I 



24 



Dale 



N. 



. ul 



Ol IS. 



y, 



lllU' 



,f/, 



/ 



i)tc. 7 





7 





w 



j:i 



-:i .i; 



:)isi2.4(i 



•? 1? 





4 





•}j 



?» 



21" 



„ .:5'.) 



ÜCC. S 





4 







.. 



:>1 .4 



.. .41) 



)) >> 





3 



3 









21 1 



2r.() 



„ .40 



I\k',iii 111' 1^1 ulisi'i'Viitiuiis. — 

 /. = .J1S-2.4'J 



Tlic ruuiaiiiiiii; tliici; JiiR'S of this uruup i^. /a; aiul />, wirci iiieasmcil iiiicin- 

 iiifliifiiliy, Ly rclL'ning tlieiii to 6,. Twenty iiieiisures were made on each line 

 willi the t'cillowiiig- results; — 



line value ut" i 



. I,., 44^ 46' ts .9 :)17l.47 



6, 44^ 43 4'.J .8 r)U;7.73 



/y, 44' 42 38 .(; .OltiC.Ki 



Uiie of the irhjcets in nie.iMuini;- the line.s in this gi'oU[i was tu M-ilt'y their 

 ilistatices from ea^-h other that thi'v might he used as standards of i'i>niii,u'i-nu in 

 tliat part «if the si)eL;truiii, as the D lines are so frei[tieiitly. In one or two 

 instances, in well known works on spectrum analysis, the mistake lias lieen made 

 oFa]>iilying the 1' lines as a unit of measure in the violet end of thi' s]iectrum 

 and failing to eoiisider tin; ahnornial exteiision of that end of the ]iiismatie 

 spectrum. The use of the i lines may ]irevent so great an error as well as he 

 e.Nceediiigly eoiivenient on aecoiint of their ]irominence and the convenient jior- 

 lioii of the speclnim in which they are I'miud. fur lle-ir dilferences of wave- 

 Iciigtli we have the following: 



/>, -* 6, — 111.'.):'. 

 b., — b., - 3.74 

 />., — /«, = I .'■>'.) 



I'lohi Ang.stidms results the same iiieasuri's appear as fullows". 



6, _ b., — l(i.:i4 

 b, — b, = 3.(J8 

 b.. ~ b, — 1 .CO 



d the two Series a'j;ree exceediiis;iv well. 



