310 ANNUAL OF SCIENTIFIC DISCOVERY. 



ness if the illumination of the atmosphere by direct solar light 

 were in question. As to the atmospheric glare due to the chromo- 

 sphere and prominences, he argued that it must be relatively 

 small, because it could bear no higher proportion to the actual 

 light of the chromosphere than ordinary atmospheric glare bears 

 to actual sunlight, and we know this proportion is very small in- 

 deed. Again, as to light reflected from the atmosphere outside 

 the shadow-cone, or from the surface of the earth, he urged that 

 that also must be small, since not any part of the atmosphere 

 above the observer's horizon was illuminated by more than a 

 half-sun, while all the parts near the shadow-cone were in nearly 

 total shadow. But a fatal objection to the view that the corona 

 could be due either to the glare from the prominences or to light 

 reflected from the surrounding air, consisted in the fact that such 

 glare ought to cover the moon's disc. He then referred to a num- 

 ber of observations confirming the view that the coronal light is 

 not terrestrial ; as the appearance of glare during partial eclipses, 

 this glare always trending on the moon's disc; the relatively 

 greater darkness of the central part of the moon's disc in annular 

 eclipses ; the visibility of that part of the moon's disc which lies 

 beyond the sun in partial eclipses, the limb being seen dark on 

 the background of the sky ; the visibility of the corona in partial 

 eclipses, even its most distinctive peculiarities having been recog- 

 nized when the sun's disc is not wholly covered ; and several other 

 phenomena. He then adduced evidences to show that a solar 

 appendage, which one would expect to appear during total 

 eclipses, actually does exist. First, the zodiacal light shows 

 that the sun is surrounded by such an appendage. Dr. Balfour 

 Stewart's theory of this object, however physically sound, was 

 opposed, he urged, by too many astronomical objections to be 

 accepted for a moment. An object which exhibits no appreciable 

 parallax, which rises and sets as the celestial objects do, and 

 maintains a position in the heavens having a nearly constant rela- 

 tion to the ecliptic, cannot by any possibility be due to any pecu- 

 liarity of the earth's atmosphere. Then Leverrier has shown that 

 the motion of Mercury's perihelion indicates the presence of a ring 

 of bodies in the sun's neighborhood ; and Mr. Baxeudall has 

 drawn a similar conclusion from the meteorological records of 

 well-known observatories. Lastly, judging of the meteor systems 

 according to the laws of probability, we have every reason to be- 

 lieve that for each one our earth encounters there must be millions 

 whose perihelia lie within the earth's orbit. Since the earth en- 

 counters 56 such systems, it will be seen how enormous must be 

 the total number. These should be visible during total eclipses, 

 and since they would shine in part by reflected light, and in part 

 through their intrinsic light (for those which come as near the sun 

 as some comets have been observed to do must be melted or even 

 vaporized by the sun's heat), we have an explanation of the 

 contradictoiy accounts given by those who have applied the po- 

 lariscope and the spectroscope to the solar corona. Mr. Stone 

 remarked that there ought to be 3 sets of observations made with 



