These facts appear to favor the hypothesis that the nature of the sun's spots may be somewhat like that of terrestrial tornadoes. But the atmospheric phenomena of our planet are not solely or even chiefly due to its rotation. They depend, to a great extent, on the peculiarities of its surface, in consequence of which different portions absorb and reflect the sun's rays very unequally. Disturbances in the solar atmosphere, however, cannot be referred to the operation of any known extraneous cause. What then is their probable origin?

Certain phenomena observed during total eclipses of the sun have led astronomers to assume the existence of a third solar envelope exterior to the photosphere. This assumption is sustained, moreover, by Arago's discovery of chromatic polarization. "Now granting the existence of such an atmosphere, its form, in obedience to the laws of equilibrium, must be that of an oblate spheroid, the ellipticities of whose strata differ from each other and from that of the nucleus. Consequently the equatorial portions of this envelope must be of a thickness different from that of the polar, density for density, so that a different obstacle must be thereby opposed to the escape of heat from the equatorial and the polar regions of the sun."* According to this hypothesis the phenomena of the spots result from currents and violent agitations due to the elevation of temperature in the underlying strata of the equatorial atmosphere.

It remains to describe those remarkable phenomena observed during total eclipses of the sun. When the solar disk is entirely obscured, a corona or luminous ring of variable extent may generally be seen, apparently surrounding the dark body of the moon, and resembling the circle of rays placed by painters round the head of the Saviour. This phenomenon was at first generally noticed during the total solar eclipse of May 3, 1715; a particular description of which was given by Dr. Halley and others. It was observed more critically, however, during the eclipse of July 7, 1842, than at any previous time. The descriptions of the corona by Schumacher, Littrow, Schwabe,

* Sir J. Herschel's Astron. Observ. made at the Cape of Good Hope, p. 434.

Bohm, Airy, Baily, and others, are somewhat various both in regard to its color and extent. The last mentioned astronomer, whose place of observation was Pavia, describes it as follows:

"The breadth of the crown of light from the outer rim of the moon, I estimated at one-half of the moon's diameter. The light was more intense near the moon and grew fainter as the distance increased, taking the form of diverging rays, which radiated in straight lines of unequal length and of gradually declining brightness, so that in no part of the crown could I find the well-defined figure of an exact circular ring. The sun appeared to be in the center, yet I had no means of fixing accurately the position of the radiant point. The color was uncommonly white, without any appearance of pearl, or yellow or red; and the rays had a lively and flickering appearance, such as is exhibited by gas-light."

The corona was again particularly observed during the eclipse of July 28, 1851. According to Mr. Lassell, whose place of observation was Trolhatta Falls, Sweden, it gave as much light as the full moon. Another observer, Mr. Adie, says, "The corona was brightest near the sun, and extended about onethird of the moon's diameter, of a soft, silvery white, with brighter coruscations shooting through it beyond the general light, which gave it a flickering appearance. In one place these coruscations extended to about two-thirds of the moon's diameter." This eclipse was also observed under very favorable circumstances by Mr. George P. Bond, of the Cambridge observatory, Massachusetts, who visited Sweden for the express purpose of witnessing the phenomena. Mr. Bond selected Lilla Edet as his place of observation, "to be as near as possible to the central line." His description of the corona is as follows: "It proceeded, without an intervening space, directly from the moon's edge, where it was by far the brightest, its radiations extending to a distance of perhaps the sun's semidiameter; but it is impossible to define precisely its limits, as it had no definite outline externally; internally it was cut off by the moon's circumference.

"Its light was pure white. I saw no external rings, but did

not look expressly for them, nor whether it was uniform in all its parts. The testimony of intelligent persons in the vicinity is strong to the effect that there was a 'bush' of light, half way between the upper and the left hand limb. Its light I did not consider uniform throughout, but rather of a radiating character.'

The corona was also observed during the total eclipse of July 18th, 1860.

The descriptions which we have quoted render it extremely probable, if not absolutely certain, that the corona is nothing else than the sun's exterior atmospheric envelope, of which we have already spoken, and of whose existence we have independent evidence. Now, since this atmosphere is visible under favorable circumstances at an elevation from the surface equal to the solar radius, its volume is at least seven or eight times greater than that of the sun himself. Consequently, with a mean density of 0.0002, or one-sixth of the density of the atmosphere at the earth's surface, its mass would be equal to that of all the known planets in our system.

The results obtained by different astronomers in regard to the relative intensity of the sun's light at the center and near the borders of his disk, are somewhat discordant. It was asserted by Lucas Valerius, as long ago as the beginning of the seventeenth century, that the circumferential rays were less powerful than those from the center. This opinion was supported by the photometrical experiments of Bouguer, in the eighteenth century, and is now maintained by Sir John Herschel and Mr. Airy. Mr. Arago, on the other hand, claimed to have shown by means of his chromatic polariscope that the intensity is uniform from the center to the margin. All observers agree, however, that the intensity does not increase from the center outwards. What conclusion may we draw from this fact in regard to the constitution of the solar atmosphere?

On account of the sun's globular form, a given visual angle comprises a greater extent of surface near the edge than at the center of the disk. Now, in the case of a solid body-a ball

• Gould's Astron. Journ., No. 31.

of metal, for instance-heated to incandescence, while the area embraced within a given angle of vision increases with the obliquity, the intensity of light for any point of the radiating surface decreases in a corresponding ratio. All portions, therefore, of the visible hemisphere are equally luminousBut in the case of a gaseous sphere, the obliquity does not produce the same diminution of the illuminating power; and as the solar light is known to originate in a gaseous envelope, either its intensity must be greater near the circumference than at the center of the disk, or the radiation must be modified by some extraneous cause. As the former supposition has been disproved by observation, we are led to the conclusion that the radiation is sensibly obstructed by the imperfectly transparent, exterior atmosphere. The marginal rays, having a greater distance to pass through this non-luminous envelope, undergo a greater diminution of intensity than those perpendicular to the surface.

Another interesting phenomenon very generally observed during recent total eclipses of the sun, is that of flame-like protuberances, apparently projecting from the moon's disk. These "solar clouds," as seen by different astronomers, have appeared of various hues, and in some cases a change of color has been noticed by the same observer. They are most frequently described, however, as red, or rose-colored. They are of various altitudes; the greatest elevation yet observed somewhat exceeding one hundred thousand miles, or one-fourth of the sun's radius. They appear generally to rest upon the moon's disk; some, however, have been certainly seen entirely separated, as if suspended in the solar atmosphere. It has been observed, moreover, that those prominences towards which the moon was moving gradually diminished in hight-the moon passing over their bases-while the altitude of those on the opposite side increased. These phenomena, therefore, evidently belong to the sun, and not to the moon. Of their true nature, however, we are entirely ignorant. According to the hypothesis most favorably received, such luminous masses are immense solar clouds, of extreme tenuity, floating in the sun's outer atmospheric envelope. This theory is supported by the names of Herschel, Arago, and other eminent astronomers.

Different explanations have been proposed; but none, perhaps, are attended with fewer difficulties. Some have supposed both the corona and the red protuberances to be produced in some way by the earth's atmosphere. This opinion seems incompatible with the appearances which we have described, and may be regarded, moreover, as completely disproved by the observations of Colonel Choctzko, a Russian astronomer, who observed the eclipse of July 28th, 1851, from a station on the Caucasus, more than twelve thousand feet above the level of the sea. Instead of Instead of any sensible diminution of the extent or intensity of the halo or luminous ring, its appearance at this great elevation was extremely brilliant. Sir David Brewster admits that the flame-like protuberances, are solar phenomena, or, in other words, that they are real appendages of the sun, but rejects the hypothesis which regards them as clouds in the solar atmosphere. "The simplest. of all conceptions," he maintains, "respecting the red cusps, is, that they are outbursts of flame modified by the exhalations which may accidentally accompany them." A similar hypothesis has been advanced in this country by Professor Charles W. Hackley,* who thinks the corona is not the sun's atmosphere, but consists of elastic matter thrown out in continuous jets, and accompanied at times by portions of matter in the solid or liquid form, constituting the red or rose-colored protuberances. Another hypothesis has been advocated by Dr. Feilitzsch, a German astronomer, who carefully observed the eclipse of July 28th, 1851. The point selected for his observations was Carlscrona, in Sweden, a place included within the moon's absolute shadow. His conclusions, deduced from a discussion of his own and of other observations, are, (1,) that the corona is not the atmosphere of the sun, and, indeed, that it has no physical existence, but is simply an appearance produced by the dif fraction of the solar rays at the moon's margin; (2,) that the red, cuspated prominences are merely optical phenomena, due to diffraction by the lunar mountains outside of the line which

*In a paper read before the American Association for the Advancement of Science, at the Newport meeting, August, 1860.