IN STUDYING QUESTIONS OF CHEMICAL MECHANICS, 299 
bubble, where the surfaces of the acting substance are curved 
by capillary action; and supposing also that the feeble portion 
of light is neglected which may be depolarized, or polarized in a 
different direction, by the ordinary refraction at its entrance and 
emergence, if the plane of introduction were oblique to the plane 
of primitive polarization*. With these exceptions, the experi- 
ment therefore proves that the internal mass is constituted 
symmetrically around each of these geometric points; and con- 
sequently the dissymmetrical effect undergone by the ray, around 
its plane of primitive polarization, cannot result from such a 
state, but from the individual action of the insensible particles 
which compose this mass. Similar characters demonstrate the 
existence, or rather the persistence of the same molecular action 
in certain media brought to a state of perfect solidity. But in 
this case it becomes observable only under certain conditions, 
which I shall indicate hereafter ; and I shall then relate, as ex- 
amples, cases in which it is established by these processes. 
* The homogeneity of the transmitted ray is obtained with sufficient exact- 
ness by receiving it through certain glasses coloured red by protoxide of copper, 
in which this property has been previously ascertained by analysis with a prism. 
But for comparative determinations, like those under consideration, the expe- 
riment may be advantageously facilitated either by measuring the deviations 
by the naked eye, if the medium observed is colourless, and taking for type the 
extraordinary tint of bluish-violet, which is always manifested in comparable 
azimuths, except in solutions of tartaric acid; or, if the medium is coloured, by 
equalizing its tints in different directions by the interposition of a glass of an 
analogous, but somewhat darker tint, as I have explained in the Comptes Ren- 
dus des Séances de l’ Académie des Sciences, t. xv. p. 624 and 632. I have re- 
cently applied the first process to the observation of a variety of turpentine, 
which produced a deviation toward the right. The dimensions of the bottom of 
the case which contained it were 90 millimetres and 50 millimetres; its height 
49 millimetres. The deviations in these three directions were observed by the 
naked eye, and determined by the azimuth of the evanescent bluish-violet tint, 
the unfavourable state of the weather not allowing conveniently the employment 
of a coloured glass. On reducing them to the true zero of the apparatus, that 
is to say to the observed plane of polarization of the primitive ray, they were 
found to be respectively + 35°50; + 219-00; + 20°50. To compare their 
relations to what they ought to be, in the condition of proportionality, I form 
their total sum and that of the corresponding dimensions, which gives 77° of 
mean deviation for a total thickness of 189". Then, deducting the partial 
deviations from this mean result, according to the respective thicknesses, we 
find 369-67 ; 20°-37 ; 199-96, that is to say, the first a little higher and the two 
last a little lower than they were found by observation. Now this apparent 
difference only confirms still more the law of proportionality, which applies to 
one and the same simple ray ; for the substance observed being sensibly coloured 
greenish-yellow, the transmitted ray lost a much greater proportion of its most 
refrangible, and consequently most deviable elements, through the long dimen- 
sion of the apparatus, than through the two smaller ones; so that the first de- 
viation should be greater than the observed, and the two last lower, when all 
three are calculated according to the mean sum of their observed values. 
