THIRD SERIES— REFRANGIBILITY OF HEAT. 



203 



74. In the following table I have given the me^in results of the different se- 

 ries of obsei-vations on which the above conclusions are founded, and fi'om these 

 numbers I have projected the curves exhibited in Plate XII, the dots correspond- 

 ing to the numbers here given, and the mode of projection being that already 

 explained : 



75. When we compare the preceding results, obtained with a rock-salt prism, 

 Avith those for light, we find that the received index of refraction for that sub- 

 stance would give to heat a higher degree of refrangibility than light, a result 

 contrary to all probability. This, however, is not confinned by dkect experiment. 

 Placing a bright small source of light at S (Plate XIII. Fig. 1), and a screen aij}, 

 I find the index of refraction for the most luminous rays to be higher than that 

 of any of the above kinds of heat, being at least 1.602, corresponding to a diagonal 

 ft b — 15.8 inches, as I have given it above. By two series of results derived fi-om 

 a very small oil flame (without -nick), I got 15.87 for the diagonal both tunes ; 

 and from the Locatelli-lamp (which on account of the size of the flame forms a 

 better standard of comparison with the experiments on heat) 15.76 ; so that I 

 consider 15.8 as a fan- representation of the case of hght. 



76. Yet it is quite certain that the index of refraction of the rock-salt used 

 is really much below 1.60. A single experiment with Dr Wollaston's instru- 



