The solstice has moved: Varāhamihira reads precession out of the archives (505 CE)

# The solstice has moved: Varāhamihira reads precession out of the archives (505 CE)

Every June, the Sun reaches its northernmost point and turns back south. Ancient astronomers marked *where* among the stars that turning happened — and because their records were kept for centuries, they left later readers a trap and a treasure in one: the turning point doesn't stay put.

Varāhamihira, writing around 505 CE, walks straight into the discrepancy and reports it flat:

> "When the return of the sun towards the south took place from the > middle of Āśleṣā, then the ayana was right; at the present time > the ayana begins from Punarvasu." > > — *Pañcasiddhāntikā* III.21, trans. Thibaut & Dvivedi (1889); > first clause repaired from scan artifacts

Translated out of the nakshatra idiom: the old texts put the summer solstice (*ayana*, the turning) in the middle of the lunar mansion Āśleṣā, and they were correct *when they were written* — "then the ayana was right." But look up now, says Varāhamihira, and the turning has slid back to Punarvasu, roughly 23 degrees earlier along the zodiac. The archive and the sky disagree, and both are telling the truth — about different centuries.

What causes it

The Earth's spin axis wobbles like a slowing top, tracing a cone once every ~25,800 years. As it wobbles, the equinoxes and solstices drift slowly backward through the constellations — about 50.3 arcseconds per year, one degree per ~72 years. This is the precession of the equinoxes, and it is far too slow for any observer to see in a lifetime. There is exactly one way an ancient astronomer could catch it: compare current observation against records old enough for the drift to accumulate.

That is what the verse does. The old solstice position — mid-Āśleṣā — is the one recorded in the Vedāṅga Jyotiṣa calendar tradition, conventionally dated ~1400–1200 BCE. Thibaut's commentary computes the gap to Varāhamihira's reported position at about 23°20'. At the true precession rate, 23°20' takes roughly 1,670 years to accumulate — which, counted back from 505 CE, lands squarely in the Vedāṅga Jyotiṣa's era. The two data points in this one verse are mutually consistent with the real celestial mechanism to within the precision the nakshatra system allows. The archive was right; the sky moved.

The honest comparison

Priority belongs to Greece, and the claim says so plainly: Hipparchus of Rhodes discovered precession around 130 BCE, by exactly the same archival method — comparing his own star positions against Greek observations ~150 years older and Babylonian records older still. He even estimated the rate. Ptolemy consolidated the result three centuries later. Varāhamihira writes six centuries after Hipparchus, and Greek astronomical material was demonstrably circulating in India by his day — [his own treatise reviews two Greco-Roman-derived schools](/c/558dc8b7-6602-5f65-b034-af6d6da3857c) — so independence cannot be assumed for the *concept*.

What the verse documents is nonetheless distinct and valuable: the *Indian archive itself* registering the drift. The solstice positions being compared are Indian records, in Indian coordinates (nakshatras), spanning an Indian textual tradition — the gap between the Vedāṅga Jyotiṣa's sky and Varāhamihira's is visible entirely within Sanskrit sources. And note what Varāhamihira does *not* do: he does not emend the old texts, accuse them of error, or paper over the conflict. "Then the ayana was right" — the ancients observed correctly; the sky has since moved. That sentence quietly contains the whole method of historical astronomy: treat old observations as data, not as scripture.

What the verse also does not contain, honesty requires adding, is a theory. No rate, no mechanism, no prediction. Whether Varāhamihira understood the shift as continuous motion or as a one-time displacement is not stated. Later Indian astronomy adopted the oscillating "trepidation" model — [the Sūrya-Siddhānta's libration of the equinoxes](/c/0874239f-e319-5272-8457-c2ab61b17876) — which handles the drift arithmetically but gets the long-term physics wrong (as, for what it is worth, did Ptolemy's constant but underestimated rate). The clean modern understanding took Newton.

Legacy

The verse became load-bearing twice over. Within the tradition, solstice statements like this one — and the Āśleṣā/Dhaniṣṭhā records behind it — anchored the *ayanāṁśa*, the accumulated precession correction that Indian calendars and astronomical tables still apply today: every Indian festival date computed in the sidereal calendar carries a descendant of this correction inside it.

For modern scholarship, the verse (Thibaut calls it "the famous, often quoted, statement") became a dating instrument: because precession is a clock, a text that records where the solstice fell can be assigned an epoch. The Vedāṅga Jyotiṣa's conventional date is itself derived largely from its solstice statement — the same arithmetic run in reverse.

There is something fitting in that. A verse about the archive disagreeing with the sky became, for historians, a way to date the archive *by* the sky. Varāhamihira, who [graded his sources by accuracy rather than ancestry](/c/558dc8b7-6602-5f65-b034-af6d6da3857c), would presumably approve.

---

Sources

- [Thibaut & Dvivedi, *The Panchasiddhantika: The Astronomical Work of Varaha Mihira*, 1889](https://archive.org/details/wg1078) — III.21 cited; Thibaut's introduction (the "famous, often quoted" remark) and commentary (the 23°20' computation). - Plofker, *Mathematics in India*, 2009 — precession and ayanāṁśa in the Indian tradition (secondary synthesis, for context). - Evans, *The History and Practice of Ancient Astronomy*, 1998 — Hipparchus's discovery of precession (referenced for the Greek comparison only).

Related claims

- [The Roman Siddhānta: Greek astronomy inside the Indian canon (505 CE)](/c/558dc8b7-6602-5f65-b034-af6d6da3857c) - [The Sūrya-Siddhānta's trepidation model of precession](/c/0874239f-e319-5272-8457-c2ab61b17876) - [Aryabhata's ecliptic obliquity (499 CE)](/c/4c1dad4d-d658-5b7e-8653-34593b458382)