Lord Rayleigh's Investigations in Optics. 51 



With regard to the material for prisms, the choice lies prin- 

 cipally between various kinds of glass and fluids. It is not 

 without difficulty that glass is prepared free from striae and 

 well annealed; but solid prisms have the great advantage 

 that, when once good ones are obtained, there is no further 

 trouble. In consequence probably of the practice of using 

 in all cases a standard angle of 60°, an exaggerated idea is 

 often entertained of the advantage of great density. Accord 

 ing to Hopkinson, the difference of indices relative to the lines 

 D and B for dense flint, extra-dense flint, and double-extra- 

 dense flint are respectively '0067, -0075, and '0091, which 

 numbers are inversely as the equivalent thicknesses. 



Of fluids, bisulphide of carbon has the merits of cheap- 

 ness and very high dispersive power, the difference of in- 

 dices for D and B being -0126. When pure, it is also in a high 

 degree transparent. On the other hand, its sensitiveness to 

 variations of temperature is so great that 1° 0. makes about 

 nine times as great a change of refrangibility as a passage 

 from Di to D 2 . Great precautions are therefore required to 

 prevent inequalities of temperature from destroying definition ; 

 and frequent shaking is generally necessary. Some observers 

 have thought that, apart from inequalities of temperature, or- 

 dinary bisulphide of carbon tends to arrange itself in strata 

 of different refracting-power ; but this does not seem to be 

 established satisfactorily. In some recent experiments with a 

 rotating stirrer, introduced with the view of promoting uni- 

 formity of temperature, I obtained evidence of a thin layer 

 of moisture floating on the surface. Under the action of the 

 stirrer this layer was broken up and the liquid rendered very 

 irregular. A few lumps of chloride of calcium introduced into 

 the liquid absorbed most of the layer ; but an arrangement 

 such that the free surface remained undisturbed would have 

 been preferable. 



Within the last few months Prof. Liveing * has proposed 

 the use of a solution of iodide of mercury, which is consider- 

 ably more dispersive than even bisulphide of carbon, the dif- 

 erence of indices between D and B amounting to *017. This 

 liquid is of a yellowish colour, and is hardly sufficiently trans- 

 parent, even at the lower end of the spectrum, to make its 

 use advisable in very powerful instruments. But for some 

 purposes its great dispersion is an important recommendation. 

 Using a single prism of 60° with an available thickness of 

 1J inch, I have obtained results which many pretentious in- 

 struments could not surpass. 



* " On the Dispersion of a Solution of Mercuric Iodide/' Cambridge 

 Proceedings, May 19, 1879. 



E2 



