If the Roman calendar was originally lunar, its conversion to the Republican
calendar scheme, early in the history of Rome, needs to be explained. Thefixed
lengths of its months—including 28 and 31 days, but strangely not 30 days—
and the intercalation of only 22 or 23 days (albeit in the form of a whole
intercalary‘month’) are structurally bizarre and difficult to understand. The
Republican calendar appears to have attempted a compromise between lunar
and solar schemes,^139 but this would have failed on both counts: the months
lost their lunar character, whilst the average year-length, assuming intercala-
tion in alternate years, exceeded somewhat the solar year.^140 The inaccuracies
of this idiosyncratic lunar and solar calendar may have arisen partly from
ignorance (Bickerman 1968: 44–5), but more likely, they were the outcome of
political and religious processes that had little concern for astronomical or
calendrical accuracy.^141
The nature of these processes, and indeed, how and when the lunar calendar
gave way to the Republican scheme, are largely unknown. Two laws (or sets of
laws) regarding the intercalation—that of the Decemviri in 450BCE, and that of
Acilius in 191BCE—suggest that formal changes were made to the Roman
calendar at these times; however, these laws are only briefly mentioned in late
sources, and nothing is known about their specific contents.^142 The date of the
lex Acilia(to begin with the latter) suggests that it was related somehow to the
disruption of the calendar around 190BCE(mentioned above); it certainly
seems to have had the effect of rectifying the calendar, at least partially, by 168
BCE.^143 Some see thelex Acilia, accordingly, as a regularization of the interca-
lary month-length;^144 others interpret it as a law investing the pontiffs with the
(^139) As suggested already by Plutarch,Numa18 and Macrobius,Saturnalia1. 13. 8–13 (the
latter’s theory, in this context, that the original Roman lunar calendar was based on an eight-year
cycle—the octaeteris—is excessively learned and historically unlikely; see Brind’Amour 1983)
28 – 140 30. See also Michels (1967) 16–22.
Since 355 + half of 22.5 exceed the solar year (c.365¼days) by one day (see Rüpke 1995:
294 – 5).
(^141) Macrobius,Saturnalia1. 13. 5, explains the avoidance of 30-day months and 354-day
years as due to a traditional preference for odd numbers (cf. Vergil,Buc. 8. 75). For a cogent,
though speculative attempt to explain some of the Republican calendar’s peculiarities, see Rüpke
(1995) 230–4 with the suggestion that they resulted from an attempt to compromise between the
sometimes conflicting demands of economic time divisions (the 8-day week ofnundinae,
regulating markets) and religious cultic dates. I do not think, however, that this alone could
account for either the end of the Roman lunar calendar, or the failure to institute an accurately
solar calendar.
(^142) Macrobius,Saturnalia1. 13. 21 (partially cited below, n. 148).
(^143) Michels (1967) 101–3, Samuel (1972) 167, Rüpke (1995) 289–92.
(^144) Bennett (2005) argues that thelex Aciliadetermined that intercalations should always
follow the 24th day of Februarius, i.e. an addition of 23 days, except when they occurred in pairs
(i.e. in two consecutive years), in which case thefirst intercalation would be after the 23rd of
Februarius, i.e. an addition of 22 days. This rule would imply that intercalations were determined
at least one year in advance, which in Cicero’s period seems not to have been the case (see below,
n. 181). According to Bennett, this rule was consistently followed from 191BCEonwards, except
in 55BCEwith a single (unpaired) intercalation of 22 days.
TheRise of the Fixed Calendars 209