1817.] Royal Institute of France. 157 
made of the refraction of a great number of gases have shown them 
that this law does not always agree with the results of observation. 
We see, then, that the refracting power of a body, far from 
being constant, as. the Newtonian theory would seem to prove, in 
the most natural hypothesis which could be made relative to refrac- 
tion, undergoes, on the contrary, variations either from the effect 
of changes of density, or from the state of combination in which 
the body is. ‘To determine the influence of each of these causes in 
particular, it is necessary to measure with accuracy the refracting 
powers of a great number of substances, and those of the com- 
pounds which they form. The work undertaken with this view 
comprehends already a considerable number of bodies. But the 
authors have felt the necessity of extending it still further before 
endeavouring to connect by any general law the different results 
which they have obtained. 
To those who may start some doubts about what they consider as 
the most natural hypothesis which can le made respecting attraction, 
the authors oppose two passages, in which the author of the Mecha- 
nique Celeste has shown himself of the same opinion, expressing 
himself thus, book x. p. 274:—‘* I shall suppose that the value 
K jis in the state of liquidity and of vapour. It is, in fact, the 
nz 
most natural hypothesis which we can admit.’ And in p. 23 of the 
preface to vol. iv. :— 1 endeavour to supply it by supposing that 
the refracting forces of water and vapour are proportional to their 
respective densities ; in this probable hypothesis,” &c. But what 
appears at one time the most natural and probable, may cease to 
seem so when new facts have brought. new light. And we may 
remark the reserve of the author of the Mechanique Celeste, who 
speaks of his calculation only as an essay which leaves him at 
liberty to try other means, if what presented itself at first was 
attacked by some fact or experiment which obliged him to reject it. 
The facts established by MM. Arago and Petit appeared to them 
of such importance that they have been anxious to follow the con- 
sequences in the different phenomena which have a more or less 
direct connexion with that of refraction, 
The coloured rays of which white light is composed are unequally 
separated from each other by their refraction in bodies of different 
natures, and this is what constitutes the difference in the dispersive 
force of bodies. It is most natural to measure the dispersive power 
by the different refractive power relative to the extreme colours of 
the spectrum ;, and in the theory of Newton, this difference ought 
to be constant for the same body, as well as the refracting power of 
the mean rays, 
Experience having shown that this last power diminishes with the 
intensity, it was easy to foresee that the dispersive power would 
diminish also. But it was important to examine if these variations 
would follow the same law. To determine the point, it was neces« 
sary to ascertain the dispersive power of the liquids and vapours, 
