408 Dr. Richardson: Effect of a Luminous Discharge 
the ions is reversible the constant 6, with certain restrictions, 
measures the change of energy when an ion is set free. 
Since the emission of the imparted ettect follows a formula 
of this type, it is evident that there is no true temperature at 
which the leak commences. With the comparatively rough 
arrangement used in these experiments, the leak was not 
perceptible till about 350° C. Using a much more sensitive 
arrangement, Strutt* was able to detect the ordinary 
positive leak from hot bodies at temperatures 150° lower 
than this. 
It is only at very low temperatures that the imparted lerk 
can be said to be a function of the temperature at all. At 
higher temperatures, owing to the imparted eftect being driven 
off, the current at constant temperature falls off very rapidly 
with time. As the theory mentioned previously would 
lead us to expect, the magnitude of the current and its 
rate of decay both vary with the temperature in a similar 
manner. 
An interesting way of investigating these phenomena is 
to maintain a pr reviously ex xposed wire for equal intervals of 
time at successively increasing temperatures whilst the leak 
is being measured. As has “been ponies out already, for 
the first hundred degrees after the leak is capable of detection 
it rises rapidly with ‘the temperature, of which it is a fairly 
definite function. As the temperature increases, so does the 
rate of decay of the current at constant temperature ; so that 
finally we reach a point where the increase in the value of 
the current due to raising the temperature is not great 
enough to make up for the “loss of the imparted emissibility 
in the previous interval, We thus obtain a maximum in the 
current-temperature curve measured in this way. Provided 
the imparted effect is due to one substance only, the current 
will now fall off as the temperature is successiv ely increased 
—the time the wire is heated at any one temperature being 
always constant—so that the curve will not rise again until 
we come to the normal leak. If, however, there are two 
separate sukstances producing the imparted emissibility, we 
might expect two maxima and so on. In this way we have 
a method of examining the eftect which is analogous to 
fractional distillation. 
The curve with points marked thus: x in fig. 3, repre- 
sents the current at successive equidistant temperatures when 
the wire had been allowed to remain at the previous tempe- 
rature for four minutes. The curve through the points 
marked thus: O represents the leak which this wire gave 
before exposure to the luminous discharge. It will be seen 
) * Phil. Mag. [6] vol. iv, p. 98. 
