4 Fred J. Bates 



frame with a stationary glass cell ( loo mm. X 50 mm. ' 12 mm.), 

 fig. I, filled with the solvent. If, now, the tube containing the 

 solution be rotated into the path of the light, it displaces solvent 

 equal to its own length. The tube containing the solvent serves 

 merely as a carrier for the cover-glasses, which compensate for 

 those closing the tube containing the solution. If the cover- 

 glasses are not of the same thickness the only change that can 

 occur in the optical path will be caused by replacing of the extra 

 thickness of glass by the solvent when the next tube is thrown 

 in. A difference of 0.02 mm. in the thickness of the cover-glasses 

 would give an error of 0.004° i^i the rotation. Thus it is evident 

 no mechanical errors can appreciably affect the measurement. 



In the usual methods for measuring the rotation of plane- 

 polarized light there appear in the observing telescope, when the 

 magnet is excited, several partially overlapping images of the 

 field. This is due to the successive internal reflections of the 

 beam of light within the tube. When the analyzing nicol is set 

 for the extinction of the ray that has passed through the liquid 



268 



