Guides
March birthdays: the record, kept honest
March birthdays: why a calendar drifts against the seasons, what the Gregorian reform changed, and why a birthday is a civil count rather than a sky event.
Every date on this site uses a calendar redesigned for one reason: the spring equinox would not stay put. That sounds like an astronomer's concern.
It also means a date from an old record needs more care than its digits suggest. A birthday is a civil count, not a fact about the sky.
What to take away
- A calendar year is a rounded number of days, and every rounding rule drifts against the seasons at its own rate.
- The Gregorian reform corrected that drift by dropping days and changing the leap rule, so that the equinox returned to roughly where it had sat in the fourth century.
- A birthday marks a count of civil days, not a return of the sun to the same place. The two agree closely and never exactly.
Why a calendar drifts at all
The tropical year, from one spring equinox to the next, is not a whole number of days. It is a little under 365.25.
Calendar drift and the 1582 reform
- Julian ruleleap day every four years
- Average year365.25 days, slightly too long
- By 1500sequinox crept about ten days
- 1582 reformdropped accumulated days
- New leap rulecentury years need divisibility by 400
- Average year365.2425 days, drift now slow
A calendar uses whole days, so it approximates. The Julian calendar approximated with a leap day every four years, making its average year exactly 365.25 days. That is slightly too long. The excess is small, about three days in four centuries, but it accumulates one way and never cancels.
Over the centuries, the calendar date of the equinox crept earlier. By the sixteenth century, it had moved about ten days from where it stood when Easter's dating rules were fixed.
The reform of 1582 did two things: it dropped the accumulated days and changed the leap rule so century years are leap years only when divisible by 400.
That brought the average year down to 365.2425 days, close enough to the real figure that the remaining drift takes thousands of years to reach a single day.
What the equinox has to do with a birthday
Nothing directly, and that is the point worth keeping. A birthday repeats a date, and a date is a position in a civil count.
The March equinox moves about two days from year to year because calendar and sun are only roughly in step. Someone born at the equinox was born on a calendar date, and that date returns.
The distinction matters because many claims about birth dates quietly assume the date is a natural fact. It is a recorded one. The January index sets out what a recorded entry consists of, and none of its four fields is astronomical.
Three calendars that mostly agree
| Calendar | Leap rule | Where it still shapes dates |
|---|---|---|
| Julian | Every fourth year, without exception | Records made before a country switched; some churches' liturgical calendars |
| Gregorian | Every fourth year, except century years not divisible by 400 | Civil use almost everywhere |
| Revised Julian | A different century rule with the same practical result for several centuries yet | Adopted by a number of Orthodox churches in the twentieth century |
The third row earns its place because the Revised Julian calendar matches Gregorian dates now but not forever.
Three calendars, three leap rules
Julian
- Leap rule
- Every fourth year
- Where it shapes dates
- Old records, some liturgies
- Matches Gregorian now?
- No
- Matches Gregorian forever?
- No
Gregorian
- Leap rule
- Except century non-400s
- Where it shapes dates
- Civil use almost everywhere
- Matches Gregorian now?
- Yes
- Matches Gregorian forever?
- Yes
Revised Julian
- Leap rule
- Different century rule
- Where it shapes dates
- Some Orthodox churches
- Matches Gregorian now?
- Yes
- Matches Gregorian forever?
- No
A date recorded under it today is safe to treat as Gregorian, though a converter that assumes they are identical will one day be wrong. Note which calendar a record used, since it costs nothing now.
What this means for an entry in this month
- Modern records need no correction. The drift argument is about which calendar a record was kept in, not about which month it names.
- Records kept under the Julian calendar after 1582 carry a day offset that depends on the century. February entries from the same period carry it too, and the site's historical pages deal with the conversion itself.
- The month boundary is a civil line. No source can show that a person born in late March shares anything with anyone else born in late March, beyond the digits.
- A claim that a birthday falls on the equinox is true of a calendar date in some years and false in others. It is not a stable property of the person.
Common questions
Was the Julian calendar wrong?
It was a good approximation that ran slightly long. Any whole-day calendar is an approximation. The only questions are how fast it drifts and whether anyone corrects it.
Does the Gregorian calendar drift too?
Yes, far more slowly. Its residual error is small enough that no correction has been scheduled, and none will be needed for any date a birthday index will ever hold.
Should a modern index convert old dates from this month?
It should state which calendar each date is in and keep the original wording recoverable. A silent conversion is the one choice that cannot be undone later.
