Recurring Changes in the Universe. 153 



say the least, appear to have something incongruous and un- 

 natural about it, when regarded from a philosophical point of 

 view; and however well grounded this aspect of the case might 

 appear, still, from the vastness of the subject and the limited 

 range of observation, it remains always conceivable that some 

 physical link may have been left out, which affects the final 

 conclusion. 



As previous attempts to explain the phenomena of nature 

 as the result of the action in the past of existing physical prin- 

 ciples have been invariably welcomed, I venture to submit the 

 following conclusion which has presented itself to me. I will 

 commence at once by an illustrative case, noticing the diffi- 

 culties as they arise. Let us imagine (for mere sake of illus- 

 tration) a cubical envelope, which permits neither change of 

 volume nor passage of heat, to enclose a space of diameter 

 say 10 10 times the distance between the Sun and Sirius. 

 First, let the matter within this space be at the zero of 

 temperature. Second, let all the matter within our enve- 

 lope be at such a temperature that it is entirely dissociated 

 into discrete molecules*. Between those two extremes there 

 is room for any number of mean states in which matter 

 might be more or less aggregated or discrete. Might not 

 the universe actually be in one of these intermediate states ? 

 i. e. consisting of portions of matter in various stages of 

 aggregation, moving among each other according to the prin- 

 ciples of the kinetic theory, but not sufficiently rapidly to 

 prevent gravity from producing a certain degree of aggrega- 

 tion. It should be scarcely necessary to observe that we have 

 limited our space merely for the sake of fixing our ideas. All 

 we require is gained if, instead of using the impermeable en- 

 velope, we surround our cube with infinite space filled with 

 matter in a similar condition to that which the cube enclosed. 

 There might perhaps be some who would find a difficulty at 

 first sight in conceiving how two masses in translatory motion 

 could collide without gravity making them coalesce and so 

 the degree of aggregation continually becoming greater and 

 greater. But it is to be noted that, if we imagine the masses 

 to collide at such a limiting speed that the velocity with which 



* The discrete molecules would of course "be in motion, rebounding 

 from each other in all directions, according to the principles of the kinetic 

 theory of gases, and pervading the cubical unit of volume uniformly. 

 Obviously we must take into account the neutralization of gravity within 

 the cubical unit of volume, by realizing the space outside the cube filled 

 with matter in a similar state to that which the cube encloses. The 

 known tendency of the kinetic theory is automatically to produce a similar 

 distribution of matter per unit of volume. Gravity acting from one unit 

 of volume to another is thus neutralized. 



