rationale, as I shall now explain. The Republican 31-day months were left
unchanged, as there was no need to increase them. The 28-day month of
Februarius, formerly designated for intercalations and now for leap years, was
also left unchanged (except in leap years), perhaps to indicate its special status.
Changes were only made to the remaining 29-day months. Most of these were
increased by just one day; but since there were only seven 29-day months in
the Republican calendar, whereas a total of ten days had to be added to convert
the Republican 355-day year into the Julian 365-day year, three of these
months had to be increased by two days. Not unreasonably, these two-day
increases were spread evenly through the months of the year, i.e. at the
beginning, near the middle,^157 and at the end of the year (see Table 4.1).
Although the sequence of month-lengths in the Julian calendar was primar-
ily derived, in this way, from the Republican calendar, it may also have been
viewed as a structural improvement over the Egyptian calendar: for the use of
31-day months—a survival from the Republican calendar—eliminated the
need for the anomaly, in the Egyptian calendar, of thefive epagomenal days.
But a much clearer improvement on the Egyptian calendar was the leap year,
with the insertion every four years of an extra day (called‘bissextile’) after 24
Februarius.^158 The leap year maintained the Julian calendar in alignment with
the solar year and the seasons, thus making it a truly solar calendar—thefirst
of its kind in the ancient world. But this improvement is likely to have been
Table 4.1.Republican and Julian month lengths
Republican month Days added Julian monthJanuarius 29 2 31
Februarius 28 0/1 28/9
Martius 31 0 31
Aprilis 29 1 30
Maius 31 0 31
Junius 29 1 30
Quintilis/Julius 31 0 31
Sextilis 29 2 31
September 29 1 30
October 31 0 31
November 29 1 30
December 29 2 31(^157) The month of Sextilis (increased by two days) was not exactly in the middle of the year,
but perhaps we should not expect too much precision. It is also possible that Junius, more in the
middle of the year, was not given the two-day increase so as to avoid a run of three 31-day
months (from Maius to Julius).
(^158) In the Julian calendar (as in its Republican antecedent) 24 Februarius wasante diem VI
kalendas Martias(the sixth day before the Kalends of March, inclusive count); the intercalated
day was accordinglya.d. bis VIkal. Mart., or simplybissextus(i.e. the‘second sixth’day): see
Ulpian inDigest4. 4. 3. 3, Celsus inDigest50. 16. 98, Augustine,On the Trinity,4.8.
TheRise of the Fixed Calendars 213