688 Proceedings of Royal Society of Edinburgh. [sess. 
potential heat — proportional to cubical space for their particles to 
get squeezed into — but they part with it more slowly because of 
their greater resistance to compression ; in other w r ords, because 
of their great elasticity.* In fact, there is nothing to prevent us 
from imagining an atom whose elasticity is so great that it would 
take a practical eternity of time for its molecules to radiate their 
heat away. How, then, can the highly elastic kind he neutral ? 
To answer this, let us imagine an infinitely elastic atom. Mole- 
cules constituted of such atoms could vibrate for ever and yet 
send no heat-wave away into space. This seems a paradox, but 
it is easily answered. At each inward phase of the vibrations, 
pressure, electricity, or energy, whatever you like to call it, would 
be withdrawn from the ether in the immediate neighbourhood of 
the molecules — it is this energy which presses the atoms together 
— and upon each outward phase exactly the same quantity would 
be restored, and would proceed no farther than the locality from 
which it was withdrawn. As an illustration, suppose a perfectly 
elastic ball to be dropped from a height, in an air vacuum, on to a 
perfectly elastic pavement — it would rebound to the height 
from which it fell ; to fall again, and rebound, and so on con- 
tinually. On falling it would gain its energy of motion from 
gravity (a force the genesis of which I have attempted to explain 
in a former paper), and upon rising it would restore to the medium 
the energy it abstracted in falling. In this illustration none of 
the energy would be transformed into heat any more than in the 
case of our hypothetical perfectly elastic molecules, but in the 
latter case note what would be happening — at each outward phase 
energy would be added to the ether, and which at each inward 
phase would be withdrawn, these opposite actions alternating so 
rapidly that the average state of the medium would be normal, or 
neutral. In fact we might say that at one moment the state of 
the ether would be positive, the next moment equally negative, 
* “In the usual notation of dynamics, assuming simple harmonic motion, 
we have x= - n 2 x. Therefore, when the vibration frequency is increased n 
times, the displacement produced by the same force becomes l/?i 2 of its former 
value ; and so, per vibration, under the action of the same force, the energy 
involved becomes 1/ri 2 of its former value. But the number of vibrations per 
unit time is increased n times, so that the rate of emission of energy, under 
the usual law of damped vibrations, would become ljn th of the former rate.” 
