IN STUDYING QUESTIONS OF CHEMICAL MECHANICS. 303 
elementary liquid threads of equal diameter, formed under the 
different densities 9. To connect them then, by a common ex- 
pression, let [«] be the deviation which the homogeneous liquid 
under consideration would impress on the plane of polarization 
of the simple ray chosen as a type, by acting upon it under the 
same circumstances through the unit of thickness, and with an 
imaginary density equalalso to unit. Then, the density becom- 
ing 6, without the energy of the molecular action changing, the 
deviation through the unit of thickness will be [a] 8, in the pro- 
portion of the mass of active matter contained in the two fluid 
columns of sensible dimension and of equal length. Moreover, 
this length becoming / for the same density, the total deviation 
will be [«]/78. Having then designated by « its magnitude, as 
it is observed under these new circumstances, we should have* 
[aj l8 =a; 
lel=7 De Lhd ee pay 
The second member of this equation (1.), which expresses [«], 
should therefore always present the same numerical value for the 
different values of 8 and of /, when the conditions on which the 
formula is based are realized; that is to say, when the diversity 
of deviation assigned to the active groups of the liquid does not 
change the individual energy of their action on polarized light. 
The quantity designated here by [«] constitutes what I have 
called the rotatory molecular or specific power of bodies. 
9, The only means by which the density of a free liquid can 
be modified without changing the nature of its ponderable ele- 
ments, is to subject it to momentaneous variations of temperature 
which contract or dilate it within sufficiently restricted limits 
for it to return spontaneously to its primitive state, when it is 
replaced in its original condition. However, these precautions, 
which determine the identity of the restitution, give no assurance 
consequently 
* I have presented the demonstration of this result several times, and under 
different forms, in the Wém. de l’ Acad. des Sciences, particularly t. xiii. p. 116; 
t. xv. p. 100; t. xvi. pp. 241 and following. But the reasoning upon which I 
found it here appears to me infinitely preferable, as being independent of all 
hypothesis, and even of any theoretical notion on the mode of action by which 
the active molecules influence the luminous principle. This mode of reasoning, 
established on the distinction of physical and geometrical infinitely small quan- 
tities, is the same as Poisson employed to reduce to means the elementary ac- 
9 which produce magnetic phenomena, capillary effects, and the equilibrium 
of heat. 
y2 
“~ 
