198 Newton's Letters, Hypothesis and Experiments 



towards helping the two last difficulties, but nothing which I 

 see not insufficient. 



Fifthly, it is to be supposed that light and aether mutually 

 act upon one another, aether in refracting light, and light in 

 warming aether, and that the densest aether acts most strongly. 

 When a ray therefore moves through aether of uneven den- 

 sity, I suppose it is most pressed, urged, or acted upon by the 

 medium on that side towards the denser aether, and receives 

 a continual impulse or ply from that side to recede towards 

 the rarer, and so is accelerated if it move that way, or re- 

 tarded if the contrary. On this ground, if a ray move ob- 

 liquely through such an unevenly dense medium (that is, ob- 

 liquely to those imaginary superficies which run through the 

 equally dense parts of the medium, and may be called the re- 

 fracting superficies), it must be incurved, as it is found to be 

 by observation in water*, whose lower parts were made gra- 

 dually more salt, and so more dense than the upper. And 

 this may be the ground of all refraction and reflexion. For 

 as the rarer air within a small glass pipe, and the denser with- 

 out, are not distinguished by a mere mathematical superficies, 

 but have air between them at the orifice of the pipe running 

 through all intermediate degrees of density ; so I suppose the 

 refracting superficies of aether between unequally dense me- 

 diums to be not a mathematical one, but of some breadth, the 

 aether therein at the orifices of the pores of the solid body 

 being of all intermediate degrees of density between the rarer 

 and denser aetherial mediums; and the refraction I conceive 

 to proceed from the continual incurvation of the ray all the 

 while it is passing the physical superficies. Now if the mo- 

 tion of the ray be supposed in this passage to be increased or 

 diminished in a certain proportion, according to the difference 

 of the densities of the aetherial mediums, and the addition or 

 detraction of the motion be reckoned in the perpendicular 

 from the refracting superficies, as it ought to be, the sines of 

 incidence and refraction will be proportional according to what 

 Descartes has demonstrated. 



The ray therefore in passing out of the rarer medium into 

 the denser, inclines continually more and more towards paral- 

 lelism with the refracting superficies; and if the differing 

 densities of the mediums be not so great, nor the incidence of 

 the ray so oblique as to make it parallel to that superficies 

 before it gets through, then it goes through and is refracted ; 

 but if through the aforesaid causes the ray becomes parallel 

 to that superficies before it can get through, then it must turn 

 back and be reflected. Thus, for instance, it may be observed 



* Mr. Hook's Micrographia where he speaks of the inflexion of rays. 



