Ours is a _tropical_ year, that is to say, according to the Julian calendar (afterwards amended by Pope Gregory) it is bound to the _seasons_, and its months maintain a constant relation to the four great divisions of the ecliptic, _i.e._ the solstices and the equinoxes. The winter solstice always falls about the 22nd of December, the spring equinox about the 21st of March, the summer solstice about the 21st of June, and the autumnal equinox about the 23rd of September. But (as has been suggested) the Accadian year was a _sidereal_ year, and its months maintained a constant relation to the twelve _star-marked_ divisions of the ecliptic, or, as they are called, the constellations of the Zodiac. Nisan always corresponded (as closely as a lunar month might) to the time during which the sun traversed the constellation Aries; Airu to the time during which it traversed the constellation Taurus; and so on through the twelve months of the year. The equinoctial points are, however, always, though slowly, changing their position amongst the twelve constellations of the ecliptic. The months, therefore, which in 3,800 B.C., and still in the second century B.C., corresponded to the same star-groups, as above noted, must have held in different ages very different positions in regard to the four great divisions or _seasons_ of the year. We find in the tablets translated by Epping and Strassmaier the year “_beginning with Nisan, hence in the spring_,” and this seems a more or less natural season from which to count the year; but when, taking the precession of the equinoxes into account, we find that the year in Ḫammurabi’s time (2,200 B.C.) must have commenced one month, and in Sargon’s time (3,800 B.C.) two months before the spring equinox, we feel surprised and perplexed to find that the year must then have begun without any reference to the seasons--the four great and most easily observed divisions of the ecliptic. It is difficult to imagine that the astronomers who so skilfully divided the ecliptic into its twelve parts, and who originated the wonderful Accadian calendar--a calendar so well thought out that, as we have seen reason to believe, it resisted all the shocks of time for nearly four thousand years--it is difficult to imagine that such astronomers should have taken no note of the four prominent divisions of the year and of the ecliptic, _i.e._ the solstices and the equinoxes. [Illustration: PLATE I. FIG. 1. FIG. 2. The first and last months of the Accadian, sidereal, year, compared with the months of the Gregorian, tropical, year: at 6,000 B.C. and at 600 A.D. [_To face p. 13._] There is, however, a way to account for this anomaly, or, rather, there is a supposition which, if adopted, will allow these astronomers of old to have taken note, not only of the _months_, but also of the _seasons_ of the year, when first they drew up their mighty scheme.