The most widespread view of Julius Caesar’s calendar reform is that after 45BC, the leap years were misapplied, namely every 3rd year instead of every 4th. Brutus assassinated Caesar in March of 44BC, so this seems believable.
This is the so-called fencepost error; see the image above. For a 10m long fence, 11 posts are needed, not 10. It is today a typical “software cycle bug” in programming. Some explain the leap year error as an “inclusive count” mistake of the leap year cycle. This is hardly plausible, as it would suggest that they may have been counting in the current “exclusive way” when the leap year cycle of the new calendar was set and promulgated.
The leaping failure lasted rather long, 36 years. Instead of nine, 12 leap years were added. Emperor Augustus had to restore the calendar after 9BC, omitting three leap years. Because emperor Augustus restored the calendar, it only started to function correctly (according to the original plan) again from AD8. Some calendar experts calculate that the restoration was finished already in AD4.
The above erroneous use of leap years is described by only one early reference and one later recollection (Solinus, 3rd century AD and Macrobius, 5th century AD). However, at least nine conflicting theories explain the misuse of leap years, and the different recalculations show unreliable results.
According to recent research on the other side, it is known that in Alexandria, the correct roman calendar leaping was applied during these years. Emperor Augustus seems to change the old civil calendar in Alexandria according to the 4-year leap-year cycle of the Julian calendar as early as 24BC.
In my opinion, if Augustus knew already in 24BC that the Roman 3-year leap year application was wrong, he would not have waited until 9BC to start correcting the error in Rome. (Even though that year he was not yet the “lord of time”, Augustus was Pontifex Maximus only from 12BC). And if in 24BC he had not yet known that the leap year was wrong, he would have introduced the bad leap year of 3 years in Alexandria, too.
The correct decreed and applied Roman leap year in Egypt may suggest that Rome’s calendar was not corrupted for 36 years. As far as I see, the later historical recollection of the misunderstanding and misuse of leap years may be wrong.
On the other side, it also appears that the Gregorian Dilemma would have meant six days instead of 3 if there had indeed been a leap year error, but Emperor Augustus would not have restored it.
For a long time, many calendar scholars assumed that Pope Gregory XIII had to delete ten days because the First Council of Nicaea (325 AD) had already removed three days from the Julian calendar earlier. The omission of 3 days by the First Council of Nicaea is now rejected by most scholars because there is no evidence. There are no sources of skipped days or changing the order of leap years either in the year of the First Council of Nicaea or in the decades followed.
However, the final solution to this dilemma has been missing until now.
All the sources I know claim that Emperor Augustus has fully restored the Julian calendar, considering all the days passed. So, as a total of inserted and cancelled leap years of the Julian calendar, no days were added or deleted before AD8. Consequently, how leap years were applied up to AD8 is only relevant if we look for a specific date between 45BC and AD8.
The significant point for the following is that between AD8 to AD1582 (from finishing the restoration by Augustus until the calendar reform of Pope Gregory XIII), the Julian calendar functioned continuously and unchanged as planned initially by Caesar and Sosigenes. (Whether or not the leap year error and the restoration took place.)
We come back to the leap year issue in a later post concerning the birthday of Emperor Augustus.
The RECOUNTING of TIME 