190C.] on Electric Discharge and Spectroscopy. 203 



Thus the maximum vahie of tlic product of the current and the 

 potential difference, i.e. rate at whicli the electric forces are doing 



work in the tube, is C y,r|?, or ^|' Vo^ and is thus proportional to 



the square root of the capacity and inversely proportional to the 

 square root of the self-induction. Thus, increasing the capacity in- 

 creases the maxinmm rate of work and therefore increases the brilhancy 

 of the lines coiTesponding to systems of the type B relatively to those 

 of type A, while inserting self-induction in the circuit increases the 

 brilliancy of those of type A as compared with those of type B. If 

 we suppose that the " blue " spectrum of argon corresponds to a 

 system of type B, the red to a system of type A, we have an explana- 

 tion of the changes in the spectrum of this gas, for by inserting 

 capacity in the circuit we can change from the red to the blue spec- 

 trum, while having got the blue we can get back to the red by inserting 

 self-induction as well as capacity. I have here a little model which 

 is intended to illustrate the way in which the red and blue spectra of 

 argon originate. It is based on the fact that when we send a current 

 of electricity through a circuit, the current does not rise to its steady 

 value instantaneously, but, starting from zero, increases with the time 

 in exactly the same way as we have supposed the intrinsic energy in 

 the atom, i.e. the way represented by the curve in Fig. 4. The quantity 

 in the electrical case corresponding to the radiation ^ is the resist- 

 ance of the circuit divided by the self-induction, while the quantity a 

 is inversely proportional to the self-induction. Thus, a circuit with 

 large self-induction and small resistance is analogous to the system A, 

 while one with small self-induction and large resistance is analogous 

 to a system of type B. Now my model of the argon atom consists 

 of two circuits, C and D, placed in parallel. C has large self-induc- 

 tion and small resistance, D has little self-induction but large resist- 

 ance. An electric lamp is placed in each circuit. If I supply energy 

 in one way, i.e. by continuous current to the system, the red lamp in 

 C Ughts up and the blue lamp in D is dark, while if fed by an 

 alternating current, the blue lamp shines and the red is dark. It 

 would be interesting to see whether, as we gradually diminish the 

 self-induction, we get the whole of the lines in the blue spectrum at 

 once, or whether the Hnes of this spectrum enter in groups one 

 after the other. I have tried somewhat similar experiments with the 

 hot lime cathode to see in a mixture of gases (mercury vapour and 

 air) which spectrum first appeared as the rate of doing work in the 

 gas was gradually increased. The great difficulty in this determination 

 is that when once the luminosity begins there is such a rapid increase 

 in the ionization that the current through the gas and the rate of 

 doing work increase in an exceedingly short time through a wide 

 range of values, and thus a gradual increase in the rate of work is 

 exceedingly difficult to obtain ; on several occasions, how^ever, I was 

 convinced that on gradually increasing the rate of work, the mercury 



