412 Prof. Rudberg on the Variations lohich Temperature 



The elasticity of the internal and the external air being 

 the same, and almost equal to 0™'76, we have when t the 

 mean temperature of the external air is +16'^', the index of 

 refraction 



= V l+0'000554<, and 

 at +80% or the mean temperature of the internal air, the index 



will be 



= VI +0-0004529 

 Whence, if v is the ratio between these two elasticities, 

 V = 1-000051. 



I,, If A is the deviation produced by the prism, and 8 A the 

 small angle through which the i-ay deviates still more in pass- 

 ing througli the plate of mica, we have 



Sin A = V . sin ( A — 8A), or 



"-1 

 8A = . tang A 



which correction ought always to be added to the observed 

 deviation. Calling, in short, +rfA the variation produced 

 in the deviation corrected by 8 A, and putting s = the re- 

 fractuig angle of the prism + d s = the variation of this angle, 

 and n = the index of refraction, we shall have with a suffi- 

 cient approximation, for crystals with one optic axis, 



sin ^(A+dA+e) 



11 = =-^ — .=^ % 



V . sm y e 



and for crystals with two optic axes, 



sin i (A +rfA+s + ^s) 



71 ^ =— - — :-^= ^= -. 



V . sin y (e + ds) 



The results of the observations were as follows : 



1. Calcareous Spar. — Refracting angle of prism =59° 55' 9". 

 a. The Ordinary Spectrum. — For this spectrum I found the 

 i-emarkable property, that the crossed xvires of the telescope be- 

 ing jdaced in the ray F, at a low temperature^ they always re- 

 mained fixed there at the highest temperature., notwithstanding a 

 difference of 64;°. 



The variation which sometimes presented itself during the 

 repetition of the observation was too small to be measured. 

 It is besides evident that the slightest change in the refract- 

 ing power, would have produced a remarkable change in 

 the deviation, which was = 52° 53' 43". This appareiit in- 

 variability of the deviation proves a small decrease in the re- 

 fracting power; because if tlie latter had been perfectly con- 

 stant, the observation would have shown an augmentation of 



