156 On the Affinities of Bodies for Light ; 
be ascertained, in some measure, from their purity; but the 
authors did not content themselves with this, and the gases 
were always submitted to chemical analysis. Messrs. The- 
ard, and Berthollet junior, assisted their friends in this part 
of their labour; and their names inspire confidence. It be- 
came still mecessary that these gases, although produced 
very pure, should be introduced with all their purity into 
the prism of the experiment. The authors succeeded in this 
essential requisite by a particular disposition of the cocks of 
communication, difficult to be understood without the as- 
sistance of drawings. 
« Knowing,” they say, “ the specific weights of our 
gases, and the refractions exercised by them upon the light, 
we concluded, by calculation, their refractive power com- 
pared with that of the atmospheric air. 
«© What we mean here by the term refractive power, is 
not only the deviation produced upon the luminous ray, nor 
is it the angle measured by-this deviation, If the function 
of the distance which expresses the action of bodies upon 
the light was of the same form for all bodies, and did not. 
differ relatively to each of them but by the produce of their 
density, and of a constant coefficient dependent on their na- 
ture, the quantity, which we call the refractive power of a 
body, would be proportional to the intensity of its attractive 
force for the light; but in every case it is the sum of all the 
actions exercised by the body multiplied by the element of | 
space and by the density. These rigorously exact notions 
are conformable with the principles given by Newton and 
by the author of the Mecanique Celeste: it appeared neces- 
gary to recall them to observation, because it is only by at- 
taching precise ideas to principles that we can employ them, 
and follow with certainty the consequences which may he 
deduced from them,” 
By applying these principles to their experiments the au- 
thors of the memoir have established the refractive power of 
yarious agriform fluids. Oxygen of all these fluids, and in- 
deed of all natural bodies yet observed, refracts the least, 
and hydrogen refracts the most,—it refracts 6} times more: 
than common air, 
Newton 
