for it is not annual, except possibly in the case of the August sprinkle. How the sun's unequal attraction for the part of a group acts as a dispersive force to draw it out into a stream, those most beautiful and most fruitful discussions of Signor Schiaparelli have shown. The groups that we meet are certainly in the shape of thin streams.
It has been assumed that the cometic fragments go continuously away from the parent mass so as to form in due time a ring-like stream of varying density, but stretched along the entire elliptic orbit of the comet. The epochs of the Leonid star-showers in November, which have been coming at intervals of thirty-three years since the year 902, have led us to believe that this departure of the fragments from Tempel's comet (1866, 1) and the formation of the ring were very slow processes. The meteors which we met near 1866 were therefore thought to have left the comet many thousands of years ago. The extension of the group was presumed to go on until at some time in the future, perhaps tens of thousands of years hence, the earth was to meet the stream every year.
Whatever may be the case with Tempel's comet and its meteors, this slow development is not found to be true for the fragments of Biela's comet. It is quite certain that the meteors of the splendid displays of 1872 and 1885 left the immediate vicinity of that comet later than 1840, although at the time of those showers they had become separated two hundred millions of miles from the computed place of the comet. The process, then, has been an exceedingly rapid one, requiring, if continued at the same rate, only a small part of a millennium for the completion of an entire ring, if a ring is to be the finished form of the group.
It may be thought reasonable, in view of this fact about Biela's comet, established by the star-showers of 1872 and 1885, to revise our conception of the process of disintegration of Tempel's comet also. The more brilliant star-showers from this comet have always occurred very near the end of the thirty-three-year period. Instead of there being a slow progress which is ultimately to produce a ring along the orbit of the comet, it certainly seems more reasonable to suppose that the compact lines of meteors which we met in 1866, 1867, and 1868, left the comet at a recent date. A thousand years ago this shower occurred in the midle of October. By the precession of the equinoxes and the action of the planets the shower has moved to the middle of November. One half of this motion is due to the precession of the equinoxes, the other half to the perturbing action of the planets. Did the planets act upon the comet before the meteoroids left it; or upon the meteoroid stream? Until one has reduced the forces to numerical values, he may not give to this question a positive answer. But I strongly suspect that computations of the forces will show that the perturbations of Jupiter and Saturn upon the group of meteoroids hundreds of millions of miles in length—perturbations strong enough to change the node of
