ON THE SPECTRUM OF THE OXY-HYDROGEN FLAME. 



41 



GROUP with second differences of wave-length equal to 2 -4. 



GROUP with second differences of wave-length equal to '522. 



Now, considering that the dispersion employed was that of a single prism, and 

 that many of the lines are weak and many diffuse, the difference between the 

 observed and calculated values falls in most of the cases above given within the limits 

 of probable error. Indeed, one cause of the difficulty of obtaining very exact 

 measurements of some of the lines arises from the overlapping of different groups, 



Mix V.I. \\.\\ 111. A. ft 



