Page:Optics.djvu/195

 It is necesarynecessary [sic] to make the edge of the crystal prism as sharp as possible, in order that the corrections made for its thickness be inconsiderable. In fact the best way of making the observation, when it can be done, is to let the rays pass actually through the edge, for then the two refracted pencils have but an infinitely small space to pass through, before they emerge together. For a similar reason, the pillar should, in the experiments, not be placed very near the vertical scale, on which the coincidences are observed, because the corrections for thickness, which are nearly insensible at moderate distances, might become more considerable.

Besides these precautions, the faces of the prisms should be ground very smooth, and plane, and their inclinations should be accurately determined, by the reflecting geniometer. Moreover, it is necessary that the direction in which the prism is cut, relatively to the axis, or axes of the crystal, should be accurately known; in order to which these axes should be previously determined, either by immediate observation of the directions in which the reflexion is single, or by inferences drawn from the experiments themselves, or by other processes that will be hereafter detailed. By following these rules, the observer will be, I believe, perfectly satisfied, as to the nicety and accuracy of the mode of experiment. These advantages are derived from the multiplicity of the coincidences, seen on the doubly-refracted scale. The alternate superpositions and separations of the lines of division produce, if I may so express myself, the effect of verniers, and enable one to judge with extreme precision, of the point where the coincidence is most perfect.

Suppose then, that by this, or some analogous process, we have determined for some given crystal, the deviation of the rays in different directions round the axis, it remains to find out the general law, which regulates the phænomenon in all cases. This Huyghens has done, as has been before mentioned, for crystals with one single axis, by means of a remarkable law that he connected with the system of undulations: but this same law has since been deduced by M. Laplace, from the principle of material attraction.

If light is to be considered as a material substance, the refraction of its rays must be produced by attractive forces, exerted by the particles of other bodies on the luminous molecules, forces which