KENT. — THE SPECTRA OF TITANIUM, ZINC, AND IRON. 297 



and the discharge is quite explosive. It was thought that this explosive 

 effect might produce a small shift toward the red.* Evidently the dis- 

 ruptive effect is of too great a period and not sufficiently violent. 



16. A plate was taken under conditions similar to those of Nos. 83 

 and 84, except that the slit-length was extended to about 1 cm., the 

 whole vertical aperture of the grating was used, and the exposure 



TABLE VII. Titanium. 

 Data transferred from Table III. 



lengthened to ten minutes. Under these conditions there results a 

 greatly over-exposed spark spectrum. Upon careful measurement, set- 

 ting on what appears to be the maximum of intensity of the body of the 

 line, shifts as great as 0.031 and 0.025 t.-m. appear; whereas set 10 

 gives values of 0.017 and 0.017. It is evident, then, as Eder and 

 Valenta suggest, that by over-exposure a false displacement may be 

 introduced in a line marked by unsymmetrical broadening. 



17. By referring to Table VII, the data given in wdiich are taken 

 from Table III, it will be seen that the energy involved is not the only 



* The suggestion that a comparison of these two spectra be made is due to 

 Professor Crew. 



