Recurving Changes in the Universe. 155 



masses ; for frequently one of the minute masses (molecular 

 clusters) may acquire such a velocity in the accidents of colli- 

 sion as to break it up into discrete molecules at an encounter, 

 these discrete molecules grouping together again in another 

 part of the gas, the mean state of aggregation (per unit of 

 volume) remaining unchanged. Of course it is obvious that 

 there are differences of detail in considering the case of the 

 masses of the universe. For example, in a compound gas, 

 where the central agency producing aggregation is " chemical 

 action," the fluctuation in size of a mass (molecular cluster) 

 under the collisions is limited, whereas in the case of the masses 

 of the universe, where the central agency producing aggre- 

 gation is mainly "gravific action"*, the fluctuation in size 



* The agency producing aggregation may even be found (when recog- 

 nized) to be of the same kind in all cases (" gravific action," " chemical 

 action/' &c). In former papers published in the Philosophical Magazine 

 (Sept. and Not. 1877, and Feb. 1878), I have called attention to the fact 

 that, if the kinetic theory be appplied to the motion of the particles of 

 aether [in addition to that of the stellar masses suggested here], the gravi- 

 tation of the molecules of gross matter may be accounted for under the 

 ingenious fundamental idea contained in Le Sage's well-known theory, 

 with the removal of all his postulates ; or gravitation maybe shown to be 

 the necessary consequence of the mere immersion of the universe in finely 

 subdivided matter moving automatically according to the principles of 

 the kinetic theory. By the application of this theo-w to the stellar masses, 

 molecules of gross matter, and particles of aether, as a vast whole, consist- 

 ing of matter of different dimensions moving under its own dyuamics, a 

 grand dynamical generalization would evidently seem to suggest itself. 

 It may be noted that, if we reject the now practically defunct conception 

 of "force " in the sense of an " action at a distance " (without the inter- 

 vention of matter), no other than a dynamical view of the universe is in 

 principle conceivable. 



Another consideration would seem to have an important bearing on this 

 subject. The pa?'ts of the molecules of matter are known to possess a con- 

 siderable capacity for motion. When, for example, the molecules of a 

 gas are exposed to the pulsatory movement of the waves of heat (from 

 some radiant source), the parts of the molecules are thrown into motion 

 (vibration) ; and this development of motion in the parts is found to pro- 

 duce an accession of translatory motion in the molecules as wholes. This 

 principle holds independently of scale ; and thus it appears that the deve- 

 lopment of motion (by any means) in the constituent parts of a mass tends 

 (under certain conditions) to produce translatory motion in the mass as a 

 whole. If the aether, in which gross matter is immersed, be itself in a 

 state of internal motion, this motion must inevitably communicate itself 

 (to a certain extent) to the constituent parts of masses of matter immersed 

 in the aether, and, accordingly, serve as a supplementary means to the 

 development of translatory motion in the masses as wholes. 



We may observe that, under the kinetic theory of gravity (based upon 

 Le Sage's fundamental idea), both the range and intensity of gravity have 

 a limit. The range (limited by the length of path of the particles) need 

 not be necessarily much greater than that of which we have proof by ob- 

 servation, which is but a relatively small range compared with the stellar 

 distances. The intensity of gravity therefore (for this reason) would not 



