Open main menu

Popular Science Monthly/Volume 61/July 1902/The Antillean Volcanoes

By. Dr. W J McGEE,


IN all ages volcanoes have played a prominent role in human thought The Vulcan of classic mythology was but the head of a family of earth-gods born of the polytechnic Mediterranean mind fertilized by the "burning mountains" of continents conjoined in the Levant; and in the still lower stages of human development represented by scores of surviving tribes, Fire-Earth deities head the primitive pantheons—indeed, the Vulcanean notion seems to run back to a pristine stage in which the forerunners of living races first stole Vulcan's torch, tamed capricious and ferocious fire even as other [to them] beasts were tamed, and thus took the initial step in that nature-conquest by which man rose above lower life. Certain it is that Vulcanean myths are most dominant in lower savagery, feebler albeit sharper-cut about the birth-time of writing, and decadent during the period of written history. Naturally a factor in the eclipse of Vulcan was the dispersion of mankind, largely in accordance with preconceived plans, from the volcanic centers in which fire was first enslaved over volcano-free regions in which the new servant was thralled by new devices; and it is significant that not only the beast-gods of the prime but the later nature-deities, like Jove and Pluto, Thor and Odin, were swept away before the tide of self-confidence raised by nature-conquest. So Vulcan and the rest lost their terrors; they even fell into oblivion like the beast-gods before, save as the chaff of concepts caught in the meshes of scripture. Most of the mythic monsters are gone utterly; yet Vulcan dies hard—and now and then the nature power for which he stands rises above all human might, and tempts men to return to that early stage of thought marked by the personification of powers.

The latest Vulcanean throes have caught the attention of the reading world. Measured by volume of material cast out, or by force of explosions, the recent Antillean outbreaks rank below many others on record—far below the stupendous outbursts of later geologic periods; yet measured by mortality, the eruption of Mont Pelee on the morning of May 8, 1902, ranks among the most appalling catastrophes of history. And never before was news of disaster so quickly spread; quick thinkers jotted the details, and cables and swift ships carried them to every country within a few hours—yet not so speedily but that history's brightest example of practical sympathy overtook the echoes of calamity. The prompt charity was not emotional merely, but a material outpouring of national substance; and it was no less rational, as attested by the presence on the relief ships of a corps of scientific students whose aim was to dispense knowledge with food and apparel, and to acquire better knowledge against future emergencies. Measured merely by mortality, Mont Pelee marks one of the darkest chapters in human history; measured by the upwelling of human sympathy, it stands for one of the brightest chapters; and measured by the prompt and effective inauguration of research, it must be said to open a new chapter in vulcanology.


The destruction of St. Pierre on the morning of May 18 was one of a series of episodes closely connected in time if not in cause. The chain began with eruptions from the long quiescent crater of Colima (western Mexico) late last year; then followed a series of earthquakes in Mexico and Central America, culminating in the shocks which wrecked Guatemala City and ruined Quetzaltenango on April 18; this shock was apparently followed immediately by steam spits from the crater of Mont Pelee (the culminating mountain mass of Martinique), which had been quiet since 1851, and also from La Souffriere, the commanding crater of the island of St. Vincent, which had rested since 1812. The steam puffs grew in magnitude, and before the end of April there were several explosions, accompanied by rumblings and tremblings of the earth, in which jets of mud were shot into the air and swept far a-sea by the trade-winds; while the warm springs and solfataras (or souffrieres) on Martinique and other islands displayed unwonted activity. The manifestations increased during the early days of May, and Professor Landes, of the Lycée de St. Pierre, detected gases of subterranean origin in the air and erupted material, and sent a note of danger to the Colonial Governor of Martinique. Most unhappily lulled by the memory of the innocent eruption of 1851, the official enjoined secrecy, and appointed a commission of inquiry, whose report did much to allay apprehension and keep St. Pierre crowded for the approaching holocaust. On May 5 there was a destructive eruption from Mont Pelee; clouds of rock-powder were blown high in the air, the rivers overflowed with scalding floods and a stream of hot mud engulfed a sugar factory—L'Usée Guerin—with a score of attachés. Still the Governor forbade the evacuation of St. Pierre; and when, on May 7, the great Souffriere on St. Vincent exploded with a violence exceeding all record, cables were broken and other means of communication interrupted, so that St. Pierre continued to await her doom with hardly ruffled composure. On the morning of May 8, Mont Pelee burst at top and side with a terrific detonation; the earth trembled, and for a few moments the people of the city and the sailors on the shipping in the roadstead began, to seek safety; then a gale of burning gas and a rain of red-hot rock fell, and all St. Pierre, with fifteen of the seventeen vessels at anchor, were destroyed in an instant. Neighboring villages were blotted out; all Martinique was shaken; the detonation rolled out over the waves two hundred miles in every direction; the column of steam and rock-powder rose miles in height, and the dust was scattered over more than a hundred thousand square miles of isle-dotted ocean. This outbreak was not the end, nor even the culmination; La Souffriere continued to belch steam and mud, and Mont Pelee to erupt daily if not hourly, the explosion of May 20 exceeding in violence those of all earlier dates; and the magnetic disturbances accompanying the severer shocks were recorded in Maryland and Kansas, in Paris, and according to one report in Honolulu. The long quiescent crater of Tacana (Guatemala) was stirred into explosive activity; the warm springs of New Mexico resumed the long-past geyser form, and the crater near Grant, dead for five centuries, steamed anew; Mount Redoubt and neighboring craters in Washington State resumed alarming activity, and rumors of fresh outbreaks in Alaskan volcanoes gained currency; and the latest reports indicate that Kilauea, in Hawaii, has joined the concert. These are but the best-known links in the chain of sequence; it may not yet be affirmed that this succession is more than one of time, yet it is significant that the localities swept by the chain are all within reach of the magnetic disturbances extending for thousands of miles from Mont Pelee.

The eruption of Colima cost a number of lives not yet counted; the mortality at Quetzaltenango and neighboring places is estimated at 2,000 or more; the outbreak of Tacana is reported to have cost over a thousand lives; La Souffriere slew some two thousand; St. Pierre lost all of her 25,000, and other villages on Martinique some seven thousand more; so that Vulcan's victims in a single province during April and May, 1902, must approach 40,000. The number will never be accurately known; for there are no censuses covering certain Central American districts and some others in which the fatalities were numerous.


The lesson of Mont Pelee and St. Pierre is especially instructive. The researches of Hill, Russell, Heilprin, Jaggar, Borchgrevink, Hovey and Kennan have already made fairly clear the external aspects of the great eruption. Northern Martinique, like other West Indian islands, is a labyrinth of mornes and pitons, i. e., of singularly steep peaks and ridges (partly volcanic cones, partly erosional forms), densely clothed with forests and herbage; it culminates in the crater rim of Mont Pelee a little less than 5,000 feet in altitude, with a sinuous divide extending southward and minor aretes stretching seaward between the gorges of radiating waterways. In general, the mass of northern Martinique bulges upward and outward, a great dome convex toward the sky; but west and southwest of the summit crater there is an irregularly triangular amphitheater, of a form suggesting that a segment of the mountain has dropped several hundred feet below the general contour. At the coast line this depression forms the shallow bay of St. Pierre roadstead; and thence to the crest the depressed segment is bounded by rocky walls—sharp divides, cliffs, and lines of mornes, in different places. Almost in the center of this triangular trough sloping down from Pelee crater to the sea stood the city of St. Pierre, the metropolis of Martinique—that gem of the Antilles which must always live in history as the birthplace of Josephine, the early home of Bernardin de St. Pierre and the real scene (through youthful associations) of the epic of Paul and Virginia. St. Pierre was, indeed, an Antillean metropolis, the abode of culture and refinement, a mart of trade and shipping, the site of educational institutions of no mean grade, a city whose strong and distinctive characters were well worthy the gifted pen of its best chronicler, Lafcadio Hearn. Up the steep slopes toward the idly gaping crater four miles away, ran well beaten footpaths (for horses and wheels are alien to the precipitous slopes of the Antilles) leading to plantations and on to suburban villages, Carbet on the south, Precheur on the northern boundary of the amphitheater, Morne Rouge on the divide stretching away from the crater. The colonial Jardin Botanique lay in the rear, some hundred feet above the city; luxuriant cane-fields covered every available spot, groves and rows of palms skirted streets and pathways, legion brooks carried living water from the upper slopes to the sea, and thick native verdure mantled all the surface save fields and paths. Some half-way down the trough from Pelee to St. Pierre stood a minor crater, with traces of fumaroles; but they were covered from sight and memory by the prevailing verdure. A picturesque pond lay at the bottom of the great crater; and much of the water flowing seaward from the verdure-clad hills of the trough gathered into a central stream, La Riviere Blanche. Such was the area of destruction.

It is probable that the explosion of May 5 vaporized the water of the crater-set pond, and blew it into the air; it is also probable that the minor crater half-way down the slope toward St. Pierre was at least partially opened. More certain it is that on the morning of May 8 a mass of molten rock in the throat of the great crater exploded by the flashing into gas of the water and other volatile substances approaching the surface and so escaping subterranean pressure; and that immediately afterward (probably timed by the disturbed air pressure due to the first discharge) the minor crater fired a smaller blast. In any event, steam, smoke, rock-dust, pumice bombs, and great blobs of half-liquid rock shot skyward from the main crater, with an earth-tremor of startling but not wrecking severity—and this discharge, extending miles into the upper air, must have initiated a series of atmospheric pressure waves; and about the time the heavier dust and bombs began to fall, or just in time to meet the recurrent wave of atmospheric pressure, the discharge of steam and other gases from the minor crater occurred. Some at least of these gases were heavier than air, and formed a black cloud which rolled down the amphitheater toward St. Pierre; according to several witnesses cross-examined by Hill, it was dense and black in front, aflame in the rear; and under the shock of the recurrent air-waves above, it was driven down on St. Pierre with such velocity that roofs flew before it like chaff, the lighthouse tower was twisted and rent into debris, heavy cannon were lifted from their carriages, and a 7-ton metal monument was blown forty yards; every vessel at anchor was careened and most of them capsized, anchor-chains were broken, and an off-shore wave was driven out of the roadstead to return in a destructive debacle. Disturbed by the initial quake, the people of the city fled to the cathedral or local shrines, sought refuge in fancied strongholds, or ran aimlessly about; when caught by the black cyclone, they were thrown against walls and amidst wreckage, bruised and burned by the red-hot rocks pouring from above, and suffocated by sulphurous fumes; then, when the gas-cloud caught fire from its own lightning or from molten rock, every living thing was scorched, seared or baked according to the local conditions of the burning. Such is the picture painted by Hill from the testimony of survivors of the Roddam and the Roraima, and of the parish priest and others who looked down on the holocaust from the cliffs flanking the St. Pierre amphitheater; Russell ascribes less effect to burning gas and more to scorching rock-powder; Borchgrevink emphasizes the evidence of electro-magnetic disturbance; but all agree that the scourge of St. Pierre was fire rather than earthquake or Pompeiian burial. The tragedy was not absolutely instantaneous; yet within three to ten minutes, the thirty thousand of St. Pierre and environs—Professor Landes, the prophetic scientist, and the misguided Governor among the rest—were no more. Thousands of bodies cumbered the debris-strewn streets or lay in the shattered houses until May 20, when Pelee again thundered—and then buried the reeking wreckage beneath a fresh layer of rock-powder. The later eruptions, like the initial one, usually combined an explosion from the main crater with an immediately subsequent one from the subordinate crater; and they sent out clouds whose movements helped to interpret those of earlier date. Thus, on May 23 Hill was able to study from below a gas-cloud like that which fell on St. Pierre; and he was even able (after realizing the futility of attempting escape) to photograph the swirling tongue of flame by which the cloud was rent, and this picture must some day tell whether the fire was lightning or the combustion of inflammable gas.

Other external features of the Vulcanean throes, especially in Mont Pelee, were the detonations heard 200 miles away (those of Krakatoa were audible over 1,000 miles in all directions, almost 3,000 in one); the seismic tremors felt throughout most of the Lesser Antilles; the magnetic shock noted by suitably equipped laboratories from Paris to Honolulu, or a full third of the way about the globe; and especially the solid and gaseous ejectamenta discharged skyward and distributed hundreds of miles in every direction. At Fort de France it was estimated that the greater discharges rose six miles from the crater, seven miles above the level of the sea; by triangulation Bernadou determined the height of the minor discharge of May 30 at 15,500 feet, or something less than three miles (the mean estimate of the height of the Krakatoan discharge in 1883 was 17 miles). The cloud of steam and other gases, with their burden of rock-powder, spread in typical mushroom shape with such rapidity that neighboring islands up to a hundred miles away were darkened, and the dust-rain began to fall within three hours; on Barbados, 100 miles from La Souffriere and 125 from Mont Pelee, the dust-rain reached a depth of a quarter of an inch, and brought down sulphurous gases. The early analyses indicate that the greater part of this material is a crystalline hypersthene-andesite, i. e., the heavier portion of a rather acidic lava of which the more glassy portions are thought by Teall Ho have been vanned away and deposited elsewhere' (Nature, Vol. 66, p. 130); while according to Diller and Steiger some of the grains collected 275 miles southeast of St. Vincent have the astonishingly high specific gravity of 3.3, and the insoluble dust contained.11 per cent, of sulphur, while the dust collected on Barbados was notable for the abundance of magnetite (Science, Vol. XV., pp. 947-950). Nearer the crater of Mont Pelee, the dust—'ashes' of the press reports, 'lapilli' of the books—lies like snow in drifts and sheets sometimes several feet in depth, and is mingled with fragments of pumice or bombs of denser rock, perhaps torn from the throat of the crater. Naturally the rock-rain was cool at Barbados and other remote stations; on St. Vincent, after the outbreak of La Souffriere, and on Martinique under each eruption of Mont Pelee, the falling rock was warm, even hot; an officer of one of the vessels wrecked in the roadstead of St. Pierre escaped the first shock only to be smitten down by a falling mass of half-molten rock; Russell found indications that some of the falling dust was hot enough to scorch the skin of victims but not to fire cotton garments; and a correspondent of Nature reiterates the incredible report that the sand falling on St. Pierre, miles from the nearest crater, was still 'white hot.' All accounts agree as to the immensity and blackness of the clouds cast out from Mont Pelee with each explosion; and all agree in indicating that an important constituent of these clouds was gas, at least in part heavier than air, and at least in part inflammable. These eruptions are especially notable for the extravasation of material in gaseous form; but the gases have not yet been measured or even identified with any approach to precision. Thus far no quantitative estimates have been made of the aggregate amount of matter erupted from either Antillean volcano; but it seems probable that the total from both will not exceed one or two cubic miles, i. e., probably less than a third of that thrown out by Krakatoa alone in the memorable outbreak of 1883. No decisive indications of subsidence of the coasts or of deformation of either insular masses or sea-bottom, such as might be expected to accompany the transfer of so vast a mass of material, have yet been detected—indeed, the geographic effects of the eruptions seem to be inconsiderable. Nor were there notable tidal waves anywhere in the Antillean region, save the outflows and subsequent inrushes in the St. Pierre harbor, ascribed by Hill to the atmospheric disturbance; and even these air-waves were of but limited extent, as indicated by the absence of records at meteorologic stations more remote than that on St. Kitt's, some 200 miles north of Martinique.

The most impressive part of Pelee's lesson is the tale of terrible mortality due to the ill-chosen site of St. Pierre. The convex slopes of the great dome stretching northward and eastward from the crater are still clad in verdure; Morne Rouge, the high-lying suburb on the principal salient stretching out from the crater, suffered nothing more serious than startling tremors and disagreeable dust-showers; it was only in the topographic funnel leading from the crater to the indented roadstead that the destruction was complete. Looking back over her history, it is easy to see that St. Pierre was founded with no more foresight than that of the spider spinning her web across a frequented path; the sacrifice of the city was but the necessary price of shortsight; yet if future dwellers on the Antilles, and the folk of other volcanoridden regions, but profit by the experience of St. Pierre, the sacrifice may not be wholly vain.

So far as indicated by external manifestations, the internal mechanism of the Antillean volcanoes was in no way unprecedented or even peculiar, save, perhaps, in the high ratio of gaseous ejectamenta and the vast extent of magnetic disturbance and even these features may not be new, but only the outcome of more refined observations than those of earlier generations.

The internal mechanism of Mont Pelee and La Souffriere is fairly clear in the light of what may be called the natural history of vulcanisni; for, just as the life history of organic orders and genera is traced only by aid of fossils, so the ontogeny of the volcano may be viewed in the light of the phylogeny traced through its fossil remains—lava sheets, tuff beds, laccolites, volcanic necks like those of the Mount Taylor plateau, and other products of organic action during the ages past.

Naturally the first question concerning the volcano relates to its character; and this is answered partly by such dynamic facts as those recently observed in the Antilles, partly by the static records of the rocks. In the light of the various phenomena, it is convenient to recognize three types of eruption, viz: (1) the Stromboli type, or that of quiet outflow of highly fluent lava; (2) the type of Vesuvius, or that of explosive eruption usually followed by quieter flows of lava; and (3) the Krakatoa type, represented by violent explosions with little, if any, extravasation of lava. In some measure the types intergrade, the middle one, indeed, approaching the extremes; yet they are so connected with the character of the erupted material and other factors as to demand recognition.

The second inquiry concerning the world's volcanoes relates to geographic distribution; and this is well answered by any convenient map, such as that of Bonney (reproduced in the June number of this journal, page 187), from which it appears that nearly all of the living and recently extinct craters are arranged in lines, or zones, coinciding approximately with continental boundaries. Two of the most striking volcanic belts of the globe are those following the chain of Aleutian islands, and that of the Lesser Antilles from Porto Rico southward to the mouth of Orinoco river; several of the world's largest volcanoes occur in the interlacing zones lying off southeastern Asia; and the world's longest belt begins with the Aleutian chain, follows the coast-wise mountains of western North America, traverses Central America, takes in the great Andean volcanoes of western South America, and stretches thence to Terra del Fuego, if not across to Antarctica to end with Mounts Erebus and Terror. The volcanic belts of the globe are sometimes styled 'lines of weakness' in the earth-crust, though a sufficient number of live or recently dead craters lie apart on oceanic islands or in continental interiors to caution conservative geologists against too simple groupings; yet all the facts seem to fall into the generalization that volcanic regions coincide with zones of exceptional activity in continent-making agencies.

A third inquiry concerning volcanoes relates to their geologic distribution, or—in ultimate analysis—to their connection with other geologic agencies and processes. The observations of numberless geologists in the different countries of the globe seem to answer this inquiry in general terms, by indicating that the agency of vulcanism is decadent—that it culminated before the beginning of that definite world—growth recorded in the stratified rocks, revived locally during various periods down to the later Tertiary, and is probably less vigorous to-day than during any earlier eon of geologic time. Going further into detail, Powell has defined what may be called the normal sequence of vulcanism in any particular geologic province: The first stage in this sequence is that of loading, or accumulation of sediments in areas of deposition; the second is that of baking, compression, and metamorphosis of the lower sediments by the rise of the isogeotherms (measuring proper terrestrial heat); the third stage is that of uplift, partly by reason of the expansion and crumpling consequent on the heating from below; the fourth stage is that of unloading, or degradation by rain and rivers; and the final stage is that of vulcanism, supervening as the degradation proceeds and sometimes continuing until the province is once more submerged.

The distribution of volcanoes, both on the present earth-face and throughout the periods of earth-growth, covers essential phases of the natural history of vulcanism; yet the mechanism of the volcano remains to be traced through specific interpretation of both processes and products, so far as these lie within reach of observation. An epoch-marking step towards the interpretation of volcanic products was made by Baron von Richthofen a third of a century ago, when he recognized a natural system of volcanic rocks; and the goal was attained when Dutton, in a flash of genius, saw 'the double function of density and fusibility' ('Geology of the High Plateaus of Utah' 1880, p. 137) which conditions the extrusion of molten rock-matter. Diller, Iddings, Lawson and others have extended the interpretation; yet the later researches have but established the inference that density, or specific gravity, and fusibility (itself affected by wetness) are leading factors in determining the mechanism of volcanic action. Briefly, it may be said (1) that the seat of normal volcanic action is deep in the earth-crust, so deep that the vertical column of denser sea-bottom rocks is heavier than the longer column of mountain rocks rising thence to the cratered crest; (2) that here the rock-matter is subjected to the enormous pressure of superincumbent miles of rock, yet so highly heated as to become mobile with any relief from pressure; (3) that by reason of the variable strains due to unloading or other causes, some portion of this confined rock-matter is sufficiently relieved from pressure to become mobile, whereupon it seeks the level determined by its density, and forces itself upward through any overlying strata of greater density in channels or vents enlarged by continuous flow; (4) that the molten rock arranges itself in the vent in the order of specific gravity, the lighter above, the heavier (and generally wetter) below; (5) that sometimes the upwelling stream of molten rock reaches a hydrostatic equilibrium and spreads out in laccolites without ever reaching the surface, though normally it forces its way upward through the lightest rocks (and these are those of mountains) well toward the surface; and (6) that as the mass approaches the surface so closely as to find relief from the subterranean pressure, its volatile constituents (chiefly occluded water) flash into gas, usually with explosive violence. Now if the lava column is of exceptionally viscous material, either because exceptionally dry or because exceptionally acidic in composition, the explosively expanding steam and other gases inflate it into pumice, or even blow it into dust; while if exceptionally fluent, by reason either of wetness or of basic composition, the explosion is less violent, the steam bubbles out as from boiling liquid, and the lava flows over the crater-rim, or through some chasm rent by its own enormous weight, in streams extending perhaps for many miles—as in Kilauea in 1841, and in the New Mexican volcano near Grant shortly before the Columbian discovery. Commonly it happens that the lighter lavas first extruded are the more viscous, the later and heavier material more fluent; so that the initial manifestation is commonly more decidedly explosive, the action then running down to relatively quiescent outflow; and their other relations depending on composition of the lava, etc., too complex for ready summing.


In the light of the natural history of vulcanism, and of the mechanism traced through the phylogeny of the volcano, the Antillean eruptions may readily be placed in the general scheme of knowledge. Both Mont Pelee and La Souffriere lie in a volcanic province in which the activity culminated ages ago, so that their activity may be likened to the dying throes of a Vulcanean giant; and this fact, while by no means to be interpreted in definite prophecy, is one of some promise to future generations. Again, both volcanoes approach the Krakatoan character rather than the innocent type of Stromboli; this character is destructive in itself; moreover, in view of the normal passage from initial explosion to final outwelling of quiet lava streams, it is to be regarded as an indication either (1) that the crisis of the spasm is not yet passed, or (2) that the andesitic lavas thus far outcast are precursors of more completely differentiated matter to be erupted during coming millenniums. In either case the outlook is less roseate than the humanitarian student would wish; for the fact that the region is one in which vulcanism is decadent when it is measured by geologic ages is of far less immediate interest than the prospect measured either in days of the single vulcanean spasm, or in millenniums of the life history of particular vents.