PREDICTION or Sot.xn ECLIPSES. The period of IS Julian years 11 days, referred to in treating of the prediction of lunar eclipses, applies equally to solar eclipses; but the ancient,. who under stood that fact, could find no law of recurrence of solar eclipses within that period so as to pre dict them. The reason of the failure is obvious; for though solar eclipses recur in a fixed order within the cycle, they are not visible at the same places on their recurrence as when first observed. By modern methods, however, eclipses of the sun may be predicted, with all their circumstances of time and places of observation. with the most perfect certainty. At the time of a solar eclipse, the sun and moon are in con junction: they are also in or near the same node; and no eclipse eon happen if they are further than 18° from the node, or if the latitude of the moon, viewed from the earth, exceeds the sum of the apparent semi-diameters of the sun and moon. When within these limits. it is a problem of numbers and of spherical trigonome try to ascertain whether an eclipse will actually occur and what its circumstances will he.
The number of eclipses of the sun and moon together in a year cannot be less than two—in which case both are solar—or more than seven, five solar and two lunar, or four solar and three lunar; but total solar eclipses are extremely in frequent in any one place, compared with the actual frequency of their occurrence. Thus total eclipses were visible somewhere in the l'nited States during the nineteenth century only in the years ISOG, 1S31, 1560. I809, 1375. ISSO, 1SSh. 1900; and in the present century. such eclipses will be visible in the years 19IS, 1923, 1925, 1945, 1954, 1079. 1954, and 1994. For a very complete list of eclipse dates. consult: Op polzer, Canon der Pinsternissr (Vienna, ISS;) ; Newcomb. "On the Beeurrenee of Solar Eclipses. with Tables of Eclipses from mc. 700 to x.n. 2300," in isiromonical Papers (Washington, I3S•2) . See SOLAR SYSTEM; SUN.