TRANSACTIONS. 
at a perpendicular incidence in transparent bodies, compared 
with the large quantity reflected at the same incidence in metals, 
is sufficient to convince us that reflection has no relation to the 
densities of bodies. To remove, however, an objection which 
might be urged, that the extension of the particles of bodies 
may not bear an invariable ratio to their weight, it will be 
necessary to examine cases where a metal, by combining with a 
new element or elements, has acquired the property of transpa¬ 
rency, and thus possesses an evident refractive power. By 
knowing the comparative weights of the metal in the two states, 
it is easy to calculate the relative numbers of similar particles in 
equal surfaces, and of course to calculate the relative quantities 
of light which ought to be reflected, if caused only by the 
ponderable particles of metal. Experience is so much at vari¬ 
ance with the hypothesis under examination, that the other 
elements in the compound may be considered even as lending 
no assistance at all. 
The results obtained by Photometry show that the metals, 
with the exception, perhaps, of two or three, reflect two thirds 
and upwards of the light incident perpendicularly on them. For 
the reflective power of the transparent bodies we may use the 
analytical formula of M. Poisson, (which was admitted by M. 
Fresnel,) to calculate it from the refractive index, though it gives 
most probably, in all cases, quantities too large and of course 
proportionally favourable to the controverted hypothesis. 
Proceeding in this manner for glass of Antimony, the reflection, 
according to Fresnel’s hypothesis, should have been at least 46 
rays of every 100 incident, whilst the quantity given by the 
analytical formula is only 19"3 rays. 
In the white oxide of arsenic or arsenious acid the reflection 
should have been 31*9 (taking even the old number for the 
specific gravity of the metal,) in place of less than 8*3, which it 
really is. 
In the red silver ore it should have been at least 37*5 rays if 
the hypothesis were correct, instead of less than 19*2 as de¬ 
termined by the analytical formula. 
If the metals of the alkalies and earths might be assumed of 
equal reflective powers with the other metals, and it is most 
likely they are, the chloride of sodium would form one of the 
strongest cases which could be brought forward ; for whilst it 
really reflects only about two per cent., it ought, according to 
the controverted hypothesis, to have reflected upwards of 60 
rays of every 100 incident, from the metal in the chloride being 
almost as dense as in its proper state. 
