48 Messrs. G. J. Stoney and J. E. Reynolds on the Cause 



which assigns to the corresponding parts of the spectrum the 

 following pattern : — 



(10) 



Halfway between the regions (4) and (10) the pattern would be 



(ii) 



And in general the variation of the pattern in passing along the 

 spectrum may be represented to the eye by conceiving the system 

 of six ordinates in fig. (5) to travel sideways while the curve is 

 fixed. In the position represented by the continuous lines in 

 fig. (5), the squares of the ordinates represent intensities which 

 would give pattern (4). If they are shifted to the left into the 

 position of the dotted lines until e comes to the point of inter- 

 section o, e will entirely disappear, a, d, and / will become of 

 equal length, while b and c also become equal but shorter than 

 the other three; and in this position the squares of the lengths 

 of the ordinates will represent the intensities of lines which 

 would give pattern (10). And all other patterns which could 

 arise on this hypothesis would be represented by the other posi- 

 tions of this system of six ordinates. 



15. On the 30th of January we were able to compare these 

 anticipations with the spectrum itself; and we had the satisfac- 

 tion of finding that changes closely approximating to the pre- 

 dicted changes of pattern actually take place in the absorption- 

 spectrum of chlorochromic anhydride, but not at the uniform 

 rate of change which the simple hypothesis represented by equa- 

 tions (8) and (9) would indicate. 



16. The results of the comparison are embodied in the follow- 

 ing Table, the left-hand side of which gives the observed inten- 

 sities of the lines on an arbitrary scale in which 10 indicates a 

 very dark line, and 1 the faintest visible; while the right-hand 

 side states what the succession of intensities would be on the 

 hypothesis represented by equations (6) and (7). Notes of in- 

 terrogation are introduced when the lines were too much dilated 

 for the observation of their intensities. The lines drawn between 

 the two columns point out where the observed succession of in- 

 tensities of column I. can be most nearly matched in column II. 

 The changes are in the main the same in the two columns ; but 

 the rate of change follows a law in column I. which has not yet 

 been traced out, while in column II. it is uniform. 



