iii., s. 237.) After midnight, on the 12th of November, 1799, occurred the extraordinary fall of stars at Cumana, which Bonpland and myself have described, and which was observed over a great part of the earth. ('Relat. Hist.', t. i., p. 519-527.) Between the 12th and 13th of November, 1822, shooting stars, intermingled with fire-balls, were seen in large numbers by Kloden, at Potsdam. (Gilbert's 'Ann.', bd. lxxii., s. 291.) On the 13th of November, 1831, at 4 o'clock in the morning, a great shower of falling stars was seen by Captain Bérard, on the Spanish coast, near Carthagena del Levante. ('Annuaire', 1836, p. 297.) In the night between the 12th and 13th of November, 1833, occurred the phenomenon so admirably described by Professor Olmsted, in North America. In the night of the 13-14th of November, 1834, a similar fall of shooting stars was seen in North America, although the numbers were not quite so considerable. (Poggend., 'Annalen', bd. xxxiv., s. 129.) On the 13th of November, 1835, a barn was set on fire by the fall of a sporadic fire-ball, at Belley, in the Department de l'Ain. ('Annuaire', 1836, p. 296.) In the year 1838, the stream showed itself most decidedly on the night of the 13-14th of November. ('Astron. Nachr.', 1838, No. 372.) On the recurrence of falls of shooting stars in North America, in the month of November of the years 1834 and 1837, and in the analogous falls observed at Bremen in 1838, a like general parallelism of the orbits, and the same direction of the meteors from the constellation Leo, were again noticed. It has been supposed that a greater parallelism was observable in the direction of periodic falls of shooting stars than in those of sporadic occurrence; and it has further been remarked, that in the periodically-recurring falls in the month of August, as, for instance, in the year 1839, the meteors came principally from one point between Perseus and Taurus, toward the latter of which constellations in the Earth was then moving. This peculiarity of the phenomenon, manifested in the retrograde direction of the orbits in November and August, should be thoroughly investigated by accurate observations, in order that it may either be fully confirmed or refuted. The heights of shooting stars, that is to say, the heights of the points at which they begin and cease to be visible, vary exceedingly, fluctuating between 16 and 140 miles. This important result, and the enormous velocity of these problematical asteroids, were first ascertained by Benzenberg and Brandes, by simultaneous observations and determinations of parallax at the extremities of a base line of 49,020 feet in length.*