CHAPTER III 
: s ' UV-'A ■ > 
QUANTUM THEORY 
Interaction of Radiation and Matter. 
34 . The theory of the equilibrium of matter and radiation at constant 
temperature depends on a principle which is a generalisation of the theory 
of exchanges (§ 29). After equilibrium is reached no visible change occurs; 
the density and constitution of the radiation, the proportion of atoms in 
various states of combination and ionisation, the number of free electrons, 
the proportion of molecular velocities between given limits, all remain 
steady; but beneath this statistical changelessness there is continual 
change happening to the individual atoms, electrons, and elements of 
radiation. 
Consider the atoms of a particular element which are uncombined and 
in their normal neutral state. The number n of these atoms in the system 
will remain constant (apart from chance fluctuations) when equilibrium 
is reached. But the individuals composing this number continually change. 
New atoms appear in this state owing to the dissolution of chemical 
molecules containing them, neutralisation of ionised atoms by the capture 
of free electrons, relapse of excited atoms to the normal state. Atoms in 
the given state disappear owing to the converse processes—combination 
to form chemical molecules, ionisation by the expulsion of an electron, 
excitation by absorption of radiation or collision with electrons or atoms. 
The steadiness of n is due to an average balancing of gains and losses. 
But the principle above mentioned is not content with formulating this 
general balance of gain and loss—a mere translation of the word “equi 
librium.” It asserts that the gain by any process balances the loss by the 
converse process. The gains due to capture of a free electron balance the 
losses due to expulsion of an electron, independently of the other sources 
of gain and loss. This principle of separate balancing extends to the smallest 
details. Gains due to capture of an electron to fill a vacancy at a particular 
level in an atom balance losses due to expulsion of an electron from that 
level ; gains due to capture of an electron of particular speed balance losses 
due to expulsion of an electron with that speed. 
We may put it in this way—Any statistical enumeration, however 
detailed, of the processes of change occurring in a system in equilibrium 
at constant temperature would remain true if the direction of time were 
reversed*. For our applications we state it in the form— 
* The ultimate laws of nature (so far as known) leave the direction of time 
indeterminate and provide no test to distinguish the past from the future. The 
direction in which time is progressing can only be found by statistical tests depending