itself an Egyptian import. As the Canopus decree of 238BCEmakes evident, it
was known for quite some centuries in Egypt that the intercalation of one day
every four years would prevent the civil calendar from drifting (see Chapter 3).
An Egyptian astronomer like Sosigenes is thus perhaps the most likely person
to have suggested it. The Julian calendar was thus essentially an improved
version of the Egyptian calendar, with some idiosyncratic features (uneven
month-lengths) inherited from the Republican calendar; its main improve-
ment on the Egyptian calendar (the leap year) was a calendrical notion that
originated from Egypt itself.^159
Thefirst decades of the Julian calendarLater sources suggest that in the decades following the institution of the Julian
calendar, the pontiffs remained in charge of declaring the leap years.^160 This
survival from the Republican calendar was not a mere formality. In these early
years, when the Julian calendar was not yet fully known and established, it
could not have been left to run itself on its own: political authority was needed
to ensure that leap years, in particular, were duly made. Thus although the
Julian calendar was meant, from its inception, to be permanentlyfixed and
free from political control, in practice this did not quite happen immediately.
Unfortunately, pontifical control of the Julian calendar had an adverse effect
on its proper running: the same sources inform us that in error, the pontiffs
declared the leap year every three years (instead of four). How exactly leap
years were reckoned in these decades is not clearly known, but a model has
been recently proposed by Bennett (2003), (2004a) that plausibly accounts for
the literary, epigraphic, and papyrological evidence that is currently extant.
According to this model, thefirst leap year was set in 44BCE, followed by leap
years at three-year intervals (41, 38, etc.) until 8BCE(inclusive).^161 The error
was then recognized, and to correct, as it were, the disruption caused by the
(^159) Bickerman (1968) 47 may thus have good reason to comment that Caesar did not merely
reform the Roman calendar, but rather replaced it with an entirely new calendar. 160
This is implicit in Pliny,NaturalHistory18. 211, Suetonius,Augustus31. 2, and Solinus 1.
45 – 7, and very much explicit in Macrobius (Saturnalia1. 14), who actually suggests that the
priests were‘in charge of months and days’:(Caesar) statuitutquartoquoque anno sacerdotes
quic 161 urabant mensibus et diebusunum intercalarent diem.
See Ch. 5, Table 5.4. This differs from the common, but unsubstantiated, view that thefirst
leap year was 45BCE, followed by leap years at three-year intervals until 9BCE(e.g. Samuel 1972:
157). According to Dio 55. 6. 6 it was in 8BCEthat the month of Sextilis was renamed Augustus,
whilst Suetonius (Augustus31. 2) associates the renaming of Sextilis with Augustus’reform of
the leap years; the combination of these sources suggests that the reform was enacted during
8 BCE(Samuel 1972: 155, Hannah 2005: 119), which remains compatible with Bennett’s model.
On the renaming of Sextilis to Augustus see also Macrobius,Saturnalia1. 12. 35, and Rüpke
(1995) 185.
214 Calendars in Antiquity