be grouped together, in 406/5–400/399, 343/2–335/4, and 322/1–308/7 (all
inclusive), and particularly at the end of the Achaemenid and beginning of
the Seleucid periods (between 343/2 and 308/7); for these periods, more
evidence from economic or observational astronomical sources would there-
fore be desirable. After 218/17 there are a few missing entries, and after 178/7
the record becomes sporadic, without any attested intercalations in years
17 and 19 of the Saros Canon cycle.
The general impression is thus that the Saros Canon cycle was regularly
followed throughout the later Achaemenid and Seleucid periods, but still, we
cannot be certain about the last decades of Seleucid rule, before the fall of
Babylonia to the Parthians in 141BCE. There are, furthermore, two cases where
intercalations were made in the wrong years: in 385/4 (and possibly also at the
next intercalation, in 382/1) and in 266/5 (but sources here are inconsistent).
There is no simple explanation, e.g. political, for these deviations from the
Saros Canon cycle. These deviations suggest that intercalation was still not
completelyfixed in the Achaemenid, and possibly even early Seleucid, periods.
As noted above, the Saros Canon cycle was instituted under Darius II, when
the intercalation of month VI 2 was substituted in year 17 (thefirst time,
probably, in 408/7). The purpose of this substitution may well have been to
improve the lunisolar synchronism of the Babylonian calendar. In the 19-year
cycle of Artaxerxes I, the last months of year 17 occurred excessively early in
relation to the solar year. This was remedied by bringing forward the interca-
lation from Addaru (XII 2 ) to Ululu (VI 2 ). By the time of Darius II, the notion
of afixed cycle of intercalation may have been sufficiently well established in
the Babylonian calendar for the king and the astrologers to study this cycle and
take deliberate action to improve it.
But the improvement of the 19-year cycle could have been carried a few
steps further. As I have explained above, the Saros Canon cycle is by no means
perfect: a better, optimal 19-year cycle would have spaced out its intercalations
more evenly, so as to achieve minimal swaying from the solar year. The
institution of such a cycle would only have required the substitution of VI 2
in year 6 (instead of XII 2 ), and of VI 2 in year 9 (instead of XII 2 in year 8; see
above, Table 2.3). Such improvements, however, were never made, and the
Babylonian calendar seems not to have progressed beyond the Saros Canon
cycle. This was surely not due to ignorance—Babylonian astrologers were
perfectly capable of designing the optimal 19-year cycle—but probably rather
to political choice. At some stage after the 380s (when deviations from the
Saros Canon cycle could still occur), the Saros Canon cycle appears to have
become unchangeable andfixed.We cannot be certain that this was ever
established as a deliberate policy; but if, as we have seen in the case of Cyrus
and the Achaemenids, it was the advent of new dynasties that often led to
profound, long-term calendrical change, it makes sense to attribute a policy of
The Babylonian Calendar 111