A set of numbers written down around 499 CE gives the length of the sidereal day as 23 h 56 m 04.10 s. The modern value is 23 h 56 m 04.09 s. This chapter is about how such numbers were obtained, what they were for, and the calendar they still drive.
Indian astronomy is unusual in never having been a specialist pursuit. Its output — the pañcāṅga — was consulted daily by ordinary households, and still is.
1What it was forकालनिर्णयः
The tradition states its own purpose in one word: कालनिर्णय, kāla-nirṇaya — the determination of time.
That is a narrower brief than modern astronomy's and it explains almost everything about the shape of the subject. The questions are: when is the new moon? when does the sun enter this constellation? when is the solstice? when will the moon be eclipsed, and for how long? These are computational questions with numerical answers, and they must be answered in advance, because rites, agriculture and civic life are scheduled by them.
Not remote
Celestial bodies are treated as continuous with earthly life rather than as distant objects. Consulting the pañcāṅga is a daily domestic act, not a scientific one, and it is why the tradition survived the loss of its institutions.
Observation, then computation
The method is systematic: observe, record, codify, look for periodicity, build a model that predicts, and revise the model when prediction and observation part company. Several Indian texts explicitly instruct the astronomer to correct the parameters against observation.
The engine of the mathematics
Every mathematical advance in chapter 8 — trigonometry, indeterminate equations, series, the beginnings of calculus — was driven by a problem in this chapter. The two subjects are not neighbours; they are one subject.
2The celestial sphereगोलः
Two great circles at 23½° to each other, and everything follows. The two points where they cross are the equinoxes; the two points furthest apart are the solstices. The Sanskrit terms are krānti-vṛtta for the ecliptic and viṣuvat-vṛtta for the equator.
3Solstices and equinoxesअयनम् · विषुवत्
4The sky as a clockपुरातत्त्वज्योतिषम्
Because the earth's axis precesses with a period of about 25,800 years, the equinox drifts steadily backwards through the constellations — roughly one degree every seventy-two years. A text that records where the equinox stood therefore carries a date.
Dates follow Subhash Kak's compilation. Chapter 1's caution applies in full: the method is sound, and the disagreement among specialists is about whether these statements are observational or prescriptive — whether the text is recording where the equinox was, or stipulating where it should be reckoned. Both readings are held by serious scholars.
A second, quieter piece of evidence
The oldest lists of the nakṣatras — Taittirīya-saṃhitā 4.4.10, and parallels in the Atharvaveda — begin with Kṛttikā. Every modern list begins with Aśvinī, two positions earlier. Two nakṣatras is 26° 40′ of arc; at one degree per seventy-two years, that is about nineteen centuries of precession. The shift in where the list starts is itself a fossil of the sky moving, and it is visible without any calculation at all — simply by comparing two lists.
5The twenty-seven nakṣatrasनक्षत्राणि
The distinctively Indian division of the zodiac is not by twelve but by twenty-seven — because the unit being tracked is the moon, not the sun.
The moon's sidereal period is 27.32 days. Divide the ecliptic into twenty-seven equal arcs and the moon occupies almost exactly one of them per night. Each arc is 360°/27 = 13° 20′ — that is, 800 minutes of arc — and each is named after a conspicuous star within it.
Two independent partitions of the same circle, one lunar and one solar, maintained simultaneously. Almost every complication in the Indian calendar arises from having to reconcile them.
6Five kinds of yearपञ्च संवत्सराः
A year can be defined in at least five ways, depending on which body you follow. The Vedic corpus names all five and keeps them distinct — which is the mark of people who have noticed that they disagree.
The intercalation problem, and its Vedic solution
Twelve lunar months come to about 354 days; the solar year to about 365¼. The two drift apart by eleven days a year, and after three years the lunar calendar is a month behind the seasons. The Ṛgvedic authors were aware of this, and the solution is an intercalary month — an extra thirteenth month inserted periodically. The Yajurveda refers to the ekādaśa-rātra ceremony as the mechanism for reconciling the two counts, bringing the year to 365 days and leaving the residual quarter-day to be handled separately. A five-year cycle, the yuga of 60 solar / 61 civil / 62 lunar months, appears in the Taittirīya and Vājasaneyī saṃhitās.
7Months and fortnightsमासः · पक्षः
The twelve lunar months are Caitra, Vaiśākha, Jyeṣṭha, Āṣāḍha, Śrāvaṇa, Bhādrapada, Āśvayuja, Kārtika, Mārgaśīrṣa, Puṣya, Māgha, Phālguna — and each is named for the nakṣatra the moon occupies at the full moon of that month. In most Caitra months the full moon stands near Citrā (Spica); in most Vaiśākha months near Viśākhā. The naming scheme is a mnemonic for a fact about the sky.
8Tithiतिथिः
The tithi is the unit that catches most newcomers out, because it is not a day. It is an angle.
Thirty tithis make a lunar month: fifteen in the bright fortnight, fifteen in the dark. Because the definition is angular and the moon's motion is not uniform, the calendar has irregularities built into it that no amount of tidying can remove — they are properties of the moon.
9The siddhāntic traditionसिद्धान्तज्योतिषम्
Work or author
Period
Contribution
Sūrya-siddhānta
before the 6th c. CE
Author unknown, several recensions. An early version is summarised by Varāhamihira; a later one is still used by traditional calendar makers today
Varāhamihira — Pañca-siddhāntikā
6th c. CE
A critical summary of five earlier astronomical systems — and therefore our main evidence for what those systems said
Āryabhaṭa — Āryabhaṭīya
b. 476 CE, Kusumapura
The rotation of the earth; accurate planetary parameters; the sine function and its table; the physical explanation of eclipses
Bhāskara I — Āryabhaṭīya-bhāṣya, Mahā-bhāskarīya
7th c. CE
The indispensable commentary; develops the Āryabhaṭan system in his own texts
A complete computational system; a practical handbook; and the algebra of chapter 8
Lalla — Śiṣyadhīvṛddhida-tantra
8th–9th c. CE
A textbook of the Āryabhaṭan system with new algorithms
Mañjulācārya — Laghu-mānasa
10th c. CE
An explicit second correction to the moon's longitude; the derivative of the sine function and instantaneous velocity of sun and moon
Śrīpati — Siddhānta-śekhara
11th c. CE
Widely quoted by later astronomers
Bhāskarācārya II — Siddhānta-śiromaṇi, Karaṇa-kutūhala
b. 1114 CE
The standard algorithms corrected and generalised; a table-based computation manual; a chapter on instruments
Work or author
Period
Contribution
Mādhava — Veṇvāroha, Sphuṭa-candrāpti
1340–1425 CE
Infinite series for π, sine and cosine; a method of computing the moon's position at short intervals
Parameśvara — Dṛggaṇita
1360–1455 CE
Fifty-five years of recorded eclipse observations, used to correct the parameters. An observational programme, stated as such
Nīlakaṇṭha Somayājī — Tantra-saṅgraha
1444–1550 CE
A revised planetary model in which the inner planets orbit the sun; systematic spherical trigonometry; improved eclipse theory
Jyeṣṭhadeva — Gaṇita-yukti-bhāṣā
16th c. CE
Full demonstrations of the school's results, in Malayalam prose
Gaṇeśa Daivajña — Graha-lāghava
b. 1507 CE
Simplified planetary computation without trigonometric tables — still used for almanac work
Kamalākara — Siddhānta-tattva-viveka
b. 1616 CE
Indian parameters with elements of Ptolemy's system incorporated
Rājā Sawai Jai Singh — Zij Muhammad Shahi
1688–1743 CE
Five masonry observatories; a new set of tables
Candraśekhara Sāmanta — Siddhānta-darpaṇa
b. 1835 CE
Revised lunar theory from his own naked-eye observations; simple instruments of his own design; reform of the Odia calendar
The last of the line
Candraśekhara Sāmanta is worth pausing on. Working in rural Odisha in the second half of the nineteenth century, with instruments he built from bamboo, he detected and corrected errors in the lunar theory by observation — including the moon's second inequality — apparently without access to European astronomy. He is the demonstration that this was still a living investigative tradition when it ended, not a fossil.
10The Āryabhaṭīyaआर्यभटीयम्
One hundred and twenty-one verses, in four sections, written when the author was twenty-three. It is the founding text of Indian mathematical astronomy.
The earth turns
Āryabhaṭa holds that the apparent daily rotation of the heavens is caused by the earth's own rotation, and gives the image that makes it plausible: a person in a moving boat sees the fixed objects on the bank moving backwards. The Sanskrit is anulomagatir nausthaḥ…, Gola-pāda 9.
Later Indian astronomers, including Brahmagupta, rejected this. It is an instructive case: the tradition contained the idea, argued about it, and did not converge.
Eclipses have physical causes
Gola-pāda states that the moon is eclipsed by entering the earth's shadow, and the sun by the moon coming between. The geometry — the shadow cone of chapter 8's gnomon figure — is worked out, and the durations computed.
This is stated in a culture where the received explanation was the demon Rāhu. Āryabhaṭa's successors keep the name Rāhu as a label for the lunar node — the point where the orbits cross — which is exactly where an eclipse can occur. The mythological term is retained and given a geometric meaning.
11The parameters, checkedभगणाः
The Gītikā-pāda gives, for each body, the number of complete revolutions it makes in a mahā-yuga of 4,320,000 years. Dividing the number of civil days in the yuga by that figure gives the body's period. Here is what happens when you do the division.
The two numbers everything depends on
Civil days in a mahā-yuga (yuga-sāvana-dina): 1,577,917,500. Sidereal rotations of the earth in the same period: 1,582,237,500. The difference between them is exactly 4,320,000 — one extra rotation per year, which is precisely what the geometry requires and a good internal check on the system.
Derived by dividing 1,577,917,500 civil days by each revolution count. For Mercury and Venus the figure given is the śīghrocca — in effect the heliocentric period — which is why those two rows come out as the planets' true orbital periods rather than their apparent ones.
Why a mahā-yuga at all?
Stating periods as “so many revolutions in 4,320,000 years” looks like cosmological ornament. It is a computational device. Working in integers avoids fractions entirely: to find where a planet is on a given day, you multiply the elapsed days by the revolution count and divide by the yuga's days, all in whole numbers. The very large yuga also lets the periods be chosen so that everything returns to a common starting configuration — which gives the system a defined epoch. Āryabhaṭa's own epoch is the start of the current Kali-yuga: Friday, 18 February 3102 BCE.
Note also that Āryabhaṭa divides the mahā-yuga into four equal quarters of 1,080,000 years, rather than the 4:3:2:1 proportion of the Purāṇas. On this, as on the earth's rotation, he differs from the tradition around him.
12Nīlakaṇṭha's revisionतन्त्रसंग्रहः
Around 1500 CE, a Kerala astronomer noticed that the standard procedure for Mercury and Venus did not work properly, diagnosed why, and changed the model.
The interest here is methodological. A discrepancy between prediction and observation was traced to a specific defect in the model, and the model was changed to fix it — with the change published, argued for, and adopted by the school. That is a scientific tradition operating normally.
13The five limbsपञ्चाङ्गम्
Pañcāṅga means “five limbs”. The Indian almanac states five quantities for each day, and every one of them is computed from the true longitudes of the sun and the moon.
Notice the economy. Two observed quantities — where the sun is, where the moon is — and five derived ones, obtained by three subtractions, one addition and four divisions. Everything in the almanac follows from the planetary theory of §11.
14Computing a pañcāṅgaगणितम्
Two worked instants, each carried through all four longitude-derived limbs.
The fractional parts matter too: the 759⁄800 in the last row says the current yoga is 95% elapsed and will end shortly. An almanac gives the clock time at which each limb changes, which is what makes it usable.
Longitudes are measured on a circle, so a difference can come out negative — add 360° and continue. A sum can exceed 360° — subtract it. Both appear in this example, and forgetting either is the commonest error in doing these calculations by hand.
15Ahargaṇaअहर्गणः
The fifth limb needs something none of the others do: a continuous count of days that does not care about months or years.
Āryabhaṭa introduces exactly that, and it appears to be the first such scheme anywhere. अहर्गण, ahargaṇa, means simply “heap of days” — the number of days elapsed since the epoch. Because calendars differ between regions and eras, and because month lengths vary, a running day-count is the only stable way to identify a date and the only convenient way to compute a weekday.
Compare the Julian Day Number, introduced by Joseph Scaliger in 1583 and used by astronomers today for exactly the same reason. The ahargaṇa is the same idea, with a different epoch, roughly a thousand years earlier.
16Finding north with a stickदिङ्निर्णयः
Before any observation can be made, the observer needs a true meridian. The construction requires a vertical rod, a piece of string and one clear day.
No compass, no clock, no prior knowledge of the site. The symmetry argument in step 4 is the whole content: two shadows of equal length must lie at equal angles either side of the north–south line, whatever the latitude and whatever the date.
17The instrument catalogueयन्त्राणि
Bhāskara II devotes a chapter of the Siddhānta-śiromaṇi to instruments. The list is a working laboratory.
Gola-yantra गोलयन्त्र
The armillary sphere: nested rings representing the fixed and movable great circles of the celestial sphere. Does the work of an astrolabe, and doubles as the teaching instrument for spherical astronomy.
Cakra-yantra चक्रयन्त्र
A graduated wheel of wood or metal with a needle at the centre. Illuminated by the sun, the needle's shadow gives the sun's altitude; used more generally for the longitudes and latitudes of planets. Cāpa-yantra is half of it, Turīya-yantra a quadrant.
Nāḍīvalaya नाडीवलय
A disc set in the plane of the equator, divided into 60 ghaṭikās and also into the twelve signs with arcs proportional to their rising times at that latitude. Read directly, it gives the time elapsed since sunrise and the rising sign.
Ghaṭī-yantra घटीयन्त्र
The sinking-bowl water clock of chapter 6 — a pierced copper bowl floated on water, which fills and sinks in one unit of time.
Śaṅku शङ्कु
The gnomon, made of ivory or metal. The simplest instrument and the most used: direction, latitude, time of day and the sun's declination all follow from its shadow.
Phalaka-yantra फलकयन्त्र
A board with a circle of 30 aṅgulas radius, graduated in ghaṭikās and degrees, from which the zenith distance can be read directly. Dhī-yantra, a stick with a plumb line, gives heights and distances.
18Jantar Mantarयन्त्रमन्दिरम्
Between 1724 and 1735, Sawai Jai Singh II built five masonry observatories — at Delhi, Jaipur, Ujjain, Mathura and Varanasi — on a diagnosis that is worth understanding, because it is an engineering argument rather than an astronomical one.
Jai Singh's programme also included commissioning Sanskrit translations of Euclid, Ptolemy and Arabic and Latin astronomical works, and sending an embassy to Portugal for European tables. It is the last great synthetic effort of the tradition before the colonial reorganisation of Indian education.
19Self-checkपरीक्षा
Check your reading
1 · Why is the zodiac divided into twenty-seven parts as well as twelve?
Two grids over one circle, one lunar and one solar. Each rāśi contains exactly 2¼ nakṣatras.
2 · A tithi is —
Because the definition is angular and the moon's speed varies, a tithi runs from about 21.5 to 26 hours — which is why tithis are sometimes repeated and sometimes skipped.
3 · What does the pattern of errors in Āryabhaṭa's planetary periods indicate?
Saturn is out by nearly a week; the sidereal day by a hundredth of a second. Theory-derived numbers would not have that shape of error.
4 · What did Nīlakaṇṭha's revised model change, and what did it not?
A geo-heliocentric system, and Nīlakaṇṭha does not claim more than that. The ellipses are Kepler's, a century later.
5 · Why is the ahargaṇa necessary?
The same reasoning behind the Julian Day Number used by astronomers today — introduced about a thousand years later.
Questions worth arguing about
Why did Āryabhaṭa's rotating earth not prevail?
Brahmagupta and others rejected it on physical grounds that were reasonable at the time: if the earth turned, things not attached to it would be left behind, and there was no theory of inertia to answer that objection. Note also that rotation makes no difference to the computations — the predicted positions are identical either way — so the question could be set aside as idle. A tradition organised around predictive accuracy has no lever to settle a question that does not affect predictions.
Is the mahā-yuga framework a scientific device or a cosmological doctrine?
Both at once, and separating them is anachronistic. Computationally it is a device for integer arithmetic with a defined epoch; cosmologically it is a claim about cycles of world-ages. Āryabhaṭa's willingness to alter the doctrine — equal quarters instead of 4:3:2:1 — where computation required it suggests he treated the computational role as primary.
What was lost when the pañcāṅga tradition became purely almanac-making?
The calculations survived, and are still performed. What thinned out was the observational side that corrects the parameters — Parameśvara's fifty-five years of eclipse records, Sāmanta's bamboo instruments. A computational tradition without an observational one drifts, slowly and invisibly, and several modern pañcāṅgas still use parameters that observation would have corrected long ago.
20Glossaryशब्दकोशः
IAST
Devanāgarī
Sense
ahargaṇa
अहर्गण
“Heap of days”: the continuous count of days since the epoch.
amāvāsyā / pūrṇimā
अमावास्या / पूर्णिमा
New moon and full moon.
ayana
अयन
A solar half-year: uttarāyaṇa (northern) or dakṣiṇāyana (southern).
ghaṭikā
घटिका
A time unit; sixty make a day and night.
gola
गोल
The sphere; spherical astronomy.
kāla-nirṇaya
कालनिर्णय
The determination of time — the tradition's statement of its own purpose.
karaṇa
करण
Half a tithi; also, a genre of practical computation manual.
krānti-vṛtta
क्रान्तिवृत्त
The ecliptic.
mahā-yuga
महायुग
4,320,000 years; the period over which revolution numbers are stated.
nakṣatra
नक्षत्र
One of twenty-seven arcs of 13° 20′; also the star naming it.
pakṣa
पक्ष
A fortnight: śukla (bright, waxing) or kṛṣṇa (dark, waning).
pañcāṅga
पञ्चाङ्ग
“Five limbs”: the almanac, and the five quantities it states.
rāśi
राशि
One of twelve 30° signs of the zodiac.
śaṅku
शङ्कु
The gnomon.
śīghrocca
शीघ्रोच्च
The “fast apex” — a parameter which for the inner planets amounts to the heliocentric period.
siddhānta
सिद्धान्त
A complete astronomical system; the genre of the major treatises.
tithi
तिथि
A 12° increment of the moon–sun elongation.
viṣuvat
विषुवत्
The equinox.
yantra
यन्त्र
An instrument.
yoga
योग
A pañcāṅga limb: the sum of the two longitudes in units of 13° 20′.