354·62 days, 376·5 days, or 33·25 years. He makes the inclination of the ring to the ecliptic about 17°. The five periods specified, he remarks, "are not all equally probable. Some of the members of the group which visited us last November [1863] gave us the means of locating approximately the central point of the region from which the paths diverge. Mr. G. A. Nolen has, by graphical processes specially devised for the purpose, found its longitude to be 142°, and its latitude 8° 30′. This longitude is very nearly that of the point in the ecliptic toward which the earth is moving. Hence the point from which the absolute motion of the bodies is directed (being in a great circle through the other two points) has the same longitude. The absolute motion of each meteor, then, is directed very nearly at right angles to a line from it to the sun, the deviation being probably not more than two or three degrees.

"Now, if in one year the group make 2 ± 1/33·25 revolutions, there is only a small portion of the orbit near the aphelion which fulfills the above condition. In like manner, if the periodic time is 33·25 years, only a small portion of the orbit near the perihelion fulfills it. On the other hand, if the annual motion is 1 ± 1/33·25 revolutions, the required condition is answered through a large part of the orbit. Inasmuch as no reason appears why the earth should meet a group near its apsides rather than elsewhere, we must regard it as more probable that the group makes in one year either 1 + 1/33·25, or 1 - 1/33·25 revolutions."

Professor Newton concludes that the third of the above-mentioned periods, viz., 354·62 days, combines the greatest amount of probability of being the true one. We grant the force of the reasons assigned for its adoption. At least one consideration, however, in favor of the long period of 33·25 years is by no means destitute of weight: of nearly 100 known bodies which revolve about the sun in orbits of small eccentricity, not one has a retrograde motion. Now if this striking fact has resulted from a general cause, how shall we account for the backward motion of a meteoric ring, in an orbit almost circular, and but little inclined to the plane of the ecliptic? In such a case, is not the preponderance of probability in favor of the longer period?

A revolution in 33·25 years corresponds to an ellipse whose major axis is 20·6. Consequently the aphelion distance would be somewhat greater than the mean distance of Uranus. It may also be worthy of note, that five periods of the ring would be very nearly equal to two of Uranus.

The Monthly Notices of the Royal Astronomical Society for December, 1866, and January, 1867, contain numerous articles on the star shower of November 13th–14th, 1866. Sir John Herschel carefully observed the phenomena, and his conclusions in regard to the orbit are confirmatory of those of Professor Newton. "We are constrained to conclude," he remarks, "that the true line of direction, in space of each meteor's flight, lay in a plane at right angles to the earth's radius vector at the moment; and that therefore, except in the improbable assumption that the meteor was at that moment in perihelio or in aphelio, its orbit would not deviate greatly from the circular form." The question is one to be decided by observation, and the only meteor whose track and time of flight seem to have been well observed, is that described by Professor Newton in Silliman's Journal for January, 1867, p. 86. The velocity in this case, if the estimated time of flight was nearly correct, was inconsistent with the theory of a circular orbit.

It is also worthy of notice that Dr. Oppolzer's elements of the first comet of 1866 resemble, in a remarkable manner, those of the meteoric ring, supposing the latter to have a period of about 33¼ years. Schiaparelli's elements of the November ring, and Oppolzer's elements of the comet of 1866, are as follows:

It seems very improbable that these coincidences should be accidental. Leverrier and other astronomers have found elements of the meteoric orbit agreeing closely with those given by Schiaparelli. Should the identity of the orbits be fully confirmed, it will follow that the comet of 1866 is a very large meteor of the November stream.

The researches of Professor C. Bruhns, of Leipzig, in regard to this group of meteors afford a probable explanation of the division of Biela's comet—a phenomenon which has greatly perplexed astronomers for the last twenty years. Adopting the period of 33¼ years, Professor Bruhns finds that the comet passed extremely near, and probably through the meteoric ring near the last of December, 1845. It is easy to perceive that such a collision might produce the separation soon afterward observed.

As the comet of Biela makes three revolutions in twenty years, it was again at this intersection, or approximate intersection of orbits about the end of 1865. But although the comet's position, with respect to the earth, was the same as in 1845–6, and although astronomers watched eagerly for its appearance, their search was unsuccessful. In short, the comet is lost. The denser portion of the meteoric stream was then approaching its perihelion. A portion of the arc had even passed that point, as a meteoric shower was observed at Greenwich on the 13th of November, 1865.7 The motion of the meteoric stream is retrograde; that of the comet, direct. Did the latter plunge into the former, and was its non appearance the result of such collision and entanglement?

CHAPTER II.
OTHER METEORIC RINGS.

II. The Meteors of August 6th–11th.

Muschenbroek, in his Introduction to Natural Philosophy, published in 1762, called attention to the fact that shooting-stars are more abundant in August than in any other part of the year. The annual periodicity of the maximum on the