Until their dispersion after the Jewish war in A.D. 70, and even much later, the Jews reckoned their new moons and leap years, and also the beginning of each year, not by strictly astronomical data, but by the method just described. The rule was that the month began with the day on the evening of which the new moon first became visible, and also that the passover should be kept when the sun was in Aries.[107] Maimonides, agreeing with what we have said above, informs us that a second Adar was interposed if the vernal equinox fell on the 16th Nisan or later. But it would be a great mistake to think that a scientifically accurate system, founded on these principles, was employed for calculating the new moons and leap years, such as would make it possible to bring the dates of the Jewish year into certain correspondence with the Julian Calendar. Still we must not think no attempts were made to reduce the Calendar to order on the basis of some cyclic system, but the caprice of the Sanhedrin always succeeded in rendering these attempts unavailing. Ideler (i. 512) shows how the new moons were treated, and Maimonides tells us that the Sanhedrin was influenced by many considerations in the choice of leap years. The Talmud preserves a remarkable letter written by Rabbi Gamaliel, the teacher of St Paul, to the Jews of Babylon and Media, which may appositely be quoted here. “We herewith inform you that we, in conjunction with our colleagues, have deemed it necessary to add thirty days to the year, since the doves (to be offered in sacrifice) are still too tender, and the lambs (for the passover) too young, and the time of Abib (the barley harvest) has not arrived.” This passage may well serve as a warning to those who, whenever they find a fixed date in ancient Jewish writings, forthwith, with the aid of lunar tables, transpose it into a date according to the Julian method of reckoning, and possibly flatter themselves they have found a fixed point which will form a basis for further calculations.