If the temple is pointed nearly solstitially the two dates at which the sun appears in it will be near the solstice; similarly, for a temple pointed nearly equinoctially the dates will be near the equinox; but if the ancients wished to divide the ninety-one days’ interval between the solstice and equinox, a convenient method of doing this would be to observe the sun at the half-time interval, such that the same temple would serve on both occasions. This could be done by orienting the temple to the sun’s place on the horizon when it had the declination 16° 20′ on its upward and downward journey, or, in other words, was, _in days_, half-way between the equinox and solstice. Thus, for the 45 days ( 91 days ) ( = ------- ) ( 2 ) from March 22, we have in-- March 9 April 30 May 6 --- 45 What, then, are the non-equinoctial, non-solstitial days of the year when the sun has this declination? They are, in the sun’s journey from the vernal equinox to the summer solstice and back again, May 6 and August 8 Sun’s decl. N. 16° 20′. Similarly, for the journey to the winter solstice and return we have November 8 and February 4 Sun’s decl. S. 16° 20′. We get, then, a year symmetrical with the astronomical year, which can be indicated with it as in Fig. 7; a year roughly halving the intervals between the chief dates of the astronomical year. With regard to the dates shown I have already pointed out that farming operations would not occur at the same time in different lands; that ploughing and seed time and harvest would vary with crops and latitudes; and I must now add that when we wish to determine the exact days of the month we have to struggle with all the difficulties introduced by the various systems adopted by different ancient nations to bring together the reckoning of months by the moon and of years by the sun. [Illustration: FIG. 7.--The astronomical and vegetation divisions of the year.] In more recent times there is an additional difficulty owing to the incomplete reconstruction of the calendar by Julius Cæsar, who gave us the Julian year. Thus, while the spring equinox occurred on March 21 at the time of the Council of Nice, in 325 A.D., by the year 1751 the dating of the year on which it took place had slipped back to the 10th. Hence the Act 24 George II. c. 23, by which September 2, 1752, was followed by September 14 instead of by the 3rd, thus regaining the eleven days lost. This change from the so-called “old style” to the “new style” is responsible for a great deal of confusion.