On the hifltKiin, Rejlexhn, and Conrs (/Light. jya 



refniaed. I faw feveral images of the candle, all highly coloured, and the colours were, in 

 order from the candle outwards, red, orange, and fo on to violet. 1 then filled the tube whh 

 clear diluted fulpliuric acid, and dropped a fmall piece of challt to die bottom, when imme- 

 diately an effervefcence tooK place by the efcape of fixed air, which role in bubbles thro«j*h 

 the tube; and looking at the candle through one of thefe, I faw the images formed with 

 the colours flill in the fame order, but a little larger than before. 



We are now to fee to what conclufions thefe experiments lead us. The firft' experiment 

 (hews, that all forts of lights, whether dire£l or reflefted or refrafted, produce colours by 

 reflexion from a curve furface. From the fecond we learn th.-.t thefe colours are diftinct 

 images orfpeara of the luminous body, much dilated in length, but not at all in brea*th, 

 and that the angle of incidence being changed, the dilatation of t!ie images is alfo changed -, 

 and from the third experiment it appears, that each full image is compofed of feven co- 

 lours, red, orange, yellow, green, blue,, indigo, and violet; and that the proper order is red 

 cutermoft, and violet innermoft, the reft being in their order. The fourth experiment 

 ftews, that thefe images are produced, not by any accidental or new modification Impreflcd 

 on the rays, but by the white light being decompofed by refieJcion ; that the mean rays> or 

 thofe at the confine of the green and blue, are refleaed at an angle equal to that of inci^ 

 dence, and the red at a lefs, the violet at a greater angle. Experiments 5th and 6th prore^ 

 beyond a doubt, the decompofition and feparation of the r.-;ys by reflexion ; for in both we 

 fee that the colours in the images are thofe, and thofe only,which were mixed in the ray bv 

 reflexion or refradion before and at incidence, whilft the 6th is (in addition) a proof that- 

 all the rays of any one image, if mixed together, compound a beam exadly fimilar to the 

 beam that was at firft decompounded. The feventh experiment fliews, that the colours 

 into which the rays are feparated by reflexion are homogeneous and unchangeable ; that thejr 

 dinger in flexibility and refrangibility; that they bear the fame part in forming im.ig'es by- 

 reflexion, and fringes by flexion, and colours from thin plates, which the rays feparated by 

 the prifm do. And in the 8th experiment we fee, that when the rays are placed in tlie 

 fame fituation withrefpea to refraaion, whether out of a rarer into a denfer or a deiifef 

 into a rarer medium, in which they before were with refpeft to reflexion, the pofition of 

 the colours produced is diametrically oppofite in the two cafes. Seeing, then, l^at in all 

 forts of light, dirca, refraaed, refleaed, fimple and homogeneous or heterogeneous, and 

 compounded, and in whatever way the feparation and mixture may have been made fome 

 of the rays, at equal or the fame incidences, are conftantly refleaed nearer the perpendicu- 

 lar than the mean rays, and others not fo near ; and feeing that by fuch reflexion the com- 

 pound ray of whatever kind is feparated into parts fo fimple that they can never more be 

 changed ; and confidering the difierent pl.iccs to which thefe parts are refletled ■ it is evi- 

 dent, that the fun's light confifts of parts di/Tcrent in reflexibihty, and that thofe which arc 

 lead refrangible are molt reflcxible. By reflexibillty I here mean a difpofition to be re- 

 fleaed fiear to the perpendicular in any degree. 



Although 1 have given what I take to be fulTicicnt proof of this property of light, yet 

 1 am aware that fomething more is requifite. It will be aflced, why does ncitlicr a plane a 

 common convex, nor a common concave mirror feparate the rays by reflexion ? This is 

 what has always hindered us from even fufpeaing fuch a thing ns different rcflexibility. I 

 {hall, however, take an opportunity of rcmoTing this obftaclc in the fecond part of the plnn^ 



when 



