298 BIOT ON THE EMPLOYMENT OF POLARIZED LIGHT 
tional to their mean diameter. To effect this test with complete 
exactness, boxes of a rectangular or cubical form should be con- 
structed, the sides of which are of thin glass with parallel faces, 
the upper one of which may be removed and replaced at will, 
the fixity of their apposition being moreover secured by metallic 
bars attached with screws. After having measured exactly the 
imner interval of the glasses in various directions in which they 
are opposed, the flask, containing the filtered and viscous tur- 
pentine to be tested, is slowly heated in a water-bath, gradually 
to boiling. When it is thus brought to an entire state of lique- 
faction, which renders it completely homogeneous, it is poured 
gently into the hot box, avoiding the formation of internal bub- 
bles of air, or waiting till they are disengaged: the upper glass 
is then replaced, which is perforated with a circular orifice, 
through which the excess of the liquid escapes; and, after the 
mass is coagulated by spontaneous cooling, the upper glass, which 
the viscosity of the matter has already rendered adherent, is 
fixed by the metallic bars. We thus obtain a limpid homoge- 
neous mass, of known dimension and of uniform constitution, ex- 
cept around the orifice, where the contraction caused by the 
cooling produces an empty space, which is filled with air. But 
this, far from being an inconvenience, is an advantage, because 
the regularity of the constitution of the medium is not at all in- 
terfered with when the observation is made at some distance from 
the empty bubble, and a space is left, in which the matter may 
extend freely, if it be desired to raise it anew by heat to pro- 
gressively higher temperatures, or even to complete liquefaction, 
in order to observe it in this state without changing its dimen- 
sions. I shall hereafter recur to the modifications which this di- 
versity of temperature produces in the optical effects. For the 
present I limit myself to the consideration of those which are 
obtained when the inclosed matter is in a state of coagulation. 
In this case, if a homogeneous ray is transmitted, not only under 
the normal incidence, but under any equal incidence whatever, 
through the various pairs of parallel faces of the apparatus, the 
deviations produced in its plane of polarization are equal among 
themselves in the cubic case, and proportional to the lengths of 
the course in the rectangular case, whatever parts of the 
mass be traversed, and whatever be the directions in which the 
transmission is effected; provided always that the observation 
be made at a sufficient distance from the limits of the empty 
