252 ON LIGHT 



tions is easily tested on any one ray of the spectrum 

 insulated from the rest by intercepting all the others. 

 The ray so insulated, whatever its tint, is no longer 

 separated or " dispersed" by subsequent refraction into 

 a new spectrum. It preserves its tint unaltered, and 

 conforms to the " rule of the sines" in its flexure, as if 

 no other colour or refrangibility existed. Hence we 

 might be led to conclude, as Newton himself did, that 

 between these two qualities refrangibility and colour 

 an absolute and invariable connexion exists. This, 

 however, is not the case. The propositions in question 

 cannot be generalized. When difl"erent media are 

 examined, we find that not only does the same colour 

 correspond to different degrees of refrangibility, or to 

 different absohife values of the refractive index in each, 

 but that the same change of colour does not correspond 

 in different media to the same proportionate change of 

 the refractive index ; and that, in short, taking the 

 " scale of colour" in all its gradations, from red, through 

 orange, yellow, green, blue, and indigo, to the least per- 

 ceptible violet, and that feeble tint beyond the violet 

 which can hardly be called a colour, but which is most 

 nearly expressed by the term lavender^ as a guide, each 

 particular medium distributes these rays through its 

 spectrum, though always in the same order of succes- 

 sion, yet in other respects according to a law peculiar to 

 itself: thus indicating both a total amount of dispersion 

 2X^A z. scale of action dependent on the physical proper- 

 ties of the medium, and in some sort as it were personal 

 to each. This power which a transparent medium has 



