above), Balasî confirms that the year should be intercalated, but suggests that
they wait and see when to add the intercalary month.^112 Similarly, it is possible
that under Darius I a cycle was used that determined which years to interca-
late, but without specifying which months to be intercalary. This‘loose
19-year cycle’would have enabled the king to retain some measure of control
over the intercalation, as he could decide, on each intercalated year, which
month to designate as intercalary. In this way, the institution of a cycle would
not have compromised entirely the authority that the king had exercised over
the calendar in previous centuries. It is quite plausible that at this stage in
history aflexible, loose cycle of intercalations was therefore preferred.
From Table 2.4 it is evident that the sequence of intercalations in this period
does not fully conform, at any point, to the rigid 19-year cycle as later known
from the Saros Canon texts. The Saros Canon cycle could have run briefly
from 525/3 to 514/13, but the next intercalation in 511/0 should have been
XII 2 , whereas it was VI 2. It could have run from 514/13, but in 503/2 it would
have failed again for the same reason. It could have run from 498/7—when
many scholars deem the 19-year cycle to have been instituted—but it failed in
479/8 when there should have been an intercalation of XII 2 (or possibly earlier
than 479/8, as the evidence for the 480s is not complete). It could have run
continuously from 476/5 (or somewhat later—the evidence for the 470s is not
satisfactory) until the end of our period (and indeed beyond), although it
failed at two points: in 446/5 and in 427/6 the intercalary month should have
been VI 2. If we ignore these last two deviations—commonly attributed to a
policy, under Artaxerxes I, of intercalating only XII 2 —it may be possible to
conclude that a rigid 19-year cycle was not instituted before the 470s.
Before the 470s, it may be argued that the eight-year cycle was used
continuously between 525/4 and 503/2, and then resumed between 495/4
and 482/1 (see Table 2.4).^113 But the possibility of a loose 19-year cycle also
deserves serious consideration. Table 2.4 demonstrates that a cycle of this kind
could have run without interruption from 525/4 to the mid 480s. This pattern
was disrupted by the next two intercalations (in 482/1 and 478/7),^114 where-
upon a new, more rigid 19-year sequence would have begun, with intercala-
tions of almost only XII 2 (except for one VI 2 , in 465/5).
The notion that a rigid 19-year cycle was‘instituted’in the 470s needs,
however, to be further refined. As I have argued above (in the context of the
eight-year cycle), the institution of a cycle does not become evident until its
(^112) Above, near n. 82.
(^113) It has often been argued that the eight-year cycle was used three times in succession
between 525/4 and 503/2, followed by the institution of the 19-year cycle not long after (see
above, n. 108); however, the data in Table 2.4 do not exactly support this view.
(^114) The intercalation of 478/7 cannot be reconciled with any eight-year or 19-year cycle. The
disruption of intercalation in this period has no clear explanation; it is unlikely to be related to
the short-lived Babylonian revolt of 481BCE(on which see Briant 1996: 541).
The Babylonian Calendar 107