396 TRANSACTIONS OF SECTION A. 
not produce ions directly, but dislodge either y rays or 6 rays from atoms by 
direct collisions, which are comparatively rare. The f£ rays alone, as C. T. R. 
Wilson’s photographs show, are responsible for the ionisation. Personally, 1 
have long been a convert to Professor Bragg’s views on the nature of X rays, but 
even if we regard the existence of neutral corpuscles as not yet definitely proved, 
it is, I think, permissible to assume their existence for purposes of argument, in 
order to see whether the conception may not be useful in the interpretation of 
physical phenomena. 
If, for instance, we assume that these neutral corpuscles or molecules of caloric 
exist in conductors and metallic bodies in a comparatively free state of solution, 
and are readily dissociated into positive and negative electrons owing to the 
high specific inductive capacity of the medium, the whole theory of metallic con- 
duction follows directly on the analogy of conduction in electrolytic solutions. 
But, whereas in electrolytes the ions are material atoms moving through a 
viscous medium with comparatively low velocities, the ions in metallic conductors 
are electric corpuscles moving with high velocities more after the manner postu- 
lated in the kinetic theory of gases. 1t is easy to see that this theory will give 
similar numerical results to the electronic theory when similar assumptions are 
made in the course of the work. But it has the advantage of greater latitude in 
explaining the vagaries of sign of the Hall effect, and many other peculiarities in 
the variation of resistance and thermo-electric power with temperature. For good 
conductors, like the pure metals, we may suppose, on the electrolytic analogy, 
that the dissociation is practically complete, so that the ratio of the conductivities 
will approach the value calculated on the assumption that all the carriers of heat 
are also carriers of electricity. But in bad conductors the dissociation will be 
far from complete, and it is possible to see why, for instance, the electric resist- 
ance of cast iron should be nearly ten times that of pure iron, although there is 
comparatively little difference in their thermal conductivities. The numerical 
magnitude of the thermo-electric effect, which is commonly quoted in explana- 
tion of the deviation of alloys from the electronic theory, is far too small to 
produce the required result; and there is little or no correspondence between 
the thermo-electric properties of the constituents of alloys and the variations of 
their electric conductivities. 
One of the oldest difficulties of the material theory of heat is to explain the 
process of the production of heat by friction. The application of the general 
principle of the conservation of energy leads to the undoubted conclusion that 
the thermal energy generated is the equivalent of the mechanical work spent in 
friction, but throws little or no light on the steps of the process, and gives no 
information with regard to the actual nature of the energy produced in the form 
of heat. It follows from the energy principle that the quantity of caloric 
generated in the process is such that its total energy at the final temperature 
is equal to the work spent. If a quantity of caloric represents so many neutral 
molecules of electricity, one cannot help asking where they came from, and 
how they were produced. It is certain that in most cases of friction, wherever 
slip occurs, some molecules are torn apart, and the work spent is represented 
in the first instance by the separation ‘of electric ions. Some of these ions are 
permanently separated as frictional electricity, and can be made to perform 
useful work ; but the majority recombine before they can be effectively separated, 
leaving only their equivalent in thermal energy. The recombination of two ions 
is generally regarded simply as reconstituting the original molecule at a high 
temperature, but in the light of recent discoveries we may perhaps go a step 
further. It is generally admitted that X or y rays are produced by the sudden 
stoppage of a charged corpuscle, and Lorentz, in his electron theory of radia- 
tion, has assumed that such is the case however low the velocity of the electron. 
A similar effect must occur in the sudden stoppage of a pair of ions rushing 
together under the influence of their mutual attraction. Rays produced in this 
way would be of an exceedingly soft or absorbable character, but they would 
not differ in kind from those produced by electrons, except that their energy, 
not exceeding that of a pair of ions, would be too small to produce ionisation, 
so that they could not be detected in the usual way. If the X rays are cor- 
puscular in their nature, we cannot logicaliy deny the corpuscular character even 
to the slowest moving rays. We know that X rays continually produce other 
X rays of lower velocity. The final stage is probably reached when the average 
> 
