MR. A. FOWLER ON THE SPECTRUM OF MAGNESIUM HYDRIDE. 459 



The modification employed by HIGGS,* or its equivalent 



A,,, = a-b(m+p.y, 



where /u, is an additional constant, usually gives only small errors when applied to 

 about a dozen lines, and may, therefore, be employed for the shorter series. By 

 expanding this equation, it assumes the form used by LESTER! for the oxygen bands 

 of the solar spectrum, namely, 



A,,, = a bmcm 2 . 



This is somewhat more convenient for purposes of calculation, and the following 

 examples will indicate the accuracy with which it represents the four short series 

 proceeding from the head of the green fluting. The magnitudes of C (observed 

 minus computed values) are expressed in thousandths of a tenth-metre. 



a series: A m = 5211-115-0-141< -0'1067( 2 , 

 ft A m = 5210-405-0-3673m-0-10603m a , 

 y A, ;l = 5210'582-0-5435m-0'100<)4m 2 , 

 S A,,, = 5209-940-0-4891m-0-11317m 2 . 



Apart from the probably composite lines, the differences between the observed 

 wave-lengths and those calculated from the formulae, although small, are largely 

 systematic and are apparently due chiefly to the impossibility of accurately repre- 



* ' Roy. Soc. Proc.,' vol.* 54, p. 200 (1893). 

 t ' Astrophys. Jour.,' vol. 20, p. 81 (1904). 



3 N 2 



