418 Mr. R. W. Cheshire on a New Method of Measuring 



mercury- vapour lamp used in conjunction with the W ratten 

 mercury monochromatic filter provides all the light that is 

 required for the purpose. 



Experimental readings have been taken on a small prism 

 with a view to determining the sensitivity of the method 

 for the D line. The glass cell was first cemented on to the 

 Pulfrich prism in the usual fashion with a solution of barium 

 mercuric iodide (7i D = l'79) and adjusted so that the fringes 

 observed in monochromatic light ran parallel to the direction 

 of the incident grazing light. The prism under test was 

 then temporarily cemented to the under surface of the cover- 

 plate with its refracting edge vertical, and the cell was filled 

 with Thoulet solution to such a height that the prism was 

 immersed in the liquid to a depth of about 5 mm. The 

 vaselined cover-plate being appreciably larger than the open 

 top of the cell could then be moved to and fro, and the 

 mixture within the cell well stirred by the prism without 

 admitting air and thereby disturbing the equilibrium and 

 concentration of the immersion solution. A glance through 

 the observing telescope was sufficient to discover whether 

 the stirring had been carried far enough and whether the 

 solution was perfectly homogeneous. The cover- plate was 

 then cautiously withdrawn a little distance to allow a few 

 drops of the concentrated Thoulet solution to be dropped 

 into the mixture contained in the cell, the mixing process 

 was repeated, and observations were then taken through the 

 observing telescope to determine whether the difference 

 between the indices had been increased or decreased. It 

 was found that after four or five trials the indices could be 

 brought so nearly to an equality that one drop of Thoulet 

 solution, or one drop of water, was sufficient to change the 

 sign of the difference. The final adjustment was then made 

 with two solutions, one with an index slightly above, and 

 the other with an index slightly below that of the mixture 

 within the cell. As soon as the equality had been obtained 

 a reading was taken at once on the refractometer for the 

 refractive index of the solution, using the second sodium 

 flame mentioned above in the ordinary manner. A series of 

 determinations of the refractive index of the equalizing 

 Thoulet solution was made in this way, and then the cell 

 was removed from the Pulfrich prism, and the glass prism 

 examined on the refractometer by the usual process and its 

 index determined directly. In this way it was possible to- 

 investigate the sensitivity of the apparatus and to check the- 

 mean value obtained by a direct reading on the prism. 



