312 



SCIENCE. 



[Vol. XXII. No. 566 



follow that the Moon's atmosphere would be of as perma- 

 nent a nature as the Earth's, for the gain and loss of 

 molecules would only take jslace near the upper limits of 

 the atmospheres, where collisions rarely occur; hence the 

 question of permanency would largely depend upon the 

 extent of the atmospheres surrounding the two bodies. 



The figures tend to show that the Earth would lose its 

 atmosiDhere very slowly, even if plunged in vacuo, and 

 that the Sun's atmosphere may be regarded as practically 

 permanent, even independently of the hypothesis of an 

 interstellar atmosphere. But the impossibility of assum- 

 ing losses to be taking place from the atmospheres of 

 planets without a compensating accession of molecules 

 from the surrounding space is at once evident when we 

 endeavor to trace the past history of the solar system. 



If the Moon ever had an atmosj)here which has now 

 flown off into space, losses of a similar nature must neces- 

 sarily have taken place in the atmospheres of all the plan- 

 ets at a time when they were much hotter than they are 

 at present, especially in the case of so small a planet as 

 Mars. And if we trace the history of the solar system 

 further and further back, we find that, if the planets 

 were hotter and hotter, they must therefore have been 

 parting with their gaseous envelopes at a greater and 

 greater rate, — a condition of things which would render it 

 impossible to account for the initial existence of plane- 

 tary atmospheres. 



i'he nebular hypothesis supposes the Sun and planets 

 to have been evolved by the gradual contraction and con- 

 densation of a nebulous mass of gas. This j)roeess would 

 be exactly the reverse of the flying-off process suggested 

 by a perusal of Dr. Eobert Ball's paper. 



It is only necessary to assume the existence of a distri- 

 bution of matter of excessive tenuity pervading interplan- 

 etary space, in order to account for the permanence of 

 the planetary atmospheres at all temperatures; and such 

 an assumption, taken in conjunction with the kinetic 

 theory, is quite compatible with the absence of any percep- 

 tible atmosphere surrounding the Moon. 



The kinetic theory enables us to compare the densities 

 at different points of a mass of gas in equilibrium under 

 fixed central forces, such as the attractions of the celes- 

 tial bodies. If we apply the theory to the system consist- 

 ing of the Sun, Moon and Earth, we shall find the rela- 

 tive densities given in Table 2, the density of the corres- 

 ponding gas in the atmosphere at the Earth's surface be- 

 ing taken as unity. If we take the density at an infinite 

 distance from the Sun to be unity, the corresponding re- 

 sults will be given by Table 3. 



The assumption on which these results are calculated 

 may be called an "equilibrium theory," isince it takes no 

 account of the motions of the bodies in question, and it 

 assumes a permanent distribution to have been attained, 

 so that the whole of the mass is at a uniform tempera- 

 ture. 



AVhen every allowance is made for the artificial charac- 

 ter of the assumptions, it is still highly unreasonable to 

 suppose that the Moon could have an atmosphere so far 

 in excess of that required by the equilibrium theory that 

 its presence could be detected even by the most careful 

 observations. 



And so far from its being necessary to assume the 

 density of the interplanetary atmosphere to be a millionth 

 of a millionth of the densitj' at the Earth's surface, we 

 should, on the assumption of a uniform temperature of 

 0°C, have to divide the latter density by a million over 

 and over a^'ain fifty-five times, before we had reached the 

 degree of tenuity required by the equilibrium theory for 

 the interplanetary atmosphere in the neighborhood of 

 the Earth's orbit. Taking the number of molecules in 

 one cubic centimetre of air as a million million million 



and employing the figures calculated for oxygen, we 

 should have to construct a cube, each of whose sides was 

 10 ' °° kilometres Ion g, in order to enclose a hundred molecules 

 of a gas of this degree of tenuity. Thus, if we multiply a 

 million by a million and repeat the process sixteen times 

 and then multiply by ten thousand, and take this number 

 of kilometres as the side of a cube and place one hundred 

 molecules of gas inside it and the Earth in the middle, 

 that hundred molecules would be sufficient to make up 

 for any loss that is going on at the surface of the Earth's 

 atmosphere. It is similarly evident from the figures in 

 Table 1 that countless ages must elapse before a single 

 molecule leaves the Earth's atmosphere, and that no per- 

 ceptible equalization is taking place between the atmos- 

 pheres of different planets. 



If we try to compare the atmospheres of different plan- 

 ets, such as the Earth and Mars, the "equilibrium theory" 

 breaks down completely. But it would be highly un- 

 reasonable to suppose that anything like a permanent law 

 of distribution existed between two bodies at such vast 

 distances apart, separated by a medium of such extreme 

 tenuity, and subject to solar radiation and so many other 

 disturbing causes. The molecules of gas flying about 

 in interplanetary space are so few and far between 

 that collisions can only rarely take place between them, 

 whereas any tendency of approach towards a permanent 

 state of distribution must necessarily depend on fre- 

 quency of collisions between the molecules. Hence the 

 rate of equalization of energy among the molecules of so 

 diffuse a medium must be infinitesimally slow, so slow in- 

 deed that practically no such equalization is taking place 

 at all. It is different in the case of two bodies so near 

 one another as the Earth and Moon. Among the mole- 

 cules of gas which at any time might find themselves in 

 the neighborhood of the Moon and Earth, the greater 

 number would be drawn in by the more attractive body, 

 and the moon would not, therefore, be likely to obtain 

 more than her fair share of air, which, as we have seen, is 

 very small in comparison with that allotted by the equi- 

 librium theory to the Earth. 



Table 3 affords some idea of how the density of the 

 Earth's atmosphere would increase with the gradual 

 cooling of the solar system. According to this theory, a 

 similar increase has been taking place in what little at- 

 mosphere there is surrounding the Moon, and at no 

 period of its history has it possessed an atmosphere of 

 oxygen and nitrogen comparable in density with that of 

 the Earth. A decrease of density in a planet's atmos- 

 phere could onlj' take place by the condensation in liquid 

 form of vapors present in it, not by matter leaving the 

 planet. 



The figures given in Table 3 are more than sufficient to 

 account for the comparative rarity of hydrogen in the 

 Earth's atmosphere, but a similar argument would also, 

 of course, require a considerable preponderance of oxygen 

 over nitrogen, which is contrary to experience. But here 

 again we have pushed the equilibrium theory too far. It 

 is highly probable that the number of molecules flying 

 about both in interplanetary and interstellar space is far 

 greater than that given by the accompanying tables, and 

 the inference is that the atmospheres of the planets are in- 

 creasing in density at a rate far greater than that due to 

 cooling alone. Even so, however, the few molecules 

 picked up by the Earth in the course of a year or even a 

 million years may have no appreciable effect on the density 

 or composition of the atmosphere. Hence, while, as Pro- 

 fessor Li vein g asserts, the same chemical elements may 

 be expected to enter into the constitution of all the celes- 

 tial bodies, there appears to be no warranty for suppos- 

 ing them to be in any way regularly distributed as re- 

 gards their relative propoTtions; and on the other hand 



