M. MELLONI ON THE REFLECTION OF RADIANT HEAT. 3S& 



whence we derive 



R = 1 + \/0-923 = 1 + 0-9607. 



As the first sign of the radical would lead to an absurd result in the 

 case under consideration, it must be rejected. The reflection at the 

 surface of the plate will then be 1 — '09607= 0'0393 to the unit of inci- 

 dence, and such also Avill be the ratio of the second reflection to the quan- 

 titj'^ of heat which reaches the further surface of the rock salt. But if 

 it be desired to have the absolute value of the second reflection, it will 

 be found by substituting 0-0393 for R in the expression R (1 — R), or 

 yet more simply by taking the difference between the numbers 0*077 

 and 0*0393, which gives in each case 0*0377. 



It now remains to be seen whether the quantities of heat reflected by 

 other transparent substances be equal or not equal to those produced by 

 the surfaces of rock salt. For the solution of this question it is suf- 

 ficient to observe, that a thick plate of glass, of rock crj'stal, or any other 

 diaphanous substance gives a calorific transmission sensibly equal to that 

 of another plate, the same in substance though diflf'erent in thickness. If 

 we take, for instance, one plate of glass 8 millimetres * thick, and an- 

 other 8|, and expose them separately to the radiation of a Locatelli 

 lamp, Ave shall find no sensible difference between the two quantities of 

 heat transmitted. From this experiment it is obvious that the layer of 

 half a millimetre, which constitutes the difference of thickness between 

 the two plates, causes no appreciable absorption of the calorific rays, 

 which have already traversed 8 millimetres of the same substance. Let 

 us now detach this thin layer from the thicker plate and expose it (thus 

 separated) to the rays emerging from the plate of 8 millimetres ; it Avill 

 reflect part of them and transmit all the rest. The quantity lost will 

 therefore represent the effect, and onli/ the effect, of the two reflections. If 

 this experiment be made with care, the number found as the quantity 

 transmitted Avill be verj' nearly 0-923 f, and therefore the number repre- 

 senting the part reflected will still be 0-077. This is true not only with 



mingle with the bundle of calorific rays directly transmitted, but will be thrown 

 laterally. If their action on the thermomultiplier be appreciable, it must be 

 perceived in the diminution of the effect. But the galvanometer always ex- 

 hibits the same deviation, both when the plate stands perpendicularly and when 

 it stands obliquely. These multiplied reflections have, therefore, no sensible 

 influence whatsoever on the measure of the transmission which is represented 

 exactly by the number 0-923. 



• 0-315 in. 



•f Are these small variations, which do not amount even to Inindredth parts, 

 to be attributed to a difference of polish in the surfaces, or to a difference of 

 energy in their reflecting powers ? This question seems not to admit of an 

 easy solution by experiment. However, if we were to judge by the complete 

 analogy which these phEenomena bear to those that occur in the reflection of 

 light, we sliould think it highly probable tliat the trifling differences observed 

 partly depend on the indices of refraction in the different substances of which 

 the plates are comjwsed. 



