The remains of the Fasti Praenestini
Prehistoric lunar calendar
The original Roman calendar is believed to have been an observational lunar calendar whose months began from the first signs of a new crescent moon. Because a lunar cycle is about 29 1⁄2 days long, such months would have varied between 29 and 30 days. Twelve such months would have fallen 10 or 11 days short of the solar year; without adjustment, such a year would have quickly rotated out of alignment with the seasons in the manner of the present-day Islamic calendar. Given the seasonal aspects of the later calendar and its associated religious festivals, this was presumably avoided through some form of intercalation or through the suspension of the calendar during winter.
Rome's 8-day week, the nundinal cycle, was shared with the Etruscans, who used it as the schedule of royal audiences. It was presumably a feature of the early calendar and was credited in Roman legend variously to Romulus and Servius Tullius.
Legendary 10-month calendar
The Romans themselves described their first organized year as one with ten fixed months, each of 30 or 31 days. Such a decimal division fitted general Roman practice. The four 31-day months were called "full" (pleni) and the others "hollow" (cavi).[a] Its 304 days made up exactly 38 nundinal cycles. The system is usually said to have left the remaining 50-odd days of the year as an unorganized "winter", although Licinius Macer's lost history apparently stated the earliest Roman calendar employed intercalation instead and Macrobius claims the 10-month calendar was allowed to shift until the summer and winter months were completely misplaced, at which time additional days belonging to no month were simply inserted into the calendar until it seemed things were restored to their proper place.
Later Roman writers credited this calendar to Romulus, their legendary first king and culture hero, although this was common with other practices and traditions whose origin had been lost to them. Some scholars doubt the existence of this calendar at all, as it is only attested in late Republican and Imperial sources and apparently supported only by the misplaced names of the months from September to December. Rüpke also finds the coincidence of the length of the supposed "Romulan" year with the length of the first ten months of the Julian calendar to be suspicious.
Other traditions existed alongside this one, however. Plutarch's Parallel Lives recounts that Romulus's calendar had been solar but adhered to the general principle that the year should last for 360 days. Months were employed secondarily and haphazardly, with some counted as 20 days and others as 35 or more.
The attested calendar of the Roman Republic was quite different. It followed Greek calendars in assuming a lunar cycle of 29 1⁄2 days and a solar year of 12 1⁄2 synodic months (368 3⁄4 days), which align every fourth year after the addition of two intercalary months. The additional two months of the year were January and February; the intercalary month was sometimes known as Mercedonius.
The Romans did not follow the usual Greek practice in alternating 29- and 30-day months and a 29- or 30-day intercalary month every other year. Instead, their 3rd, 5th, 7th, and 10th months[b] had 31 days each; all the other months had 29 days except February, which had 28 days for three years and then 29 every fourth year. The total of these months over a 4-year span differed from the Greeks by 5 days, meaning the Roman intercalary month always had 27 days. Similarly, within each month, the weeks did not vary in the Greek fashion between 7 and 8 days; instead, the full months had two additional days in their first week and the other three weeks of every month ran for 8 days ("nine" by Roman reckoning). Still more unusually, the intercalary month was not placed at the end of the year but within the month of February after the Terminalia on the 23rd (a.d. VII Kal. Mart.); the remaining days of February followed its completion. This seems to have arisen from Roman superstitions concerning the numbering and order of the months. The arrangement of the Roman calendar similarly seems to have arisen from Pythagorean superstitions concerning the luckiness of odd numbers.
These Pythagorean-based changes to the Roman calendar were generally credited by the Romans to Numa Pompilius, Romulus's successor and the second of Rome's seven kings, as were the two new months of the calendar.[c] Most sources thought he had established intercalation with the rest of his calendar. Although Livy's Numa instituted a lunar calendar, the author claimed the king had instituted a 19-year system of intercalation equivalent to the Metonic cycle centuries before its development by Babylonian and Greek astronomers.[d] Plutarch's account claims he ended the former chaos of the calendar by employing 12 months totaling 354 days—the length of the lunar and Greek years—and biennial intercalary months of 22 days.
Plutarch believed Numa was responsible for placing January and February first in the calendar; Ovid states January began as the first month and February the last, with its present order owing to the Decemvirs. W. Warde Fowler believed the Roman priests continued to treat January and February as the last months of the calendar throughout the Republican period.
The consuls' terms of office were not always a modern calendar year, but ordinary consuls were elected or appointed annually. The traditional list of Roman consuls used by the Romans to date their years began in 509 BC.
Gnaeus Flavius, a secretary to the pontifex maximus, introduced a series of reforms in 304 BC. Their exact nature is uncertain, although he is thought to have begun the custom of publishing the calendar in advance of the month, depriving the priests of some of their power but allowing for a more consistent calendar for official business.
Julius Caesar, following his victory in his civil war and in his role as pontifex maximus, ordered a reformation of the calendar in 46 BC. This was undertaken by a group of scholars apparently including the Alexandrian Sosigenes and the Roman M. Flavius. Its main lines involved the insertion of ten additional days throughout the calendar and regular intercalation of a single leap day every fourth year in order to bring the Roman calendar into close agreement with the solar year. The year 46 BC was the last of the old system and included 3 intercalary months, the first inserted in February and two more—Intercalaris Prior and Posterior—before the kalends of December.
After Caesar's assassination, Mark Antony had Caesar's birth month Quintilis renamed July (Iulius) in his honor. After Antony's defeat at Actium, Augustus assumed control of Rome and, finding the priests had (owing to their inclusive counting) been intercalating every third year instead of every fourth, suspended the addition of leap days to the calendar for one or two decades until its proper position had been restored. See Julian calendar: Leap year error. In 8 BC, the plebiscite Lex Pacuvia de Mense Augusto caused Sextilis to be renamed August (Augustus) in his honor.[e]
In large part, this calendar continued unchanged under the Roman Empire. (Egyptians used the related Alexandrian calendar, which Augustus had adapted from their wandering ancient calendar to maintain its alignment with Rome's.) A few emperors altered the names of the months after themselves or their family, but such changes were abandoned by their successors. Diocletian began the 15-year indiction cycles beginning from the AD 297
census; these became the required format for official dating under Justinian. Constantine formally established the 7-day week by making Sunday an official holiday in 321. Consular dating became obsolete following the abandonment of appointing nonimperial consuls in AD 541. The Roman method of numbering the days of the month never became widespread in the Hellenized eastern provinces and was eventually abandoned by the Byzantine Empire in its calendar.