158 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1924 



oXlO-^ gra..; our calculation shows that if these satellites came into 

 being as gaseous condensations in a filament, the gas in this filament 

 must have been anything from one to a million times as dense as 

 lead. Such a conclusion is, of course, preposterous. The only proper 

 conclusion is that these satellites can not have originated as gaseous 

 condensations. 



This conclusion is not surprising, or even unexpected. Even now 

 these satellites, on account of the smallness of their mass, are in- 

 capable of retaining a gaseous atmosphere, whence it follows that 

 they can never have existed in the gaseous state. They must have 

 been born either liquid or solid. 



In this way we come upon the practical limitation to the possibil- 

 ity of endless generations of satellites being born. Primarily it is 

 that after a time the satellites would be too small for their gravita- 

 tion to hold them together. A brief reprieve from the operation of 

 this law is afforded by the possibility of the matter liquefying or 

 even solidifying before it scatters into space, and it is probably 

 owing to the operation of this reprieve that all the satellites of the 

 planets, and probably also the smaller planets themselves, owe their 

 existence. 



What of our earth, which interests us above all other planets ? Its 

 present mass is rather too small to have been born out of a purely 

 gaseous filament, but we must remember that if it were born gaseous 

 a large part of its mass might be immediately dissipated away into 

 space, the present earth representing only a remnant of a once much 

 more massive planet. This line of investigation leads nowhere. A 

 more promising line of attack is through a consideration of our satel- 

 lite, the moon. The more liquid a planet was at its birth the less 

 likely was it to be broken up tidally by the still gaseous sun, but, in 

 the event of this breaking up taking place, the ratio of mass between 

 satellite and primary would be much nearer to unity than in the case 

 of a wholly gaseous planet. Thus, as we pass from planets which 

 were wholly gaseous at birth to planets which were wholly liquid, we 

 ought to start from planets with large numbers of relatively small 

 satellites, and after passing through the boundary cases of planets 

 with a small number of relatively large satellites reach planets hav- 

 ing no satellites at all. This is precisely what we find in the solar 

 system. Leaving Jupiter and Saturn, each with their nine relatively 

 small satellites, we pass through Mars. Avith its two satellites, to the 

 earth, Avith one relativel}'^ very large satellite, and after this come to 

 Venus and Mercury, with no satellites at all. Proceeding in the 

 other direction from Jupiter and Saturn, we pass through Uranus, 

 with four small satellites, to Neptune, with one comparatively big 

 satellite. Looked at from this point of view, the earth-moon system 

 figures as the obvious boundary case betAveen the planets which were 



