but still before Augustus’reform, would have been 8BCE; the decree is thus to
be dated to this year.^116
Confirmation of this dating (and hence, of the model assumed in Table 5.4)
can be drawn from the fact—generally ignored by earlier scholars—that only
8 BCEfits the dates of the old calendar, which as a Macedonian calendar, was
presumably lunar. According to the decree, indeed, 14 Peritios of the old
calendar occurred in that year on 23 January; thus the (old) month of Peritios
began on 10 January. Only in 8BCEwas the Roman date of 10 January close to
the new moon: the conjunction occurred on that day (in the evening), and the
new moon crescent would have been visible two days later, in the evening.^117
The sequence of months laid down in thekoinon’s decree differs from what
the proconsul had originally proposed. The fragmentary passage from the
Latin section (or version) of the proconsul’s edict indicates that the last four
months of the year, according to him, should have followed the sequence of
30 – 31 – 30 – 31 (days);^118 whereas according to thekoinon’s decree, the se-
quence of the last four months was 30– 31 – 31 – 30 (see above). There is no
need to assume a scribal (or other) error.^119 At most, the scheme of the
proconsul was not well thought out; and this would justify thekoinon’s
decision to modify it. For the month-lengths proposed by thekoinon—unlike
those of the proconsul—had the distinctive advantage of corresponding exact-
ly to those of the Julian calendar (from October to September), with the result
that all Asian months began on the same day of the Roman, Julian month (the
9th before the kalends: see above).
(^116) Bennett (2003) 226–7. It is unlikely that the decree dates from 5BCE, which thekoinon
would have taken to be a leap year because they were unaware of Augustus’reform; see Bennett
in http://www.tyndalehouse.com/egypt/ptolemies/chron/roman/008bc_fr.htm, and Hannah
(2005) 131–5. Mommsen (cited in Sherk 336), followed by other scholars (Magie 1950: 480–1,
Laffi1967: 27–34, Jones 2000b: 159 n. 3), preferred the date of 9BCE, but only on the assumption
that 9BCEwas the Roman leap year, which according to Table 5.4 we should no longer accept;
this date, moreover, is completely incompatible with the lunar date of the older calendar (see
next n.). Samuel (1972) 182 n. 1 is unnecessarily cautious.
(^117) This follows the assumption that the Roman date of 10 January, in 8BCE, was equivalent to
11 January of the theoretical, retrojected Julian calendar: see above, Table 5.4 and Bennett
(2004a) 167. (In 5BCE, the Roman date of 10 January (then equivalent to Julian 12 January)
was also fairly close to the new moon, as the new crescent would have beenfirst visible two
evenings earlier.) The year 9BCEcan befirmly ruled out, as it would entail a discrepancy from the
moon of well over ten days. The lack of an exact correspondence, in 8BCE, withfirst visibility of
the new moon should not be regarded as problematic: it only demonstrates that post-Seleucid
calendars were not as astronomically exact as the Babylonian calendar. But it is interesting to
note the assumption, implicit in thekoinon’s decree, that all the cities of the province of Asia
reckoned the lunar month in the same way (for a similar assumption in thekoinonof Macedonia,
see Ch. 6 n. 8).
(^118) Laffi(1967) 19 and 86 (Doc. II). The text translates as:‘9th 30, 10th 31, 11th 30, 12th 31; an
intercalary (day) will be added...’. It is unclear to which month the intercalary was to be added.
(^119) As Mommsen (followed by Laffi1967: 19, 48) unnecessarily conjectured with regard to the
text of the proconsul’s edict (see Sherk 1969: 333).
276 Calendars in Antiquity