Littell's Living Age/Volume 180/Issue 2332/Krakatoa

Accounts of the great eruption ascribed to Skaptá Jökull in Iceland, in the year 1783, and of subsequent atmospheric appearances, bring before us, with some degree of detail, the most obvious character of those phenomena, but a comparison with the recently issued volumes dealing with the outburst of Krakatoa serves to mark the wide stretches of intellectual territory which the energy of scientific research has, within a century, added to human knowledge. To the poet Cowper the strange aspect of the heavens was "portentous, unexampled, unexplained;" the present generation has learnt that the recent phenomena of 1883 were neither unexampled, except in magnitude, nor portentous, nor, except in a remote sense, unexplained. Both of these tremendous catastrophes occurred within the most active volcanic regions of the earth's crust; both were preceded by manifestations of strong activity, but insufficient to produce alarm, and both arose from places which had long been more or less quiescent and undreaded. The submarine volcano off the south-west cape of Iceland, which had been burning for weeks before the outburst on the mainland in the following June, corresponds with the Strombolian condition of Krakatoa in the months of June and July, 1883, when columns of vapor were rising from two craters on the island, with occasional violent detonations. In the case of Krakatoa, however, the grand explosion of the end of August had been preceded by a considerable eruption on May 20 and three following days. The terrible earthquakes which desolated Calabria in February, 1783, had no parallel in the eastern hemisphere in 1883, though there does appear to be evidence of an unusual prevalence of earthquakes in the neighborhood of Sunda Strait, which might, after the event, be regarded as premonitory of the approaching destruction.

The islands of Java and Iceland have throughout historic times been remarkable for the number and activity of their volcanoes, and for the calamities which have overtaken their inhabitants. The Tenger Mountain in Java, one of the largest volcanoes in the world, measures four and a half by three and a half miles in diameter, and, like a lunar crater, contains volcanic peaks within its arena, a plain covered with shifting sand. In 1772, the volcano Papandayang threw out an immense quantity of scoriæ and ashes in one night, and covered an area of seven miles in diameter with a layer nearly fifty feet thick. But perhaps the most suddenly violent eruption on record was that of Galungoon, a few miles from Papandayang, on October 8, 1822. At noon all was peaceful and quiet in the thriving districts around; soon after midday a dense mass rapidly rose into the air with appalling noise, and in a few minutes the whole landscape was plunged in darkness, pierced only by incessant flashes of lightning. Stones and sand, which had been projected to an enormous height, covered up and destroyed almost everything within a radius of twenty miles. On the twelfth, another eruption of equal intensity followed, a large part of the mountain was broken off, and blocks of basalt were thrown to a distance of seven miles. By such manifestations, and the great number of craters within its area, Java came to be regarded as the chief focus of volcanic activity on the surface of the globe.

In our own hemisphere Iceland has probably no equal in the frequency and violence of its eruptions, and it may well be doubted whether a land so fatally subject to the worst influences of frost and fire should be allowed to retain its present struggling and dwindling population.

The eruption of 1783, above alluded to, is stated in most geological treatises to have belonged to the frozen mountain of Skaptá, but in reality issued from a large number of craters to the south-west, north, and east of Mont Laki.^[2] Immense masses of pumice and lava were thrown out; some of the stones fell at a distance of about seventy miles. The lava streams were more extensive than any single eruption had before produced in historic times, and their volume, according to Lyell, considerably exceeded that of Mont Blanc. Pumice covered the sea for long distances, and ashes fell in the north of Scotland. For months afterwards a thick dry mist spread over Europe. In Italy objects at a distance of three miles could not be distinguished, the sun was invisible near the horizon, and red or pale like the moon during the daytime, and the nights were strangely luminous.

The great eruption of Krakatoa in 1883 entered the violent stage on August 26, producing effects in the neighborhood which must have been quite appalling. The sky presented the most terrible appearance, fierce flashes of lightning penetrating the dense masses of cloud over the island, clouds of black matter were rushing across the sky, rapidly recurring detonations like discharges of artillery, with a crackling noise in the atmosphere, were heard continuously, and large pieces of pumice, quite warm, rained down at a distance of ten miles. At a point seventy-six miles from Krakatoa, the height of the black cloud projected from the volcano was estimated at seventeen miles. At forty miles distance this cloud looked "like an immense wall with bursts of forked lightning at times like large serpents rushing through the air." Balls of fire (corposants) rested on the mastheads and on the extremities of the yard-arms. During the night the intense darkness was relieved by a "peculiar pinky flame" which seemed to come from clouds and touched the ship, chains of fire seemed to be ascending from the volcano to the sky, while balls of fire rolled on its sides, and lightning flashed so far and frequently that the mainmast conductor of the G. G. Loudoun, forty or fifty miles N.W. of the volcano, was struck five or six times. The natives on board were busily engaged in putting out the corposants with their hands, for fear the "evil spirits" would scuttle the ship. At Anjer on the twenty-sixth it was pitch dark early in the afternoon, and as far as one hundred and eighty miles south of Krakatoa ashes were already falling on the evening of that day from a densely overcast sky. The noise of the explosions during the afternoon was heard all over the island of Java, in Borneo, Celebes, New Guinea, Burmah, the Andaman Islands, and Ceylon. Westwards as far as fourteen hundred miles from Java the sky was "all of a flare" at sunset. The force of the explosions accumulated during the night, and they were actually heard at Rodriguez, 2,968 miles distant, and on board a ship about twelve hundred and eighty miles eastwards. This is equal to an explosion on the north coast of Africa being heard far north of the Shetland Islands.

At Batavia, ninety-four miles distant, on the night of the 26th—27th many of the inhabitants did not dare to go to bed and walked on the promenade. In the early morning of the twenty-seventh the noise was "simply deafening," and about 2 a.m., and again at 3 a.m., many of the gas-lamps were extinguished, and panes of glass were broken. About 8.25 a.m. there was a most violent detonation which cracked the walls of houses. After this almost nothing was heard till after 8 p.m., when the noise recommenced and continued to a late hour. This interval of stillness is very remarkable, for at 9.58 a.m. took place that great culminating explosion which wrote its mark on all the chief barometric recorders of the world. At Serang a loud detonation occurred at 10 a.m., and the noise only ceased after the air was filled with ashes; when they cleared off the noise began again. The sound of this explosion does not appear to have been generally noted at distant places as exceeding the reports which had puzzled the inhabitants for many hours previously. Conjectures as to the cause of the strange thunderings were various. At Carimon, Java, 355 miles distant, native boats were despatched to assist an imaginary vessel in distress; at Acheen, 1,073 miles distant, it was supposed that a fort was being attacked, and the troops were put under arms; at Singapore, 512 miles, two steamers were sent to look out for a vessel in distress; at Penang, 868 miles, the sounds were supposed to be a salute from an American corvette; at Elopura, 1,210 miles, it seemed as if heavy guns were being fired at a distance of not more than four or five miles away; at Manila, Luzon, 1,804 miles, preparations were made to render assistance to a supposed vessel in distress; at Timor, 1,351 miles, a government steamer was despatched to ascertain the cause; at Dorey, New Guinea, 2,014 miles, sounds were heard like distant cannonading; at Perth, 1,902 miles, and other places in western Australia, sounds like guns firing were heard; in South Australia, at places over two thousand miles distant, sounds like the blasting of a rock were heard; at many places in Ceylon, nearly two thousand miles, and at the Andaman Islands, sounds as of a vessel in distress or of blasting were frequent; also at Diego Garcia, 2,267 miles, and Rodriguez, 2,968 miles. Never before have sounds been known to reach any distance approaching three thousand or even two thousand miles, and the area covered by audible vibrations probably fairly represents the measure of intensity of the explosions. That area exceeded twice the surface of Europe.

The smaller vibrations in the vast concussions which reverberated above Krakatoa, being the only waves which would sensibly affect the ear, encountered so dense and wide a cloud of pumice, dust, and ashes in the air beneath that they were largely stopped and softened, and the detonations in the neighborhood of the volcano were thus rendered bearable, and, it seems, even less appalling than at greater distances. A similar dulling of sound has been observed during thick snowstorms, and must be partially due to the multiplied reflection and absorption, involved in the passage from gas to solid, and solid to gas, through a heterogeneous mixture. Tyndall, in experiments made off the South Foreland and in the Alps, was unable to discover any enfeeblement of sound during storms of rain, hail, and snow; indeed, the effect of a heavy shower of rain was to increase audibility.^[3] This unexpected result was attributed by him to the condensation of water, which, in the state of vapor so mixed with air as to form non-homogeneous parcels, acted powerfully in wasting sound. Under the action of a strong sun prior to the rain the air had been in this flocculent condition, but the descent of a shower restored in part the homogeneity of the atmosphere and augmented its transmissive power. With regard to fog, a similar effect follows from the condensation of vapor into visible particles. These appear to have "no more influence upon the waves of sound than the suspended particles stirred up over the banks of Newfoundland have upon the waves of the Atlantic." There can be no doubt, however, that, other things being equal, the passage of sound must be more or less stopped by interposed solid or liquid surfaces. A belt of trees such as oaks or pines will give an echo almost as sharply as a wall, and the noise of a train in passing by dense foliage is converted into a rushing sound of higher pitch, through the breaking up of large waves into smaller ones by numberless closely following reflections from the leaves successively encountered. If leaves reflect so much sound from a narrow strip of coppice, the effect of miles of air packed with scoriæ fine and coarse must be considerable; moreover, the variations in temperature and humidity would be quite abnormally great where hot ashes were raining through the cold upper strata, and the whole air near the volcano was in violent commotion. These variations probably constituted the chief obstacle to the propagation of sound downwards near the volcano. As a matter of fact the explosions were not much noticed in the neighborhood of Krakatoa soon after 10 a.m. — that is, just after the great paroxysm — although they were heard for some hours later at greater distances.

The quantity of foreign matter in the air on the twenty-seventh may be realized when we hear of the mud, which succeeded a heavy fall of pumice, accumulating on the deck of the G. G. Loudoun at the rate of six inches in ten minutes, of dust reaching a depth of seven inches on board a vessel three hundred and seventy miles distant, and of a vast area of the ocean being thickly covered with pumice, sufficient in some parts to impede navigation. On board the Sir R. Sale pumice stones are reported to have fallen of the size of a pumpkin, and the crews of several vessels were employed for hours in shovelling the sand from their decks.

The sounds which thus called forth wondering inquiries over one-fourteenth of the entire surface of the globe within four hours of their emission, were, in fact, announcing, not the "salute of a corvette," but the blowing to pieces of a mountain by the hidden artillery of nature.

The expulsion of two-thirds of the Krakatoa mountain has left a magnificent section of the volcano by which to study its internal structure. Two drawings, reproduced from Verbeek's atlas, accompany Professor Judd's article in the volume issued by the Royal Society. An examination of the remaining solid portion of Krakatoa, and of the ejecta which have been collected from various places, has led this author to a theory of volcanic action differing considerably from the views formerly held by geologists. Both the older and more recent lavas have been subjected to careful microscopic study, and the results have thrown much light on the history of this mountain and of volcanic action in general. The ultimate chemical composition of recent lavas and the nature of certain crystals in them indicate the re-fusion of earlier lavas before ejection. But in the eruption of 1883, from May 20 to the final paroxysm, it is calculated by Verbeek that at least ninety-five per cent. of the materials thrown out consisted of pumice and dust, and not more than five per cent. of compact lava and of fragments torn from the side of the vent. This opinion is qualified by the English theory of re-fusion. The lava of 1883 presented itself in two different forms, porphyritic pitch and porphyritic obsidian. In each of these, crystalline elements constitute only about ten per cent. of the whole bulk. The obsidian has been found to be possessed of very remarkable properties, which, in the opinion of the author, go far to explain both the energy of volcanic action and the celestial appearances which astonished the world in 1883. This mineral, which in thin sections is almost colorless, has a strikingly vitreous lustre, is easily fused in a gas-flame, and during fusion bubbles and swells up into cauliflower-like masses which will float on water. The masses in appearance and structure exactly resemble the pumice ejected from Krakatoa. After fusion they are found to have lost from one to six per cent. of their weight. In examining the common pumice of Krakatoa it was found to have undergone a dilatation to five and a half times its original bulk, although something like one-tenth of the original lava consisted of undilatable crystals which remained to weight the mass. The obsidian or glassy rock has only to be heated in order to give off its volatile ingredients; these, like carbonic acid in dough, swell out the mass to five or six times its former bulk, and the melting glass is thus converted into true pumice, penetrated throughout with the vesicles produced by the escape of its original gaseous constituents. It seems probable that the water and volatile substances given off by such rocks at a white heat were in actual combination, and caused the rock to be fusible at comparatively low temperatures. The pumice of Krakatoa exhibits plates and threads of glass drawn out to the smallest dimensions visible under the microscope. The rapidity with which it cooled is shown by its extreme brittleness and by the depolarization of light.

The volcanic dust consisted chiefly of this pumice reduced to the finest powder by being carried up by the gases escaping from the interior with explosive violence, and by the grinding together in the air of fragments rendered brittle by intense strain. The heaviest particles would fall near the volcano, the very light and friable glassy dust would be carried to great distances. This dust would be composed of the ultra-microscopical, the elongated, and the very thin particles, and being less basic in composition would be the most transparent. Much of it must certainly have been carried by upper currents to distant parts of the world, and have reached the earth after long wanderings; but the sediments found in rain-gauges and on snow give no evidence by which such extremely minute and perhaps chiefly ultra-microscopic dust could be recognized.

The process by which this great eruption was brought about is considered to be typical of the physical action of volcanoes all over the world. Sea and surface-water obtain access to the vent or to the heated rocks below it, and if brought suddenly into contact may give rise, by the development of steam, to earthquakes or eruptions of moderate strength, but it is to the slow percolation of water into rocks in a certain condition that the author attributes the principal part in cataclysmal outbreaks. The water combines with the material of the rock, and by this combination the melting-point of the rock is reduced; it only requires the subjection of the hydrated compound to such heat as would be supplied by the anhydrous lavas in a fluid condition to disengage steam and other gases in enormous quantities, and to produce outbursts proportionate to the pressure and the strength of the enclosing walls. If, while this process is going on, water in large quantities gains access to the surface of the heated mass, solidification might take place and the escape of gases through the crater would be temporarily checked. When at last the accumulated force bursts the newly formed crust, this and other obstacles would be speedily removed by the tremendous violence of the blast, and the sides of the crater might either be blown away or fall into the seething lava. Such appears to have been the working of the final and self-destructive eruption of Krakatoa. The objection that water could not percolate to great depths, owing to the upward pressure of steam already formed, is met by recent experiments which show that the capillary action continues in spite of such pressure.

But, as if to confound the most ingenious explanation of terrestrial volcanoes, the moon looks down in scorn at the minute cones and craters of earth, and seems to declare in plain language that her mighty array of huge volcanic mountains, her hundreds of extinct Etnas, built themselves up in fire and fury without the aid of any water at all. There are craters fifteen times as large as the largest on our globe; there is the whole surface studded with cones as large as Vesuvius, a piled record of eruptions of tremendous force, and of internal energy so great that enormous circles, representing the walls of craters, overlap each other, and cracks extend for hundreds of miles from the volcanic centres. It is true that the largest circular walls on the moon's surface have been supposed to have been formed, not in the manner of the sides of terrestrial volcanoes, but by the sinking of the area within them; but the difficulties of this supposition have not been overcome. The abundance and size of craters testify to an effectual power of lunar volcanic action greatly exceeding anything with which we are familiar on the earth, but it must be borne in mind that, the force of gravity on the moon being only one-sixth of that of the earth, the height to which rocky matter would be thrown would be six times as great, and the crater walls proportionately extensive. Moreover, the ancient crust of the earth, denuded of its stratified and earthy deposits, would exhibit some very large crater rings, many being now well known, and fissures hundreds of miles long seem to correspond with the far more conspicuous cracks and bright lines of the moon. Vast lakes of lava, too, seem to have extended over hundreds of square miles in Europe and America, as a consequence not of violent eruption, but of quiet extrusion. Possibly these deposits may resemble the so-called "seas" on the moon. There can be no doubt that most of the volcanoes of the earth are arranged on certain lines of weakness, but the pressure of solidified matter being much greater than on the moon, eruptive action has been more confined to particular areas. It is surprising to find that an ancient Krakatoa has been traced which might be compared with many of the rather large lunar craters, having a circumference of something like twenty-five miles, and a height of ten or twelve thousand feet. Some great outburst, far exceeding that of 1883, seems, at a remote period, to have eviscerated the whole volcano, and left only a basal wreck, of which one portion was the recent Krakatoa. We can hardly accept either the "steam-engines' theory of some vulcanologists, or the hydrated-lava theory of Professor Judd, without admitting the former existence on the moon of a large volume of water. It is improbable that the chief agency of paroxysmal eruptions differed in the two cases. In each of the two globes the expansion of fluid rock in the process of cooling would bring to bear an enormous pressure, resulting in outwellings of lava, and violent eruptions would be accounted for by the development of steam on a large scale. That communication frequently exists between reservoirs of molten rock at great distances from each other on lines of fissure appears to be certain. Heated rocks which have long been subject to the hydration and aeration of infiltrated water would probably occupy more space in a solid than in a pasty or liquid condition, and would melt, as Professor Judd points out, at a lower temperature. Solid iron and solid bismuth will float on the melted metals, and solid lava floats on the liquid lake of a crater.^[4] It is true that the contraction by cooling of the solidified part of the globe works in the opposite direction; but while this process is fairly regular and even, solidification may take place unequally, rapidly, and by local causes such as cooling by extensive aqueous percolation. Another cause of periodic increases of pressure would be the shrinkage of the earth's crust upon the cooling interior, the percolation of water through fissures, and the closure of these fissures by changes of level, so that steam developed at some miles below the surface would force the fluid lava through the nearest volcanic vent. The apparent objection, however, to Professor Judd's theory, arising from a consideration of the non-aqueous surface of the moon, is disposed of if we admit, what seems not at all improbable, that the water previously existing in the moon in a free state has been entirely absorbed by the rocky substance. The intermittent character of most eruptions, their sudden violence, and the nature of the matter ejected are very well explained by the new theory; problems at least as difficult remain for solution.

The destruction caused by ashes and stones was slight in comparison with that which was brought about by sea-waves. These waves seemed to have started at the same time as the heaviest air-waves, and to have been connected with the culminating explosions. By successive waves, the largest of which occurred soon after ten o'clock, the towns of Anjer, Telokbetong, Tyringin, Merak, and many villages, were swept away. The height of the great wave was about one hundred feet at Merak, about eighty feet at Katinbang, seventy-two feet at Telokbetong, where the man-of-war Berouw was carried nearly two miles inland up the valley, and left about thirty feet above the level of the sea. The actual height of the wave before reaching the shore appears to have been about fifty feet. The travels of the principal sea-waves, and many details respecting them, are given with great elaboration by Captain Wharton. Eastwards of Krakatoa, the water is not deep, the narrow channel opens into the Java Sea, encumbered with reefs and shoals, and hemmed in by numerous islands. On the west, the water is clear of such obstructions. Consequently, at Sourabaya, four hundred and forty miles east of Krakatoa, a maximum rise of only ten inches was noted, and at Singapore and Hong Kong no disturbance was remarked, while towards the west the wave was propagated to greater distances than have hitherto been recorded of any such disturbance. Tide-gauges on the coast of India recorded waves of a varying height according to local conditions. At Karachi the height was twelve inches; near Calcutta, on the river, three inches; at Batticaloa and other places in Ceylon a rise of eight feet was noticed, representing probably a short wave superposed on one of the large ones. The waves were observed at Mauritius, and lasted for several hours, creating considerable commotion, and driving coasters from their anchorage. They were also conspicuous at Rodriguez and the Seychelles. At Port Alfred, in South Africa, the rise of the sea was one foot four inches, and at Table Bay eighteen inches; even at Orange Bay, Cape Horn, one of the waves was as high as seven inches. The coasts of France give indications of the arrival of several waves in succession, and at Havre, a distance of ten thousand seven hundred and eighty miles, undulations up to one inch are taken to represent the same disturbance.

The seismic flows and ebbs which thus covered a very large part of the globe were composed of long undulations, with periods of over an hour, and of shorter superposed irregular waves at brief intervals. The rate of propagation was in all cases less than theory would demand for the supposed depth of water. The average speed seems to have been something between three hundred and thirty and three hundred and eighty miles per hour. The mean depths deduced by the usual formula from this speed are less than those given by actual soundings. The cause of this discrepancy is not clear; but if the tide-gauges can be relied upon, and the disturbances recorded are due to identical original waves, it seems probable that submarine elevations and ridges, hitherto unknown, retarded the progress of the disturbance. The period of the long wave was originally about two hours, but at distant stations, such as Orange Bay and the ports of the English Channel, the period seems to have been reduced to about one-fourth, and throughout the course of the undulations its original character appears to have undergone considerable modification. The cause of an undulation with a period of two hours remains a mystery, but of the correspondence between the water and air waves in point of time at starting there can be no question. An upheaval of the sea bottom must have been very slow to account for the length of the wave; no earthquake was observed, and the evidence generally is against earth disturbance as a cause. The author of the geological section observes that the bulk of the fragments thrown out during the explosions must have fallen into the sea, and by their impact, almost coinciding with the violent evisceration of the crater, must have contributed to the rush of the destructive waves, and Captain Wharton calculates that a fiftieth part of the missing mass of Krakatoa, which was estimated to be at least 200,000,000,000 cubic feet, would, by dropping suddenly into the water, form a wave circle of one hundred miles in circumference, twenty feet high, and three hundred and fifty feet wide. But this is clearly totally inadequate to account for the long wave, and he therefore believes that the destructive waves in the Strait of Sunda were mainly due to masses falling into the sea, or to sudden explosions under the sea, but that the long wave recorded by distant tide-gauges had its origin in upheaval of the bottom. No consideration appears to have been given in any part of the report to a possible cause of some portion of the sea disturbance in the great barometric alternations in the air caused by the principal explosions. Already, on August 26, barometers were observed to fall nearly an inch at short intervals at about two hundred and forty miles from the valcano, and at a distance of about eleven hundred miles a fall and rise of .3 of an inch occurred in half an hour. The movements of pressure on the twenty-seventh must have been much greater. Like the great sea-wave, the barometric disturbance caused by the explosion of 10 a.m. extended over a period of nearly two hours, beginning with a rapid rise, passing to a deep depression and other less conspicuous alternations.

Although we do not find in the report any barometric observation in the neighborhood of the volcano during the passage of this great air-wave, we have the strongest evidence of an undulation of unique magnitude in the record of barometers all over the world, and in the tracing of the pencil of the Batavia gasometer, which was carried beyond the scale. If any recording instrument had existed in close proximity to the island, the probability is that a rise and fall of several inches would have been indicated between 10 a.m. and noon, and this would entail a change of several feet in the level of the sea, for the air-wave was long enough to allow of a large movement of water following variations of pressure. The researches of Sir W. Thomson and Mr. G. H. Darwin lead to the inference that the earth is not only solid throughout, but possesses at least the rigidity of an equal bulk of steel. Yet an increase in atmospheric pressure of only one inch is calculated to cause a sinking of several inches in the area of the earth's surface over which it extends. On the waters of the sea, and especially in confined channels, such a difference of pressure must lead to a dangerous disturbance. The great rise in air pressure which undoubtedly took place above Krakatoa at the time of the ten o'clock explosion may have been due both to an actual wave of compression, of the nature of a sound-wave, and to the enormous quantity of gases and vapors projected to an immense height, and taking a considerable time to spread over the surrounding space. The increased pressure of the atmosphere on the sea in the Strait of Sunda would, in effect, combine with the falling matter to produce an outrush in all directions, and it must have been many minutes, as shown by the barograms, before the wave of rarefaction ensued. These long airwaves are not easily understood, and more information is needed on several points; for instance, the approximate actual rises and falls of the barometer in parts of an inch at different places, a statement as to the number of minutes during which the reading was above or below the mean in each wave, an explanation of the apparently nearly equal barometric oscillations at places near and far, particulars of the effects, at the time erroneously ascribed to earthquakes, of the air-vibrations in Java and Sumatra, and a theoretical value for the amplitude and density of the air-wave near its source, calculated from the barometric indications at long distances. But the plain story of the progress of this wonderful wave, and the elaborate diagrams, beautiful and interesting in themselves, which illustrate the section, undoubtedly form one of the most valuable contributions in the whole inquiry. Never before has so vast an atmospheric disturbance been recorded by the barometers of the world; never, we believe we may add, have the diurnal tracings been thought of as likely to be sought for in connection with the activity of a distant volcano. They have emerged from their quiet recesses with one accord to bear testimony to the truth of a scarcely credible tale. From forty-seven stations, fairly representing the whole civilized world, we learn that the wave spread out from Krakatoa as a centre, expanding in a circular form till half round the globe, concentrated again towards the Antipodes, whence it started afresh and travelled back to Krakatoa, occupying in the double journey thirty-six hours, rebounded, and set off again on the same revolution, and repeated the movement at least three times sufficiently strongly to be recorded. At some stations no less than seven passages, going and returning, are indicated by the diagrams. The whole process was almost exactly similar to the alternate expansions and contractions of a wave of water caused by dropping a stone at the centre of a circular pool. Certainly, without the most general and impassive testimony in its favor, the startling induction represented with calm precision in these four fascinating plates would have been contemptuously rejected. But the sensitive paper of the barograms has no theories and no prepossessions, no personal equation and no love of the marvellous, no credulity, and, above all, no incredulity. In a matter of human observation, nothing stands so much in the way of progress as the indolent habit of explaining the new and unknown by the old and familiar, the unreadiness to derive new ideas from new facts; and a quick imagination, though sometimes mistaken, proves itself more productive in the end than the mind which either rejects the fact for its novelty, or insists on saddling it uncomfortably on an old hypothesis. But the safe groundwork of facts always repays close attention. All through this inquiry we are reminded of the large results to be obtained by small but accurate instruments, and by a few careful measurements, rather than by numerous casual observations. From the barograms, then, we have tidings of atmospheric movements comparable to gigantic waves of sound, starting from a small area and encompassing the globe, not only once, but several times in succession, completing each circuit in about thirty-six hours. The mean speed of propagation was about seven hundred miles an hour, which is less, by twenty-three miles, than the velocity of sound at zero Fahrenheit; the velocity, in fact, seems to have corresponded to that of sound in air at twenty or thirty degrees below zero. No explanation is given of this deficiency. It is believed, though perhaps not established, that the rate of propagation of sound diminishes with diminishing intensity, and since this air-wave must have become very greatly reduced in its circuit of the earth, we should find that a longer time was occupied in the second and third circuits than in the first. The diminution actually occurred; the rate for the first passage in one direction was 10'23º per hour, for the last passage 9'77º per hour, and in the other direction 10'47º to 10'27º respectively. But, considering the wave as a sonorous vibration of great intensity, it is remarkable that the rate to distances of two or three thousand miles in the tropics, where high temperature would favor rapidity of advance, did not much exceed the rate to much greater distances and to places in higher latitudes. One other factor would tend to increase velocity. Low notes are supposed to travel faster than high notes, and this wave might be considered as of a note far below the range of hearing. Yet its maximum rate was only slightly above that of sound in air at zero Fahrenheit.

One result revealed by the tables seems especially noteworthy, the difference of the velocities of the waves which travelled with and against the direction of the earth's rotation, amounting to about twenty-eight miles an hour; this is accounted for by the direction of the winds along the paths of the waves which passed over the majority of the stations being on the whole westerly. A current of fourteen miles an hour would, it is stated, cause a corresponding acceleration, or retardation, in the wave, according as the wave were advancing with or against it resulting in the observed difference of twenty-eight miles. From Krakatoa to Mauritius the rate of the wave was comparatively unaffected; in the opposite direction from east to west round the earth to Loanda on the west coast of Africa it was retarded. Speaking generally, in the extra tropics the wave from west to east was accelerated, that from east to west retarded, while within the tropics the eastward passage was retarded. So far as can be gathered therefore from the data, a general movement of the air within the tropics from east to west may be inferred, and without the tropics from west to east. Those waves which passed near the north and south poles give unaccountable results, for the direct wave from Krakatoa viâ the north pole does not seem to have been sensibly retarded by the low temperature, and the velocity of that which passed close to the south pole was only very decidedly reduced after the first passage, and in the next circuit was greatly increased. The barometer curves of forty stations, mostly European, are given on a much reduced scale, and copies of barograms from eight selected stations show the character of the first four oscillations; elaborate tables showing the intervals between successive waves are appended.

The pressure-gauge from the Batavia gasworks supplies an interesting narrative of the various air-waves passing over the town on August 26 and 27; this is reproduced both in M. Verbeek's and in the English report. Very strong outbursts appear to have taken place about 5 p.m. on August 26, and from midnight to 10 a.m. on the twenty-seventh, the hour of the culminating explosion which so far exceeded all the rest and drove the pencil against the stops of the scale. The differences of pressure within a short space of time exceeded 0'4 in. of mercury, if we correctly read the diagram. But the accompanying detonation does not seem to have corresponded in excessive intensity with the amplitude of the wave on which it was borne.

The immediate consequences of the great explosion were that a wave fifty feet high and of great breadth swept along the strait and with diminishing height traversed the Southern Ocean; the sea for hundreds of miles was covered with masses of pumice descended from the darkened sky, an air-wave of unexampled grandeur was circling round the globe, impenetrable darkness extended for scores of miles in many directions, ashes and dust fell in great quantities on ships hundreds of miles distant, and within a circle of two thousand miles people of many nations and languages were unsuccessfully puzzling at the riddle of strange noises. This was not all. Not only were earth and sea disturbed and the air darkened near the Sunda Strait, but on the same day the blue sky was almost covered with a thin white mantle a thousand miles and more westwards, and the sun himself was almost extinguished, struggling through the mist either like a dull red lamp or a ball of fire, or like a weak moon, or, as at Batavia, emerging from the dust-cloud transformed to green. The rapidity of these events is surprising. Within twenty-four hours of the explosion strangely colored suns were seen at enormous distances, up to two thousand miles, at such widely sundered places as Labuan, Ceylon, and Diego Garcia. The Ceylon observation indeed is open to question, being a native report from the northern part of the island, and referring to sunrise of the twenty-seventh, that is before the major eruption occurred, and unsupported by further testimony from Ceylon and India. It appears certain that already on the twenty-sixth vessels one thousand miles westwards of Java experienced some very singular phenomena, showing the passage overhead of a broad stream of dust from the eruptions of that day, and we may fairly infer that some of the heavier matter composing that dust-cloud fell into a strong southerly blowing towards the coast of Ceylon and traversed the distance of about eleven hundred miles in about twenty hours. Thus the Ceylon observation, and perhaps Captain Vereker's near Labuan as well, would refer to cloud-streams of dust and steam, of no great magnitude, the one filtered out from a current going rapidly westwards, and the other carried by the S.W. monsoon towards Japan. The early arrival (twenty-eighth) in Japan of matter causing a coppery sun would be similarly due to the S.W. monsoon bearing the products of the twenty-sixth.

The mass of the powdery matter thrown out by the explosion of the twenty-seventh seems to have spread out at such an enormous altitude that the finer particles were forthwith conveyed by a full fair easterly gale steadily and without pause on a great circle of the globe. This lofty, unresting hurricane has been hitherto unsuspected. No means of ascertaining the winds of inaccessible altitudes in the tropics had been devised by human ingenuity. Nothing but a great natural experiment such as Jules Verne would have hardly dared to dream of would have disclosed the circulation of the upper atmosphere over the greater part of the world, and the disclosure has been made by particles on which, till lately, the "eternal hills" reposed or floated.

The principal celestial phenomena in the Indian Ocean from August 27 to 30 were a peculiar soft haze, a very strange appearance of the sun, and a wonderful red glow long before sunrise and after sunset. The general list of first appearances gives, as far as possible, the words of observers used at the time, and we thus get a very interesting impression both of the various features of the phenomena and of the way in which they struck various minds. The captain of the Barbarossa, nearly one thousand miles from Krakatoa, saw the "whole sky of a peculiar red, like bright polished copper," and this color suddenly changed to uniform grey. This appearance was followed by "frequent, but strikingly short, thunder," in reality, the noise of the eruptions of the night of the twenty-sixth. The same evening, still further west, the sky was "all of a flare." On the twenty-eighth, at twelve hundred miles due west of the volcano, "the sky was very hazy, and a fine white powder fell in a constant shower like snow, covering the whole ship." Many other ships had similar experiences. The sun was nearly obscured by a pale yellowish haze on the wind twenty-ninth and thirtieth. At about eighteen hundred miles west a quantity of light dust, like Portland cement, fell at the same time. So late as September 8 a deposit of sand occurred on board the Scotia in 10º N. 53º E.; at the same time a partial halo formed round the sun, and the moon was green before setting; on the following morning the sun was green, and the sky for several days was covered with haze. In the Atlantic, at St. Helena, on August 30, a red light like a distant fire surprised one of the inhabitants at 4 a.m., and on the same day a remarkable glare and leaden sky were noticed in other parts of the Atlantic within the tropics. On the following day, so far as 13º 30' N. 31º 20' W., a "curious electric-light appearance" and other phenomena were noted; and near the equator the sun was like copper, with a metallic haze over the sky. On September 2, at 18º S. of the equator, the sun was like polished lead, and the whole sky grey, and on the same day the whole of the northern part of South America was astonished with blue suns, or red skies. These phenomena continued their rapid course westwards, and by September 7 seem to have covered nearly the whole of the Pacific within the same latitudes. On September 9 and 10 green and blue suns were observed over nearly the whole of India; the dust cloud was already well advanced on its second circuit of the globe. On the twenty-second the green suns returned in force to India; the stream of matter was now on its third circuit, and can be traced to the western Atlantic on September 28; after this, its increasing tenuity prevented further observations definite enough to be used in the tables of velocity. During the whole of its rapid and wonderfully even revolution round the earth, the great cloud was extending itself less conspicuously towards the north and south, and many scattered observations in the temperate zones afford evidence that the sifting out of heavier particles continued without interruption, and that these, in sufficient quantity to produce moderate afterglow, were carried by the anti-trades and other elevated currents to great distances. During October the spread of the immense stratum of particles of extreme tenuitv which gave rise to most of the phenomena was slow and gradual. There are many indications, and the authors conclude that they represent the fact, that while a continuous current, with a speed of between seventy-two and eighty-three miles an hour, prevails between 16º N. and 16º S., and probably somewhat beyond these limits, the circulation be-comes less rapid towards north and south and at some latitudes not very far from 35º becomes converted into a flow from S.W. and N.W., and in still higher latitudes to a direction from nearly W. These directions are understood to apply to altitudes of about one hundred thousand feet at the equator, and sixty to ninety thousand feet in the temperate zones. The arguments by which the height of the glow stratum has been calculated are most elaborate, and, from an exhaustive analysis of many observations, it is stated that the altitude progressively diminished from one hundred and twenty-one thousand feet in August to about sixty-four thousand feet in the following January. The width of the particles which caused the corona known as Bishop's ring, after its first observer, is found to be about 1/16888 of an inch in average diameter, and since most of the particles were probably thin plates, their thickness would be very much less. From a formula given in 1851 by Professor Stokes, relating to the viscosity of the air, we find that such particles would take more than two years to fall fifty thousand feet, so that at the end of that time they would still be above the ordinary level of cirrus. This estimate applies to the smaller particles; those which were most effective in the sunset glows may have been larger, and may have reached the lower atmosphere within a year. It is surprising to find that, theoretically, the rarity of the atmosphere at such a height as twenty miles would little affect the rate of fall of very small particles, such as those of which smoke consists.

The twilight skies in northern latitudes in November and December, 1883, were grand in the extreme, and in the southern hemisphere they were similar in every detail. From New Caledonia we hear of a western sky after sunset

like white hot steel, with an exquisite green eastward. At 7 p.m., or a little after, nearly the entire western half of the horizon has changed to a fiery crimson; as time goes on, the northern and southern areas lose their glory, and the greys of night contract from the northern end first most rapidly; the east is of the normal grey. The south now closes in, and presently, about 8 p.m., there is only a glare in the sky, just over the sun's path, as of a distant conflagration, till the fire in the west dies out. I have been attempting to describe one of our cloudless evenings, of which we have had only too many, having just come through a fearful drought that has lasted all this while; but who shall paint the glory of the heavens when flecked with clouds? burnished gold, copper, brass, silver, such as Turner in his wildest dreams never saw, and of such fantastic forms!

At Worcester in England the twilight scenery was remarked on as follows in December —

On the 5th inst. the southern heavens were resplendent with the richest and most brilliant colors, to attempt the description of which would be somewhat puzzling. It seems as if of late the grandest displays occur before sunrise. The afternoon effects were remarkable less for richness of coloration than for the lustre of the light which arose in the west after sunset, and for the predominance over the whole sky of opalescent white colors. The reflection of the light on church towers and buildings brought the architecture in strong and startling relief; there was, however, at 4.15 a colorless display, and on this occasion the moon for a short time was again changed to a hue of emerald green. On December 15

the sunrise was of a most impressive character. … The room in which the observations were made had two windows, one facing east and the other south, and the marvellous spectacle was witnessed of a flood of crimson glare filling the east window, while through the south window poured a volume of green light.

At half past five in the morning of November 30, that is, two hours before sunrise, persons crossing London Bridge were startled by a red glare in the eastern sky, and attributed it to a great fire in the City. Three days earlier the fire-engines had been called out in the morning at Poughkeepsie, on the Hudson, and in the evening at New Haven, Connecticut; and on November 28 in Austria the red glow in the darkness was also attributed to a conflagration. These alarms fixed very well the time of the first appearance of the grand displays. On November 30 a correspondent telegraphed from Rome —

Yesterday evening the population of Rome was struck with admiration, mingled with awe, at the sight of a splendid phenomenon. From fifteen minutes after sunset until more than an hour later the north-western hemisphere was tinged with crimson, gradually increasing in intensity until it had the appearance of the reflection of an extensive conflagration in front of which the tower of the castle of St-Angelo, the cupola of St. Peter's, and the outline of Monte Mario, as seen from the Poncio, stood out in prominent relief. Immediately above the horizon there was a broad belt of orange red, and above that another of green, surmounted by the crimson glare of the aurora. The sky of the eastern hemisphere presented a uniform sea-green tint. The phenomenon was repeated again this morning and again this evening.

Professor Riccò describes the sunset on December 3 at Palermo: —

The sky is bright yellowish pink and bronze color at 4.45 p.m.; higher up it is rose color, dividing into shafts of an intense purple, separated by spaces of violet; this very brilliant light extended to the zenith and strongly illuminated the city and environs, which assumed a new and strange aspect; the crescent moon appeared greenish blue by contrast. At 5.28 the whole sky was invaded by another light, uniform purple, and rather intense, especially towards W.; at 6.4 the purple light was low down, at 6.29 only a trace of reddish haze remained.

Professor Riccò supplies tables of the weather conditions during the eruption of Graham's Island in 1831, and from these it appears that there was a dense mist for several days from July 23, that red twilights, unusually prolonged, occurred from August 4 to October, and that the sun was dim and bluish white on August 8. Excellent plates are given illustrating the various phenomena. He is inclined to attribute the blue coloration of the sun in 1883 to vapors produced from the volcano, and the red twilights to the rapid precipitation of vapor on small dust, if it is admitted that the dust could be projected to an adequate height, and could remain suspended for so long as three years.

The comparison to the glare of a fire was made in almost every country where the fore-glows and after-glows appeared. At London, in Canada, for instance, the following language was used: On November 22 "most extraordinary sunset, pitch dark in east and zenith, a blaze of red lurid fire in west;" at Baltimore there was an appearance of a tremendous fire along the horizon, and at an altitude of 40º;" at Victoria, British Columbia, the glows were most magnificent on November 23, "as if the country were ablaze with flame," and their duration was two hours. But in many places, and especially in France, the red skies were attributed to aurora; indeed, the theory of aurora was held very persistently. On board the Sunbeam, near the Canary Islands, Lady Brassey noted "an indescribably splendid sunset, sky colored purple, orange, yellow, green, and blue." The colors were not only indescribable, but apparently incapable of being depicted on paper, for no artist, so far as we know, succeeded in representing an after-glow in a sky free from clouds. Many of the displays, if correctly represented, would have appeared too theatrical, metallic, and unearthly; the effect was of too lurid and awful a nature, too much wanting in repose. But many, on the other hand, were at once too delicate and too magnificent for imitation. We may here remark, as a matter of experience, that the neglect by the public of those grand natural spectacles, presented gratis, is quite astonishing, and that during one of the most striking of all the evening displays not one inhabitant of a large town on the south coast ventured on the beach to behold it. Everywhere shutters were closed at the customary hour, before the development of a scene which, if artificial, would have attracted thousands from distant nations, and which could not be expected to occur twice in a lifetime.

It is said that in some parts of the world the sunsets are habitually beautiful. In Italy and Egypt the rosy after-glow of the western sky in certain seasons is well known, but the coast of Peru and the ocean westward seem to surpass all other localities in their celestial scenery. Stewart Ellis in his voyage to the Sandwich Islands describes them as follows: —

We are now (15º S. 96º W.) off the coast of Peru, and have been delighted with the beauty of the sky and clouds, which is here very peculiar, and I should think unrivalled in any part of the world. Towards evening and in the morning I have seen at the same time clouds of almost every color in different parts of the heavens, and of hues I never beheld there before; for instance, a rich and perfect green, amber, and carmine, while the hemisphere around the rising or setting sun has been one blaze of glory. Last night the tinge on the ocean was of a perfect blood-color, occasioned by the reflection of a fleecy veil of crimson clouds stretched over the greater part of the heavens; the appearance was so singular as to cause us almost to shrink from it, as from something supernatural.

Proximity to the volcanoes of the Andes, which are always to some degree active, gives this pre-eminence to the Peruvian twilights.

As the rainbow appearing after a storm arises from the refractive power of raindrops, being different for different waves of light, so the grand procession of rainbow colors in the twilight displays of 1883 has been attributed by some, notably by Professor Kiessling, to the diffraction of light by very minute particles. The authors of the optical portion of the English report hold another view. They believe that although diffraction through both the stratum of foreign matter, which was composed mainly of microscopic or ultra-microscopic pumice particles, and through the lower atmosphere, had much to do with the phenomena, the chief part in the brilliant glows was played by reflection. It is shown that small transparent glassy particles are competent powerfully to reflect the sun's rays, and that the height of the stratum would cause the reflection of the beams of the setting sun to take place when the intervening air, including the greater part of what we call the blue sky, had been darkened by the shadow of the earth. The colors reflected by the particles would be those which had traversed with least loss the length of the stratum through which the sun shone, and later in the evening, in the case of the after-glow, those which had traversed the lower air — that is, the red and orange parts of the spectrum. The secondary glow, in their opinion, was caused by reflection of the rays of the first glow, as it sank, viewed from the high level, on the horizon. It is shown that an extremely small quantity of matter is sufficient to produce striking effects; for instance, the tails of comets have been calculated to be of so great a tenuity, that the matter contained in a tail of one hundred million miles in length, and fifty thousand miles in diameter, would, if compressed, scarcely amount to a cartload.

Professor Kiessling has succeeded in experimentally producing, by means of the formation of a cloud of sulphate of ammonia and other fine powders chemically produced, in air, absorptive or rather diffractive effects on the sun's rays, which may be compared to the blue and green suns of 1883. The color of the sun's rays changed rapidly in passing though this cloud, from dark copper-color, through violet and crimson, to a brilliant azure blue. By experimenting under a variety of conditions a number of interesting changes of color were produced, and a very fair imitation of the remarkable colored rings, which were observed for nearly three years after the Krakatoa eruption. These rings, or coronæ, both in their extent and persistence, seem to have been new to meteorology, and it is from their size that we derive the best approximation to the average size of the particles which composed them. The coronæ, like the haze canopy and twilight glows, were little, if at all, affected by weather conditions near the surface of the earth, and only required a clear atmosphere in order to become visible. It is interesting to observe that the visibility of the coronæ increased as the wonderful sunsets decreased. This was owing to the gradual descent of the larger particles and the increasing homogeneity of the constituents of the remaining cloud.

Professor Kiessling, like the authors of the Royal Society report, gives maps of the progress of the dust cloud from August 26 to September 30, and his general conclusions are similar with regard to its velocity and character.

The year 1831 was very remarkable for the number of its eruptions and for concurrent phenomena, such as blue and green suns, dry fogs, light sufficient for reading at midnight, and very fine red twilights. A volcano had formed a new island in the Mediterranean in July, and the height of the column of dust was found, by the measurements of Professor Hoffmann and Dr. Schultz, to be about thirteen miles. Arago explained the prolonged twilights of 1831 by the great height of the dry fog and the multiple reflections of the sun's rays.

Incidentally, an eclipse of the moon in 1885, in which the earth's shadow was very much darker than usual, has been explained by the absorption, or reflection, by the layer of dust enveloping the earth, of the red rays which are usually refracted and reach the moon's surface. The amount of extra matter in the air was undoubtedly sufficient to interfere seriously with astronomical definition for one or two years.

By a very complete chain of evidence, due in great part to the observations which happened to have been undertaken by the captains of ships for the Meteorological Office in 1883, and to the excellent observations of the captains of German vessels, the connection between the phenomena which were observed in all parts of the world, viz., the haze in the sky, the colored sun, the coronæ, and the twilight illuminations, and the derivation of all these from Krakatoa, is established. To corroborate the conclusion and extend its application, the authors bring forward evidence of the production of persistent dry fogs, and red twilights, in former years distinguished by great eruptions, and of blue suns, observed not only through volcanic dust clouds, but through the dusty atmosphere of the Loess in China, of the Sahara Desert, and of the neighborhood of stone works at Eastbourne, where large quantities of sea-beach are crushed by machinery. The coronæ, as we have seen, have been artificially produced by particles corresponding in size to those of the pumice cloud. From the conclusions reached by Professor Archibald and Mr. Russell in the optical sections, we learn that a cloud of fine dust may remain suspended at a height of from twenty-four down to thirteen miles without being sensibly affected by the weather of the lower atmosphere; that such a cloud in the tropics revolves round the earth from east to west in thirteen days; that it does not apparently condense vapor upon itself; that it interrupts the red more than the blue solar rays; that, like Tyndall's actinic cloud produced in the laboratory, it may strikingly reflect rays falling upon it without sensible hindrance to transparency, and, like it, may consist of an almost incredibly small quantity of matter.

It is remarkable, bearing in mind the strong electric effects occurring at such altitudes as the summit of Mount Washington, that the elevated dust cloud exhibited no forms suggesting electric arrangement, as in the case of cirrus. It was either uniform and featureless, or covered the sky with parallel streaks resembling the long rollers of an unruffled ocean.

A study of the dates and particulars furnished in the English and German reports should lead to a knowledge of atmospheric movements above the cirrus region which has hitherto been inaccessible. According to theoretic views stated by Mr. Archibald, the system of circulation indicated by the dust stratum might have been, and to some extent was, though rather heretically, anticipated. No opportunity seems to exist at present of testing the validity of the theory enunciated, for balloons have never reached a height exceeding one-third of that of the stratum, and clouds are not formed above a comparatively moderate altitude. The motion of the highest cirrus, moreover, can only be learned in the somewhat disturbed condition which their presence betokens. If small test-balloons could be constructed to remain for a definite time at heights from seventy to one hundred thousand feet, and to be brought down at will, interesting information would be gained respecting the eternally unclouded region within twenty miles of the habitable surface, and its system of regular currents of high velocity. In the distant future, when ballooning has attained a much higher stage of development, such knowledge may possibly be of practical value.

The section of the English report dealing with opinions and hypotheses expressed is very instructive, and shows extraordinary differences among scientific men on their first acquaintance with the facts. In America the meteoric-dust theory was much in favor, and in all countries the Krakatoa origin of the phenomena was widely discredited. The tendency of the observer everywhere was to connect them with the particular branch or twig of science with which he was best acquainted. Only the minute investigation of the whole range of sensible consequences of this great natural experiment could have led to the establishment of the truth respecting their origin and their relation to one another. More than one conclusion which has been arrived at will be the starting-point for fresh discovery and, we may hope, the means of practical advantage to mankind.

Notes

1. 1. Krakatau. Par M. Verbeek. Publié par ordre de Son Excellence le Gouverneur Général des Indes Néerlandaises. Batavia: 1884 and 5885. Paris: 1885 and 1886. 2. The Eruption of Krakatoa and Subsequent Phenomena. Report of a committee appointed by the Royal Society. 4t0. 1883. 3. Untersuchungen über Dämmerungserscheinungen zur Erklärung der nach dem Krakatau-Ausbruch beobachteten atmosphärischoptischen Störung. Von J. Kiessling. Hamburg and Leipzig: 1888. 4. Osservazionie Studii dei Crepuscoli Rossi, 1883-6. Dal Professore Riccò. Estratto degli "Annali della Meteorologia Italiana," Parte I. 1885.
2. Smithsonian Report, 1885, part i.
3. On the Atmosphere as a Vehicle of Sound, Proc. Roy. Soc., 1874.
4. Nasmyth, The Moon