CONTEMPORARY ADVANCES IN PHYSICS 345 



It was the beautiful experiment of Rouse and Giddings which con- 

 firmed that the first of the steps is the entry of the atom into the 

 2'P] state; for they showed that ionization of the gas occurs only when 

 the impinging quanta have just the energy required for that transition, 

 not when they have a little less or even a little more. This they were 

 able to show because of the phenomenon of "self-reversal." When a 

 luminous gas becomes dense and hot, the lines of its emission-spectrum 

 broaden ; for the atoms perturb one another, the energy- values of their 

 stationary states are changed by various amounts, and the frequencies 

 of many of the quanta which emerge are appreciably shifted upwards 

 or downwards from the original or "standard" values appropriate to 

 isolated atoms. If in addition the region of density and heat is sur- 

 rounded by another where the gas is cooler and more rarefied, the 

 atoms in this outer zone, being relatively unperturbed, will be able 

 to absorb the quanta having the standard frequencies, but not those 

 others of which the frequencies are shifted. In technical language, the 

 "core" of the line is absorbed; only the "wings" pass through; the Hne 

 exhibits "self -reversal." In the spectrum of the ordinary mercury- 

 vapor lamp, the line 2537 is notably self -reversed. Cooling the lamp 

 with flowing water or an air-blast, however, abolishes the effect; the 

 line shrinks to its normal narrowness, the wings disappear, but the 

 photons of the core are able to escape from the tube. Any action there- 

 fore which is performed by the light of a cooled mercury-vapor lamp, 

 but ceases when the cooling is suspended, must be due to quanta having 

 energies adjusted exactly to the values which are able to excite isolated 

 atoms of mercury. Rouse and Giddings found that ionization of 

 gaseous mercury is precisely such an action.^ 



We cannot so readily conclude that the second step in the ionization- 

 process is the absorption of a second 4.9-volt quantum. It would be 

 rather of an odd coincidence, if there were an excited state of the 

 mercury atom differing in energy from the 2^P] state by just so much as 

 this latter differs from the normal state — not, however, an impossible 

 coincidence. Another and a stronger argument is furnished by the 

 fact that when quanta of various wavelengths shorter than 2537 — in- 

 cluding some which could transfer the atom from the 2^Pi into other 

 known excited states — are projected into the gas along with 2537, 

 the rate of ionization is not augmented. If none of these can help the 

 electron to escape, it is not so likely that a second quantum of precisely 

 the wavelength 2537 can achieve it. Moreover the duration of the 



' There was still a residual current in the irradiated tube when the cooling of the 

 lamp was discountinued; but it depended on the size of the cathode in such a way as 

 to suggest that it was due to light falling on that electrode (cf. page 343). Houter- 

 mans later verified this result. 



