also in the spring of 1829 in Murcia, between Orihuela and the sea-shore, upon a space of scarcely sixteen square miles. When the cultivated surface of Jorullo upon the western declivity of the plateau of Mechoacan in the interior of Mexico was shaken uninterruptedly for 90 days, the volcano rose with many thousand cones of 5-7 feet in height (los hornitos) surrounding it, and poured forth a short but vast stream of lava. In Piedmont and Spain, on the contrary, the concussions of the arth gradually ceased, without the production of any other phenomenon.

I have considered it expedient to enumerate the perfectly distinct kinds of manifestation of the same volcanic activity (the reaction of the interior of the earth upon its surface) in order to guide the observer, and bring together materials which may lead to fruitful results with regard to the causal connexion of the phenomena. Sometimes the volcanic activity embraces at one time or within short periods sc large a portion of the earth, that the commotions of the soil excited may be ascribed simultaneously to many causes related to each other. The years 1796 and 1811 present particularly memorable examples of such a grouping of the phenomena.

30

30 The great phenomena of 1796 and 1797, and 1811 and 1812, occurred in the following order :

27th of September, 1796. Eruption of the volcano of the island of Guadaloupe, in the Leeward Islands, after a repose of many years November, 1796. The volcano on the plateau of Pasto, between the small rivers Guaytara and Juanambu, became ignited and began to smoke permanently;