E. E. Lawton — Bands in the Spectrum of Nitrogen. 107 



3d line of 



the triplet. 





Calculated 





1 



m = 14 



2508-021 



15 



•778 



16 



2509-565 



17 



2510-412 



18 



2511-308 



19 



2512-294 



20 



2513-244 



21 



2514-296 



22 



2515-391 



23 



2516-537 



24 



2517-732 



25 



2518-977 



26 



2520-272 



27 



2521-617 



28 



2523-011 



29 



2524-455 



30 



2525-949 



31 



2527-493 



32 



2529-083 



33 



2530-730 



34 



2532-423 



35 



2534-160 



N = 



Difl 



'erence 



-0-052 



+ 



•024 



+ 



■030 



+ 



•005 



+ 



•008 



+ 



•042 



— 



■on 



— 



•021 



— 



•003 



— 



•023 



— 



•031 



— 



•036 



— 



•015 



— 



•014 



— 



•019 



— 



•025 



— 



•024 



— 



•005 



+ 



•010 



+ 



•017 



+ 



•025 



+ 



•059 



Observed 

 1 

 1 



2507-969 



2508-754 



2509-535 



2510-407 



2511-300 



2512-252 



2513-255 



2514-317 



2515-394 



2516-560 



2517-763 



2519-006 



2520-287 



2521-631 

 •2523-030 



2524-480 



2525-973 



2527-498 



2529-073 



2530*713 * 



2532-398 



2534-101 



It is seen from these results that the formula applies fairly 

 well to the third line of the triplet, but the agreement between 

 the calculated and observed values is not so good in the case of 

 the other two lines. Another thing to be noticed is that values 

 of N for small values of in do not exist. The same result was 

 obtained by the writer when Deslandres'* formula was applied 

 to the measurements of Hermesdorf. If the residuals are 

 plotted, they show in each case that the curve approximates a 

 parabolic form. Considering the complex structure of the 

 nitrogen spectrum, the agreement is as close as could be antici- 

 pated, and it is evident that it will require a more complicated 

 formula, than has as yet been proposed, in order that the resid- 

 uals may be brought within the limits of observation. 



The points of chief interest which have been brought out by 

 the foregoing are : 



1. The wave-lengths of the lines of the bands beginning 

 X4059 and X3998 have been measured, nearly half of them for 

 the first time ; and all the lines have been measured with a 

 degree of precision hitherto not attained for this portion of the 

 spectrum. 



2. Application has been made of Deslandres' formula to the 

 band X 3998, and the constants determined giving as close an 

 agreement between the calculated and the observed values as can 

 be expected, considering the complex nature of the spectrum. 



* Loc. cit. 



