246 CARNEGIE INSTITUTION OF WASHINGTON. 



present, they were reduced to about one-third of the amount measured during 

 April. The half-period displacements, which would correspond to the ether 

 drift if present, were also observed, but were reduced to about two-fifths of 

 their previous value, or 0.08 fringe on the average. While it seems impossible 

 at present to draw definitive conclusions from these results, it appears to be 

 established that the displacements observed on Mount Wilson can not 

 exceed one-tenth of the amount predicted for the ether drift. It should 

 be remembered, however, that the factors taken into account in computing 

 the predicted amount are not complete. 



The interferometer and steel frame have now been returned to Cleveland 

 and Professor Miller will continue his observations with a view to detecting 

 possible differences from the Mount Wilson results. It is also hoped to dis- 

 cover the source of the full-period displacement and to learn whether the half- 

 period effect may be observed independently of the other. 



PHYSICAL LABORATORY. 



ELECTRIC-FURNACE INVESTIGATIONS. 



Tests of Ionization Phenomena. 



Certain features of Saha's theory of ionization processes admit of tests by 

 the electric furnace, which have been carried out by Mr. King. The phenom- 

 ena in question are based on the emission of enhanced lines by atoms which 

 have each lost one electron, these free electrons forming a gas which under 

 given conditions of temperature and pressure has a definite concentration. A 

 further supply of electrons through the addition of a more easily ionized 

 element should then retard the formation of ionized atoms of the first element, 

 with a resultant weakening of its enhanced lines, while the arc lines radiated 

 by the normal atom remain unchanged. Observations for the detection of 

 this effect were made on the spectra of calcium, strontium, barium, and 

 scandium. In each case the spectrum of the element alone was compared 

 with that obtained when a more easily ionized element such as potassium or 

 caesium was added, and in each case the effect was found to be a weakening 

 of the enhanced lines. Mixture with a substance having about the same 

 ionization potential produced no effect. 



In addition to the confirmation of Saha's theory afforded in this way, the 

 results in the case of scandium, whose series relationships are now known, 

 demonstrated that its enhanced lines arise from the ionized atom. Further, 

 they confirmed the selection of enhanced lines obtained by other methods, 

 and showed the approximate magnitude of the ionization potential of scan- 

 dium. The scandium enhanced lines are decidedly weakened in the spectrum 

 from a mixture with potassium and barium, having ionization potentials of 

 4.3 and 5.1 volts, respectively, but are unaffected by calcium, which requires 

 6.1 volts. This indicates an ionization potential for scandium close to that 

 of calcium. The method can be used with other elements whose enhanced 

 lines are within the range of the furnace excitation. 



Another feature of Saha's theory is that lines of a given type should appear 

 in absorption at a lower temperature than in emission. While more extensive 

 observations on this point are needed, tests which have been made for the 

 strong low-temperature lines of calcium, strontium, and barium have yielded 

 positive results, a temperature being found in each case which gave the lines 



