[HOME]

Roman Dates

Intro page      |      How to Read the Tables      |      The Roman Calendars       |       Sources      |      Analysis
Conversion table: (Excel) (HTML) (CSV)                                                   Fasti consulares: (Excel) (HTML)

The Roman Year

Structure of the Year

The Roman civil year was normally divided into 12 months: Ianuarius, Februarius, Martius, Aprilis, Maius, Iunius, Quintilis (later Iulius), Sextilis (later Augustus), September (later briefly Germanicus), October, November and December. The lengths of the months and the means by which the calendar was aligned to the tropical year changed during Roman history, most notably before and after the reforms introduced by Julius Caesar in A.U.C. 708 = 46 B.C.

Varronian Years

The primary method of identifying a Roman year used today is by the number of years from the year Rome was founded -- A.U.C. (ab urbe condita). The festival of the Parilia, held on a.d. X Kal. Mai., marked the supposed anniversary of the founding of Rome. However, in practice, years A.U.C. after the foundation of the Republic are equated to the eponymous years of the consular fasti, i.e. they run from the start to the end of the consular term, which did not necessarily start at the beginning of the calendar year and which changed from time to time.

The convention used by modern scholars is that which was introduced by the late Republican scholar M. Terentius Varro, in which A.U.C. 1 corresponds approximately to 753 B.C. By convention, since Roman years and Julian years are approximately synchronised, Varronian years (and by extension eponymous years, to which they are equated) are mapped to the Julian year which covers the bulk of the Roman civil year. In the text here, this dual dating is usually given explicitly in the form A.U.C. nnn = mmm B.C. or A.U.C. nnn = mmm. This convention should be read as "the Roman year of the eponymous consuls, not named here, which corresponds to A.U.C. nnn, and which mostly overlaps the Julian year mmm B.C." On the few occasions where years A.D. are referred to, the form is always A.U.C. nnn = A.D. mmm.

While the Varronian date of A.U.C. 1 = 753 B.C. is by far the most well-known equation today, it was not the only A.U.C. system on offer, nor even the system most widely used in classical times. The Fasti Capitolini, which explicitly record the A.U.C. date of every 10th year in the Fasti Consulares and of every triumph in the Fasti Triumphales, use a system corresponding to A.U.C. 1 = 752 B.C. P. Brind'Amour, Le calendrier romain 211, notes that recorded ancient Roman estimates for A.U.C. 1 varied from 758 B.C. (Calpurnius Piso) to 728 B.C. (L. Cincius Alimentus). In De Re Publica 2.17, Cicero follows Polybius 3.22 in dating the foundation of Rome to Ol. 7.2 = 751/0; the same view was held by Livy. But the Greek historian Timaeus dated the foundation of Rome and Carthage to the same year, 814 B.C., and the early annalist Ennius seems to have held an even more radical view. Varro, De Re Rustica 3.1, quotes some lines of Ennius which refer to Rome as having been founded around 700 years earlier. While we do not know when these lines were written, they suggest that Ennius himself may have dated the foundation of Rome seven centuries before his own adult life, i.e. to the period c. 920-870 B.C.; alternately, based on what little we know about his work, he may have dated the foundation a few years after the fall of Troy.

Eponymous Years

In the period covered by these tables, the Romans normally identified their civil years eponymously, by the names of the consuls who were elected annually. This system was believed to have been used from the start of the Roman republic and continued to be used in the empire. A large number of eponymous tables (fasti consulares) are known. The earliest such tables known to us were compiled in late republican times and canonised under Augustus, but are generally agreed to be accurate after 300 B.C., which covers the period of interest here. As well as the internal evidence of the fasti, some confirmation of their accuracy comes from the calculated dates for consuls named by Polybius, who wrote in the mid second century B.C., which correspond with those in the fasti consulares; see further discussion here.

The consular eponyms for a year would normally be known in advance, though in times of disorder (e.g. A.U.C. 677 = 77, the year of the tumultus Lepidianus) this would not be the case. If a consul died or was otherwise replaced by another consul (a "suffect" consul), the replacement did not usually affect the annual eponym, though exceptions are known. If a consul designate died before taking office, or was removed in disgrace, the situation is more ambiguous; the available sources break both ways.

This system effectively ties the start of the eponymous year to the nominal start of the consular term. According to Livy, Periochae 47, the Roman consular year started on Kal. Ian. after A.U.C. 600 = 154. However, earlier consular years started on Id. Mart. (A.U.C 600 was a short year running from Id. Mart. to prid. Kal. Ian.). There is also strong, if controversial, evidence that the consular year started on Kal. Mai. before A.U.C. 532 = 222. Before the period covered by these tables, other dates are recorded: Kal. Sex. (Livy 3.6), Id. Mai. (Livy 3.36), Id. Dec. (Livy 4.37, Livy 5.9), Kal. Oct. (Livy 5.9), Kal. Quin. (Livy 5.32, 8.20)

The Roman Day and Month

According to our literary sources, the Roman day began at midnight (e.g. Pliny, Hist. Nat. 2.79), although in the few horoscopes from Egypt that give Roman dates the start of the day is set at sunset, in the Greek fashion.

The Romans had what is for us a very unusual method for counting days in a month, although similar methods were used in several Greek calendars. Each month contained three special days: the Kalends, which was always the first of the month; the Ides, which was on the 13th or 15th day of the month, depending on the month; and the Nones, which was 8 days before the Ides. Other days were identified by counting the days up to and including the next reference point, in an abbreviated form ("a.d." = "ante diem" = "days before"). Thus, a typical date would be:

a.d. V Id. Mart. = the fifth day before the Ides (15th day) of Martius, inclusive = the 11th day of Martius

This system was modified for the day before a reference point, which was simply known as "the day before" (pridie), so:

prid. Id. Mart. = the day before the Ides of Martius = the 14th day of Martius

This system has an effect that seems rather odd to us. Days after the Ides are given with reference to the Kalends of the next month. Thus, in the Julian calendar:

a.d. XVII Kal. Apr. = the 17th day before the Kalends of Aprilis, inclusive = the 16th day of Martius

Additionally, in the Julian calendar, at least after the Augustan reform of A.U.C. 746 = 8, the leap day was intercalated, in effect, by making a.d. VI Kal. Mart. a 48-hour day, the so-called "bissextile" day (a.d. bis. VI Kal. Mart.).

Many days were also festival days and were known by their festival name (Quirinalia, Lupercalia, Saturnalia etc). A major issue in the design of the Caesarian reform of A.U.C. 709 = 45 was to insert extra days in such a way that the relationship between the festival days and their day numbers were unchanged. For a detailed worked example of the Roman festivals (in a Julian calendar), see here.

The Julian Calendar

Month Lengths

The structure of the Roman calendar after the reform of A.U.C. 709 = 45 is given in detail by the fifth century author Macrobius, and is confirmed by a number of fasti (calendars) surviving from the reigns of Augustus and Tiberius. It is essentially that of the modern Western calendar. The year was normally 365 days long, and the months had the modern lengths. The following table shows the day number for the Nones and Ides of each month, and the length of the month, with lengths in leap years shown in brackets:

Ian.

Feb.

Mart.

Apr.

Mai.

Iun.

Quin. (Iul.)

Sex.
(Aug.)

Sept. (Germ.)

Oct.

Nov.

Dec.

Nones

5

5

7

5

7

5

7

5

5

7

5

5

Ides

13

13

15

13

15

13

15

13

13

15

13

13

Length

31

28 (29)

31

30

31

30

31

31

30

31

30

31

Month Names

When the Julian reform was introduced, the months retained the names they had had under the republican calendar. However, several months changed names under the early Empire, most only for a few years.

Only the first three of these changes fall within the scope of the tables presented here.

Intercalation

Leap days in the Julian calendar were intercalated every fourth year. The day after the Terminalia, a.d. VI Kal. Mart. = 24 February, was repeated, the so-called "bissextile" day (a.d. bis VI Kal. Mart.).

The intended position of the bissextile day -- before or after a.d. VI Kal. Mart. -- is not entirely clear, and may not have originally been specified. The earliest evidence is Celsus 39, written late in the first or early in the second century AD, cited in Justinian, Digest 50.16.98. Celsus notes that the two days were treated as a single date (biduum) with prior and posterior days, and states it was the posterior day that was considered to be intercalated for purposes of reckoning birthdays. The inscription CIL VIII 6979, dated A.D. 168, notes that a.d. V Kal. Mart. was the day after the bissextile day ("V K. Mart. qui dies post bis VI K. fuit"). However, Censorinus, De Die Natali, 20.10, writing in A.D. 238, states that Caesar intercalated a single day, "now" called the bis sextum, after the Terminalia ("dies unus, ubi mensis quondam solebat, post Terminalia intercalaretur, quod nunc bis sextum vocatur"), and Macrobius, Saturnalia 1.14.6, explicitly states that it was followed by the last 5 days of February, i.e. a.d. VI Kal Mart. was after the bissextile day. This is the position that it held throughout the Middle Ages, until the Roman system of numbering days was discarded.

C. L. Ideler, Handbuch der mathematischen und technischen Chronologie II 621, held that Celsus' use of "prior" and "posterior" was with reference to Kal. Mart., so that the "posterior" day was the day which was furthest from Kal. Mart., i.e. the first of the two days. He cites no other examples or justification for this interpretation, but it does allow Celsus' statement to reconciled with later practice. He was unaware of CIL VIII 6979.

Ideler's interpretation has led some scholars to infer that there were actual dates "a.d. VI Kal. Mart. priorem" and "a.d. VI Kal. Mart. posteriorem", but Celsus does not say so and, to my knowledge, no such dates are known in the record.

T. Mommsen, Römische Chronologie 278, held that CIL VIII 6979 shows that a.d. V Kal. Mart. followed the bissextile day, and that Celsus' text should be interpreted according to the ordinary usage of the terms, with the word "posterior" meaning that the bissextile day was considered by Celsus to be the second half of a single 48-hour day, as in ordinary usage.

W. Sternkopf, JCP 41 (1895) 718 at 721, suggests that originally neither day was specified to be intercalary, so that the date "bis VI K." in CIL VIII 6979 merely indicated that a.d. VI Kal. Mart was twice as long as normal, i.e. that it was referring to the biduum.

This explains Celsus' statement that the intercalary month was accounted to be 28 days long (Mensis autem intercalaris constat ex diebus viginti octo) in response to the opinion of Cato and Quintus Mucius that the length of the (pre-Julian) intercalary month is added to that of February. That the compilers of Justinian's Digest understood Celsus to be talking about the Julian intercalary month (February), not the pre-Julian intercalary month (Intercalaris), is clear from the context and the word autem. Mommsen noted that this interpretation is confirmed by the ninth century Byzantine law code, the Basilika 2.2.95, which incorporated and updated Justinian's code, and explicitly restates, on this issue, that "February" has 28 days.

Sternkopf supposed that the concept of a 48-hour day dropped out of ordinary use, at which time the first of the two days came to be considered as the bissextile day, hence Censorinus' statement that the intercalary was now (nunc) called the bis sextum. He pointed out that Caesar had gone to some lengths to avoid changing the calendar position of various religious festivals and that the concept of a biduum allowed the length of Februarius to remain at 28 "days" even in a leap year. However, one aspect of intercalation could not be avoided: the festival of the Regifugium. Sternkopf held that in the pre-Julian calendar this was celebrated on the sixth (inclusive) day before Kal. Mart., at the end of the intercalary month. Therefore, by analogy, it must have been celebrated on the second day of the biduum after Caesar's reform. He concluded that the intercalary day, to the extent that it was considered to have a separate existence, was the first day of the biduum.

As to Celsus, Sternkopf rejected Ideler's proposed interpretation of his use of "prior" and "posterior" as specious. He noted that Ulpian, who lived under Septimius Severus and Caracalla, cited Celsus' view, using the terms in their ordinary sense. Ulpian explained that the issue was to determine the precise moment a man reached his majority (Justinian, Digest 4.4.3), i.e. his 25th birthday, if he was born on one of the two days of a.d. VI Kal. Mart. in a leap year. Sternkopf argued that Celsus was making a legal ruling without regard to the historical origins of the biduum.

Much of what Sternkopf says is quite plausible. However, I think his argument about the Regifugium is weak, because one could equally well argue that the notion that Februarius always had 28 days means the festival should be celebrated at its regular time every year. In support of this position, Macrobius, Saturnalia 1.14.9 explains that Caesar took great care not to disturb the position of feriae (such as the Regifugium) when creating the Julian calendar.

There is an additional point that Sternkopf did not consider. The biduum concept implies it had a single nundinal letter. The analysis of the imperial nundinal cycle presented here, based on evidence not available to Sternkopf, shows that the market day fell on the first day of the biduum in A.D. 44 if it is accounted as a single day in the cycle. Dio Cassius 60.24.7 notes that the market day was moved to another day in A.D. 44 due to "religious rites". Combining this with Celsus' ruling, the "religious rite" may be identified as the Regifugium. This means that the market day was moved from the first to the second day of the biduum. It follows that the Regifugium was celebrated on the day after a.d. VII Kal. Mart. and that the following day was the intercalated day.

As the Regifugium became an archaic festival that was no longer celebrated, and as the nundinal cycle was replaced by the sabbatical week, this reason for making the bissextile the second day would have disappeared. The nundinal cycle was still in use in the early third century A.D., but the Regifugium seems to have fallen away earlier. The vagueness of the position of the intercalary day which Sternkopf argues is implicit in the concept of the biduum could well have led to the first day becoming the customary intercalary day.

The disappearance of the biduum concept may have been gradual. The fact that Celsus is cited by Justinian suggests that it still had meaning in the 6th century A.D. Ammianus Marcellinus 26.1.7 and 26.2.1 explains that the accession of Valentinian I in A.D. 364 was postponed to the day after both the bissextile day, considered unlucky, and the following day, reasons that are not clearly explained. The Fasti Idatiani confirms that his accession date was a.d. V Kal. Mart. (26 February), i.e. that the bissextile day preceded a.d. VI Kal. Mart. in A.D. 364.

Initially Julian intercalation was not as automatic as it later became. According to Macrobius, Saturnalia 1.14.6, Caesar intended leap days to be intercalated every fourth year, though we are not told what phase he intended (i.e. which year of a four cycle). However, Macrobius, Saturnalia 1.14.14 informs us that in the years immediately following his death, leap days were intercalated every third year, until 12 days had been intercalated in the time that should have seen 9 intercalations. After this, Augustus reformed the calendar by suspending intercalation for 12 years, to compensate for the incorrect intercalations, and then resumed it according to the schedule we know today. The primary difficulty for Roman/Julian conversion in these years is to determine exactly when these events occurred.

The Republican Calendar

Month Lengths

According to the account of Macrobius, Saturnalia 1.14.7, the ordinary year was 355 days long before the Caesarian reform. The year had the same twelve months as the later Julian calendar, but with different lengths, as shown by the following table:

Ian.

Feb.

Int.

Mart.

Apr.

Mai.

Iun.

Quin.

Sex.

Sept.

Oct.

Nov.

Dec.

Nones

5

5

- (5)

7

5

7

5

7

5

5

7

5

5

Ides

13

13

- (13)

15

13

15

13

15

13

13

15

13

13

Length

29

28 (23 or 24)

0 (27)

31

29

31

29

31

29

29

31

29

29

This account is confirmed and refined by the surviving fragments of the contemporary calendar known today as the Fasti Antiates Maiores.

Intercalation

The Republican calendar maintained an approximate alignment to the solar year through a complex process of intercalation. 22 or 23 days -- as determined by the pontifex maximus to be appropriate -- were inserted at intervals determined by him. The central problem of Republican chronology in the period covered here is to determine which years had intercalary months, and how long they were.

The exact method by which intercalation was done is controversial. It was long held that the intercalary month ("Intercalaris") itself was of variable length, and was inserted on the day after the Terminalia. 22 or 23 days were then combined with the final 5 days of Februarius to form an intercalary month of 27 or 28 days. However, the Fasti Antiates Maiores shows a fixed Intercalaris of 27 days, with the Kalends being assigned a nundinal letter of G, and ending with a nundinal letter of A, as in a standard Februarius, so it appears that the intercalary month itself was fixed, and the variability was handled in a different way.

The solution appears to be given by Livy 43.11.13, which states that in A.U.C 584 = 170 an intercalation was inserted on the third day (counted inclusively) after the Terminalia, and Livy 45.44.33, which states that in A.U.C 587 = 167 an intercalation was inserted on the day after the Terminalia. These passages suggest that it was effectively Februarius that was made variable in length. The simplest model is that proposed by A. K. Michels, The Calendar of the Roman Republic 160ff. In this model, a fixed length 27-day Intercalaris started either on the day after the Terminalia (23rd day of Februarius) for a 22 day intercalation, or on the following day, for a 23-day one. The remaining days of Februarius were dropped, and the religious festivals that normally followed the Terminalia, the Regifugium and the Equirria, were held on the last four days of Intercalaris, days which had the same date as the the last four days of an ordinary Februarius (a.d. V Kal. Mart. etc).

Certain formal events, e.g. triumphs and laws, are sometimes dated with respect to festivals in late Februarius (Quirinalia, Feralia, Terminalia -- no example of Lupercalia-based dating is yet known), rather than being antedated to Kal. Mart. While the exact festival day is used as early as the fourth century, the earliest known example of antedating with respect to a festival is a date a.d. III Feralia from A.U.C. 654 = 100. This custom ensures that the date in question is understood to be in Februarius rather than in Intercalaris. However, that need not imply that the year was intercalary. A.U.C. 704 = 50 is a case where such dates were used; the year is known to have been a candidate for intercalation but intercalation did not occur. In these pages it is assumed that such dates indicate a candidate intercalary year. However, it may simply be that this was a customary way to date things in late Februarius, regardless of whether the year was a candidate intercalary year, and that it existed alongside antedating to Martius.

It would not be possible to name the day between the Terminalia and Kal. Int. in a 378-day year using this technique. It is unknown how this was done. There are two possibilities. In a 377-day year, the day following the Ides of Februarius was a.d. XI Kal. Int. In a 378-day year, that day could have become a.d. XII Kal. Int; alternately, it could have been accounted as a.d. X Term. in both types of year, as is documented for A.U.C. 610 = 94 B.C. In any case, the day after the Terminalia could simply have been prid. Kal. Int. Alternatively, there may have been a separate name for the day after the Terminalia, as is suggested by Macrobius' account of a separate intercalary day.

Finally, we are told by Censorinus 20.8 that Caesar inserted two extraordinary intercalary months, totalling 67 days, between November and December A.U.C. 708 = 46. The individual lengths of each month are not known; for Rüpke's plausible proposal see discussion under A.U.C. 708 = 46. The total matches three regular intercalations (67 = 22+22+23), so it is possible that that they were intended to compensate for intercalary months that had been omitted earlier in Caesar's pontificate. In the 17 years from his accession in A.U.C. 691 = 63 to A.U.C. 708 = 46, only 5 regular intercalations can be identified. Interpreting this extraordinary intercalation as three additional intercalary months would bring the total up to 8, which matches the average frequency of intercalation for earlier years. However, it is unclear why they should be represented as two months, not three.

Nundinal Letters

An important feature of the Roman calendar for chronological reconstruction is the 8-day nundinal cycle, which was roughly similar to the modern 7-day week. The cycle is not explicitly described in any surviving literary source, but its operation is generally clear from surviving fasti. The cycle operated in both the Republican and Julian calendars. Macrobius, Saturnalia 1.16.34, tells us that the Roman market day occurred every 8 days. Since Dio Cassius 48.33.4 records that a day was added to A.U.C. 713 = 41 in order to avoid a market day on the first of the following year, it seems certain that this market cycle was continuous throughout Republican times. However, Dio Cassius 60.24.7 notes that the market day was changed in A.U.C. 797 = A.D. 44 and not for the first time, from which we can conclude that it was interrupted in the early imperial period. By this time, it also coexisted with the modern sabbatical week, which eventually came to replace it.

Each day in the civil year was associated with a nundinal letter from A to H, in a cycle that was reset to A on Kal. Ian. every year, at least after A.U.C. 600 = 154, and repeated every 8 days thereafter, except in intercalary years.

After A.U.C. 600 = 154, the civil year was the same as the consular year. From A.U.C. 532 = 222 to A.U.C. 600 = 154, the consular year began on Id. Mart. Before then it most likely began on Kal. Mai. in the period covered here. Surviving annals are organised by the consular year. As noted above, it is not known for sure whether the civil year tracked the consular year or whether it already started on Kal. Ian., although the latter seems far more likely. Hence it is not clear exactly how the nundinal cycle operated at this time, e.g. whether it was reset to A on Id. Mart. instead of Kal. Ian. However, the consular reform did not change the phase relationship between the nundinal cycle and the calendar, so for chronological purposes we may assume a (possibly proleptic) reset of the nundinal letters on Kal. Ian. throughout the Republican period.

Website © Chris Bennett, 2001-2012 -- All rights reserved

25 Jan 2012: Various corrections and clarifications; expand discussion of start of calendar year; note Rüpke's theory of intercalary announcement.