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



seen that the section represented by BO only is effective- 

 It would o£ course be possible to utilize the full aperture AH 

 of the refractometer prism by making the width of the cell 

 appreciably greater than the full aperture which it is desired 

 to all, but, on the other hand, this change would increase 

 the thickness of the layer of immersion fluid through which 

 light has to pass. Unfortunately all the suitable immersion 

 fluids absorb very strongly light at the blue end of the 

 spectrum, and it is therefore of importance to keep the width 

 of the cell as small as possible. The refractive index of the 

 glass forming the base-plate must also be greater than that 

 of any liquid which it is proposed to examine in the cell. 

 Practical constructional difficulties rendered it inadvisable 

 to diminish the thickness of the base-plate of the cell below 

 about 2 mm., and this was therefore the dimension finally 

 adopted. With this thickness and a base-plate of extra 

 dense flint for which 72 D = 1'74, no trouble was experienced 

 in obtaining sufficient light and aperture to give good 

 settings with the refractometer up to a value w D = l'66 for 

 the fluid contained within the cell. The deposition of dirt 

 in the right angle formed by the base-plate and the sides, 

 or the presence there of a slight trace of the cement used in 

 building up the cell, would ruin the sharp definition at the 

 edge of a line seen through the telescope of the refracto- 

 meter and it was therefore considered desirable to chamfer 

 the inner edges of the base-plate over a small fraction of a 

 millimetre, as shown in fig. 2. The upper and lower faces 

 of the base-plate must be as nearly as possible plane parallel, 

 and in the cell built up for the purposes of this investigation 

 by Messrs. Adam Hilger, the plane parallelism of these faces 

 was guaranteed to an accuracy of 5". 



The side faces of the cell through which passes the light 

 used to determine the equality or otherwise of the indices of 

 the fluid and the glass under test should be plane polished, 

 but it is not necessary that they should also be parallel ; and 

 in order to secure good optical contact between the lower 

 surface of the base-plate and the Pulfrich prism, it is 

 desirable that the former should project very slightly below 

 the level of the side faces of the cell, as shown in the figure. 

 The inside dimensions of the cell as actually constructed 

 were 5 X 2 x 1*5 cm. 



The immersion fluid finally selected for use in the cell 

 was an aqueous solution of mercury potassium iodide, more 

 commonly known as Thoulet's solution. An excellent 

 account of the preparation and properties of this solution is 

 given in Johann sen's ' Manual of Petrographic Methods/ 

 where also will be found a very comprehensive table of 



