486 



Mr. G. AV. Walker on the 



The method of verification must thus be changed, and we 

 have to rely on the temperature-variation of yu, and the 

 dispersion for the determination of the constants in the 

 formula. Considerable uncertainty is thus inevitable. A 

 rough trial to explain the measured temperature-effect and 

 dispersion showed that the visible spectrum must be rather 

 close to the point A, so that I had to tabulate the numerical 

 values of the function in the vicinity of the point A with 

 great care. The results (which we also require for CO2 and 

 H2) are given in the following tables. 



Table I. 



a*. 



E(^). 



dE d'E 



d'E d'E 



3 



4 



5 



1-48372 

 1-43824 

 1-35383 



-h -01086 - -17209 

 - -07868 - -03088 



-•08306 +-1229 



1 



j^..99999 _ -19828 



+•07794 ' -09375 

 +•01847 - -03324 



Table I. was calculated by aid of the formulae 



Table II. was computed by did of Table I. and the formula 



E(.t' + ?)=E. 



■^ -2 -d^i- "^ (5 -dcv' 



+ U -d^v' ' 



Table II. 



X. 



E. ! 



a?. 



lEx. 



X. 



Eo-. 



2-7 



1-47165 



3-7 



1-46007 



4-7 



1-37920 



2-8 



1-47780 



3-8 



1-45325 



4-8 



1-37006 



2-9 



1-48175 



3-9 



1-44594 



4-9 



1-36219 



3 



1-48372 



4 



1-43824 



5 



1-35383 



3-1 



1-48398 



4-1 



1-43024 



5-1 



1-34559 



3-2 



1-48274 



4-2 



1-42199 



5-2 



1-33749 



3-3 



1-48017 



i 4-3 



1-41357 



5-3 



1-32954 



3-4 



1-47822 











