tively about 382 miles; 7 miles; 3 miles; and rather less than 13 mile per second.

Apart from any resistance of a surrounding atmosphere, it follows that a particle projected vertically upwards (i.e. in the reverse direction) from any of the above-named surfaces, with a speed exceeding that just stated as corresponding to the globe in question, would go on and on, gradually moving more and more slowly, but never coming to a stop. It would run right away and never return.

The average velocity of the molecules, even of hydrogen (in the cold outer regions of the Earth's atmosphere), being only about one mile per second, and those of oxygen and nitrogen and vapour of water from one-fourth to one-third of that value, it would at first sight seem as if there were no chance that the molecules of any one of these gases could thus decamp from the Earth, since it can control a runaway velocity of nearly 7 miles per second. But it must be remembered that the velocities just stated are only average ones. Some of the molecules would at any time be moving much faster than others. They would all be constantly altering their speeds by mutual interchanges of velocity through the medium of their knocks.

It is probable that, at any ordinary temperature, some molecules of every gas would be moving at any moment with sufficient speed to run away from the atmosphere of any body, however great its attraction might be. But the permanence of the atmosphere of such a body would be practically secured if the proportion that might thus escape should be excessively small. And investigations in thermodynamics indicate that no knocks of molecule against molecule, no succession of interchanges of velocity between one and another of the molecules in oxygen or nitrogen or vapour of water would enable those gases to escape, either from the Earth, or from Mars, through one molecule after another decamping in the course of ages. But it appears that all hydrogen may have thus departed from the Earth, and, a fortiori, from Mars. The hypothesis also suggests that no atmosphere at all could be retained by the hundreds of little minor planets, whose power of control is very much less.

In the case of the Moon it seems that the whole of an atmosphere of a composition similar to that of the Earth might have been lost in this way. But it is, perhaps, more probable that the Moon may never have had any appreciable atmosphere at all, owing to the Earth (through its proximity and its greater attractive power) having taken into its own atmosphere any gaseous molecules from surrounding space which might otherwise have gone to the Moon. There are also other ways in which the present non-existence of any appreciable atmosphere upon the Moon or of any free hydrogen in the Earth's atmosphere may be accounted for.

We consider that the hypothesis to which we have thus referred is

interesting, but that it needs further discussion. As regards, however, its application to Mars, there are probable indications of the presence of the vapour of water upon that planet, whether it be there because its molecules are unable to run away, or for other reasons. Those indications have been given by the spectroscope,3 or by telescopic views of the apparent formation of clouds (such as seemed to obscure a region as large as Europe for several days last October), as well as by other effects which may be due to aqueous vapour. At any rate, we may say that the molecular knocks, of which we have been speaking, do not militate against the habitability of Mars, so far as that habitability may depend upon the existence of the vapour of water in its atmosphere.

We may, however, remark in passing that there are many reasons in favour of the supposition that Mars is more likely to have been inhabited in past ages, than at the present time, in spite of its atmosphere, or water, or clouds. Mr. Proctor, for instance, has pointed out that a globe of the size of Mars would cool rather more than two and a half times as quickly as one of the size of the Earth. If the Earth and Mars were in a similar condition 18,000,000 of years ago, Mars would have attained (according to that rate of cooling) to the Earth's present condition in 7,000,000 years, i.e. 11,000,000 years ago; and the Earth would now require 28,000,000 future years in which to cool as much as Mars has cooled during the last 11,000,000 years. So far as regards that consideration, therefore, the probability of the present habitability of Mars must be compared with the probability of the Earth's being inhabited when 28,000,000 more years shall be past and gone

Let us next consider some greater astronomical knocks, to which we will pass on by an illustration connected with those which we have already discussed. We will suppose a small sphere of gas, perhaps an inch or two in diameter, in which the gas has been reduced by a Sprengel air-pump, or otherwise, to the most extreme rarefaction attainable. Its density would then be much less than one-millionth of the ordinary density of air (as in some of Mr. Crookes's experiments with the Radiometer), while the number of gaseous molecules in it would be exceedingly reduced. These molecules would, however, still fly about with great velocities, but their free paths from one successive mutual knock to another would be greatly lengthened. Their knocks would at the same time be far less frequent, inasmuch as the molecules might only be a few hundreds of millions in number.

Now let us imagine this globe of gas to be allowed to expand in vacuous space until it should attain to a diameter of millions of millions of miles. If the molecules could then be much in

* See the observations of Dr. Huggins, Astrophysical Journal, March 1895, p. 208. Proctor's Old and New Astronomy, p. 542.

creased in size and also become of various sizes and very bright, the result might represent so much of the Stellar Universe as the utmost telescopic power reveals. For it is a Universe in which every star is hurrying onwards, like these molecules, with its own proper motion; a motion which is, however, dwarfed in many cases almost to apparent rest by the remoteness of the stars from us. The stars are all alive with movement, ever changing their positions, their mutual relationship and influence, their configuration, their attraction upon each other. This fact vastly increases our interest in these glorious orbs. Each at the same time may have its train of attendant planets. The past and future of our own Sun and its planets may have depended, and may yet depend, upon the Sun's onward travel. Our health, our life, our warmth and cold, may be determined by the locality which the Sun may reach in his unceasing journey of about half a million of miles per day.

These onward movements of the stars are of no small account. The Telescope only shows to us that portion of any such motion which is athwart, or at right angles to, our line of sight as we look at a star. But in 50,000 years to come, which is only as a moment compared with the millions of years which astronomy and geology call upon us to contemplate, those partial movements would abolish the belt of Orion and bring Sirius to be directly under one of Orion's feet instead of far away to the East, while Procyon would be nearer to Orion than Sirius now is. 50,000 years ago the seven chief stars of the Great Bear, instead of forming a plough or wain, appeared from this cause as an elongated cross. At the same time it should be remembered that the Spectroscope also shows the existence in the stars of additional movements, upon an equal scale, directly towards or from us.

And besides all the myriads of bright stars thus seen in Telescope or Spectroscope, there are probably many which have so far cooled as to be invisible. In certain cases (as, for instance, in the periodic occultation of the greater part of the light of Algol) we have decisive evidence of the existence of huge dark bodies, which, for aught that we know, may be very numerous, but which, by their cooling, would not in anywise have lost their onward velocities through space.

If so, it is only natural to ask whether knocks may not occasionally occur between some of these brighter or darker orbs? Possibly. But such knocks would certainly be very rare. However vigorously two bodies might draw each other together by their mutual attraction, they could not finally knock unless they had almost exactly the same velocity, or almost no velocity, in the direction perpendicular to that of the line of their mutual approach. Otherwise, their near approach could only result in their whirling once past each other in sharply curved paths (their outer parts possibly grazing), just as a comet hurries round the Sun. Or it might result

in their revolving round one another, as the components of a double star revolve, in elliptic orbits.

But we cannot deny that more or less direct knocks might, from time to time, take place. And the occurrence of so-called New or Temporary Stars, which involve a tremendous and sudden development of light and heat, may indicate that something of the kind has occasionally happened, and been seen by us. We refer to observations of outbursts of stellar light such as Tycho Brahe saw in A.D. 1572, or Kepler in 1604; or to those more recently seen in the New Stars of 1848, 1866, 1876, 1885; and in the very remarkable Nova, or New Star, in Auriga in 1892. In such cases there may possibly have been some terrific knock or series of knocks.

If so it must, however, be allowed that it is very difficult to account for the very rapid falling away of the light after its first outburst. The direct knock of two dark masses might certainly produce a wondrous brightness, invisible before, as the result of the conversion of the energy of their movements into heat and light. But it would not seem probable that the light of the united mass, thus rendered so brilliantly luminous, would fall in about a couple of months, as that of the New Star of 1892 fell, from the 4th to the 15th magnitude, which means a one-hundred-thousand-fold diminution of light; or that the star of Tycho Brahe could diminish its light a thousand-fold, as it did, in little more than a year.

In 1892 it was thought that two brilliant bodies indicated their presence in the New Star by a double spectrum, and that the displacement of the lines in the spectra, which were considered to be of two distinct classes, indicated that both bodies were in exceedingly rapid motion. If so, they might have so disturbed one another, even without actual contact, as to have produced very important eruptive effects. The spectroscopic observations have, however, received other interpretations. There may, perhaps, have been only a near approach of some dark orb to a bright sun, and a less rapidity of movement.

In any case, the explanation of the observations of all New Stars, including that of 1892, is undoubtedly difficult. Whatever the truth may be, even if in most cases no actual knock of two great globes has occurred, such outbursts of light, owing to the violence of the disturbance produced, must almost inevitably involve a considerable number of what may be termed greater knocks. But all the phenomena exhibited would also ultimately depend upon countless molecular knocks in the light and heat evolved.

If, however, in especially rare instances (of which, perhaps, that of Tycho Brahe's surpassingly bright star of 1572 may be one), it be allowed that an actual knock of two huge bodies, meeting more or less directly, may have taken place, such an event would be very suggestive. It would carry us in thought to an epoch

in the evolution of the Solar System earlier than the nebular epoch imagined by Laplace (and recently illustrated by Dr. Roberts' photograph of the Nebula in Andromeda), when that system may have been a vast Nebula from which the planets and the Sun were subsequently cast off or condensed. It would suggest that such a Nebula may have originated in the terrific knock of two great globes, the joint mass of the two being about equal to that of the Sun.

We think it possible that the light evolved in such a case might be comparatively transient (as with Temporary Stars) if the knock were strong enough to vaporise the whole of the two masses. Certain solid portions, while still unvolatilised, might give forth a great light, but only for a short time, by their incandescence, like the particles of carbon in the flame of a gas, or candle. But subsequently the highly heated gas resulting, when the whole was vaporised, would not necessarily be very luminous; just as the very hot flame of a Bunsen gas-burner gives out little light. When, however, the greater part of the Nebula should contract into a Central Sun, it might become much brighter again.

The great compression of its mass as it contracted would conduce to an increase of its luminosity; and it is a very interesting fact that such a gaseous mass, by the very act of contraction, would necessarily for a long time increase its own temperature, even though that contraction should all the while result from the radiation of heat from its outer surface. Mr. J. Homer Lane, of Washington, some years ago showed that a globular mass of gas contracting by the radiation of its heat to one-half of its original diameter would double its temperature. Otherwise the eightfold increase of pressure outwards, due to the compression in volume, would only be one-half great enough to resist a sixteenfold increase of the inward pressure. We say sixteenfold because the inward pressure would be increased fourfold by the increased gravitating attraction inwards due to the lessened distance of the surface from the centre of the sphere, and fourfold more (i.e. sixteenfold altogether) by the smaller surface over which that attraction would be spread.

Consequently it can be shown in the case of such a body as our own Sun that, at the same time that it contracts through becoming cooler in any region where the density remains the same, it may become hotter where the density is increased through that contraction. It may rise in temperature at a given depth below its surface so as to radiate more heat to the Earth, while the process of cooling from its outer surface continues.

Such an increase of effective temperature in a contracting globe is really due to the attraction of the mass of the globe upon the millions of millions of minute molecules in its gases. The Sun is doubtless in the main a great globe of gas, although so intensely compressed in its central parts that they may be in a thickened or semi