When we were attempting to describe in some measure the region of space from which the sun obtained the nebulous matter out of which it was formed, we found that it would produce a nebula somewhat resembling a most gigantic starfish, with arms or legs stretching out from it in every direction, which might be likened to mountain-peaks rising from a tableland or range of mountains; and when we began to condense the nebula we concluded that these peaks would very soon, comparatively, be left behind the main condensation, owing to their being more under the influence of the attraction of surrounding suns. And we might then have added less under the attraction of the main body, on account of its gradually increasing distance arising from its greater rapidity of contraction. Now, we propose to return to these portions of the sun's property so long left out in the cold, to think of what in all probability became of them, seeing that they must all have had somehow a part of some kind to take in the formation of the solar system.

First of all, we have to form some idea, however vague, of their number, which may be divined to a very limited extent from the following considerations: We see, from Table VIII., that the sun's sphere of attraction extends to more than 4000 Neptune distances in the direction of a Centauri, the star nearest to the earth, which corresponds to 11 billions of miles. Then, although we have said, in Chapter XV., that instead of there being a peak on the nebula in that direction there would be a deep hollow in it, we shall proceed to find out what might be the diameter of the base of a peak at that distance supposing it to be somewhat in the form of a cone. We know that the moon does more than eclipse the sun, which is 867,000 miles in diameter; so, for facility of calculation, we may suppose that it eclipses a portion of space at its distance of 1,000,000 miles in diameter. Consequently, the base of a peak such as we are measuring would be eclipsed were it 129,000 millions of miles in diameter, and then only. Moreover, we have deduced the diameter of the base of such a peak from one diameter of the moon; so that wherever we see two stars only one breadth of the moon from each other, there we have room for at least one peak with a base of the above diameter. Last of all, when we come to think that there are as many as six to seven thousand stars visible to the naked eye, and of the intervening spaces between them, we have to conclude that the number of peaks surrounding the original nebula before they began to be left behind, or cut off, must have been almost beyond our conception; more especially if we look at Table VII., where we see that the star Canopus is 25 times farther from the sun than a Centauri. We are accustomed to look with wonder on the volcanic peaks of the moon, but they can do nothing more than give us an exceedingly faint representation of the original nebula seen from an appropriate distance outside, when it had begun to contract more rapidly than the peaks could follow it; seeing that we are comparing a diameter of 2,160 miles with one really almost infinitely greater.

Finding ourselves, then, with an innumerable host of peaks, or cones, of cosmic matter on our hands, we have to think of what can be done with them, and we begin by saying that the use to be made of them was suggested to us when we discovered the jagged nature of the domains of the sun. Some of them have been most probably swallowed up in the formation of the sun, and could we believe in the plenum of meteorites in all space, that has been fancied to exist by some physicists, we might derive its origin from a part of these peaks; but if there can be such a plenum in space, its origin might be much more naturally derived from a suggestion made in a former chapter, at page 258, to which we shall refer presently. In the meantime, looking upon the multitude of comets, meteor-swarms, etc., which revolve around the sun, or are supposed to exist somehow in its neighbourhood, it is very natural to entertain the belief that they have been made out of some of the most important peaks—or the refuse from them—that must have formed part of the original nebula. To deal with all of them when we cannot number them, or even with the six of Table VIII., about which we actually know something, is out of the question, so we shall only try to show what could be made out of one of them.

Confining ourselves, then, to the peak of a Geminorum, whose collecting ground had originally reached to 24,000 Neptune distances, or 67 billions of miles—this being the point of space where the attractions of the sun and that star balance each other—if we suppose it to have been contracted till its base was of the same diameter, and its distance the same from the sun, as that of the base of the peak we measured not many minutes ago, 129,000 million miles, and 11 billions of miles, respectively, we can easily conceive that its height may have been 10 times as great as the diameter of the base, or more than 1¼ billions of miles. Here then we have in the direction of only one star a mass of cosmic matter out of which something more than a comet, even of the grandest known to modern astronomy, could be made. Of its tenuity, all that we have any necessity to think is, that it would be much less—i.e. more dense—than that of the original nebula.

Beginning then with the dimensions we have just stated, we know that the attraction of the nebula would draw the matter of the base-end of the peak more rapidly towards itself than that of the apex-end; we know also that there would be different rates of contraction going on in different parts of the length of the peak—for the same reason we have given for the peaks being cut off from the nebula; so that the condensation throughout its whole height, or length, would be proceeding at different rates at different places, which would certainly divide the peak into several parts, perhaps into many. If now we suppose that the leading part of it—the one nearest to the nebula or sun—or even the whole of it, formed itself into a comet, it is not difficult to see that it might have a tail infinitely longer than any comet the length of whose tail has been measured.

There can be no doubt that in the whole length of the peak the action of attraction would be exactly the same as we have found it to be in the nebula itself; that is to say there is no reason why it should not come to be a hollow cone—comets are reported to be hollow in most cases—condensed into layers, and to revolve on their axes throughout a great part at least of where their diameters are greatest. This mode of formation seems to throw light on some of the phenomena that have been observed in comets. We have just said that our peak would be divided into several parts, so if we suppose the leading part of it to have been made into a comet, we can see why its tail should have the appearance of a hollow cylinder; and there might be no reason why the second division, or even the third, should not become a comet also. Then for further divisions, where the diameter came to be too small to make a comet, its matter might have formed itself into a meteor-swarm, and account for the fact of some comets and meteor-swarms revolving round the sun in the same orbits; perhaps even for some of the observed meteor-swarms being denser at one part than another, owing to two or more of the sections of the peak following each other at some distance. We have to notice, after what we have just said, that it is quite possible that if the different sections of our peak did come to revolve round the sun, their perihelion distances might be so different that it would be impossible to trace any connection between them and the peak from which they were derived. But if we were to attempt to set forth all the explanations of the phenomena of comets and meteor-swarms that have occurred to us, there would be no end to our labour.

Passing now from one to the whole host of peaks, we have seen that at one time they projected from all sides of the nebula; it is clear, therefore, that the bodies formed from them must have fallen in towards the sun from all directions, which is exactly what they have been found to do. Then, if we think of the multitude of them there would be, we have also to think that there would most certainly be collisions among them, which would smash them to atoms, and thus help to make the plenum, or host of independent meteorites that are supposed to exist, or would be swallowed up by the sun in mouthfuls. Others might coalesce, which they could only do through coming in from slightly different directions and with nearly similar velocities; and they would thus account to us for comets with a plurality of tails. Again, looking back to what we have just said of the form that might be assumed by the leading end of the peak a Geminorum, which was suggested by Donati's comet, we could imagine another, the same in almost all respects, coalescing with it, and between the two showing us how Coggia's comet was formed. Furthermore, with respect to one of the gigantic comets with endless tails: If we suppose it to rotate on its axis, and to be not so smooth on its outside as a cone formed in a turning lathe, we could account for the light from the sun reflected from it having an appearance of flickering; and, were the outside very rough, for the reflected light flashing from millions of miles of its length in a few seconds.

All this about nebular peaks, comets, etc. formed from them, will, far more than likely, be looked upon as imagination or speculation run mad; but if it is looked into properly, it will be found that no part of it is based on assumption; farther than that, the universe has been formed out of cosmic matter of some kind. There is no step in the whole process, from cosmic matter to the sun—even myriads of suns—that does not conform to what are generally called the laws of nature; whereas it is not difficult to show that some other speculations on the same subject have never been carried beyond the stage of conception.

When thinking of how comets might be formed, we could not help thinking of their orbits and periods of revolution. It was easy to see that their orbits depended on where, and how far, they came from; that the where might be from any and every direction, and that the how far would be the principal element in their greater or lesser ellipticity, which could only be determined by measurement; but their periods of revolution, as far as we can see, could only be determined by observation, which would involve the study of several revolutions. On these points the data we have been able to collect are not very satisfying, neither are they given to us as very reliable, except as to those whose orbits have been often observed and measured; and even among these the orbits are said to vary, and some of the comets to disappear altogether. Again, some of them are said to have a disposition to become associated with particular planets; and yet again, some people have gone the length of supposing that they have been ejected from some of the planets. To us it seems much more rational to suppose that the known periodical comets have been made out of part of the multitude of peaks which must have surrounded the nebula at one time, if the sun was formed out of nebulous matter, subject to the attraction of similar matter surrounding it on all sides. It seems to be only a way of getting out of a difficulty to suppose that matter ejected from, say the earth, with a velocity of 7 miles per second would be freed from its attraction, that it would be involved somehow in the sun's attraction, and that it would revolve thenceforth round the sun like any other wanderer; because we cannot see what would stop its progress upwards, so to speak, from the earth after getting beyond its control, or communicate to it at the right height and time, the exact velocity required to make it revolve for ever afterwards round the sun; nor, supposing the sun would have nothing to do with it, where it would go. When it left the earth, it might have a direct motion of near one-third of a mile per second derived from its rotation, and also one of 18 miles per second due to the revolution of the earth round the sun. It might also be ejected in a direction exactly away from, or directly towards the sun; so we should have two very different cases to reconcile in order to set up the theory of ejection of comets from planets, and of their being involved somehow in the sun's attraction. It presents us with a very strong case for calling for either the immediate intervention of some power other than what we conceive attraction to be, and of which we know nothing physically, or we have to trust in manipulation of which we have no very exalted idea. We prefer to look upon the formation of all comets as derived from the peaks we have been treating of, or, if that is inadmissible, from shreds and patches of the original nebula; where no immediate intervention, or instant application, of supernatural power is required, but only the even and tranquil operation of original design.

For comets larger, and which travel to greater distances, than those alluded to above, it is very difficult to get data on which we can form satisfactory calculations of the lengths of their orbits and mean velocities of revolution, for there is almost always awanting some one or more of their elements, or totally different statements given of their value; but we think we have found a few from which we can collect data sufficiently accurate to enable us to show that there is no necessity for going beyond the domains of the sun, as described by us in a former chapter, to account for any one of the comets which have been taken notice of in astronomical history; and still less necessity to suppose that any of them have wandered, or been shot forth, from some neighbouring star into the solar system.

From the data we have been able to collect it would appear that when a comet comes to have a period of over 70 years, it is either too far removed from the sun at its aphelion passage, or its mass is too great for it to be perturbed by the attraction of any of the planets. For instance, we have Halley's comet, which has been observed for not far from 2000 years, whose period has averaged very close upon 77 years during the whole of that time, showing that it has not been perturbed to any appreciable extent when near its perihelion passage. No doubt 2000 years is a very small period of time to judge from, and its aphelion distance being only 3,258,000,000 miles, it might be influenced to some extent by some planet, so we can hardly count upon its being permanently exempt from perturbation. Indeed, Halley himself supposed that its velocity of revolution had been considerably increased when it was in the neighbourhood of Jupiter in the interval between 1607 and 1682; but if it was so, there must be some counter-perturbation which restores the balance so as to make the average period of 77 years. Looking over the register of its appearances, we find that in its re-appearances of the years 66 and 1758, the period was about 75 years, and that in those of 451 and 1066 it was 79 years; so that if there are perturbations, we must claim that there are also compensations. Seeing, then, that we can find no evidence to the contrary, we may suppose that when the periods of comets, and, perhaps more especially, when their aphelion distances reach to beyond—and the farther the more so—the orbit of the most distant planet, they may be looked upon as not being liable to be seriously perturbed by any of the members of the solar system, until something to the contrary had been proved. Following this idea, then, it occurs to us that something may be learnt from their mean velocities in their orbits, as will be seen from the following very small list of those we have been able to submit to calculation, which form the accompanying

TABLE XI.— Showing the Mean Velocities in Orbit of several Comets.

These orbital mean velocities per second have been calculated from aphelion distances as diameters and from circular orbits, which probably give results rather lower than would be derived from elliptical orbits—were they known—but on the other hand, the perihelion distances have not been taken into account in fixing the diameters—because they were unknown—so the error will be so far compensated, if not altogether.

We know that the mean velocities in orbit of the planets decrease as their distances from the sun increase, and our table, as far as it goes, leads us to believe that the same holds good with comets whose aphelion distances are comparable to those of the planets, in being measured by hundreds of years or less of revolution; but with those whose periods are measured by thousands of years, the same rule seems to fail. One thing, however, that we seem entitled to believe is that, generally speaking, the greater the period of revolution of a comet is, the less will be its mean velocity per second in its orbit. It will be observed that the average mean velocity of the three remote comets in the table is only 0·83 mile per second, and it is by no means unreasonable to suppose that the average mean velocity per second of any number of comets whose aphelion distances are greater than the highest of those in the table, is not likely to be so great as the average of the three; on this understanding, then, let us take, or suppose, one whose mean velocity in orbit per second is only one mile, and look into what may be learnt from it.

Going back to the peak of a Geminorum which we supposed, at page 321, to be condensed to 129,000 million miles in diameter of base, its height 1¼ billion miles, and distance from the sun 11 billion miles, we may take a comet formed from it as an example. If, then, we suppose the leading part of it to have been formed into a comet with that aphelion distance—11 billion miles—and other dimensions suitable to its new condition; taking its mean velocity in orbit at 1 mile per second, we find that its period of revolution might be 1,200,000 years, or three times greater than that of the comet of 1882, namely 400,000 years, mentioned by Mr. Chambers as being not very reliable, probably because its angles in orbit could not be measured with sufficient accuracy. Then, when we think that the sphere of the sun's attraction in that direction—of a Geminorum—extends to 67 billions of miles, and that there are stars more than 6 times farther off, e.g. Canopus, see Table VII., we see that a supposed comet might have an aphelion distance equal to that; and were we further to consider that were its major axis 67 billion miles long, including aphelion and perihelion distances, and that it went straight from the one end of it to the other and back again, its period of revolution, if it could be so called, would be 8,500,000 years; that is 20 times greater than Mr. Chambers's doubtful 400,000 years for the comet of 1882. There seems, therefore, to be no necessity for the solar system sending its cometary produce to a foreign market; and our mechanical imagination is not sufficiently vivid to allow us to conceive what kind of potential energy even Jupiter can have to give an impetus to a comet, great enough to send it flying to so great a distance. What velocity would it have when it left the sun? And what would remain in it to carry it over the debatable land between the sun and a distant neighbour? Or are we to believe that all the solar system's produce of that kind is only sent over the channel, as it were, to our nearest neighbour, a Centauri? Conceptions of that kind are too elevated for us, and we must leave them alone. Mr. Chambers expresses doubts as to the determination of whether the orbit of a comet is elliptical or parabolic when its period of revolution is measured by hundreds of thousands of years, and we think we are safe in following him until actual proofs are presented. If the comet of 1882 never comes back, we may then believe it has gone elsewhere.

Having used up all the nebulous matter in the sun's domains, as described at the beginning of Chapter XV., or at least shown how it may have been, or may yet be, used up, we have now only to make a few remarks to prove that our description of the said domains is not by any means fanciful. It matters very little whether the solar system was begun to be brought into existence at the same time as the surrounding systems or before or after them. What is certain is that the sun's sphere of attraction among its neighbours is bounded, at the present time, just in the way we have taken to describe its domains. How they were filled with cosmic matter may be disputed, but filled they must have been somehow, if the solar system was formed out of a nebula; and the way adopted by us was the only one that occurred to us when we began to reconstruct the original nebula. Since then we have had time to reflect on our work, and to see how it points out the simplest way that can be conceived, which may be expressed in the few following words. We may suppose that the ether was the primitive matter, as we have done at page 258, and that the whole material universe has been formed from it and through it. This idea will assist physicists in forming their theory of a plenum of meteorites or meteoric matter, if such they choose to call it. It will also enable us to complete the circle of our notions with respect to matter. We believe that we can neither destroy nor produce the smallest portion of it, although we can change its form. Thus, looking upon the ether as primitive matter, we can understand how the solar system could be elaborated from it; and how, after having accomplished the purposes for which it was brought into existence, it may again be resolved into the primitive element out of which it was made, ready to take its part in the evolution of some other system with, perhaps, a new earth "without form and void."

We have now to direct our thoughts, as far as we can, to the mass, which furnishes the really effective power of the sun as the ruler of the system; and, first of all, we have to think of what are the real active elements which form that mass. Hitherto we have looked upon them as all included within a diameter of 867,000 miles, but now we have to take notice of the clouds of meteoric matter which have been supposed by some astronomers and physicists to be revolving round the sun and continually raining into it; and of the enormous atmosphere which surrounds it. With regard to the former of these two elements, we shall compound our ignorance by looking upon it as a merchant does on his account of Bills Receivable, as not being available in the case of a sudden demand for cash, and therefore as not forming a part of the mass, any more than as the attraction of the earth aids the sun in its management of the planet Neptune; the same as the bills receivable strengthen the credit of the merchant. But with regard to the second element of the two, we must recognise that it forms part of the mass and power over the whole of the system, and from all that is known about it we are not authorised to look upon it as a negligible quantity. It so happens that the only thing we have to which we can compare it is the atmosphere of the earth, and we immediately find that there is absolutely nothing to be learnt from such a comparison. We know that one-half of the weight or mass of the earth's atmosphere is contained in a belt of 3½ miles high above its surface, so that double the volume of that belt estimated at atmospheric pressure gives us the true measure of its mass. This mass, when reduced to the density of water, and compared to that of the earth as we have dealt with it all along, turns out to be about 1/824,000th part of it; and were we now to add that to the earth's mass we have been using, its mean density would be 5·66065 instead 5·66 times that of water.

Now, let us suppose the sun to have an atmosphere of the same kind as the earth's: Seeing that the force of gravity at its surface is about 28 times greater than it is at the surface of the earth, a belt around it which would contain one-half of its mass would be 28 × 3½ = 98 miles, or say 100 miles thick. Dealing then with this dimension in the same manner as we have done in the case of the earth, we find that its supposed atmosphere would be 1/836,000th part of its mass, which, if added to the mass we have used for it, would make its mean density 1·413016 instead of 1·413 times that of water. Then again, if we suppose the earth's atmosphere to extend to 100 or 200 miles above its surface, the supposed atmosphere of the sun would extend to 2800 or 5600 miles above its surface, according to which of the above heights on the earth is adopted; whereas the highest of our authorities say that the corona, or apparent atmosphere, extends to at least 350,000 miles from its surface.

It would appear then that there is no analogy whatever between the atmospheres of the sun and the earth; but there must be some analogy, because the law of attraction cannot be suppressed at the surface of the sun; neither can any vaporous matter near it cease to be attracted in the same proportion as it is at the surface. Our atmosphere causes a pressure of 29½ inches of mercury at the earth's surface, and the attraction of the sun at its surface must cause a pressure equal to nearly 28 times that without fail, i.e. 420 lb. per square inch instead of the 15 lb. of the earth. We know that some spectroscopists believe that the pressure at the surface of the sun is sometimes as low as it is at the surface of the earth, even lower; but we require an explanation of why it is so. At the surface of the sun one second of arc corresponds to a height of 450 miles above its surface, and Mr. Proctor states in his "Sun," page 295, that if even "two or three hundred miles separated the lower limit of chromatosphere from the photosphere, no telescopes we possess could suffice (when supplied with suitable spectroscopic appliances) to reveal any trace of this space. A width of two hundred miles at the sun's distance subtends an arc of less than half a second; and telescopists, who know the difficulty of separating a double star whose components lie so close as this, will readily understand that a corresponding arc upon the sun would be altogether unrecognisable." We can understand this, and perhaps find an explanation for ourselves.

According to our supposition that the sun may have an atmosphere similar to the earth's, at one hundred miles in height it would be reduced in pressure to 14 atmospheres, and, extending the analogy, at 2800 miles high the pressure would still be equal to one-eighth of 28 atmospheres, or equal to something less than 2 lb. per square inch at the surface of the earth; so that if spectroscopists have measured the sun's atmosphere at the disk, and found it to be lower than the earth's at its surface, their results must have been caused by some fortuitous circumstance which they did not notice at the time; because the force of attraction at the surface of the sun can never be overcome except by some counteracting force, which, if in the form of a vapour, or what we call a gas, issuing from its interior, would increase rather than diminish the pressure. We know that in the heart of a cyclone on the earth there is sometimes a vacuum sufficient to explode (pull out the walls of) houses near which it passes; and, at the same time, we know, more or less, what heat the sun sheds upon the outer atmosphere of the earth, and also the rate of rotation of the earth in the regions where the fiercest of these cyclones occur, the only two causes which can produce them. Now, if we compare these causes in the two bodies, that is, the earth's rotation of about 16 miles per minute and the sun's of, say, 60 to 75 miles per minute, and the temperatures of the sun and the earth at their respective surfaces, we can imagine that in the heart of a cyclone on the sun there may be a vacuum much nearer absolute zero than there can be in any one on the surface of the earth. If then the spectroscopists, without knowing it, have caught the spectra of the hearts of cyclones, we can conceive them to be right, otherwise no.

Again, we know that when big guns are fired off partial vacuums are formed near them, sufficient to cause disaster to windows, doors, and even walls of houses too near them, but whatever we may have said of force sufficient to produce explosions in the sun, we have never believed that matter is ejected from the sun by explosions. We have supposed the sierra, or chromosphere, to have oozed out through its pores, sometimes to less, sometimes to greater heights, like steam from an open boiler, and the prominences to be eruptive, neither of which modes could produce anything approaching to vacua in their neighbourhoods. There can be no resemblance between the ejection of matter or gas from the sun and from a cannon, but there is between the ejection of vapours and the escape of steam from the safety-valve of a closed steam boiler; both of them continue to pour out their vapours till the pressure within falls down till it is equal to the resistance to their escape; there is no explosion, therefore no vacuum, appreciable at least, in the neighbourhood. There may be surrounding matter drawn up by the velocity of the outward current, but that is all.

Notwithstanding all this, we see no reason why the sun should not have an atmosphere of exactly the same kind as the earth's, composed of exactly the same kinds of gases, including vapour of water in some part of it, though, perhaps, far removed from the photosphere. Every other element found on the earth can be found in the sun, and so it is not unreasonable to suppose that the same kind of atmosphere may exist upon it; we have only to acknowledge that its conditions must be somewhat varied, all the difference being that the atmosphere of the sun must be heated up to the temperature of the photosphere where it comes in contact with it, while that of the earth is only of the temperature of the earth at its surface. In the case of the earth, if this were at a white heat, one-half of the weight of its atmosphere would not be comprehended in a belt around it of 3½ miles thick. That balance of mass might take place at a height of even hundreds of miles—we have no means of calculating how high—and still its pressure at the surface would be the same as now, as long as the earth's attraction remained the same; so must it be with the sun. Instead of limiting its height to 5600 miles at the utmost as we have done above, it would be no stretch of imagination to suppose that it might extend to ten, twenty, or more times that height. In addition to this we have to take into consideration that the sun's atmosphere must be swept up to something far beyond 5800 miles high by the whirlwinds created by the velocity of rotation at its surface, the same as we saw the earth's might be when we were explaining how an aurora could be made to glow at heights far beyond what we were accustomed to believe its atmosphere could reach. Adding, then, together these two motive forces for elevating the atmosphere of the sun, it would be a bold assertion to say that it cannot have one exactly similar to the earth's, reaching up to the height of 350,000 miles mentioned a few pages back. And now, having got this length, we may venture to assert that the corona of the sun is made up of this atmosphere, and of the vapours of the elements thrown out from its interior, somewhat in the manner we have described in last chapter; to which we have only to add that the bubbling up of vapours all around the sun, which produces the sierra or chromosphere, would not be interfered with in any way by the tremendous commotions which we have shown must be produced between the surfaces of the sun-spot zones and the centre; and that the projection of the high prominences would assist in elevating the aeriform atmosphere.

If then the sun has a compound atmosphere of this kind, it must be considerably more dense, proportionately, than that of the earth, and will consequently form a greater addition to its mass than we have found would be made by its airlike atmosphere. But, whatever density has to be added to it on that account has to be subtracted from the interior having been ejected from thence; because, in whatever manner its mass has been calculated in respect of the other members of the system, the total amount must turn out to be always the same. We have always estimated its mass from a diameter of 867,000 miles, which gave us a volume of 341,237,638⁹ cubic miles, so that if we now include in the diameter the 350,000 miles height of the atmosphere, we get a volume of 2,053,500¹² cubic miles, which is as near as possible six times the volume in which we had to distribute the volume of the sun. How to do this, we know not. We cannot fix the region of greatest density in the same manner we have done at page 221, but we know that it must be considerably nearer to the surface of the photosphere than we have there placed it; and of one thing we are sure, and that is, that the densities we have named for that region and the outer and inner surfaces of the shell, at page 223, must be less than those there expressed; how much we cannot calculate, but we have certainly found that the limits must be lower, and that most probably there is no matter in the sun exceeding the half of the density of water.

Whatever the composition of the sun's atmosphere, or corona if that name be preferred, may be, spectroscopists have found in it a spectral line derived from some substance totally unknown to science. Now, looking back on our work from almost the very beginning, it seems to have been gradually borne in upon us that this unknown substance is the ether. That it is a material substance we were hardly ever in doubt, and our studies of it have substantiated and confirmed our belief. In our analysis of the Nebular Hypothesis in Chapter VI., after combating the notion that the light of nebulÆ is occasioned by incandescent gas, we showed, by the example of an air furnace, that an incandescent gas is composed of two elements, one consisting of solid matter which takes up and gives out heat and has all the properties of a heated solid or liquid substance, and the other of gaseous matter which, being the element that fills up the empty spaces between the solid atoms of a gas or vapour, only performs the office of carrying the solid part into the furnace. This forced upon us the idea of the gaseous part being a carrying agent, and very naturally to think of its being really the ether, that being the only acknowledged agent for the carriage of light, heat, and electricity, two of which are easily seen and felt, and the third cannot be awanting, in an air furnace. Again, when treating in Chapter VII. of what effect the ether might have on the density of the original nebula, we concluded that its density must be much lower than what we then knew it had been estimated to be, and also that its temperature in space must be lower than -225°; which two circumstances combined showed us that if it is a gaseous substance it must be very different to any gas that had been liquefied up to that time. This we repeated in great part in Chapter XII., calling attention to the peculiarity of its being able to carry a higher temperature than its own—to all appearance—into a "hot box." Then we have dedicated two Chapters, XIII. and XIV., almost exclusively to the study of the ether, and have been led from one stage to another to look upon it as the only substance that agrees with the definition of a gas as given by science; true gas there is; as the primitive and sole element in the formation of all matter and in the evolution of the universe; and what is something more than an unfounded guess, as the mysterious and incomprehensible agent attraction, unfortunately almost universally spoken of as gravitation. And now to conclude: From what we have been able to learn, very slight differences have been found in various spectra of the position of the line representing the unknown substance, but this can cause very little doubt of its always being the same, as spectra often contain several lines of hydrogen, owing most probably to combinations with other substances; and if the ether is the primitive chemical element, there may be slight differences in the position of its line, as shown in all the phases in which we seem to have found it, but they must be slight as compared with the hydrogen lines, because even these must be in some measure, perhaps even great, influenced by the unfailing and inevitable mixture of the ether in their composition.

LONDON: PRINTED BY WILLIAM CLOWES AND SONS, LIMITED,
GREAT WINDMILL STREET, W., AND DUKE STREET, STAMFORD STREET, S.E.