' dnd Saccharine Solutions exercise on Heat. 469 



the way of its exact determiuation ; for the thermoscopie appa- 

 ratus, notwithstanding its delicacy, is ever less sensible than the 

 eye. To make the thing plain, we will observe in the first place, 

 that the magnitude of the observed deviations never exceeded 

 25 divisions under the most favourable circumstances, that is to 

 say, when the principal section of the analyser was parallel to 

 the plane of polarization of the ray. At 85 and at 95 degrees 

 from this position, we ought, according to the law of INIalus, to 

 observe a deviation equal to 35 ~ x cos^ 85, or 0"18 div., and it is 

 scarcely possible to appreciate so small a deviation. It wovild 

 not be attended with greater certainty to seek, by simple galva- 

 nometric observations, the position to be given to the analyser 

 so as to obtain the maximum effect. 



We have had recourse to other methods, which we will describe 

 in their order. 



If it be not easy to find the position of the analyser by which 

 the maximum deviation is obtained, the magnitude of the latter 

 may be detenuined without ambiguity. To obtain it, it is only 

 necessary to observe the effects which correspond to two rectan- 

 gular positions of the spar. According to the law of Mains, 

 their sum is equal to the maximum effect. In dividing by this 

 maximum deviation that which corresponds to any determinate 

 direction whatever of the principal section, we obtain the square 

 of the cosine of the angle which this direction makes with the 

 changed plane of polarization, and thus the magnitude of the 

 rotation is found. When this occu.rs between 30 and 60 degrees, 

 it is convenient to obsen^e the deviations for the positions and 

 90 degrees of the analysing spar. The square of the cosine of 

 the sought rotation is obtained immediately by dividing the 

 effect obtained at zero by the sum of this effect and that obtained 

 at 90 degrees. 



When the rotation is less than 30 degrees or greater than 60 

 degrees, the galvanometric deviation observed at zero, or that 

 which corresponds to 90 degrees, is too feeble to give a certain 

 result ; two other rectangular positions should be chosen which 

 are more convenient. 



We will now give the results obtained by experimenting in the 

 manner indicated. 



Heat of the green portion of the spectrum; the line 0° — 180° of 

 the limb of the analyser being vertical and coincident with the 

 plane of primitive polarization. 



Length of the tube filed with essence of turpentine, 0-15 millim. 

 Deviations. 

 Spar at^O degrees . . ^4-4 j Rotation 58° 54 . 



