On the Arc Spectrum of Vanadium. 



207 



.them are obviously due to other metals existing as impurities either in 

 the poles or in the compound of vanadium which was used, and 

 although several of these lines occur in the Kensington photograph, 

 they have been discarded. Attempts to trace the remaining lines to 

 other origins have been unsuccessful. 



AVith reference to the lines which are absent from Eowland's list, but 

 which appear in the other two, it seems certain that many genuine and 

 strong lines of vanadium have either not been identified by him, or 

 have for some reason been discarded from his list. In this connection, 

 it may be stated that many of the lines recorded by Eowland in his 

 " Table of Solar Wave-lengths " as being due to vanadium, do not 

 appear in his list of vanadium arc lines, though nearly all of them occur 

 as strong lines in both Hasselberg's and the Kensington records. A list 

 of these is given on the next page. Those marked with a f are taken 

 from a list of corrections which he has given* to his " Tables of Solar 

 Wave-lengths." The remainder are taken from his original tables. 



Included in this list are seven lines possibly identical with lines in 

 Eowland's arc spectrum, though the difference in his two recorded 

 wave-lengths of the possibly corresponding arc and solar lines varies 

 from ten to nineteen hundredths of a tenth-metre, a difference which is 

 greatly in excess of what he claims to be his limiting error in the 

 estimation of wave-lengths. 



In the Kensington list there are 194 lines which do not appear in 

 either Hasselberg's or Eowland's. It will serve no useful purpose to 

 enumerate these in a special table, as they can be easily referred to 

 in the general comparison table given in an earlier part of the paper. 

 An analysis of their intensities shows that seventy-seven are very weak 

 lines, of intensity designated < 1, fifty-three of intensity 1, thirty-nine 

 of intensity 2, twenty of intensity 3, three of intensity 4, and two of 

 intensity 5, the maximum intensity adopted being 10. 



No other probable origin has been found for any of them, although 

 the vanadium spectrum has been compared directly with the arc 

 spectra of the following elements : — Ag, Au, Ba, Bi, Ca, Cd, Ce, Co, 

 Cr, Cs, Cu, Di, Fe, Hg, In, Ir, K, La, Li, Mg, Mn, Mo, Na, Ni, Os, Pb, 

 Pel, Eb, Eh, Eu, Sc, Sn, Sr, Ta, Te, Th, Ti, Tl, U, W, Yt, Zn, Zr. 



As these lines appear in the spectrum when either the oxide or 

 chloride of vanadium is used, there seems to be no reason to doubt 

 that they are really due to vanadium. 



Several of them are evidently present in Hasselberg's photograph, 

 as in his comparison of certain vanadium lines with lines of equal or 

 nearly equal wave-length belonging to other metals he records the 

 following, but has left them out of his comprehensive list of vanadium 

 lines, presumably as being due to other metals which exist as impurities 

 in his vanadium. 



* ' Ast.-Pliys. Jour.,' vol. 6, p. 384, 1897. 



