When Is Holy Week: The Astronomical and Liturgical Mechanics Behind the Dates

Holy Week begins on Palm Sunday, the Sunday before Easter, and runs through Holy Saturday. Its dates shift every year because Easter is fixed to the first Sunday after the first full moon on or after the spring equinox, calculated on a traditional lunar cycle rather than the modern astronomical moon. In 2025, Holy Week runs from April 13 to April 19; in 2026, from March 29 to April 4.

That single rule, codified at the Council of Nicaea in 325 CE, explains almost everything about why Holy Week lands where it does. The rest of this explainer works through the mechanics, the edge cases, and the practical consequences of a calendar that refuses to sit still.

The lunar rule that moves everything

Easter's date is not arbitrary, but it is also not directly astronomical. The Church defines Easter as the Sunday following the "ecclesiastical full moon" that falls on or after March 21, the fixed date of the ecclesiastical spring equinox. That full moon is not the moon you see in the sky. It is a computed approximation drawn from the Metonic cycle, a 19-year pattern in which lunar phases repeat on nearly the same calendar dates.

The Metonic cycle contains 235 lunar months across 19 years. Because 235 synodic months (about 29.53 days each) total roughly 6,939 days, and 19 tropical years total roughly 6,939.6 days, the two stay locked within about two hours per cycle. That is close enough for calendar purposes, and it is why the Church could compute Easter centuries in advance without observing the sky.

Two separate Easters result. Western churches, including Catholic and most Protestant bodies, use the Gregorian calendar and its updated epact tables. Eastern Orthodox churches use the Julian calendar for the fixed feasts and the Julian computation for Easter. The two systems diverge because the Julian calendar drifts about one day every 128 years relative to the solar year. Today the Julian calendar runs 13 days behind the Gregorian.

The practical effect: Orthodox Easter usually falls one, four, or five weeks after Western Easter, and occasionally on the same day. Same-day coincidences happen when both computations land on the same Gregorian date, which occurred in 2010, 2011, 2014, 2017, and 2025. The next shared date is 2028.

Plain version: the date is set by a math formula imitating the moon, not by watching the moon.

Quantified context: how far the date can swing

The earliest possible Easter is March 22, and the latest is April 25. Those boundaries are not arbitrary; they fall out of the rule. If the ecclesiastical full moon lands on March 21 and that day is a Saturday, Easter is March 22. If the full moon lands on April 18 and that day is a Sunday, the next Sunday is April 25.

Holy Week inherits that eight-day window minus six days. Palm Sunday can therefore fall as early as March 15 and as late as April 18. The entire span of possible Holy Weeks covers about 35 days.

The cycle repeats. Because the Gregorian computus uses a 5.7-million-year cycle in theory, but practically repeats a 400-year pattern, the distribution of Easter dates is fixed. March 22 and April 25 are rare: each occurs only a handful of times per 400-year cycle. Late March and mid-April dates are common. The most frequent Easter date in the Gregorian calendar is April 19.

This matters for planning. A parish that schedules Holy Week services around a fixed spring break, a choir tour, or a school calendar is scheduling against a moving target that shifts by up to five weeks year over year.

Note that the "very late" row uses a projected date; the table shows the structural pattern, not confirmed future schedules beyond the years listed.

Edge cases: Orthodox divergence and civil holidays

The Orthodox–Western gap is the most consequential edge case. Because Orthodox Easter uses the Julian computation, its Holy Week can land in late April or even early May. In 2024, Western Easter was March 31 and Orthodox Easter was May 5, a five-week gap. In 2026, Western Easter is April 5 and Orthodox Easter is April 12, a one-week gap.

The gap follows a rough pattern: one week, then four weeks, then five weeks, then one week again across consecutive years, driven by where the two computus systems land relative to each other.

A second edge case involves civil calendars. Some countries declare public holidays tied to Holy Week rather than Easter Sunday. Spain, Mexico, and much of Latin America mark Jueves Santo and Viernes Santo, the Thursday and Friday of Holy Week, as public holidays. In those countries the practical "when is Holy Week" question is answered by the government gazette, not the almanac.

A third edge case: churches that follow a revised or local calendar. A small number of communities use alternative computus rules, and their Holy Week dates can differ by a day or more from both Western and Orthodox norms. These are rare but not negligible for anyone coordinating across jurisdictions.

Misconceptions that won't die

The most persistent misconception is that Easter is set by the astronomical full moon. It is not. The ecclesiastical full moon can differ from the real full moon by up to two days, and occasionally more. The Church's tables are approximations, and the real moon is irrelevant to the official date.

A second misconception: that Holy Week is always the week before Easter. It is, by definition, but "week before" is only true if you accept that Holy Saturday is the last day of Holy Week and Easter Sunday begins the next liturgical season. Some traditions treat the Easter Vigil on Saturday night as the first liturgy of Easter, which effectively ends Holy Week at sundown on Saturday.

A third misconception: that the date is fixed by the spring equinox. The equinox itself drifts between March 19 and March 21 in the Gregorian calendar, but the Church fixes the ecclesiastical equinox at March 21 regardless. So the "spring equinox" in the rule is a convention, not an observation.

A counterintuitive professional caveat: liturgical scholars sometimes point out that the computus was never meant to be astronomically accurate. It was meant to be stable, reproducible, and independent of local observation, so that scattered communities could agree without correspondence. The drift from the real moon is a feature, not a bug.

Practical actions: planning around a moving week

For anyone who needs to schedule around Holy Week, the actionable steps are concrete.

Look up the year's date from an authoritative computus table or a church calendar, not from a general almanac. Confirm whether you need Western or Orthodox dates before you book anything. Block the full eight-day window, from Palm Sunday through Holy Saturday, not just Easter Sunday, if your event depends on Holy Week liturgies. Check local public-holiday status in the country where the event occurs. Jueves Santo and Viernes Santo are official holidays in several jurisdictions, which affects staffing and transport. For multi-year planning, use the 400-year Gregorian pattern rather than extrapolating from two or three recent years. Cross-check any projected date against at least two independent sources before committing to contracts or travel.

A red-flag warning for coordinators: do not assume that a one-week gap between Western and Orthodox Easter will repeat. The gap sequence is not a simple alternation, and a five-week gap can follow a one-week gap in consecutive years.

Limitations: what the rule cannot tell you

The computus gives a date, not a schedule. It cannot tell you the local start time of a Vigil, whether a diocese has transferred a feast, or how a particular community handles the Saturday-to-Sunday boundary. Those are local decisions.

The rule also cannot resolve the Orthodox–Western divergence for you. If your event spans both traditions, you need to decide which calendar governs, and then reconcile the two dates explicitly rather than picking one and hoping.

Finally, the astronomical language in the rule invites over-reading. The ecclesiastical full moon is a table value, not a prediction, and the equinox is a convention date. Anyone expecting the rule to track the sky will be disappointed; anyone expecting it to be consistent and reproducible will find it works exactly as designed.