excessive leap years, the emperor Augustus suspended all leap years until 4CE,
when the leap year was resumed.^162 It is only from this year that the Julian
calendar was correctly reckoned, never to be adjusted again until the Gregori-
an reform of the later sixteenth century.
The occurrence of this error and its perpetuation over such a long period is
perhaps forgivable, inasmuch as the Julian calendar had only recently been
instituted; but still, the rule of one intercalation in four years should have been
simple enough for errors to be prevented. The confusion, which Suetonius
(Augustus31. 2) puts it down to the priests’negligence, is usually thought to
have been due to the common Roman practice of inclusive count (for example,
in the count of days of the month): thus‘every fourth year’(quartoquoque
anno) was misinterpreted as meaning, inclusively, every three years.^163 But a
more precise explanation of the error can be offered on the basis of a report of
Dio (48. 33. 4) that an additional day was intercalated in 41BCEto prevent the
coincidence of the subsequent 1 Januarius (40BCE) with thenundinae(the
market day, which recurs every eight days), which would have been inauspi-
cious; and that to compensate for this, a day was suppressed at a later stage.
We may assume that this later one-day suppression was made in 40BCE.
Although it is unknown whether the added day in 41BCEwas a bissextile
day in Februarius, and whether the suppressed day in 40BCEwas the bissextile
day that should have been due—according to Bennett’s model—in Februarius
that year, Dio clearly implies that the 366-day year occurred in 41BCEinstead
of 40BCE. This may have misled the priests to believe that the leap-year interval
was three years, as it had been from 44 to 41BCE. The subsequent leap year was
consequently declared in 38BCE, instead of 36BCE; and the practice of three-
year intervals became established thereafter.^164
(^162) This differs again from the common view that thefirst leap year after Augustus’calendar
reform was 8CE. The literary sources (above, n. 160) on the error of the priests, the intercalation
at three-year intervals, and the reform of Augustus, do not provide any dates; Augustus’reform is
only referred to as a twelve-year period without intercalation, which Bennett interprets as a
twelve-year interval (inclusive count) from 8BCEto 4CE. The rationale behind Augustus’calendar
reform remains to be fully understood. It is unclear why a suspension of leap years was needed,
when in fact the Julian calendar could have been left in the same relationship to the solar year as
it had reached in 8BCE; all Augustus really needed to do was to change the leap-year interval from
three years to four (for some not very convincing explanations, see Bennett 2003: 230–2; 2004a:
166 – 7). Had the intention been to restore the Julian calendar to the same relationship to the solar
year as it had been at the time of its original institution (in 46BCE), then three leap years at correct
four-yearly intervals should have been omitted (because three excessive leap years had been
reckoned between the institution of the Julian calendar and 8BCE), whereas according to
Bennett’s model, only two such leap years were omitted by Augustus (in 4BCEand in 1CE), or
alternatively, three leap years at three-year intervals (in 5BCE,2BCE, and 2CE). See further Ch. 5
n. 90.
(^163) Brind’Amour (1983) 12–15; Bennett (2003) 232; but for a different explanation, Rüpke
(1995) 381–3.
(^164) Bennett loc. cit. Rüpke (1995) 582–7 argues the same, with the additional suggestion that
the use of the leap year in 41BCEas a means of avoiding a clash between the New Year and the
TheRise of the Fixed Calendars 215