328 Mr. T. P. Dale on the Index of Refraction and 



of this investigation. When the line B is given, the octave 

 ray to H may be obtained from A by interpolation, and these 

 values were calculated by this method. It gives results suffici- 

 ently near the truth, but has been verified in case of CS 2 (temp. 

 1 0, 5) and mint hydrocarbon by reference to the two indices 

 A and H, so that the table depends on these two indices only. 



It will be seen that in the cases examined — — is constant, 



within the limits of observation, unaffected by temperature. 



Dr. Gladstone and myself, in our joint paper, read before 

 the Royal Society, May 6, 1863, and printed, p. 31 7, in the 

 Transactions, have pointed out that in consequence of an 

 accidental relation amongst the coefficients used in Cauchy's 

 series, that the equation 



v =^u — 3 Of— Mb) 

 holds with considerable accuracy in the cases examined. It 

 appears generally to give a fair approximation to the limit*, 

 and thus to interpolate indices within the visible spectrum. 



While, however, this investigation was in progress Prof. 

 Langley's first paper on the determination of wave-lengths in 

 the invisible spectrum appeared (Phil. Mag. March 1884). 

 He there gives the wave-lengths of lower heat-rays down to 

 length 2'030, and shows that the observed indices of refraction 

 are not only below the limit obtained by Cauchy's formula, 

 but even below those found by other formulae. The material 

 of his prism was flint glass. I found that the values derived 

 from 6 a and 2 n, while fairly accurate in the visible spectrum, 

 were increasingly discordant in the longer wave-lengths. If 

 Oh were increased the lower indices were found more accor- 

 dant ; but then the visible spectrum was not adequately 

 represented within the limits of any error which could be 

 imagined possible. This seemed to throw a doubt on the 

 existence of a limiting value. It appeared, however, that the 

 glass prism absorbed the longer waves considerably. The 

 Phil. Mag. for May 1886 contained another paper by Prof. 

 Langley, the material of the prism being rock-salt. On com- 

 paring the values found by observation with those furnished 

 by the above equations, the agreement was sufficiently close 

 to encourage investigation, and, accordingly, the calculations 

 were carried further, the approximation being to single 

 seconds in the value of 6 and every value found. 



The results are embodied in Tables I. and II. It will be 



* The divergence in case of CS 2 is, however, considerable, and v is too 

 large in comparison with that given by the above equations. Especially 

 does this occur in highly dispersive media. 



