294 On the Affinities of Bodies for Light, 
of hydrogen. Thus supposing that the composition of am- 
monia was unknown, but that barely the nature of its prin- 
ciples was known, their relative proportion might be disco- 
vered by the proof of their refractive power as well as by 
chemical analysis. 
And in water composed of hydrogen and oxygen, in which 
there is a strong, intimate, and very dense combination re- 
latively to its component parts, the same law still holds, 
with some slight difference only, which may arise either 
from the uncertainty even of the experiment which Newton 
himself made to determine the refractive power of water at 
a time when the necessary instruments were much less per- 
fect than they are at present, or rather from the influence 
of the state of condensation of the molecules of the liquid 
compared with the state of gaseous dilatation. Thus the re- 
fractive, power of water, calculated according to the propor- 
tions given by Humboldt and Gay-Lussac in their excellent 
memoir upon eudiometry, is = 1°50, that of the atmo- 
spheric air being 1. But according to Newton, and by em- 
ploying for the mean refraction that of yellow light, it would 
be = 1°73; stronger than the preceding by about one-eighth 
of the total value. Far from being astonished at this dif- 
ference, we ought to be surprised that it is not more ¢con- 
siderable, when we reflect upon the enormous condensation 
experienced by oxygen and hydrogen thus combined. 
After having proved-the exactitude of this new process of 
analysis by a comparison of its results with those already 
obtained by the ordinary chemical analysis, it became inter- 
esting to endeavour to apply it to the solution of the grand 
znigma of the composition of the muriatic acid. The au- 
thors therefore proved, not without great difficulties in the 
manipulation, its refractive power in the state of gas. They 
found it a little more considerable than that of azote ; which 
proves that the above acid is not a compound of azote and 
oxygen, since the presence of the latter principle would di- 
minish instead of augmenting the refraction of the former. 
Nor is it an oxide of hydrogen Jess oxygenated than in wa- 
ter, as has been supposed from some late Galvanic experi-— 
3 ! ments 5 
