BIAXIAL CRYSTALS. 195 
angle, opposed, no ray passes between the opposite and 
parallel faces without undergoing deviation, he rendered 
all objections vain.* 
* The paradoxical mention of proofs of refraction, where no refrac- 
tion takes place, may need a brief explanation. 
Fresnel’s experiment, here referred to, was performed by means of 
the simple interference of two rays produced by reflexion from plane 
mirrors very little inclined from the same plane, or by transmission 
through a very obtuse-angled prism. If, in the path (as explained in 
a subsequent note) of each of the two interfering rays, plates of glass 
of exactly the same thickness are interposed, the position of the stripes 
remains unaltered; but if the plates be cut from a dbiawial crystal in 
different directions with respect to its axis, but still of exactly the 
same thickness, even if we employ those rays which correspond to the 
ordinary rays in Iceland spar, there will be a displacement of the 
stripes, showing a difference of velocity or refraction, in these rays, 
on the principle hereafter explained, (see note infra.) 
The more direct experiment alluded to consists in this: Fresnel cut 
two prisms in different directions from the same crystal of topaz, 
which, being cemented together with their axes in one line, were 
ground together to exactly the same angle, and the whole achroma- 
tized by another opposed prism. On looking through the two prisms 
thus fixed side by side at a line of light, that line was seen to be bro- 
ken at the junction, indicating different refractions in the two. 
The law of Huyghens, or the construction of the sphere and sphe- 
roid, was found to hold good not only in Iceland spar, but in many 
other doubly refracting crystals. But these were all characterized by 
possessing only one axis or line along which there was no double re- 
fraction, and which, by the aid of polarized light, is easily detected as 
forming the centre of the rings. 
Sir D. Brewster, in examining a vast variety of crystals, discovered 
a class in which there was not one such axis, but two, and in which 
the rings consequently assumed new and more complex forms, being 
either arranged in two separate sets if the axes were distant, or in 
coalescing curves if they were close. 
For biaxial crystals Huyghens’s law will not apply. The incident 
ray is divided into two; but neither of them follows the law of the 
sines represented by the sphere in his construction. One of the rays 
is, indeed, usually less subject to deviation than the other, and thus, 
for convenience, is still often called the ordinary ray; but both are, in 
strictness, extraordinary rays. 
