TRANSLATIONS AND SELECTED ARTICLES. 303 
The process in order to decide whether the substance which ren- 
ders the sun visible is liquid, solid, or gaseous, will be nothing more 
than a very simple application of the preceding remarks, notwith- 
standing the difficulties which appeared to arise from the enormeus 
distance of that star. 
The rays which render visible to us the borders of the disk have 
evidently issued from the incandescent surface at a very small angle. 
If, then, the borders of the two images seen directly through tbe 
polariscope, appear colored, the light of these borders must proceed 
from a liquid body, for every supposition which would make the ex- 
terior of the sun a solid body is definitely excluded by the observa- 
tion of the rapid change of form in the spots. And if the borders 
retain in the polariscope their natural whiteness, they are of necessity 
gaseous in character. Now, observations made by viewing the sun 
directly any day of the year through large polariscopes, fail to dis- 
cover the least trace of polarization. Therefore the substance in 
combustion which defines the sun’s outline is gaseous, and we can 
generalize this conclusion, because the different points of the sun’s 
disk, by reason of the movement of rotation, come, each in its turn, 
on the border. 
This experiment removes from the region of mere hypothesis the 
theory we have above indicated of the physical constitution of the 
solar photosphere. 
We do not find any thing, properly understood, either in the arbi- 
generalise? This is my answer; according to the two only explanations that have bcen given 
of the abnormal polarization presented by the rays emitted under small angles, the results 
ought to be the same in all respects, except that of magnitude, whatever be the liquid ex- 
amined, provided that the svrface of emergence has a sensible reflecting power. The only 
case of exception might be that of an incandescent body which should be, as rezards density, 
analogous to a gas, as, for example, the fluid of an almost ideal rarity, which many geometers 
have been led to place hypothetically at the extreme limit of our atmosphere, where the 
phenomena of polarization and color might possibly disappear. I am not ignorant that I 
should add weight to the resulcs mentioned in the text, by discussing them in a photometric 
point of view. I possess all the materials for such an examination, but this is not the place 
to develope them. I will, however, here anticipate a difficulty. It ought to be remarked 
that the lights proceeding from two liquid substances may, according to the special nature 
of these substances, not be identical as regards the number and position of the dark lines of 
Fraunhofer, which occur in their prismatic spectra. These differences are of a kind to be 
» considerably augmented by the differently constituted atmospheres which the rays have 
traversed before reaching the observer.—{ Author's note.) 
The experiments spoken of in the text and note have been objected to as inconclusive, by 
M. Kirchoff. on the ground that the liquids there examined were in a state of rest. If their 
surfaces were in much agitation, as that of the sun must doubtless be, the rays would be 
emitted at all anzles, and every trace of polarization would probably disappear. In addition 
to this, it may be remarked that Arago takes no accounit of the possible effect which might 
be produced on the rays by passing through the suu’s atmosphere.—( Trans.) 
