Three Watches of the Night

Vigilance (jāgariyānuyoga)

"When, Aggivessana, the ariyan disciple is moderate in eating, the Tathagata disciplines him further, saying: "Come you, monk, abide intent on vigilance; during the day while pacing up and down, while sitting down, cleanse the mind of obstructive mental states; during the middle watch of the night, lie down on the right side in the lion posture, foot resting on foot, mindful, clearly conscious, reflecting on the thought of getting up again; during the last watch of the night, when you have arisen, while pacing up and down, while sitting down, cleanse the mind of obstructive mental states."

I wonder, does anyone here know anything about timekeeping in ancient India? If so, I have a couple of questions about the three yāmas mentioned above. They arose from a recent thread on Dhamma Wheel.

Firstly, was it a general Indian practice to divide the night this way or was it peculiar to Buddhism?

Secondly, with regard to the length of the yāmas it seems there are five possibilities:

  1. All three are of equal length and of the same length throughout the year; like the four vigilias of the Roman night.
  2. All three are of equal length but this length is subject to seasonal variation, like the Halachic hour of the Jews.
  3. The three are of unequal length and the length of each is subject to seasonal variation.
  4. The three are of unequal length but each retains the same length throughout the year.
  5. The three are of unequal length, with one/some retaining the same length throughout the year and the other(s) subject to seasonal variation.

At the moment I’m inclining to #5, based on the Sammohavinodanī (Vibh-a. 345), which states that the middle watch of the night (a bhikkhu’s sleeping time) amounts to one sixth of the day and night (rattindivassa chaṭṭhakoṭṭhāsasaṅkhāta; i.e. four hours) but doesn’t specify any fixed duration for the first and last watches. It seems the absence of any such specification is most reasonably attributed to the first and last watches being longer in the cold season, owing to the longer hours of darkness, and shorter in the hot season. But is the Sammohavinodanī’s definition supported by any more ancient source?


I’ll wager that monastics primarily used an incense clock, or oil lamps & candles (…almsbowls as water clocks?). But otherwise the only information I can find involves Vedic time cycles, from Brahma’s lifespan to the time of one eyeblink.

Wiki suggests that a Yāma = 1/4 of a day (light) or night, so eight Yāmas make half of the day (either day or night).

Here’s something about praharas (“The day is divided into eight parts: four praharas for the day, and four for the night”), in a footnote:

Some scholars correctly infer that in seasons (and regions) where days and nights are unequal in length, the praharas expand and contract in length. See Duncan Forbes’ early comment on the pahar (= prahar) in Northern India: “The first pahar of the day began at sunrise, and of the night at sunset; and since the time from sunrise to noon made exactly two pahars, it follows that in the north of India the pahar must have varied from three and a-half hours about the summer solstice, to two and a-half in winter, the pahars of the night varying inversely.” (Duncan Forbes, LL.D, transl. Bāgh O Bahār; or Tales of the Four Derwishes, by Mīr Amman of Dihli. London: W. H. Allen & Co. 1882, (p. 23, note 1). What we are encountering here is a difference between modern cultures that rely on clock time, and traditional cultures where the length of a day is observed in the sky, from sunset to sundown.


It appears to be a general practice among indian ascetics:

That Buddhist and Brahminic meditators exchanged ideas and practices in early times is not in doubt. Similar verses on meditation, found in both the early Brahminic and Buddhist literature, seem to prove this. We can compare, for example, the following verses:

Mbh XII.180.28:

The wise man, constantly disciplining himself in the earlier and
later parts of the night, taking little food, being pure, sees the self in the


The teacher has conquered this path which transcends attachment and
the fear of birth and old-age. Being diligent, discipline [yourself] in the
earlier and later parts of the night, make [your] practice firm. The Origin of Buddhist Meditation. Ch 3. Pg. 24

The Pali and Sanskrit didn’t copy and paste well so I had to delete it, but you can see it in the book in the link. Although I question whether the website I got it from, A Handful of Leaves, has permission to use the works on there.


Wikipedia, citing an article by Hartmut Scharfe, claims that at Nalanda University,

“… four hours a day and four hours at night were measured by a water clock, which consisted of a copper bowl holding two large floats in a larger bowl filled with water. The bowl was filled with water from a small hole at its bottom; it sank when completely filled and was marked by the beating of a drum at daytime. The amount of water added varied with the seasons and this clock was operated by the students of the university.”

No guarantee the Indian use of water-clock methods that were used in the Buddha’s time. But apparently, the basic water-clock method goes all the way back to Mohenjo-Daro.

See also:



My object in this paper is to bring out the point that the more ancient form of the Indian water-clock was a more simple arrangement, and one which worked in the opposite way, namely, by emptying itself in the course of a nadika. As may be seen hereafter, some of the passages of the later period mention this form also, along with the other.

So, maybe an almsbowl with a properly-sized hole in it could work.


btw: the quotation from Alexander Wynne’s book is on page 28 in the print version.

Many on the books / articles there are probably in the public domain, but some are questionable e.g.:
Analayo’s first Satipatthana book
Wynne’s book
Noa Ronkin’s Early Buddhist Metaphysics

The site is, however, an immense treasure, IMO, making available a sizable chunk of significant scholarship – old and new. For instance:
The Buddhist Philosophy of Thought” by Alexander Piatigorsky
A book out of print, available now used from $40, but when I got a copy the least expensive was ca. $120. An odd-ball writer, hardly noticed these days, but using a phenomenological approach and coming up with some notable insights. For instance the observation that, in reviewing virtually all the Western scholarship available then (prior to 1982), writers invariably attempt to shoehorn an understanding of Buddhist thought into their own favorite Western concepts, ignoring (probably mostly unaware) the fundamental notion in the Buddha’s teachings (particularly highlighted in abhidhamma analysis) that all such conceptual fabrications are to be deconstructed (phenomenologically “bracketed”) to be able to properly understand and transcend the machinations of the mind (and find release therefrom).

Researching (mainly Wikipedia) time and ways of measuring it, there appears to be no hard evidence that Indic culture of the Buddha’s time used any specific mechanisms to measure time to the hour or so. The Egyptians, already for some 3000 years by then, could measure time in daylight with sundial technology. The Greeks came up with a mechanical “clepsydra” (water clock – klepto-hydros “water stealing”) ca. 4th century BCE, and something related is mentioned already in the time of Plato (born 50 years or so after the Buddha). Water clocks are known to have been used in Babylon and Egypt more than 1000 years prior to the Buddha’s time, and documented in China in the 6th century BCE. On the other hand, earliest documentation of candle, incense or other devices doesn’t seem to appear prior to the early to mid 1stmillennium CE. The sand-clock (hour-glass) doesn’t appear documented prior to 150 BCE.

Using the stars in navigation (in terms of direction) goes back to at least the Minoan civilization (2600-1100 BCE), though I haven’t found mention of time-measurement using the stars in navigation prior to the age of European explorers (mid 2nd millennium CE). India had astronomy at least since ca. 1500 BCE, but I found no mention of using it for time-keeping (to the hour or minute) prior to mid 1st millennium CE.

Given that time measurement (also independent of sunlight) was known in multiple cultures at least a millennium before the Buddha’s time, and the fact that significant developments tended to get spread around as humans are always wandering among each-others’ cultures, are inherently curious, I suspect that some methods of relatively detailed time measurement (other in daytime) were know in 5th-century BCE India. If perhaps not from elsewhere, human ingenuity would very likely have come up with something.

On another note, there’s the phenomenon that one of the five “masterys” of jhana practice is predetermination of duration of absorption. (Yes, detailed by Buddhaghosa, but there’s little reason to believe that the Buddha and his star students knew had less mastery.) And the curious quality that absorption shares with sleep, as a qualitative change of conscious state – though the latter eclipses consciousness and the former enhances it. It’s well known today that one can exactly program sleep duration (without external alarm mechanisms); again, little reason to believe that adepts at mind training in the Buddha’s day weren’t also able to do that.

There must have been general awareness at the time (given astronomy) of the regular ebb and flow of daylight with the seasons, and some way of marking “business hours”, meal times, etc. in organized settings like trading centers, large (aristocratic) family organizations. People must have had a pretty good sense of “what time it is” in their daily routines, which, by force of habit and internal clocks, extended to some sense of measuring time at night. There could well have been some official watching and notification of time periods,with bells or such. Like even today in any city, town, even village, in Europe, where one often doesn’t need a watch or clock at hand as every 15 minutes the church bells keep one informed, with special protocols in the morning, noon, and evening. In groups of people,there were likely individuals who had a keener sense of time passage who helped orient the group to periodic demarcation.

Pardon the on-running papanca here, but I believe we’re well-advised not to fall into the modern mental trap of considering more “primitive” cultural settings (not having our gadgets, scientific prowess, etc.) to have been “lost in the dark”, so to speak.


I also think #5 is the most likely. I was reading this passage yesterday:
AN 6.20 marana-sati (death mindfulness)

:diamonds: “idha, bhikkhave, bhikkhu
"here, monks, (a) monk:
divase nikkhante
(as) day departs
rattiyā patihitāya
(and) night returns,
iti paṭisañcikkhati —
thus (he) reflects -
‘bahukā kho me
"many indeed (for) me,
paccayā maraṇassa —
(are) [potential] causes (of my) death -
ahi vā maṃ ḍaṃseyya,
(a) snake ** I might-be-bitten (by),
vicchiko vā maṃ ḍaṃseyya,
{or a} scorpion ** I might-be-stung (by),
sata-padī vā maṃ ḍaṃseyya;
{or a} hundred-footed [centipede] ** I might-be-bitten (by),
tena me assa kāla-kiriyā,
because-of-that my **** time-of-death [would arrive],
so mamassa antarāyo.
that (for) me, (would be an) obstacle.

and the marana sati practice is repeated for the night to day shift:

(repeat: death by animals for night to day shift)
:diamonds: “idha pana, bhikkhave, bhikkhu
"here furthermore, monks, (a) monk:
rattiyā nikkhantāya
(as) night departs
divase patihite
(and) day returns,
iti paṭisañcikkhati —
thus (he) reflects -

I think it’s safe to assume that sunset and sunrise are going to make the first watch and last watch vary in length by season.

And since marana sati is something practiced 24/7, it must correspond to the instructions on wakefulness for the 3 yāmas in AN 3.16

proper way to sleep

Kathañ-ca, bhikkhave, bhikkhu
jāgariyaṃ anuyutto hoti?
{devoted-to} wakefulness ****?
Idha, bhikkhave, bhikkhu
Here, monks, a-monk

(sunrise to sunset, 6am to 6pm)

[during] day [time],
caṅkamena nisajjāya
walking-[meditation]-forward-and-backward (and) sitting [meditation],
āvaraṇīyehi dhammehi cittaṃ parisodheti,
obstructive qualities (of) mind (he) purifies,

(first watch of night, 6pm to 10pm)

rattiyā paṭhamaṃ yāmaṃ
At-night-time, first watch-of-the-night,
caṅkamena nisajjāya
walking-[meditation]-forward-and-backward (and) sitting [meditation],
āvaraṇīyehi dhammehi cittaṃ parisodheti,
obstructive qualities (of) mind (he) purifies,

(middle watch of night, 10pm to 2am)

rattiyā majjhimaṃ yāmaṃ
At-night-time, middle watch-of-the-night,
dakkhiṇena passena sīhaseyyaṃ kappeti
(on his) right side, (the) lion-lying-down [posture] *******,
pāde pādaṃ accādhāya
(one) foot (on the other) foot overlapping,
sato sampajāno
mindful (and) clearly-comprehending,
uṭṭhāna-saññaṃ manasi karitvā,
rising-[from sleep]-perception (his) mind sets,

(last watch of night, 2am to 6am)

rattiyā pacchimaṃ yāmaṃ
At-night-time, last watch-of-the-night,
After-rising [from lying down],
caṅkamena nisajjāya
walking-[meditation]-forward-and-backward (and) sitting [meditation],
āvaraṇīyehi dhammehi cittaṃ parisodheti,
obstructive qualities (of) mind (he) purifies,
Evaṃ kho, bhikkhave, bhikkhu
thus indeed, monks, a-monk
jāgariyaṃ anuyutto hoti.
{is} {devoted to} wakefulness.
Imehi kho, bhikkhave,
These indeed, monks,
tīhi dhammehi samannāgato bhikkhu
three qualities possessed (by a ) monk,
Apaṇṇaka-paṭipadaṃ paṭipanno hoti,
(the) faultless-path practiced ****,
yoni cassa āraddhā hoti
{and} the basis aroused for
āsavānaṃ khayāyā”ti.
Asinine-inclinations’ destruction.
(end of sutta)

The fixed timing for all 3 watches I assigned in the comments come from Ven. Thanissaro’s translation footnotes.

Thanks for the interesting topic Ven. Does the Vibhanga say anything about when then fixed 4 hour middle watch occurs? Thanissaro had 10pm through 2am, and I know Ajahn Mun and his various famous disciples followed 11pm through 3 am in practice.

I think 11pm through 3am accord nicely with chinese medicine recommendations for essential time one should be lying down to sleep, and western science has studies about biorythms and sleep related hormones that are produced only in this “middle watch of the night” period.


Thanks for this. It does seem relevant; as does the Udāna’s Yasojasutta, Ud3.3, which seems to suggest that it was the appearance of dawn that signalled the end of the third watch, and not any other timekeeping method.

No, the Vibhaṅga itself just repeats the sutta description of ‘devotion to wakefulness’ that I quoted in my opening post. Then its commentary says:


Thanks for raising this interesting question Bhante!

I had wondered if it might be a bit like the Islamic system. Islam has some culturally unique things going on that are sometimes different (like the day starting at nightfall), and sometimes surprisingly the same (like head shaving, and wearing robes in ihrahm) but it was on my mind.

May a more knowledgeable person please correct me.

There are 3 main fixed points in the Islamic night: sundown, Islamic midnight and dawn. From memory, everything else works around those points. Maghrib prayer time begins approximately at sundown, e.g. 7pm. (=the first watch of the night). Isha is a little after that, e.g. at 8.30pm-Islamic midnight (could be prayed at 9.30pm-10pm or so before sleep). The last watch of the night possibly matches tahujjud time, which is between 2am-4am, prior to fajr at about 5.30am (seasonal).

Unfortunately, I don’t know the method of calculating when the first time for isha starts, but it’s considered Islamic good practice (sunnah) to delay it later, and in some timezones, working Muslims have to seasonally combine it with fajr because of fatigue.

ANYWAY the Islamic way of calculation seems very sensible to me. I always thought the concept of living on four hours’ sleep was a bit extreme. But if what was meant was more like the Islamic system, observing the watches of the night could mean a range of times, just like there are a range of times for tahujjud prayer approx 2am-4am. Muslims wouldn’t necessarily observe the earliest time.

E.g. a Buddhist could meditate from 7pm-9.30pm, get six or hours sleep, wake at 3.30am, meditate until 5.30am, and still be observing meditation in the first and last watches of the night. It doesn’t have to be 10pm-2am sleeping.

Which is basically what many people do anyway- even though there might be a proper way of calculating the earliest time in the original Buddhist system, what people actually do is a bit variable and idiosyncratic. If people want to push the last watch later or first watch earlier, that way it is their choice.