on Magnetic Rotatory Polarization in Gases, 395 



of magnitude as the 'latter, it was essential to determine them 

 with the greatest accuracy. 



They may be grouped in the following manner : — 

 I. Variations in electromagnetic intensity. 

 II. Want of homogeneity in the luminous source. 

 III. Magnetic rotations caused by the glasses and the 

 glass of the mirrors interposed in the course of the 

 luminous rays. 

 IY. Corrections due to variations of temperature in the 

 gases. 

 V. Displacement of the luminous point. 



I. Variations in Electromagnetic Intensity. — We have 

 already stated that the heating of the conductors during the 

 passage of the electric current and the operation of the pile 

 gave rise to a regular variation in the electromagnetic inten- 

 sity, and that the intensity of the induced current, which is 

 in proportion to that of the principal electric current, was 

 measured repeatedly. The numbers obtained enabled the 

 observed results to be reduced by a simple proportion to what 

 they would be if the intensity remained constant. The accu- 

 racy of these numbers will therefore depend on the precision 

 with which the intensities are measured. We made use of a 

 sine-compass placed at 10'50 metres from the middle of the 

 large solenoid in the apparatus. At this distance the mag- 

 netic influence on the compass was still sensible, and deflected 

 the needle about 10', either to the right or left, according to 

 the direction of the current in the bobbins. In order to 

 eliminate this disturbing element, each determination of the 

 intensity consisted of two successive measurements, made by 

 reversing the direction of the electric current in the bobbins; 

 and the average of the two deviations obtained was taken as 

 the definitive measurement. 



During the course of a series of observations these deter- 

 minations were repeated very frequently, every ten or fifteen 

 minutes ; and it is certain that, for the measurements included 

 in these intervals of time, the numbers obtained by interpola- 

 tion gave the intensity of the current with a precision equal 

 to that of direct determinations. 



The deviations of the compass varied between 25° and 22°; 

 the total variation from morning to evening for the same 

 series scarcely amounted to 2°. All the results were reduced 

 to what they would have been if the deviation had been con- 

 stantly 24°. The error in the measurements of the angles 

 did not exceed 1/; so that, with respect to the intensities, the 

 error could not amount to O00X of their value. This accuracy 



