98 SECTIONAL ADDRESSES. 
gaseous atmosphere which surrounded it. 'This suggestion carried with 
it either implicitly or explicitly the view that the source of power 
behind the emission was not the thermal energy of the source, but 
was the chemical energy of the postulated reactions. 
This: type of view has never had any success in elucidating the 
phenomena, and I do not feel it necessary at this date to weary you 
with a recital of the facts which run entirely counter to it, and, in 
fact, definitely exclude it as a possibility. They have been set forth 
at length elsewhere on more than one occasion. I shall take it to 
be established that the phenomenon is physical in its origin and 
reversible in its operation. 
Establishing the primary character of the phenomenon does not, 
however, determine its nature or its immediate cause. Originally I 
regarded it as simply kinetic, a manifestation of the fact that as the 
temperature rose the kinetic energy of some of the electrons would 
begin to exceed the work of the forces by which they are attracted 
to the parent substance. With this statement there is, I think, no 
room for anyone to quarrel, but it is permissible to inquire how the 
escaping electrons obtain the necessary energy. One answer is that 
the electrons have it already in the interior of the substance by virtue 
of their energy of thermal agitation. But thermal agitations now 
appear less simple than they used to be regarded, and in any event 
they do not exhaust the possibilities. 
We know that when light of short enough wave-length falls on 
matter it causes the ejection of electrons from it—the so-called photo- 
electric effect. Since the formula for the radiation emitted by a body 
at any given temperature contains every wave-length without limita- 
tion, there must be some emission of electrons from an incandescent 
body as the result of the photoelectric effect of its own luminosity. 
Two questions obviously put themselves. Will this photoelectric 
emission caused by the whole spectrum of the hot body vary as the 
temperature of the incandescent body is raised in the way which is 
known to characterise thermionic emission? A straightforward thermo- 
dynamic calculation shows that this is to be expected from the theo- 
retical standpoint, and the anticipation has been confirmed by the 
experiments of Professor W. Wilson. Thus the autophotoelectric 
emission has the correct behaviour to account for the thermionic 
emission. The other question is: Is it large enough? This is a 
question of fact. I have considered the data verv carefully. There is 
a little uncertainty in some of the items, but when every allowance 
is made there seems no escape from the conclusion that the photo- 
electric effect of the whole spectrum is far too small to account for 
thermionic emission. 
This question is an important one, apart from the particular case 
of thermionic emission. The same dilemma is met with when we 
seek for the actual modus operandi of evaporation, chemical action, 
and a number of other phenomena. These, so far as we know, might 
be fundamentally either kinetic or photochemical or a mixture of both. 
In my judgment the last of these particular alternatives is the most 
probable. (I am using the term photochemical here in the wide sense 
