REFRANGIBILirr of LIGHT. 35 



violet to reddifh violet, and on the other from greenifli orange 

 to greenifli blue. As the difperfion is diminifhed, the red 

 gains on the violet within the focus, and the greenifli blue up- 

 on the orange without it, and vice verfa ; and there is a consi- 

 derable latitude, within which, varying the difperfion, makes 

 little difference in the diftinctnefs." 



As this laft obfervation put it beyond doubt that an invefti- 

 gation of the caufe of thefe appearances was of the laft im- 

 portance to the improvement of optics, I now began to reafon 

 concerning them. 



The firft conjecture that offered was, that this colour might 

 fomehow proceed from the furfaces of the convex glafs lenfes, 

 and the concave lenfes of difperfive fluids, not correfponding 

 at different diftances from the centre, as the plane furfaces of 

 prifms every where do. In order to examine what effect this 

 might have, I procured two pieces of plate-brafs, with which 

 I could cover the whole of the objecT-glafs ; and out of one of 

 thefe I caufed a ring, of a quarter of an inch in breadth, to be 

 cut towards the centre, and out of the other, a ring of the 

 fame breadth, clofe to the circumference. For I perceived that, 

 if the colour arofe from the caufe above mentioned, its appear- 

 ance ought to be different through thefe two rings, when there 

 is an accurate correction of colour in that part of the object- 

 glafs, which is equi-diftant from the centre and the circumfe- 

 rence. But upon trying the experiment, the fame purple and 

 green colour appeared through both thefe rings, as through the 

 whole object-glafs, and the colours lay in the fame order in both 

 cafes. My remark upon this experiment is in the following 

 words : " Upon trying with a ring either external or internal, 

 the appearances remain the fame, as when the whole aperture 

 is ufed ; which feems to prove that this colour arifes from the 

 difperfion not being proportional, and not, as was fuppofed, 

 from the furfaces not correfponding. It is evidently the greateft 

 bar to increafing the aperture, and giving high powers 5 there 



E 2 is 



