sun, at noon, looked as blank as a clouded moon, and shed a rust-coloured ferruginous light on the ground, and floors of rooms; but was particularly lurid and blood-coloured at rising and setting. All the time the heat was so intense that butchers' meat could hardly be eaten on the day after it was killed; and the flies swarmed so in the lanes and hedges that they rendered the horses half frantic, and riding irksome. The country people began to look with a superstitious awe at the red lowering aspect of the sun; and indeed there was reason for the most enlightened person to be apprehensive, for, all the while, Calabria and part of the Isle of Sicily were torn and convulsed with earthquakes, and about that juncture a volcano sprang out of the sea on the coast of Norway.”

And Cowper in the “Task” describes the same appearances (Task, Book ii.)

“Sure there is need of social intercourse,
Benevolence, and peace, and mutual aid,
Between the nations in a world, that seems
To toll the death-bell of its own decease,
And by the voice of all its elements
To preach the gen'ral doom. When were the winds
Let slip with such a warrant to destroy?
When did the waves so haughtily o'erleap
Their ancient barriers, deluging the dry?
Fires from beneath, and meteors from above,
Portentous, unexampled, unexplain'd,
Have kindled beacons in the skies, and th' old
And crazy Earth has had her shaking fits
More frequent, and forgone her usual rest.”

From another source (“Nature,” 17th July) I hear that extraordinary sunrises and sunsets in that year lasted for eleven months, and have been attributed by Arago to volcanic dust, for besides the volcanic activity to which the Vicar of Selborne refers in the letter mentioned, there were, it appears, in 1783 tremendous eruptions of Asama Yama, in Japan, and Skaptar Jokul, in Iceland. Mrs. Somerville, Sir C. Lyell, and Sir J. Herschell all refer to the latter event, and the last named says that on that occasion 21 cubic miles of lava were ejected from the crater. Again, similar phenomena were observed in Europe and America in 1831, and at first when this fact was disinterred from historical archives or recalled by human memory, it was stated that that year was memorable for no particular eruption. But a quotation from Nile's Register, 31st October, 1831, given in “Nature” of May last, puts a different aspect on the case. It says that on 7th August preceding, there occurred a violent eruption off the coast of Sicily, during which Graham' Isle was formed: two days later the sunsets began to be very lurid and remarkable, the glow extending to the

zenith, and lasting beyond twilight. After the 11th August a blue sun was seen at the Bermudas, and on the 15th at Mobile. In October at Washington, and at Alexandria in Virginia, there was a red sun-glow, and at midday the sun was silver white. Again, Dr. Taylor (in his Australasian notes) quotes Col. Stuart-Wortley to the effect that he has seen similar skies in the Pacific, caused by the volcanic eruptions of South America: and Ed. Whymper, on the slope of Chimborazo, 3rd July, 1880, saw an eruption of Cotopaxi, 65 miles away, which produced not only a blood-red sky, changing to verdigris green and the colours of brick dust, copper, and shining brass, but also a veritable green sun.

The fact which I wish to bring out by all this evidence is that somehow or other by the discharge of smoke, dust vapour, or gas—i.e., matter in one of its three leading forms—from volcanic rents, the colours in the sky and of the heavenly bodies have been and can be for a period of time, short or long, very considerably changed. Here, therefore, we have an additional reason besides that of sequence in point of time, for associating the late manifestations with the outbreak at Krakatoa, and the very exceptional violence and extent of the operation of that eruption enables us to see why the sun-glows recently have been so remarkable and brilliant, and of such long continuance.

But that is not all. The year 1883 was as a whole wonderful in its volcanic activity. The old fiery life is still throbbing in the earth' veins, and as the Times recently remarked in a leading article, the lava catastrophe clearly shows that our little world is not near its latter end yet. I find that from March of last year almost up to the present time there has been a series of earthquakes and eruptions such as mankind has probably never before experienced in historic times, within a similarly brief period. By the microseismometer, earth-tremors of the faintest character are duly recorded; and this delicate instrument reveals the fact that for some time, particularly in certain parts of the world, earth-tremors have been more than usually continuous. I would ask you to carefully note the following list of the eruptions and earthquakes during the period indicated. I have compiled it up to Christmas last from a journalistic summary of the events of the year:—

• 22nd March. Earthquakes in Sicily, eruption of Etna.

• 26th March. Eruption of Hecla.

• 12th April. Earthquakes in Tasmania begin, and last through year.

• 20th May. First eruption of Krakatoa.

• 22nd July. Earthquake in Ischia (6,000 killed; felt even so far as Wiesbaden).

• 25th August. Great eruption of Krakatoa.

• 28th August. Earthquakes in West Australia and Tasmania.

• 29th August. Earthquakes in New South Wales and Tasmania.

• 30th August. Earthquakes in New South Wales and Queensland.

• 20th September. Earthquakes in New South Wales.

• 8th October. Eruptions in Iceland.

• 10th October. Last eruption of Krakatoa.

• 11th October. Earthquakes in California. Eruptions in Alaska.

• 15th October. Earthquakes in the Levant and Asia Minor, 200 perished.

• 22nd October. Earthquake in Lisbon.

It is indeed, though very incomplete doubtless, a formidable list, and since Christmas almost up to date, the volcanic activity has been continued. Ischia and Alaska have again suffered, and even steady Old England has felt the effect of internal fire, or shrinking, or whatever may cause earthquakes—having experienced a severe shock in the neighbourhood of Colchester.

But of all these outbreaks that at Krakatoa was by far the most formidable and gigantic. Professor Verbeek, who was at the head of the Committee of Enquiry appointed by the Dutch Government at Batavia to investigate on the spot the causes, features, and effects of the eruption, has already sent in a long report, a translation of which appeared in “Nature” of 1st May; it is a most complete and interesting account. Even a synopsis of it, however, would be too long for me to give to-night. I must content myself with mentioning a few of the particulars and estimates given, mainly with the view of showing the magnitude of the operations.

It appears that there is a rent in the earth across the Straits of Sunda, and that a slight pressure on the molten matter beneath the crust, or the entrance of salt water causing the generation of steam, produced the explosion. The Professor naturally thinks that the simultaneous occurrence of earthquakes in Australia indicates a very large area of operations. The activity of Krakatoa really began in May, 1883, but culminated on Monday, 27th August, at 5 minutes past 10 a.m., in an explosion so terrific that the noise of it was heard in Ceylon, at Perth in Western Australia (1,800 miles away), and even at Diego Garcia, near Mauritius, 3,000 miles away; in fact over nearly one-fourth of the earth' circumference. Seven or eight air-waves were also formed in all directions—violent throbbings of the entire atmosphere of the world—producing effects like those of earthquakes. Even at Pasoeroean 830 kilometres (say 520 miles) distant, walls were rent by these jerks in the air, which were very quickly communicated and—being registered at places so wide apart as Batavia, Wellington, Dunedin, Melbourne, St. Petersburg, and London—they form, as it were a register of the

successive explosions, and are exceedingly interesting. The air-wave following the 10.5 eruption seems to have travelled three and a quarter times round the world before it allowed the atmosphere to regain its normal condition; journeying at the rate of 674 or 706 miles an hour according as it passed to the east or to the west, i.e., according as it was assisted or retarded by the upper currents of wind (Col. Strachey, Royal Society). Besides this “atmospheric shudder,” as Ellery graphically calls it, there were enormous sea-waves—tidal-waves as they are improperly called—formed by the tumbling-in of the burnt-out mountain, or the falling into the sea of vast bodies of ashes, or by submarine explosions or otherwise. These waves were, as you know, awfully destructive to human life—perhaps carried off 100,000 people. To show their force and rapidity, I may state that they reached Geraldton in Western Australia so early as the 27th, Mongonui and Timaru, New Zealand, on the 29th, and Nelson on the 30th. Shortly afterwards they reached places more distant still, even the coast of France. Verbeek computes that 18 cubic kilometres—nearly twenty-five thousand million cubic yards—of solids, and more of gas, were ejected. The steam cloud rose to the height of 11,000 metres (nearly seven miles) even on 20th May, when the eruption was trifling, and probably to the height of 20,000 (over thirteen miles) on the 27th. But naturally enough, nobody in the vicinity on that exciting day was sufficiently calm to note with accuracy such phenomena. Before the eruption the Island of Krakatoa contained 33½ square kilometres (nearly thirteen square miles), now it contains only 10½ square kilometres, that is, less than one-third of the old area. It consisted of three large peaks, one of which was 2,500 feet high; the two smaller of these and a cleanly-cut half of the largest one have disappeared entirely, and the sea over the place where they were is now over 1,000 feet deep. The whole neighbourhood is changed. One island, Poelsche Hoedje, has vanished entirely. Others are trebled in size. Within a radius of 15 kilometres (say nine miles) the ashes are 20 to 40 inches deep, and an area as large as Germany, Holland, and Belgium put together is covered to a less extent. A locality subject to visitations of this kind, with sixteen active volcanoes, some of them 12,000 feet high, and many more only quiet for a time, is well called “The Lid of Hell,” and after such a visitation to any one part of the world, we need not wonder at all to see curious meteorological phenomena as a consequence, even in very distant places.

But, as I have already pointed out, this eruption of Krakatoa was by no means the only outbreak through the earth' crust during last year. That of Hecla on 26th March, and that of Alaska, were anything but insignificant. The latter especially must have resulted in the throwing up of a vast quantity of matter. The whole Alaskan peninsula was in volcanic activity.

One mountain, St. Augustine, was split into two parts; another 12,000 feet high was very violent—threw the whole district into darkness for some time and covered every foot of the ground to a long distance with a layer of ashes five or six feet deep. An island a mile and a half long and 75 feet high was suddenly formed, and tidal-waves 30 feet high broke on all the surrounding coasts. Still even this was vastly inferior to the Krakatoa catastrophe, and almost all the prominent scientific men of the day concur in connecting our sun-glows with the latter event, although they allow that the smaller eruptions of the year may have contributed to the intensity and duration of the displays.

But it remains to be shown how that connection is established. Something unusual was evidently carried into the higher air in immense quantities in the last days of August, 1883. What was that something? Dust, vapour subsequently to be frozen into ice-crystals by the cold aerial upper currents, or some gas such as hydrogen or chlorine. Each of these has its advocates—perhaps the truth is that each has contributed to the result. Let us see more closely what is or can be advanced in favour of each supposition.

Firstly, as to the hydrogen gas there really is little to say. Several authorities (Ellery and Hazen, e.g.) incidentally mention it as a possible product of the eruption, resulting from the decomposition of the ejected water through electric action, I suppose. When once released the hydrogen would readily and naturally ascend to a great height, for its specific gravity is less than that of any other gas, and we can conceive it as suspended at a great altitude in great quantities, and for a long time. But granting all this, would it produce the sun-glows? Mr. Ellery thinks not, and nobody of scientific standing contradicts him. Again the heavy yellow gas—chlorine—would probably be liberated in large quantities by the pouring of sea-water into the fiery craters of Krakatoa, and in combination with other causes may have contributed to the exceptional colouring.

Secondly, there is the aqueous vapour, or rather ice-crystal theory. There is more to be said here. But, on the threshold, there is a great difficulty. Aqueous vapour in the air is no unusual thing. Often the atmosphere is heavily charged with it. But the sun-glows are phenomenal. Is it not curious that the vapour has never before assumed the condition requisite for producing such an effect? Hazen argues in favour of this theory thus: there are three conditions necessary to produce the best results in the way of sun-glows: (1) clear sky; (2) abundant refracting material; and (3) great force to carry the same high up into the air; all these conditions have been fulfilled. The universality of water secures the fulfilment of the second condition even independently of the Krakatoa

eruption; and for the fulfilment of the third condition we have electricity, which during the past year has been in great force as shown by electrometers—incessant volcanic eruptions, a maximum development of sun-spots (Secchi and others offer testimony to this also), and the striæ gathering about the setting sun. One of the three conditions (clear sky) being occasionally absent, the glows would naturally be intermittent. The frost particles in the upper air would produce the colouring by diffraction, and being transparent, would not hide the stars from view. The persistence of the phenomena through many consecutive months is certainly against this theory; and the absence of the rain band is a very strong objection, but not so strong as it looks; because frozen water does not produce the rain band like aqueous vapour: e.g., before a light rain a rain band of 70 degrees has been observed, but before a heavy snow storm only one of 10 degrees (F. W. Cory).

All this is very plausible it must be allowed, but against it one may ask, Where has the upper-air moisture gone to—whether liquid or ice? It certainly has not been precipitated, and the glows have ceased or are ceasing. Russia (Times, 11th July) has been excessively wet, but per contra an English writer says, about July last—“we are passing through a droughty summer, the driest we have had since the famous one of 1868;” and the Times summarizes the great defects of the English climate from 1st January to 26th May as—(1) prevalence of cold searching dry air causing frequent frosts at night and (2) absence of rainfall. Nevertheless, there are many little facts which may be gathered all along the line telling in the same direction as Mr. Hazen goes. Lockyer, for example, mentions that he has seen a green sun through the mist on the Simplon Pass, and every one who has travelled in an alpine land knows the beautiful effect, which I think the Germans call Alpenglûhen, produced by the rays of the setting sun falling aslant through the keen icy air upon glaciers and masses of névé snow. The most gorgeous as well as the softest colours play about the mountain peaks. Every icy crystal coruscates and flashes like a diamond, and the sky all around is flushed with the golden glow beneath it. I once saw this to perfection, standing in the valley of Lauterbrunnen and gazing towards the glaciers of Breithom and the Tschingel Alp. Never will the picture leave my memory! Again there is the strong argument drawn from the accounts given by those who have voyaged in polar regions. There, glows and coloured suns seem to be of frequent occurrence.

Thirdly and lastly we have the volcanic dust theory of which Dr. Taylor says, or rather said, a few months ago—that “there is a general consensus of opinion, now that the microscope and polariscope have been brought into the discussion of the question, that it is the true theory.” It is stated in

“Knowledge” to have been first advanced by G. T. Symons in a letter to the Times of 1st December. But in these colonies the theory was advanced, I feel pretty sure, long before that date. It was repudiated in fact by Mr. Ellery on the 15th November, before the Royal Society of Melbourne, and had been advocated both here and in Victoria some time before that.

It will be well perhaps if we consider in the first place the various points that are in favour of this view.

1. We have already by anticipation considered the certainty that there was something unusual projected into the upper air, and that it was thrown there by the Krakatoa eruption, which, as says Professor Rees, was the greatest eruption on record.

2. The immense quantity of dust ejected from Krakatoa suggests that dust was the thing thrown into the higher atmosphere. As Langley says—“Krakatoa ejected millions of tons which would not soon sink.” Thirty miles away from Sunda the ashes fell in such quantities as to make pitch darkness at noon-day. Java was like Holland in its garb of snow. White dust fell on the decks of vessels from 300 to 400 miles away.

3. Dust in the upper air is not an inherent improbability, because the air at all heights is always more or less charged with quantities of dusty particles, as is easily seen when an isolated ray of light is admitted into a dark room. The motes are always in the sunbeam. The smoke of Chicago was seen, again, on the Pacific Coast, so that dust in the air travels far; and Nordensfeld found the fissures in Greenland ice full of fine dust. Where did this come from if not from the air? In Italy from time to time also there occur veritable showers of sand brought over by the sirocco from Africa. Piazzi Smyth, in 1872, noted such an occurrence in Palermo Bay and, if I mistake not, I have read somewhere that the red sand of the Sahara, carried by the Harmattan and other desert winds, finds its way to immense distances across the Atlantic. Dr. Taylor says that the air on the very summit of Monte Mazo, 9,000 feet high, is full of Bacteria, mould, spores, etc. In fact, though ordinary dust is not carried very high and is always denser in the lower strata of the atmosphere, which was demonstrated by Tyndall' experiments on the Bel-Alp, there is always dust in the air in greater or less quantities. This is clearly shown by the phenomena of radial polarization of light (i.e. the reflection or refraction of light so that it has new properties—sides as they are called of different intensities), and to some extent it is also shown by the beautiful effects which we call twilight and dawn, though the principal cause of these is the reflection of light from the air itself and the clouds and vapour suspended in it. Professor Langley says that there is a zone of dust to the height of three miles all round the earth. In 1878 he wintered on Mount Etna and he

found the air full of dust not local in origin. So Piazzi Smyth also found on Teneriffe at the height of a mile; and on Mount Whitney, in South California, 15,000 feet high, Langley found a sea of dust 6,000 feet deep. Tyndall says, “What mainly holds the light in our atmosphere after the sun has retired behind the earth is, I imagine, the suspended matter which produces the blue of the sky, and the morning and evening red. Through the reverberation of the rays from particle to particle, there may be at the very noon of night a certain amount of illumination. Twilight must continue with varying degrees of intensity all night long, and the visibility of the nocturnal firmanent itself may be due, not as my excellent friend Dove seems to assume to the light of the stars, but in great part to the light of the sun scattered in all directions through the atmosphere by the almost infinitely attenuated matter held there in suspense” (“Hours of Exercise in the Alps”). Incidentally, I must remark that this probably gives us the true cause of the remarkable light that has been observed at midnight in different places during the period of the most intense glows. Given exceptionally large quantities of dust in the air, and assuming the correctness of Tyndall' theory, exceptional luminosity at night-time is just what we should expect.

4. In the next place, notice particularly Mr. Lockyer' argument that the order of the first appearance of the sun-glows in different parts of the earth is, upon the whole, in proportion to the distance of those places from Krakatoa, and therefore such as we should expect if dust were the reflecting or refracting medium of which we are in search. Generally speaking the tropics first witnessed the displays, and first of all those parts of the tropics nearest to Krakatoa. The temperate zones were reached at a later time, and more irregularly—the irregularity of winds in the temperate zones accounting for this naturally enough. I believe, if observations had been carefully made at all the different centres of population, the steady onward progress of the upper-air dust, as it radiated outwards from Krakatoa, would be even more apparent and convincing than it is now,—but to understand that progress thoroughly we ought to know more than we do as to the upper currents of wind in the earth' atmosphere.

5. Dust in the upper air is sufficient to account for sun-glows, coloured suns, and all the other phenomena. In the Loes district of China, where the air is often laden with yellow dust, blue suns are constantly seen. F.A.R.R., a writer in “Nature,” 12th June, says that the weather in the upper air must have been unusual, for ordinarily whatever matter may be there assists the blue rays of light and scatters them, whereas lately the blue rays have been absorbed. Now, a stratum of larger particles than ordinary, 20 to 40 miles high and descending at the rate of 1,000 feet a day,

would produce the effects with which we have been familiar. Such dust would cause a green reflection to be seen, followed by a yellow one. Then there would be competition between the red-arresting upper dust and the blue-dispersing lower air. Lower still the yellow would pass to orange, pink, and crimson—more striking as darkness increased. Perhaps there is some analogy to this decomposition of white light and reflection of certain component rays, in the green colouring of the ocean, which Tyndall regards as resulting from the interruption of the usual green rays by impurities suspended in the water. If this matter be absent and the sea be deep as well as pure, it is of a blue colour. On Mount Whitney, Langley found the dust itself bright red, but the sky, as seen through it, violet; but near the sun quite white. He says red rays are transmitted with the greatest ease through our air, the variation of colour depending on the size of the particles of dust therein contained. Krakatoa may have charged the air, or a belt of it, with dust large enough to scatter the red rays and partly absorb the others. G. F. Chambers, at a January meeting of the Astronomical Society, mentioned the case of the crushing of seaweed by steam machinery at Eastbourne. The engineer of the works there says that he frequently sees the sun blue and green through the fine dust in the air. Dr. Buddle, in “Nature,” on the 20th December, refers to the Frenchman in Algeria, who said one day when looking at the sun, C'est la premiere fois que j'ai vu le soleil bleu, and was informed that the dust from the Sahara was the cause of the novel colour. Ranyard' explanation is this: the particles of dust, when small compared with the wave-length of light or of invisible spectrum, disperse different proportions of red and blue—the larger the wave-length the less the intensity of dispersed light. Usually the colour of the sun is not affected by dust in the atmosphere. But when that dust is much increased in amount, the intensity of the dispersed light is much increased also, and the blue colour of the light between us and the sun affects the colour of the sun itself. Lockyer at first thought that the particles floating in the air were themselves blue and red, and thus that the colours we have witnessed were simply the result of reflection, and Hardwick' “Science Gossip” recently spoke of the blue and red particles remaining suspended in mid or upper air. But no such coloured particles have fallen, and so this hypothesis has been, I believe, generally abandoned.

Ranyard in “Knowledge” refers to the blue sun in the tropics becoming green as it neared the horizon, and sinking red. At Trincomalee, in Ceylon, from the 9th to the 12th September the sun rose green, and continued to be of that colour till it reached the height of 10 degrees above the horizon, then it became blue, and at noon bright blue. During its declining, similar changes were noted, but in the reverse order. The moon was

similarly tinged. According to the size of the particles suspended in the air, would be their effect by refraction upon the rays of light falling on them. Gravitation would naturally bring the heavier particles down first, and so it is conceivable that for some months perhaps the air was filled with strata of dust, whereof the lower were composed of heavier, and the upper of lighter particles. Hence, as it seems to us, the change in the colour of the sun as it mounted in the heavens, and again in reverse order as it descended, was just what it was natural to expect. But if you ask me why the abnormal colours of the sun were not seen daily like the sun-glows, I can only suggest that the dust varied in density and character from day to day according to winds, or that for the most part some counteracting influence, moisture e.g., was at work, so that the coloured sun and moon were rare phenomena. Even the after-glow is a result which dust in the air might be expected to produce; for this species of second twilight is not by any means unusual in the Nubian Desert, and Sir T. Herschell referred it to a second reflection of solar light in the atmosphere (Chambers' Encyclopædia, “Twilight”). I do not know that this circumstance has hitherto attracted any notice in this discussion, but it seems to me that it deserves to do so.

6. Chemists and microscopists have been busy in many different places in collecting from rain- and snow-water the dust brought down from the atmosphere, and have, generally speaking, from the analysis subsequently made, come to the conclusion that the matter so collected is volcanic in its origin; further, that it corresponds to the dust ejected from Krakatoa. M. W. Beyerinck, of Wageningen, says that this is beyond doubt. That rain-and snow-water have for some time back left considerable deposits in the rain-gauge and otherwise seems pretty certain. Whether such sediment has been collected and examined here in Nelson, our local microscopists can tell us. The deposits have been collected at places very wide apart, e.g., Queenstown (Cape Colony), Launceston (Tasmania), Harrow (England), Sandhurst (Victoria), Unaluska (Alaska), and in parts of Norway. But as there is always more or less dust in the atmosphere, and as in some places from local causes, such as iron and other manufactories, dust storms, strong desert winds, etc., it occasionally becomes abnormally charged with additional matter, accurate and careful analysis alone will help us here. Such analysis would need to be made by most skilful experts, or it would not be trustworthy as the basis of argument. But when we find such men as M. Daubrèe, of Paris, and M. Renard, of Brussels, agreeing with Macpherson, Murray, and Diller, as to the analysis and identification of the dust, scepticism becomes less justifiable. The Royal Society of England has recognized the great importance of the investigation, and specially

appointed a Commission to collect all observations and documents bearing thereupon. It appears that the Krakatoa ashes contain abundance of vitreous matter and elongated gas-bubbles enclosed in pumice, volcanic sand, and ashes. The two first-mentioned are almost characteristic, and, if so, make the Krakatoa dust less difficult of identification than one might suppose. Diller says he has found glass always most abundant in Krakatoa dust collected on the Java coast; Macpherson, in dust collected from Madrid, found crystals of hypersthene, pyroxene, magnetic iron, and volcanic glass, all of which Daubrèe found in the Javan volcanic ashes. Verbeek, by microscopic examination of Krakatoa dust on the spot, found therein glass with oval vacuoles, felspar, pyroxene, and magnetite in grains, and octahedra. He also gives the chemical analysis, which I need not repeat. Unfortunately the enquiry is complicated by the circumstance that the vitreous and mineral fragments found near Krakatoa are similar to those found in deep-sea deposits, i.e., such as have fallen from time to time through countless ages from the air into the sea, and which maybe either of meteoric or cosmic origin or both. Further, Mattieu Williams on the 5th December last got a black sediment from 75 ounces of snow which fell at Harrow. In this he found much black oxide of iron readily attacked by the magnet and containing nickel. This, he says, is a characteristic of meteoric iron, and cannot possibly have come from Krakatoa, being too heavy. But too much weight must not be attached to this either, for it is evident that the recent fall of Krakatoa dust in different parts of the world would not be likely to stop the supply of meteoric or cosmic dust that the earth is continually receiving. Altogether, this sixth argument at the present date, in the absence of fuller information, is very difficult to work out, and I do not ask you to lay much stress upon it.

7. The occurrence, as we have already said, of similar phenomena after the volcanic eruptions of 1783 and 1831, furnishes a strong argument for associating our sun-glows with the Krakatoa eruption, and, of the various materials cast out by Krakatoa, dust seems more likely than water or gas to be the operative cause of the colouring.

8. The very gradual disappearance of the sun-glows is what we should expect if the volcanic dust theory were a correct one, for the dust would fall from the atmosphere very slowly—the heavier particles first, then those of medium weight, and last of all, and after perhaps a long period of time, the most minute and insignificant.

9. Where, as in this case, a proposition does not admit of positive demonstration, the occurrence of a number of arguments all tending to show its probability must be allowed to have preponderating weight—particularly if the objections thereto can be satisfactorily answered.

Let us, therefore, in the last place examine the objections that either are or can be, in my humble opinion, urged against the dust theory.

1. The amount of dust that would be requisite to form even a thin belt over so large an area of the earth' surface as is comprised, say, between the 52nd parallel of north latitude (that of London), and the 44th parallel of south latitude (that of Dunedin),—that is to say, roughly speaking, about 150 millions of square miles,—would be so enormous that it is difficult to believe that it could all have been thrown up by Krakatoa. We must allow this to be a forcible objection; but reflect at the same time on what has been said as to the gigantic scale of the operations at Krakatoa, and furthermore consider two points:—(1) That the dusty cover need not have been, as far as I know, very thick, and the third dimension, the thickness of the cover, would materially affect the total amount of its solid contents; (2) That the dust was probably not distributed equally thickly or densely over the whole of the vast area; indeed one can well understand that it hung suspended only in particles, for there were undoubtedly periods of intermission in the displays, and these intermissions may have been occasioned by the temporary absence of the causal medium as well as by other counteracting atmospheric conditions. What do we really know for certain about either the exact amount of dust ejected from Krakatoa or that requisite for producing the phenomena? Very little I imagine.

2. It is said, even if dust had been shot up to an enormous height by the Krakatoa eruption, from its inherent weight it would soon have settled down on the earth again (Professor Rees). Undoubtedly the heavier particles would have done so, and actually did so; but in proportion to the minuteness of the particles in all probability would be the distance and time they would travel. The heavier dust fell at Krakatoa, but we have read of thick dust falling on a vessel at sea many hundred miles away. One writer suggests that the air may possess some viscosity which would check the tendency of the dust to settle. Moreover, Professors Preece and Crooke say that minute particles ejected into the upper air would be negatively electrified, and therefore repelled by the earth and by one another (which, by the way, would account perhaps in part for their rapid diffusion) and so would probably remain at a high level for a long time. Besides the lower strata of the atmosphere itself might be denser than the foreign matter of the upper strata—even though not composed of exceptional gases ejected from Krakatoa. Rollo-Russel points out, too, the vesicular nature of pumice, each particle really consisting of a bubble of glass—which, when shot high and removed from the action of vapour and weather, would be easily carried to the most distant parts of the globe (“Hard. Sci. Gos.”). Very fine dust

of any kind settles from the air on a surface warmer than itself very reluctantly. Once in the higher air, therefore, dust would be perhaps able for a long time to resist gravitation. Professor Trowbridge believes that the upper currents of air would tend—by their motion, I presume—to keep the dust in suspense. Of these said upper currents our knowledge, as Lockyer remarks, is really very limited; but the investigation of these sun-glows and a careful comparison of the dates of their appearance in different places may perhaps extend our knowledge in this respect as well as in others. And this brings up

3. The actual dispersion or diffusion of the dust as a great difficulty. As Hazen says, the currents of the upper air, if they caused the diffusion, must have acted in opposite directions, whereas we have always understood that the upper air moves steadily in one current from west to east. In answer to this I can only repeat what has been said. Nothing is certainly known about the higher aerial currents. And as I have just now ventured to suggest incidentally, the electrical repugnance of the particles of dust to one another may have had something to do with their diffusion.

4. Hazen regards the rapidity of the dust dispersion as a stumbling block. The upper currents or something else must have carried the dust 12,000 miles (to Barinas, Venezuela) in 150 hours, i.e., at the rate of 80 miles an hour; whereas observations on Pike' Peak, 14,134 feet high, show the current there to be running at the uniform rate of 20 miles an hour. To this the reply at once occurs: an observation at the height of, say three miles, is not conclusive as to the rate of the wind throughout the upper regions, and we do not know at what height the line of dust was spread out. Moreover, Symonds thinks that the rapidity of the westerly dispersion may perhaps be accounted for by the rotation of the earth from west to east, so that the dust was as it were left behind by a process analogous to that which causes the trade winds in the tropics. But this would, it appears, only account for a progress of 440 miles a day, whereas that to Barinas was at the rate of 1,700 miles a day. Then, again, Symonds suggests that perhaps the dust got quite out of the earth' atmosphere, and so was left a whole hemisphere behind, which idea Cowper Ranyard will not entertain at all, and I confess that I do not understand it. That the dispersion of the unusual material in the upper air (for the difficulty is the same whether dust was the material or not) was exceedingly rapid in some directions, particularly to the west, is certain. But, perhaps, the earth' rotation, the upper-air currents, and the electrical repulsion combined, were sufficient propelling causes. It is, you will notice, in connection with this branch of the subject that accurate observations of the dates of the first appearances

of the sun-glows in different places are so important; e.g., we could tell the rate of progress of the dust to us in Nelson pretty well if we knew for certain when the glows first appeared in Port Darwin, Brisbane, Sydney, and Nelson. Unluckily people did not attach much scientific importance to the phenomena when they first appeared, and so made no notes.

5. The intermittent nature of the phenomena, also, Hazen thinks, is against the supposition that dust is the cause; but I see no force in this. It is only natural that the appearances should vary in intensity with varying meteorological and atmospherical conditions, and, as I have already said, the dust may not have been very evenly distributed.

6. Hazen' last objection is that ashes are opaque, and the appearances indicate a transparent something. Another writer (in the “Sci. Amer.”) puts this objection in this way:—He says, earthly or lunar volcanic dust cannot be the cause of the glows at the rising and the setting of the sun, or the ordinary light from sun, moon, Mars, Jupiter, etc., would be affected. But there is no diminution of this light. Therefore the cause of the glows lies far beyond the sun' orbit.

But are the premises in this argument true? Have we not already stated that a haziness of the sun has been observed at several places—e.g., at Seychelles and Tokio? And is not a change in the colour of the sun and moon virtually a change in the intensity of their light? Moreover, although individually the particles may be opaque (which their intense character leaves open to doubt), collectively they do not constitute an opaque mass, being exceedingly minute and widely scattered. Also, before we can attach much importance to this objection we must know accurately what the photometer has to say on the subject. As the writers quoted do not tell us this, we may presume they speak from conjecture only.

7. Some may say that the volcanic force of Krakatoa, however great it might be, would scarcely be great enough to force up such a vast amount of ashes to so great a height. It is generally considered that the dust cloud, or whatever it was, lay very high indeed; but what gigantic forces were at work at Krakatoa! An electrical force was in violent action also, for simultaneously with the outbreak all the telephones in Singapore were unworkable, and on one line reports like pistol shots and a mighty roar were continually heard; and again the heated air, emanating from the crater of Krakatoa, would carry with it to unknown heights vast volumes of dust, etc.; and incidentally I may mention that the heated air of the tropics ascending as it always does, to be replaced by cooler air from the north and south temperate and polar zones, would account to some extent for the dispersal of the dust.

8. Again it may be asked, if dust were the cause of the sun-glows, why were they not seen in the daytime? Whatever foreign material in the upper air caused the glows, this difficulty would remain the same. Probably the greater perpendicularity of the sun and therefore the greater intensity of his light during the daytime account for the glows then being absent. Towards night and morning, also, his rays would traverse a greater length of the dust-laden upper air.

9. Lastly, if dust caused the green and blue suns and the air is always dust-laden, why do we not have such phenomena more frequently? Ordinarily the dust in the air is not very high, and the particles near the earth are large, and aqueous and gaseous vapour counteract the effect of dust by absorbing more light at the blue end of the visible spectrum than at the red.

In conclusion, there is no reason, as far as I know, why we should confine ourselves to any one theory. Perhaps many causes have co-operated accidentally to produce one phenomenal result. If we must have only one theory, I incline to the Krakatoa dust theory, because I think the balance of evidence is in its favour. In any case, whatever the upper-air medium may have been, it certainly seems to have emanated from the Straits of Sunda during the memorable eruption of 27th August. However, when we come to speculate upon the nature of the substance thrown into the upper currents, difficulties and objections and doubts meet every suggestion or supposition. We shall, therefore, if we are wise, guard ourselves against all dogmatism on the subject, particularly as fresh facts and fresh theories are coming forward daily.

Even as I write there appears, in the notes on popular science by Dr. Taylor in the pages of the Australasian, a modification of the aqueous vapour theory which is well deserving attention. It appears that Dr. Prince, a well-known meteorologist, thinks the phenomenal sunsets are due to the crystallization of saline particles from masses of sea-water ejected in the form of vapour into the upper air by the Krakatoa eruption, and he argues that the greatest displays having occurred in the coldest weather, can only be accounted for on the theory that the crystallization of saline products is a great factor in their production. So long ago as last January, M. Beyerinck, of Wageningen, collected from rain which fell during a storm succeeding a fine after-glow a remarkable quantity of common salt. Again, I find in “Nature” of 3rd July, that M. Gay, in a paper read before the Paris Academy of Sciences on the 23rd June, not only connects the recent sun-glows with the Krakatoa eruption, but thinks that the persistent rains of this year are owing to the same cause, directing attention to the fact that after the volcanic eruptions of 1788, 1831, 1856, and 1862,