until the end of November it appears nearly stationary, ascending from the horizon to the constellation Leo, in some part of which it terminates, the vertex varying somewhat in altitude with the condition of the sky. After the 13th of November the light fades in the morning sky, contracts in dimensions, and soon becomes stationary and then retrograde with respect to the sun, proceeding eastward no farther than y Virginis, a point which it reaches by the 26th of November, having at this time an elongation of only 60°, whereas a fortnight before the elongation was 90°. As the sun advances in the ecliptic, while the light appears nearly stationary, the elongation on this side continues to diminish, as well as the dimensions and the illumination, until early in January, after which it is scarcely seen in the east until August.

The foregoing general statements are supported by observations taken at different times through the period of six years before mentioned, a few of which I extract from my records:

"Nov. 26, 1837.-This morning about daybreak saw the zodiacal light; very bright and distinct, but elongation only 60°.

"Nov. 28.-Commenced observations at 5 o'clock; zodiacal light brighter than usual in preceding years at this season, but the vertex appears nearly stationary in Gamma Virginis.

"Dec. 5.-Zodiacal light visible this morning as early as 3 o'clock; not quite so bright as on the 28th of November, but increased in brightness from 3 o'clock till daybreak; vertex still in y Virginis.

"Dec. 9.-Examined the eastern sky from 4h 30m till daybreak; very cold and clear; zodiacal light much less bright than on the 5th; width also less; when I first went out could scarcely see it; became distinct by 5 o'clock, half an hour before daybreak, yet much feebler than it was ten days ago; contracted between Spica and Theta Virginis, 4° north of Spica, whereas a few days since the border grazed this star. “Jan. 18, 1837.-Zodiacal light very diffusive and ill-defined; seen after this no more in the east."

We will now introduce the observer to the western sky. Here the zodiacal light first comes into view, so as to be distinctly defined, about the 21st of November, at which time it

lies far in the south-west, crosses the milky way, the head of Capricornus, and has its vertex near the right shoulder of Aquarius, with an elongation from the sun of full 90°. From this time it climbs rapidly upwards, until by December 2, it reaches nearly to Algenib in the equinoctial colure, having an elongation of more than 100°. By about Christmas the vertex reaches almost to Alpha Arietis, having an elongation of towards 120°. It becomes nearly stationary through the month of January, but in February and March it moves slowly onward through Taurus to Gemini, beyond which it scarcely advances. The accompanying diagram (Pl. II.) is intended to represent the general appearance of the zodiacal light, when seen under favourable circumstances. Near the time of the vernal equinox, it is seen of a pyramidal form, with a broad base resting on the horizon. Its northern border grazes the bright star Algenib in Pegasus, passes south of Alpha Arietis seven or eight degrees, and about two degrees south of the Pleiades. Along its southern boundary we recognize the stars in the mouth and neck of the Whale, and still higher, Aldebaran, the Hyades, and the horns of the Bull. The successive positions attributed to the zodiacal light, from the time of its earliest appearance in the western sky, the 21st of November, to the vernal equinox, are not absolutely uniform, but they still correspond to observations made during the six years before-mentioned, as will appear from a few extracts from my record book. I have already recited the observation of November 21, 1837, when the return of the body to the western sky was first recognized.

"Nov. 26, 1837.-Light feeble, Venus being very bright; but seen after Venus was set, reaching nearly to the Fish, south of Pegasus. Elongation 100°.

"Dec. 2.-New moon begins to interfere with observations, but this evening the zodiacal light was visible after the moon was set; covers the Pentagon in Pisces, and reaches beyond it. Elongation 110°.

"Dec. 18.-Early part of the day a violent rain and high wind; cleared off towards night; zodiacal light very bright, reaching at least to Alpha Arietis; nearly as bright as the milky way. Elongation 120°."

It ought to be remarked that the phenomena of this body were particularly striking in the autumn and winter of 1837, and the observations made this year shew a greater intensity of light and a greater elongation from the sun, than those of corresponding dates in 1835 and 1836.

"Dec. 21, 1835.-This evening atmosphere very transparent; zodiacal light very conspicuous, reaching nearly to Algenib though quite faint towards the vertex. Elongation 90°. "Dec. 28, 1837.-Night favourable; appeared to me not to reach quite so far eastward as it did a few nights since ; certainly not beyond the equinoctial colure; could not be certain much further than the pentagon of stars in Pisces. Elongation 75°.

"Feb. 7.-Zodiacal light very conspicuous since the last moon, but has advanced eastward very little since Christmas, still reaching only to Alpha Arietis. Elongation 75°.

"Feb. 24.-First night since the moon has been away; sky favourable for observation; zodiacal light bright and well defined, its axis nearly in the ecliptic; reaches to the space between Aldebaran and the Pleiades. Elongation 85°.

"March 26.-Zodiacal light very bright, reaching above the Pleiades, which are a little north of the axis. Elongation 60°. "March 29.-Light more faint. Elongation 60°. Vertex near the ecliptic.

66

April 6.-Light fading rapidly. Very diffuse.

May 1.-Last night a very plentiful rain after a series of warm days; to-day air keen and sky very clear; this evening zodiacal light remarkably distinct (for this season of the year), being discernible much nearer the horizon than common, and reaching further eastward among the stars than I ever observed it before, namely, into the neighbourhood of Castor and Pollux. Elongation 60°, but presumed to be much greater than it would be, but for the extraordinary transparency of the atmosphere.

66

May 10, 1834.-Zodiacal light seen for ten minutes after twilight ceased-say till ten minutes after nine. Reached to Castor, but very diffuse. Elongation 57°.

"Seen no more in the west till the latter part of November."

To present at one view the various elongations from the sun, observed from Nov. 21 to May 10, the result is as follows:

From this tabular view, it appears that when the body first came into view, on the 21st of November, it extended about 90° eastward of the sun; that its elongation increased rapidly from this period, being five days afterwards 100°, in six days more 110°, and in sixteen days after this 120°, which is the greatest elongation I have ever noticed; and being at the same time about 60° westward of the sun, its whole extent in longitude was 180°.

I have, in a few instances, remarked what was apparently a sudden and remarkable expansion of the zodiacal light, a circumstance more than once noted by Cassini. My record for November 21, 1838, is as follows:

"At 5 A.M., about 20 minutes before twilight, the zodiacal light was very large, extending in breadth from Corvus to Arcturus. Never saw it so broad before. More inclined towards the south than usual, its vertex passing one or two degrees to the south of Regulus."

Whether this extraordinary enlargement in breadth, implying a space of more than 40°, was owing to a change in the body itself, or to some unusual atmospheric refraction, or the accidental presence of an aurora borealis, it is impossible for me to decide.

It is well known that the great French astronomer, Dominique Cassini, was the first to direct the attention of astronomers towards the zodiacal light, and that he made numerous observations on it extending from 1683 to 1688 inclusive, which are published in the eighth volume of the Memoirs of the French Academy, together with observations on the same phenomenon made at Geneva by a friend of his, M. Fatio. An elaborate digest of these records is made by

Mairan in his celebrated Treatise on the Aurora Borealis, including also a few observations of his own, and of several other philosophers. It is interesting to compare these ancient observations with such as we have been able to make at corresponding times of the year; and having made this comparison in numerous instances, I feel able to say, that the zodiacal light, in the main, is the same thing that it was in the days of Cassini and Mairan, being subject to similar variations at different seasons of the year and in different states of the atmosphere. I shall avail myself of such aid as I can obtain from this and every other source in the remaining parts of this essay.

Nature and Constitution of the Zodiacal Light.

1. Length. The extreme portions of this body sometimes. extend beyond the earth's orbit. It is obvious that, at an elongation of 90°, it must reach a tangent drawn to the earth's orbit at the place of the spectator; and if it reaches beyond that tangent, as is sometimes the case, it must of course extend beyond the earth's path. According to one of our observations, on the 18th December 1837, its elongation was 120°. The variable apparent elongation to which this phenomenon is subject, is more or less influenced by three causes; the state of the atmosphere, the inclination of its line of direction to the horizon, and the length of the twilight. In order to eliminate the effect due to atmospheric changes we require numerous series of observations, continued through successive years, and, if possible, instituted at long intervals of time. The mean of such an assemblage of observations would exhibit results nearly free from the effects of accidental variations in the transparency of the atmosphere. Since the axis of the zodiacal light does not deviate far from the ecliptic, we may imagine it to be represented by a portion of that circle on the artificial globe, and we shall easily see that since its inclination to the horizon varies between 25° and 72°, being 25° at the vernal equinox (25° with the eastern, and 72° with the western horizon), this cause must greatly affect the degree of intensity of the zodiacal light. The same must obviously be the case