20 



Dat(! 



h oLserveil 



T 



h reduced to 18-" 



Dec. 27 



44^ r.7' 0".:? 



\:p:y 



44^ r)G' 50".S 



jj 



„ r.(i'r.r)".i 



15".;-) 



„ „ 49".6 



ji 



^)7' 1 " 



ir)".3 



„ „ r>o".7 



Mean lA' ?A nn-asurcineiifs. 

 /, = J4° '^r>' .^)()".3r> ±: 0".08 



B 



Amonsj tlio various values for tlie \vave-l(^ii2;th of this line, juililislied liy 

 diflerent antliorities, are .several that ditTer so widely from the others that it 

 seems diflicult toaecounl for the discordance in any other way than hy assnniinü; 

 tliat ililTerent lines hav(^ heen measured in ditVerent cases. This region of the 

 s|H^ctrnm ahoun Is in ahsorptioti lines and niishihes might (easily he made in the 

 .»ieleetion of tlie jiroper line. In the ]iresent instance the line selected islielieved 

 to he thai niion which .\ni;slrrini made his uieasnrements. Tt is a line, some- 

 what darker and mow. prominent than the others, on 1h<^ edge of the group in 

 the direction of least wave-length, as is indicated in his map. A veiy fine line 

 is often visihle, still further in the same direction and distant from R something 

 less than one unit iif wave-length. 



'riiis is th(^ only line in the series in which all of the measm-(>nients were 

 mad'' ill any other than the spectrum of the second order. The first order was 

 preferri'd in this case on account of the faintness of this part of 1hi> spectrum 

 on one side cif ilie normal to the grating. 



Mate 



No 



«if ..l.s 



Value of/' 



/ 



Dec. 1. 





4 



27^ r)2' 34".4 



0866.84 



, , 





•* 

 .1 



„ „ 34".3 



,. .83 



7T 





3 



„ .. 34". 7 



., .86 



IVe. 2. 





4 



„ „ 3:5".S 



„ .80 







4 



„ „ 33".7 



., .79 



M 





3 



„ „ 34".4 



.. .84 



i\Iean of 21 idiservatimis. 

 / = 6866.82 



c 



']'his lino is hroad and in preference t" attcmj)ting to set upon the middle 

 of the line in ever\- instance, the edge of the line was selected and the results 



