and her rising in the east, and, by using the days of the month to ascertain the relative positions of the moon, he found that throughout the whole course of his period this particular combination took place three times.[52] He then proceeded to examine in the same way the rest of the observations, with their dates, as supplied by the two tablets, and, by working them out in detail for the central one of his three possible periods, he obtained confirmation of his view that the observations did cover a consecutive period of twenty-one years. In order to obtain independent proof of the correctness of his figures, he proceeded to examine the dates upon contemporary legal documents, which could be brought into direct or indirect relation to the time of harvest. These dates, according to his interpretation of the calendar, offered a means of controlling his results, since he was able to show that a higher or lower estimate tended to throw out the time of harvest from the month of Nisan, which was peculiarly the harvest month. It must be admitted that the last part of the demonstration stands in a different category to the first; it does not share the simplicity of the astronomical problem. It formed, indeed, merely an additional method of testing the interpretation of the astronomical evidence, and the dates resulting from the latter were obtained in complete independence of the farming-out contracts of the period. Taking, then, the three alternative dates, there can be no doubt, if we accept the figure of the Kings' List for the Second Dynasty as approximately accurate, that the central of the three periods is the only one possible for Ammi-zaduga's reign; for either of the other two would imply too high or too low a date for the Third Dynasty of the Kings' List. We may thus accept the date of 1977 B.C. as that of Ammi-zaduga's accession, and we thereby obtain a fixed point for working out the chronology of the First Dynasty of Babylon, and, consequently, of the partly contemporaneous Dynasties of Larsa and of Nîsin, and of the still earlier Dynasty of Ur. Incidentally it assists in fixing within comparatively narrow limits the period of the Kassite conquest and of the following dynasties of Babylon.[53] Starting from this figure as a basis, and making use of the information already discussed, it would follow that the Dynasty of Nîsin was founded in the year 2339 B.C., that of Larsa only four years later in 2335 B.C., and the First Dynasty of Babylon after a further interval of a hundred and ten years in 2225 B.C.[54]