To see the law of recurrence of corresponding eclipses in the successive periods let us suppose the line of conjunction ES1 to be that at which there is a very small eclipse, visible only in high northern or southern latitudes. At the end of 18 years 11 days a second eclipse will occur along a line nearly half a degree nearer EN, the line of nodes. The successive eclipses will occur at the same interval through about ten periods, or 180 years, when the line of conjunction will pass within 11° of EN. Then the eclipse will be central, whether annular or total depending on circumstances: in the first one the central lines will pass only over the polar regions; but in successive eclipses of the series it will pass nearer and nearer to the equator until the conjunction line coincides with the node. The path of centrality will then cross in the equatorial region. During 22 or 23 more recurrences the path will continually approach to the opposite pole and finally leave the earth entirely. The entire number of central eclipses in any one series will generally be about forty-five. Then a series of continually diminishing partial eclipses will go on for about ten periods more. The whole series of eclipses will therefore extend through about sixty-five periods; and interval of time of about twelve hundred years.
Another remarkable eclipse period recurs at the end of 358 lunations. At the end of this period the line of mean conjunction ES1 falls so near its former position relative to the node that we find each central eclipse visible in our time to be one of an unbroken series extending from the earliest historic times to the present, at intervals equal to the length of the period. The recurring eclipses in this period do not, however, have the remarkable similarity of those belonging to the Saros, but may differ to any extent, owing to the different positions of the line of conjunction with respect to the moon’s perigee. Moreover, they recur alternately at the ascending and descending node. The length of the period is 10,571.95 days, or 29 Julian years less 20.3 days. Hence 18 periods make 521 years, so that at the end of this time each eclipse recurs on or about the same day of the year. As an example of this series, starting from the eclipse of Nineveh, June 15, 763 B.C., recorded on the Assyrian tablets, we find eclipses on May 27, 734 B.C., May 7, 705 B.C., and so on in an unbroken series to 1843, 1872 and 1901, the last being the 93rd of the series. Those at the ends of the 521-year intervals occurred on June 15, O.S., of each of the years 763, 242 B.C., A.D. 280, 801, 1322 and 1843. As the lunar perigee moves through 242.4° in a period, the eclipses will vary from total to annular, but at the end of 3 periods the perigee is only 7.1° in advance of its original position relative to the node. Hence in a series including every third eclipse the eclipses will be of the same character through a thousand years or more. Thus the eclipses of 1467, 1554, 1640, 1727, 1814, 1901, 1988, &c., are total.
IV. Chronological Lists of Eclipses of the Sun.
The following is a brief chronological enumeration of those total eclipses of the sun which are of interest, either from their historic celebrity or the nature of the conclusions Notable eclipses. derived from them. In numbering the years before the Christian era the astronomical nomenclature is used, in which the number of the year is one less than that used by the chronologists. The Chinese eclipses are passed over, owing to the generally doubtful character of the records pertaining to them.
—1069 June 20 and —1062 July 31; total eclipses recorded at Babylon.
—762, June 14; a total eclipse recorded at Nineveh. Computation from the modern tables shows that the path of totality passed about 100 m. or more north of Nineveh.
—647, April 6; total eclipse at or near Thasos, mentioned by Archilochus.
—584, May 28; the celebrated eclipse of Thales. For an account of this eclipse see [Thales].
—556, May 19, the eclipse of Larissa. The modern tables show that the eclipse was not total at Larissa, and the connexion of the classical record with the eclipse is doubtful.