342 M. 11. Ruhlinann on the Alteration produced by Heat 



In order to correct for the excentricity of the prism, the 

 above-mentioned distances had to be measured. For the first 

 arrangement, to which belong the observations marked with an 

 asterisk, — 



millirus. 



(1) Distance of the slit from the axis of J i~i Q i 



rotation of the universal circle 1 7 i Q o 



Mean . . . =7795 



(2) Distance of the prism from the uni- J 0/.9.-1 



versal circle I qji.y 



Mean . . . = 342-0 

 In the second arrangement there was found — 



Corresponding with 



(1) 





Corresponding with (2). 



7789 









337-4 



7791 









337-5 



7790 









337-3 



7790-5 









337-4 



7790 









337-2 



7790 









337-1 



. =7790-1 





Me 



an 



= 337-3 



Mean 



From the complete agreement between the results obtained 

 from both arrangements we can conclude as to the accuracy of 

 the method. 



The centres of the direct and of the refracted image of the slit 

 did not lie in a horizontal plane, but differed 7 ! 45" in height. 

 The correction to be introduced is easily got from the right- 

 angled spherical triangle ABB', fig. 9. Let AB' be the devia- 

 tion of the luminous ray read off on the chief circle, and B B' 

 the difference of height ; then the angle of deviation which has 

 really to be measured is A B, and cos A B= cos AB'- cos B B'. 

 If we further notice that log cosB' = 9'9999988, it is evident 

 that no correction need be made of the angle read off. 



The angle of least deviation, read directly off on the uni- 

 versal circle, had, on account of the excentricity of the prism, 

 to receive a + correction i/r. To facilitate this, small auxi- 

 liary Tables were constructed giving corrections for every five 

 minutes. 



Finally, concerning the errors of individual observations, they 

 may arise from want of accuracy (1) in the determination of <£ 



