of brilliant displays at the end of July. Then there occurred a break until eighty-three years later, when it several times reappeared with similar splendor. A wide interval of more than three hundred years brings us to the year 1243, when it seems to have been again recognized, after which, until 1709, there is only one other observation of the shower (in 1451). During the last hundred years it has, however, been frequently observed, though many of the recent displays can not be compared with those of ancient times. The intermittent and rare character of the shower, as it existed between the tenth and eighteenth centuries, proves that few returns were of a sufficiently imposing nature to be recorded, and that possibly the conditions were opposed to its appearance. If the meteors of the orbit during that period were condensed in the region of their derivative comet, then we can understand the singular paucity of observations. The earth, as it passed the node, would year after year encounter no meteors until the perihelion approach of the cluster, when possibly the display may have occurred in the daytime, and been of such brief duration as entirely to elude detection.
The entry of this stream into the solar system probably dates back to a very remote antiquity—for there are several circumstances which conspire to prove that such must have been the case, and that it preceded, by many ages, the apparition of the Leonids, Andromedes, and some of the other periodical meteor-showers. The fact that it constitutes an unbroken ring leads to the inference that it must have existed from the earliest times in order to bring about so complete a dispersion of its particles, for on its first introduction, as a comet, to the earth, it is to be assumed that it formed a condensed mass like the Leonids, and only appeared as a meteor-shower when the comet returned to perihelion. A very slight difference in the periodic times of the individual meteors following the nucleus must have eventually distributed them (by its cumulative effects) along the entire orbit. In other words, the original group must have undergone a process of lengthening out, until, at the present day, it consists of a parabolic zone of meteoric pellets, through which the earth passes annually on August 10th. Moreover, the radiant point of the shower often fails to become sharply defined. Several concentric streams of similar meteors appear to diverge from the region about η Persei, and their physical identity is unquestionable. They are merely the deflections or offshoots from the original system which must be greatly disturbed and contorted as the earth annually intersects it. The full effects of these perturbations can hardly be estimated: many of the particles must be diverted into new orbits, and one of the results upon the main stream may be a constant widening out, so that the apparent duration of the shower must go on increasing. It now actively extends over at least eight nights; hence the width must exceed 10,000,000 miles. And some diminution in its intensity must occur at each return, unless there is a
