Tithi: The Lunar Day That Is Almost Never 24 Hours Long

Every year somebody tells me their family calendar says one date for a festival and the temple down the road says another, and asks which one is wrong. Usually neither is wrong. They are applying the same rule to a quantity that does not sit still, and the rule produces different answers depending on where you are standing and which sunrise you measure from.

The quantity is the tithi. It is the first of the five limbs of the pañcāṅga, and it is the one people misunderstand most, because the word gets translated as “lunar day” and that translation makes it sound like a day. It is not a day. It is an angle.

What a tithi actually measures

Take the Moon’s longitude and subtract the Sun’s longitude. That difference, the elongation, starts at zero at the new moon and grows to 360 degrees by the next new moon. Slice that 360 degrees into thirty equal parts of 12 degrees each. Each slice is one tithi.

So Pratipadā is the interval during which the Moon is between 0 and 12 degrees ahead of the Sun. Pañcamī is 48 to 60 degrees. Pūrṇimā is the fifteenth slice of the bright half, ending at exactly 180 degrees, which is the instant of full moon. Amāvāsyā ends at exactly 360 degrees, which is the instant of new moon.

This is a purely geometric definition. There is no reference to sunrise, no reference to a calendar, no reference to where you live. At any given instant, the entire planet is in the same tithi. That single fact resolves half the confusion people carry about pañcāṅga, and it creates the other half, which I will get to.

Why a tithi is almost never twenty four hours long

If the Moon and the Sun both moved at constant speeds, every tithi would have the same length and this article would be short. They do not. The Moon’s orbit around the Earth is an ellipse, so the Moon moves faster near perigee and slower near apogee.

Its daily motion swings between roughly 11 and a half degrees and a little over 15 degrees. The Earth’s orbit is also an ellipse, so the Sun’s apparent daily motion varies too, between about 57 and 61 arc minutes.

A tithi ends when the gap between them has grown by 12 degrees. When the Moon is racing and the Sun is dawdling, that gap opens quickly and the tithi is short. When the Moon is crawling, the tithi drags.

In practice a tithi runs somewhere between about 21 and a half hours and about 26 hours. The average is roughly 23 hours 37 minutes, which is why a lunar month of thirty tithis runs about 29 and a half days instead of thirty.

So a tithi is shorter than a solar day more often than it is longer. That imbalance is the engine behind everything that follows.

The sunrise rule, and why your date differs from your neighbour’s

Festivals and vratas need a date, and a date is a solar day. The classical bridge between the two is the sunrise rule: the tithi running at sunrise gives its name to that day.

If Ṣaṣṭhī is in force when the Sun comes up on Thursday, Thursday is Ṣaṣṭhī, even if Ṣaṣṭhī ends at nine in the morning and the rest of Thursday is actually Saptamī.

Sunrise is local. It happens later in Mumbai than in Kolkata, and later in Delhi than in Chennai on some dates. A tithi that ends at 6:20 in the morning has already ended for a city whose sunrise is 6:05 and has yet to end for a city whose sunrise is 6:35. Same instant, same sky, two different named days. Nobody made an arithmetic error.

On top of that, different vratas use different rules. Some observances take the tithi at sunrise. Some take the tithi in force at moonrise, some at midday, some at pradoṣa in the evening, some at midnight. The five limbs of the pañcāṅga are computed identically by every serious almanac. The reckoning rule applied on top of them is where compilations differ.

Kṣaya tithi and adhika tithi

Because most tithis are shorter than a solar day, a tithi can occasionally begin after one sunrise and end before the next. It never touches a sunrise at all. It exists in the sky, and it gets no day named after it. That is a kṣaya tithi, a lost tithi.

The opposite also happens. A long tithi can be in force at two consecutive sunrises, so two calendar days carry the same tithi name. That is an adhika tithi, a repeated one. Over a lunar month the losses and repeats roughly cancel, which keeps the lunar month close to thirty named days.

Here is the misreading I want to correct, and it is a common one. A kṣaya tithi is routinely described online as an inauspicious omen, a day when things go wrong. It is a bookkeeping outcome of orbital mechanics. The Moon was moving quickly, the tithi was brief, and it fell between two sunrises. Nothing about that carries a verdict on your day.

A live example from 2027

This is easier to see with a real case. In 2027 the Pratipadā at the start of Āśvina Śukla is a kṣaya tithi. Āśvina Śukla begins on 1 October 2027, and the first tithi of the bright half does not survive to a sunrise in the usual way, which means the day-numbering of the Navrātra sequence cannot be stated the way it usually is.

Our own engine declines to emit a Navratri start date for 2027 for exactly that reason, and I would rather ship a refusal than a confident guess.

Compilations that assign a date anyway are each applying a defensible convention, and they will not all land on the same day. The nearby fixed points are firmer: Durga Ashtami falls across 7 to 8 October 2027 and Vijayadashami falls on 10 October 2027.

The thing that is rarely said about tithi

Tithi is the only limb of the pañcāṅga that depends on both luminaries at once. Vāra depends on the count of days. Nakṣatra and yoga and karaṇa are read from position. Tithi is a difference, and a difference between two moving bodies carries information neither one carries alone.

That is why the tithi encodes the Moon’s phase and therefore its light, and why the classical texts treat the waxing half and the waning half so differently for muhūrta.

A śukla pakṣa Ṣaṣṭhī and a kṛṣṇa pakṣa Ṣaṣṭhī sit at the same distance from their respective new or full moon, and the Moon carries opposite amounts of light in the two cases. The number is the same. The condition is not.

Where tithi shows up beyond festivals

Three places, mostly. First, karaṇa is derived directly from tithi: each tithi contains exactly two karaṇas, one for each half of the 12 degree span, which is why karaṇa is sometimes called a half tithi.

Second, several muhūrta rules are stated in tithi terms, with Rikta tithis (the fourth, ninth and fourteenth) traditionally avoided for beginnings. Third, the annual chart called Tithi Praveśa is cast for the moment the birth tithi recurs, rather than the moment the Sun returns to its birth longitude.

None of that touches Vimśottarī. The Vimśottarī daśā sequence is built from the Moon’s nakṣatra, which is an absolute position rather than an elongation, so a chart can have an early tithi and a late nakṣatra without any contradiction. People conflate the two constantly. If you are trying to work out when a Chandra mahādaśā begins, tithi has nothing to say about it.

How to read a tithi line and not be misled by it

A pañcāṅga entry that says “Ṣaṣṭhī upto 14:37” is telling you the tithi in force at that day’s sunrise and the clock time it ends. It is not telling you that Ṣaṣṭhī lasted from midnight. It began at some point on the previous day. If you need the start, look at the previous day’s line, where the tithi before it ends.

Two habits will keep you out of trouble. Always ask which city the almanac was computed for, because the sunrise rule makes that decisive. And always ask which reckoning the specific vrata uses, because a moonrise-reckoned observance and a sunrise-reckoned one can legitimately fall on different dates in the same year.

Karma Jyotisha computes tithi from the actual Moon and Sun positions for the place and instant you give it, and it will tell you plainly when a tithi is kṣaya or adhika rather than smoothing it over. That is the whole point of doing the arithmetic rather than repeating a table.