After Artaxerxes’reign, a small change was made to the cycle with the
intercalation of VI 2 (instead of XII 2 ) in year 17, thefirst time probably in
408/7, towards the end of the reign of Darius II (see Table 2.5 n.a). This
effectively established the 19-year cycle as later assumed in the Saros Canon
texts. I thus propose to differ from the common view that Artaxerxes I
‘deviated’from the Saros Canon cycle by intercalating XII 2 in year 17, and
that the Saros Canon cycle was‘restored’after his reign. In actual fact, the
Saros Canon cycle was not instituted until the reign Darius II. Before Artax-
erxes I, the 19-year cycle was loose; Artaxerxes I was thefirst to make use of a
rigid 19-year cycle; and it was only when his cycle was altered, under Darius II,
that the Saros Canon cycle was instituted.
That afixed cycle took a long time, over the course of thefifth century, to
become established, finds further support from the dated documents of
Elephantine in southern Egypt. The authors of these documents were often
unaware that an intercalation had been made, and depended on information
reaching them from Babylonia which could often be considerably delayed
(Stern 2000a). It seems clear that in this period the establishment of a rigid
cycle of intercalations was not yet common knowledge, or at least the existence
of such a cycle could not yet be taken as definitive. Even in the latefifth
century, the intercalations may still have been subject to a decision-making
process and royal decree. At the very least, kings like Artaxerxes I and Darius
II were at liberty to decide how the 19-year cycle of intercalations would run
during their reigns.
Intercalation in the late Achaemenid and Seleucid periodsThe evidence for intercalations in the late Achaemenid and Seleucid periods,
as laid out in Table 2.5, is reasonably sound. The vast majority of entries in the
Table, extending from 422/1 to 178/7, are attested either in economic/admin-
istrative sources, or in observational astronomical sources. The latter are likely
to have been drawn from authentic records of astronomical observations (such
as astronomical diaries) and thus can be treated as historically reliable.^116
Some entries, however, are attested only in theoretical astronomical sources
(such as the Saros Canon texts) which may have assumed, for convenience, a
purely theoretical intercalation scheme which was not necessarily followed in
historical reality. Entries that are based only on theoretical sources tend to
months (the dynasties ofŠulgi, the Amorites, and the Kassites cover a continuous period from
the late third to the late second millenniumBCE).
(^116) This table, largely based onWalker’s subsequent research, dispels his earlier concern (Walker
1997: 24) that intercalations in the late Achaemenid and Seleucid periods might be attested mainly in
theoretical astronomical sources which could not be relied upon as historically accurate.
110 Calendars in Antiquity