Chase.] ^-^ |ApriM; 



cremental coefficients for lines 4 and 3. The difference between lines 1 

 and 2 is the same as that between lines 4 and 5. 



The greatest difference between the theoretical and measured values 



(5022.7 — 5017 = 5.7) is only of an inch. The closeness of the 



40000000 



accordance may be more readily seen by dividing each of the theoretical 



values by 1.00028. 



Reduced Theoretical. Measured. 



5416 5416 



5268 5269 



5268 5268 



5235 5235 



5021 5017 



4215 4215 



In some respects this symmetry seems even more remarkable than those 

 which I found, more than eighteen months ago, in many of the chemical 

 elements. They were, however, directly harmonic, being based on cen- 

 trifugal relations to the centres of wave systems (Law 5). These are 

 reciprocally harmonic, being based on centripetal relations to the surface 

 of Sun's chromosphere. 



Multiples of the fundamental increment often appear in the differences 

 between the wave lengths of elementary spectra. The following instances, 

 in elements for which I have already shown harmonic relations,* will 

 serve as examples. The left hand columns contain exact multiple differ- 

 ences ; the rightdiand columns, measured wavedengths : 



Mercury. 

 546.09 546.13 



542.80 542.80 



Ruthenium and Iridium. 

 545.34 545.44 



530.52 530.52 



In the copper lines, the first theoretical difference is 30 X 1.6468 ; the 

 second is I as much ; the third is the sum of the other two. In arsenic, 

 the second line is 2 2 X 3 1 X 1.6468. 



* Ante xvii, 297; 1878. 



