264 SENARMONT ON THE OPTICAL CHARACTERS OF 



by crystalline laminae, perpendicular to a bundle of parallel 

 polarized rays, and the angle described by the lamina in its own 

 plane, when the original polarization was no longer disturbed, 

 was observed at first in one half of the plate, and then in the 

 other. 



This method appears to give tolerably accurate results when 

 the observations are repeated with different plates. The section 

 of the crystals may be verified by means of the goniometer ; it 

 is moreover evident, that when the surfaces in juxtaposition coin- 

 cide perfectly, there must be a considerable error in the section 

 to produce any appreciable inaccuracy in the results. 



The other method, which I shall call the second, is applicable 

 only when the plane of the optical axes is perpendicular to the 

 horizontal diagonal of the base. It is very nearly the same as 

 that adopted by Mr. Miller in his researches*. 



I selected from among the faces of the crystal two which were 

 parallel to each other and normal to the plane of the optical 

 axes, through which might be seen, in convergent polarized light, 

 at least one of the systems of rings. When such faces did not 

 exist, I cut them artificially, taking care at the same time to 

 leave some of the natural faces untouched, in order that the 

 position of the former ones might be accurately determined by 

 means of the goniometer. I then measured the apparent in- 

 clination of each optical axis to the normals of these artificial or 

 natural faces in the following manner : — 



The plate was fastened upon a Wollaston^s goniometer with a 

 little soft wax, in such a way that its plane was perpendicular 

 to that arm of the carrier which is moveable about its axis 

 parallel to the limb. The adjustment is known to be exact 

 when a reflected image remains immoveable during that revolu- 

 tion. On bringing the plane of the optical axes parallel to the 

 plane of the limb, the rings were observed, making their centre 

 coincide with a very remote object seen directly, then rotating 

 the plate in its plane 1 80°, and again bringing the centre of the 

 rings to coincide wdth the same object. Between these two 

 positions the limb had passed over an angle twice as great as 

 that which the optical axis makes with the normal. In order to 



• Transactions of the Cambridge Pliilosophical Society, vols, v, and vii. 



