(Ovid, _Metam._ lib. 15-340,)

are the memorable words which are put into the mouth of Pythagoras by the Roman poet, and they are followed by speculations as to the cause of volcanic vents shifting their positions. Whatever doubts the philosopher expresses as to the nature of these causes, it is a.s.sumed, as incontrovertible, that the points of eruption will hereafter vary, _because they have formerly done so_; a principle of reasoning which, as I have endeavored to show in former chapters, has been too much set at naught by some of the earlier schools of geology, which refused to conclude that great revolutions in the earth's surface are now in progress, or that they will take place hereafter, _because_ they have often been repeated in former ages.

_Division of the subject._--Volcanic action may be defined to be "the influence exerted by the heated interior of the earth on its external covering." If we adopt this definition, without connecting it, as Humboldt has done, with the theory of secular refrigeration, or the cooling down of an original heated and fluid nucleus, we may then cla.s.s under a general head all the subterranean phenomena, whether of volcanoes, or earthquakes, and those insensible movements of the land, by which, as will afterwards appear, large districts may be depressed or elevated, without convulsions. According to this view, I shall consider first, the volcano; secondly, the earthquake; thirdly, the rising or sinking of land in countries where there are no volcanoes or earthquakes; fourthly, the probable _causes_ of the changes which result from subterranean agency.

It is a very general opinion that earthquakes and volcanoes have a common origin; for both are confined to certain regions, although the subterranean movements are least violent in the immediate proximity of volcanic vents, especially where the discharge of aeriform fluids and melted rock is made constantly from the same crater. But as there are particular regions, to which both the points of eruption and the movements of great earthquakes are confined, I shall begin by tracing out the geographical boundaries of some of these, that the reader may be aware of the magnificent scale on which the agency of subterranean fire is now simultaneously developed. Over the whole of the vast tracts alluded to, active volcanic vents are distributed at intervals, and most commonly arranged in a linear direction. Throughout the intermediate s.p.a.ces there is often abundant evidence that the subterranean fire is at work continuously, for the ground is convulsed from time to time by earthquakes; gaseous vapors, especially carbonic acid gas, are disengaged plentifully from the soil; springs often issue at a very high temperature, and their waters are usually impregnated with the same mineral matters as are discharged by volcanoes during eruptions.

VOLCANIC REGIONS.

_Region of the Andes._--Of these great regions, that of the Andes of South America is one of the best defined, extending from the southward of Chili to the northward of Quito, from about lat. 43 S. to about 2 N. of the equator. In this range, however, comprehending forty-five degrees of lat.i.tude, there is an alternation on a grand scale of districts of active with those of extinct volcanoes, or which, if not spent, have at least been dormant for the last three centuries. How long an interval of rest may ent.i.tle us to consider a volcano as entirely extinct is not easily determined; but we know that in Ischia there intervened between two consecutive eruptions a pause of seventeen centuries; and the discovery of America is an event of far too recent a date to allow us even to conjecture whether different portions of the Andes, nearly the whole of which are subject to earthquakes, may not experience alternately a cessation and renewal of eruptions.

The first line of active vents which have been seen in eruption in the Andes extends from lat. 43 28' S.; or, from Yantales, opposite the isle of Chiloe, to Coquimbo, in lat. 30 S.; to these thirteen degrees of lat.i.tude succeed more than eight degrees in which no recent volcanic eruptions have been observed. We then come to the volcanoes of Bolivia and Peru, reaching six degrees from S. to N., or from lat. 21 S. to lat. 15 S. Between the Peruvian volcanoes and those of Quito, another s.p.a.ce intervenes of no less than fourteen degrees of lat.i.tude, said to be free from volcanic action so far as yet known. The volcanoes of Quito then succeed, beginning about 100 geographical miles south of the equator, and continuing for about 130 miles north of the line, when there occurs another undisturbed interval of more than six degrees of lat.i.tude, after which we arrive at the volcanoes of Guatemala or Central America, north of the Isthmus of Panama.[469]

Having thus traced out the line from south to north, I may first state, in regard to the numerous vents of Chili, that the volcanoes of Yantales and Osorno were in eruption during the great earthquake of 1835, at the same moment that the land was shaken in Chiloe, and in some parts of the Chilian coast permanently upheaved; whilst at Juan Fernandez, at the distance of no less than 720 geographical miles from Yantales, an eruption took place beneath the sea. Some of the volcanoes of Chili are of great height, as that of Antuco, in lat. 37 40' S., the summit of which is at least 16,000 feet above the sea. From the flanks of this volcano, at a great height, immense currents of lava have issued, one of which flowed in the year 1828. This event is said to be an exception in the general rule; few volcanoes in the Andes, and none of those in Quito, having been seen in modern times to pour out lava, but having merely ejected vapor or scoriae.

Both the basaltic (or augitic) lavas, and those of the felspathic cla.s.s, occur in Chili and other parts of the Andes; but the volcanic rocks of the felspathic family are said by Von Buch to be generally not trachyte, but a rock which has been called andesite, or a mixture of augite and albite. The last-mentioned mineral contains soda instead of the potash found in common felspar.

The volcano of Rancagua, lat. 34 15' S., is said to be always throwing out ashes and vapors like Stromboli, a proof of the permanently heated state of certain parts of the interior of the earth below. A year rarely pa.s.ses in Chili without some slight shocks of earthquakes, and in certain districts not a month. Those shocks which come from the side of the ocean are the most violent, and the same is said to be the case in Peru. The town of Copiapo was laid waste by this terrible scourge in the years 1773, 1796, and 1819, or in both cases after regular intervals of twenty-three years. There have, however, been other shocks in that country in the periods intervening between the dates above mentioned, although probably all less severe, at least on the exact site of Copiapo. The evidence against a regular recurrence of volcanic convulsions at stated periods is so strong as a general fact, that we must be on our guard against attaching too much importance to a few striking but probably accidental coincidences. Among these last might be adduced the case of Lima, violently shaken by an earthquake on the 17th of June, 1578, and again on the very same day, 1678; or the eruptions of Coseguina in the year 1709 and 1809, which are the only two recorded of that volcano previous to that of 1835.[470]

Of the permanent upheaval of land after earthquakes in Chili, I shall have occasion to speak in the next chapter, when it will also be seen that great shocks often coincide with eruptions, either submarine or from the cones of the Andes, showing the ident.i.ty of the force which elevates continents with that which causes volcanic outbursts.[471]

The s.p.a.ce between Chili and Peru, in which no volcanic action has been observed, is 160 nautical leagues from south to north. It is, however, as Von Buch observes, that part of the Andes which is least known, being thinly peopled, and in some parts entirely desert. The volcanoes of Peru rise from a lofty platform to vast heights above the level of the sea, from 17,000 to 20,000 feet. The lava which has issued from Viejo, lat. 16 55' S., accompanied by pumice, is composed of a mixture of crystals of albitic felspar, hornblende, and mica, a rock which has been considered as one of the varieties of andesite. Some tremendous earthquakes which have visited Peru in modern times will be mentioned in a subsequent chapter.

The volcanoes of Quito, occurring between the second degree of south and the third degree of north lat.i.tude, rise to vast elevations above the sea, many of them being between 14,000 and 18,000 feet high. The Indians of Lican have a tradition that the mountain called L'Altar, or Capac Urcu, which means "the chief," was once the highest of those near the equator, being higher than Chimborazo; but in the reign of Ouainia Abomatha, before the discovery of America, a prodigious eruption took place, which lasted eight years, and broke it down. The fragments of trachyte, says M. Boussingault, which once formed the conical summit of this celebrated mountain, are at this day spread over the plain.[472]

Cotopaxi is the most lofty of all the South American volcanoes which have been in a state of activity in modern times, its height being 18,858 feet; and its eruptions have been more frequent and destructive than those of any other mountain. It is a perfect cone, usually covered with an enormous bed of snow, which has, however, been sometimes melted suddenly during an eruption; as in January, 1803, for example, when the snows were dissolved in one night.

Deluges are often caused in the Andes by the liquefaction of great ma.s.ses of snow, and sometimes by the rending open, during earthquakes, of subterranean cavities filled with water. In these inundations fine volcanic sand, loose stones, and other materials which the water meets with in its descent, are swept away, and a vast quant.i.ty of mud, called "moya," is thus formed and carried down into the lower regions. Mud derived from this source descended, in 1797, from the sides of Tunguragua in Quito, and filled valleys a thousand feet wide to the depth of six hundred feet, damming up rivers and causing lakes. In these currents and lakes of moya, thousands of small fish are sometimes enveloped, which, according to Humboldt, have lived and multiplied in subterranean cavities. So great a quant.i.ty of these fish were ejected from the volcano of Imbaburu in 1691, that fevers, which prevailed at the period, were attributed to the effluvia arising from the putrid animal matter.

In Quito, many important revolutions in the physical features of the country are said to have resulted, within the memory of man, from the earthquakes by which it has been convulsed. M. Boussingault declares his belief, that if a full register had been kept of all the convulsions experienced here and in other populous districts of the Andes, it would be found that the trembling of the earth had been incessant. The frequency of the movement, he thinks, is not due to volcanic explosions, but to the continual falling in of ma.s.ses of rock which have been fractured and upheaved in a solid form at a comparatively recent epoch; but a longer series of observations would be requisite to confirm this opinion. According to the same author, the height of several mountains of the Andes has diminished in modern times.[473]

The great crest or cordillera of the Andes is depressed at the Isthmus of Panama to a height of about 1000 feet, and at the lowest point of separation between the two seas near the Gulf of San Miguel, to 150 feet. What some geographers regard as a continuation of that chain in Central America lies to the east of a series of volcanoes, many of which are active in the provinces of Pasto, Popayan, and Guatemala. Coseguina, on the south side of the Gulf of Fonseca, was in eruption in January, 1835, and some of its ashes fell at Truxillo, on the sh.o.r.es of the Gulf of Mexico. What is still more remarkable, on the same day, at Kingston, in Jamaica, the same shower of ashes fell, having been carried by an upper counter-current against the regular east wind which was then blowing. Kingston is about 700 miles distant from Coseguina, and these ashes must have been more than four days in the air, having travelled 170 miles a day. Eight leagues to the southward of the crater, the ashes covered the ground to the depth of three yards and a half, destroying the woods and dwellings. Thousands of cattle perished, their bodies being in many instances one ma.s.s of scorched flesh. Deer and other wild animals sought the towns for protection; many birds and quadrupeds were found suffocated in the ashes, and the neighboring streams were strewed with dead fish.[474] Such facts throw light on geological monuments, for in the ashes thrown out at remote periods from the volcanoes of Auvergne, now extinct, we find the bones and skeletons of lost species of quadrupeds.

_Mexico._--The great volcanic chain, after having thus pursued its course for several thousand miles from south to north, sends off a branch in a new direction in Mexico, in the parallel of the city of that name, and is prolonged in a great platform between the eighteenth and twenty-second degrees of north lat.i.tude. Five active volcanoes traverse Mexico from west to east--Tu[x]tla, Orizaba, Popocatepetl, Jorullo, and Colima. Jorullo, which is in the centre of the great platform, is no less than 120 miles from the nearest ocean--an important circ.u.mstance, as showing that the proximity of the sea is not a necessary condition, although certainly a very general characteristic of the position of active volcanoes. The extraordinary eruption of this mountain, in 1759, will be described in the sequel. If the line which connects these five vents be prolonged in a westerly direction, it cuts the volcanic group of islands called the Isles of Revillagigedo.

To the north of Mexico there are said to be three, or according to some, five volcanoes in the peninsula of California; and a volcano is reported to have been in eruption in the N. W. coast of America, near the Colombia river, lat. 45 37' N.

_West Indies._--To return to the Andes of Quito: Von Buch inclines to the belief that if we were better acquainted with the region to the east of the Madalena, and with New Granada and the Caraccas, we might find the volcanic chain of the Andes to be connected with that of the West Indian or Carribee Islands. The truth of this conjecture has almost been set at rest by the eruption, in 1848, of the volcano of Zamba, in New Grenada, at the mouth of the river Madalena.[475]

Of the West Indian islands there are two parallel series: the one to the west, which are all volcanic, and which rise to the height of several thousand feet; the others to the east, for the most part composed of calcareous rocks, and very low. In the former or volcanic series, are Granada, St. Vincent, St. Lucia, Martinique, Dominica, Guadaloupe, Montserrat, Nevis, and St. Eustace. In the calcareous chain are Tobago, Barbadoes, Mariegallante, Grandeterre, Desirade, Antigua, Barbuda, St.

Bartholomew, and St. Martin. The most considerable eruptions in modern times have been those of St. Vincent. Great earthquakes have agitated St. Domingo, as will be seen in the twenty-ninth chapter.

I have before mentioned (p. 270) the violent earthquake which in 1812 convulsed the valley of the Mississippi at New Madrid, for the s.p.a.ce of 300 miles in length, of which more will be said in the twenty-seventh chapter. This happened exactly at the same time as the great earthquake of Caraccas, so that it is possible that these two points are parts of one subterranean volcanic region. The island of Jamaica, with a tract of the contiguous sea, has often experienced tremendous shocks; and these are frequent along a line extending from Jamaica to St. Domingo and Porto Rico.

Thus it will be seen that, without taking account of the West Indian and Mexican branches, a linear train of volcanoes and tracts shaken by earthquakes may be traced from the island of Chiloe and opposite coast to Mexico, or even perhaps to the mouth of the Colombia river--a distance upon the whole as great as from the pole to the equator. In regard to the western limits of the region, they lie deep beneath the waves of the Pacific, and must continue unknown to us. On the east they are not prolonged, except where they include the West Indian Islands, to a great distance; for there seem to be no indications of volcanic disturbances in Buenos Ayres, Brazil, and the United States of North America.

[Ill.u.s.tration: Fig. 89.

MAP OF ACTIVE VOLCANOES AND ATOLLS of THE INDIAN ARCHIPELAGO, and Part of the adjoining PACIFIC OCEAN.]

_Volcanic region from the Aleutian Isles to the Moluccas and Isles of Sunda._--On a scale which equals or surpa.s.ses that of the Andes, is another line of volcanic action, which commences, on the north, with the Aleutian Isles in Russian America, and extends, first in a westerly direction for nearly 200 geographical miles, and then southwards, with few interruptions, throughout a s.p.a.ce of between sixty and seventy degrees of lat.i.tude to the Moluccas, where it sends off a branch to the southeast while the princ.i.p.al train continues westerly through Sumbawa and Java to Sumatra, and then in a northwesterly direction to the Bay of Bengal.[476] This volcanic line, observes Von Buch, may be said to follow throughout its course the external border of the continent of Asia; while the branch which has been alluded to as striking southeast from the Moluccas, pa.s.ses from New Guinea to New Zealand, conforming, though somewhat rudely, to the outline of Australia.[477]

The connection, however, of the New Guinea volcanoes with the line in Java (as laid down in Von Buch's map) is not clearly made out. By consulting Darwin's map of coral reefs and active volcanoes,[478] the reader will see that we might almost with equal propriety include the Mariana and Bonin volcanoes in a band with New Guinea. Or if we allow so much lat.i.tude in framing zones of volcanic action, we must also suppose the New Hebrides, Solomon Isles, and New Ireland to const.i.tute one line (see map, fig. 39, p. 351).

The northern extremity of the volcanic region of Asia, as described by Von Buch, is on the borders of Cook's Inlet, northeast of the Peninsula of Alaska, where one volcano, in about the sixtieth degree of lat.i.tude, is said to be 14,000 feet high. In Alaska itself are cones of vast height, which have been seen in eruption, and which are covered for two-thirds of their height downwards with perpetual snow. The summit of the loftiest peak is truncated, and is said to have fallen in during an eruption in 1786. From Alaska the line is continued through the Aleutian or Fox Islands to Kamtschatka. In the Aleutian Archipelago eruptions are frequent, and about thirty miles to the north of Unalaska, near the Isle of Umnack, a new island was formed in 1796. It was first observed after a storm, at a point in the sea from which a column of smoke had been seen to rise. Flames then issued from the new islet which illuminated the country for ten miles round; a frightful earthquake shook the new-formed cone, and showers of stones were thrown as far as Umnack. The eruption continued for several months, and eight years afterwards, in 1804, when it was explored by some hunters, the soil was so hot in some places that they could not walk on it. According to Langsdorf and others, this new island, which is now several thousand feet high, and two or three miles in circ.u.mference, has been continually found to have increased in size when successively visited by different travellers; but we have no accurate means of determining how much of its growth, if any, has been due to upheaval, or how far it has been exclusively formed by the ejection of ashes and streams of lava. It seems, however, to be well attested that earthquakes of the most terrific description agitate and alter the bed of the sea and surface of the land throughout this tract.

The line is continued in the southern extremity of the Peninsula of Kamtschatka, where there are many active volcanoes, which, in some eruptions, have scattered ashes to immense distances. The largest and most active of these is Klutschew, lat. 56 3' N., which rises at once from the sea to the prodigious height of 15,000 feet. Within 700 feet of the summit, Erman saw, in 1829, a current of lava, emitting a vivid light, flow down the northwest side to the foot of the cone. A flow of lava from the summit of Mont Blanc to its base in the valley of Chamouni would afford but an inadequate idea of the declivity down which this current descended. Large quant.i.ties of ice and snow opposed for a time a barrier to the lava, until at length the fiery torrent overcame, by its heat and pressure, this obstacle, and poured down the mountain side with a frightful noise, which was heard for a distance of more than fifty miles.[479]

The Kurile chain of islands const.i.tutes the prolongation of the Kamtschatka range, where a train of volcanic mountains, nine of which are known to have been in eruption, trends in a southerly direction. The line is then continued to the southwest in the great island of Jesso, and again in Nipon, the princ.i.p.al of the j.a.panese group. It then extends by Loo Choo and Formosa to the Philippine Islands, and thence by Sangir and the northeastern extremity of Celebes to the Moluccas (see map, fig.

39). Afterwards it pa.s.ses westward through Sumbawa to Java.

There are said to be thirty-eight considerable volcanoes in Java, some of which are more than 10,000 feet high. They are remarkable for the quant.i.ty of sulphur and sulphureous vapors which they discharge. They rarely emit lava, but rivers of mud issue from them, like the moya of the Andes of Quito. The memorable eruption of Galongoon, in 1822, will be described in the twenty-fifth chapter. The crater of Taschem, at the eastern extremity of Java, contains a lake strongly impregnated with sulphuric acid, a quarter of a mile long, from which a river of acid water issues, which supports no living creature, nor can fish live in the sea near its confluence. There is an extinct crater near Batur, called Guevo Upas, or the Valley of Poison, about half a mile in circ.u.mference, which is justly an object of terror to the inhabitants of the country. Every living being which penetrates into this valley falls down dead, and the soil is covered with the carca.s.ses of tigers, deer, birds, and even the bones of men; all killed by the abundant emanations of carbonic acid gas, by which the bottom of the valley is filled.

In another crater in this land of wonders, near the volcano of Talaga Bodas, we learn from M. Reinwardt, that the sulphureous exhalations have killed tigers, birds, and innumerable insects; and the soft parts of these animals, such as the fibres, muscles, nails, hair, and skin, are very well preserved, while the bones are corroded, and entirely destroyed.

We learn from observations made in 1844, by Mr. Jukes, that a recent tertiary formation composed of limestone and resembling the coral rock of a fringing reef, clings to the flanks of all the volcanic islands from the east end of Timor to the west end of Java. These modern calcareous strata are often white and chalk-like, sometimes 1000 feet and upwards above the sea, regularly stratified in thick horizontal beds, and they show that there has been a general elevation of these islands at a comparatively modern period.[480]

The same linear arrangement which is observed in Java holds good in the volcanoes of Sumatra, some of which are of great height, as Berapi, which is more than 12,000 feet above the sea, and is continually smoking. Hot springs are abundant at its base. The volcanic line then inclines slightly to the northwest, and points to Barren Island, lat.

12 15' N., in the Bay of Bengal. This volcano was in eruption in 1792, and will be described in the twenty-sixth chapter. The volcanic train then extends, according to Dr. Macclelland, to the island of Narcondam, lat. 13 22' N., which is a cone seven or eight hundred feet high, rising from deep water, and said to present signs of lava currents descending from the crater to the base. Afterwards the train stretches in the same direction to the volcanic island of Ramree, about lat. 19 N., and the adjoining island of Cheduba, which is represented in old charts as a burning mountain. Thus we arrive at the Chittagong coast, which in 1762 was convulsed by a tremendous earthquake (see chap.

29).[481]

To enumerate all the volcanic regions of the Indian and Pacific oceans would lead me far beyond the proper limits of this treatise; but it will appear in the last chapter of this volume, when coral reefs are treated of, that the islands of the Pacific consist alternately of linear groups of two cla.s.ses, the one lofty, and containing active volcanoes, and marine strata above the sea-level, and which have been undergoing upheaval in modern times; the other very low, consisting of reefs of coral, usually with lagoons in their centres, and in which there is evidence of a gradual subsidence of the ground. The extent and direction of these parallel volcanic bands have been depicted with great care by Darwin in his map before cited (p. 351).

The most remarkable theatre of volcanic activity in the Northern Pacific--or, perhaps, in the whole world--occurs in the Sandwich Islands, which have been admirably treated of in a recent work by Mr.

Dana.[482]

_Volcanic region from central Asia to the Azores._--Another great region of subterranean disturbance is that which has been imagined to extend through a large part of Central Asia to the Azores, that is to say, from China and Tartary through Lake Aral and the Caspian to the Caucasus, and the countries bordering the Black Sea, then again through part of Asia Minor to Syria, and westward to the Grecian Islands, Greece, Naples, Sicily, the southern part of Spain, Portugal and the Azores. Respecting the eastern extremity of this line in China, we have little information, but many violent earthquakes are known to have occurred there. The volcano said to have been in eruption in the seventh century in Central Tartary is situated on the northern declivity of the Celestial Mountains, not far distant from the large lake called Issikoul; and Humboldt mentions other vents and solfataras in the same quarter, which are all worthy of notice, as being far more distant from the ocean (260 geographical miles) than any other known points of eruption.

We find on the western sh.o.r.es of the Caspian, in the country round Baku, a tract called the Field of Fire, which continually emits inflammable gas, while springs of naphtha and petroleum occur in the same vicinity, as also mud volcanoes. Syria and Palestine abound in volcanic appearances, and very extensive areas have been shaken, at different periods, with great destruction of cities and loss of lives. Continual mention is made in history of the ravages committed by earthquakes in Sidon, Tyre, Berytus, Laodicea, and Antioch, and in the Island of Cyprus. The country around the Dead Sea appears evidently, from the accounts of modern travellers, to be volcanic. A district near Smyrna, in Asia Minor, was termed by the Greeks Catacecaumene, or "the burnt up," where there is a large arid territory, without trees, and with a cindery soil.[483] This country was visited in 1841 by Mr. W. J.

Hamilton, who found in the valley of the Hermus perfect cones of scoriae, with lava-streams, like those of Auvergne, conforming to the existing river-channels, and with their surface undecomposed.[484]

_Grecian Archipelago._--Proceeding westwards, we reach the Grecian Archipelago, where Santorin, afterwards to be described, is the grand centre of volcanic action.

It was Von Buch's opinion that the volcanoes of Greece were arranged in a line running N. N. W. and S. S. E., and that they afforded the only example in Europe of active volcanoes having a linear direction; but M.

Virlet, on the contrary, announces as the result of his investigations, made during the French expedition to the Morea in 1829, that there is no one determinate line of direction for the volcanic phenomena in Greece, whether we follow the points of eruptions, or the earthquakes, or any other signs of igneous agency.[485]

Macedonia, Thrace, and Epirus, have always been subject to earthquakes, and the Ionian Isles are continually convulsed.

Respecting Southern Italy, Sicily, and the Lipari Isles, it is unnecessary to enlarge here, as I shall have occasion again to allude to them. I may mention, however, that a band of volcanic action has been traced by Dr. Daubeny across the Italian Peninsula, from Ischia to Mount Vultur, in Apulia, the commencement of the line being found in the hot springs of Ischia, after which it is prolonged through Vesuvius to the Lago d'Ansanto, where gases similar to those of Vesuvius are evolved.

Its farther extension strikes Mount Vultur, a lofty cone composed of tuff and lava, from one side of which carbonic acid and sulphuretted hydrogen are emitted.[486]

_Traditions of deluges._--The traditions which have come down to us from remote ages of great inundations said to have happened in Greece and on the confines of the Grecian settlements, had doubtless their origin in a series of local catastrophes, caused princ.i.p.ally by earthquakes. The frequent migrations of the earlier inhabitants, and the total want of written annals long after the first settlement of each country, make it impossible for us at this distance of time to fix either the true localities or probable dates of these events. The first philosophical writers of Greece were, therefore, as much at a loss as ourselves to offer a reasonable conjecture on these points, or to decide how many catastrophes might sometimes have become confounded in one tale, or how much this tale may have been amplified, in after times, or obscured by mythological fiction. The floods of Ogyges and Deucalion are commonly said to have happened before the Trojan war; that of Ogyges more than seventeen, and that of Deucalion more than fifteen centuries before our era. As to the Ogygian flood, it is generally described as having laid waste Attica, and was referred by some writers to a great overflowing of rivers, to which cause Aristotle also attributed the deluge of Deucalion, which, he says, affected h.e.l.las only, or the central part of Thessaly. Others imagined the same event to have been due to an earthquake, which drew down ma.s.ses of rock, and stopped up the course of the Peneus in the narrow defile between mounts Ossa and Olympus.

As to the deluge of Samothrace, which is generally referred to a distinct date, it appears that the sh.o.r.es of that small island and the adjoining mainland of Asia were inundated by the sea. Diodorus Siculus says that the inhabitants had time to take refuge in the mountains, and save themselves by flight; he also relates, that long after the event the fishermen of the island drew up in their nets the capitals of columns, which were the remains of cities submerged by that terrible catastrophe.[487] These statements scarcely leave any doubt that there occurred, at the period alluded to, a subsidence of the coast, accompanied by earthquakes and inroads of the sea. It is not impossible that the story of the bursting of the Black Sea through the Thracian Bosphorus into the Grecian Archipelago, which accompanied, and, as some say, caused the Samothracian deluge, may have reference to a wave, or succession of waves, raised in the Euxine by the same convulsion.

We know that subterranean movements and volcanic eruptions are often attended not only by incursions of the sea, but also by violent rains, and the complete derangement of the river drainage of the inland country, and by the damming up of the outlets of lakes by landslips, or obstructions in the courses of subterranean rivers, such as abound in Thessaly and the Morea. We need not therefore be surprised at the variety of causes a.s.signed for the traditional floods of Greece, by Herodotus, Aristotle, Diodorus, Strabo, and others. As to the area embraced, had all the Grecian deluges occurred simultaneously, instead of being spread over many centuries, and had they, instead of being extremely local, reached at once from the Euxine to the southwestern limit of the Peloponnese, and from Macedonia to Rhodes, the devastation would still have been more limited than that which visited Chili in 1835, when a volcanic eruption broke out in the Andes, opposite Chiloe, and another at Juan Fernandez, distant 720 geographical miles, at the same time that several lofty cones, in the Cordillera, 400 miles to the eastward of that island, threw out vapor and ignited matter. Throughout a great part of the s.p.a.ce thus recently shaken in South America, cities were laid in ruins, or the land was permanently upheaved, or mountainous waves rolled inland from the Pacific.

_Periodical alternation of Earthquakes in Syria and Southern Italy._--It has been remarked by Von Hoff, that from the commencement of the thirteenth to the latter half of the seventeenth century, there was an almost entire cessation of earthquakes in Syria and Judea; and, during this interval of quiescence, the Archipelago, together with part of the adjacent coast of Lesser Asia, as also Southern Italy and Sicily, suffered greatly from earthquakes; while volcanic eruptions were unusually frequent in the same regions. A more extended comparison, also, of the history of the subterranean convulsions of these tracts seems to confirm the opinion, that a violent crisis of commotion never visits both at the same time. It is impossible for us to declare, as yet, whether this phenomenon is constant in this and other regions, because we can rarely trace back a connected series of events farther than a few centuries; but it is well known that, where numerous vents are cl.u.s.tered together within a small area, as in many archipelagoes for instance, two of them are never in violent eruption at once. If the action of one becomes very great for a century or more, the others a.s.sume the appearance of spent volcanoes. It is, therefore, not improbable that separate provinces of the same great range of volcanic fires may hold a relation to one deep-seated focus, a.n.a.logous to that which the apertures of a small group bear to some more superficial rent or cavity. Thus, for example, we may conjecture that, at a comparatively small distance from the surface, Ischia and Vesuvius mutually communicate with certain fissures, and that each affords relief alternately to elastic fluids and lava there generated. So we may suppose Southern Italy and Syria to be connected, at a much greater depth, with a lower part of the very same system of fissures; in which case any obstruction occurring in one duct may have the effect of causing almost all the vapor and melted matter to be forced up the other, and if they cannot get vent, they may be the cause of violent earthquakes. Some objections advanced against this doctrine that "volcanoes act as safety-valves," will be considered in the sequel.[488]