Past Changes in the Universe . 
365 
1878.] 
the adtion of gravity into states of aggregation, or groups, 
on the same principle (and perhaps even much on the same 
process) as the constituents of the molecules of a compound 
gas group together under “ chemical ” adtion. The range 
of “ chemical ” adtion, like the range of “ gravific ” adtion, 
is limited; the constituents of the stellar masses, which 
group together under the limited range of “gravity,” being 
in this respedt comparable to the constituents of the mole- 
cules of a compound gas which group together under the 
limited range of “chemical” adtion. Just as in the case 
of a compound gas (on account of some molecules possessing 
much higher velocities than others) the components of two 
molecules frequently become dissociated by a collision, and 
group together again in a different part of the gas; so in 
the Universe, where some stars (no doubt for analogous dy- 
namical reasons) possess a higher proper motion than 
others, we have complete disintegration by collision in some 
parts, aggregation in others, continual change, but the mean 
aggregation remaining unchanged. 
It may be observed that in principle, in order to explain 
the continued working of physical phenomena, or the con- 
tinuance of change in the Universe, the existence of some 
process of recurrence is absolutely essential. The particular 
material that is utilised for the development of fresh suns 
or centres of heat must be the cooled down or dead material 
of former suns. For if this were not the fadt there would 
be a continual accumulation of the matter of extindt or 
useless suns in the Universe, and processes of renewal and 
maintenance of the energy of the Universe would come to 
a deadlock in the absence of matter to operate upon. It 
therefore becomes absolutely necessary to show, in any 
attempt to explain the continued working of physical phe- 
nomena under present causes, how the material of extindt 
suns can — in accordance with recognised dynamical prin- 
ciples — be made available for the development of fresh suns 
(or centres of heat). This conclusion would seem to evolve 
itself naturally on the basis of the dedudtions following 
from the physical theory of gravitation — in regard to the 
j motion of the stars taking place in straight lines, which 
involves collisions and an alternating renewal and loss of 
j heat. This appears on broad principle the only conceivable 
j way in which there should be recurrence, under the condition 
, that the same matter should be used again. On account of 
the relatively enormous area of free space, compared with 
the relatively very minute portion of space occupied by each 
star, a stellar sun would, in the course of its proper motion 
