On this page
Venus is never seen at midnight. It always stays within a fixed angle of the Sun, as Mercury does. Any model of the sky has to explain this, and for over a thousand years Indian astronomy did so with a rule that worked for positions but needed a separate, awkward rule for the planets' height above or below the Sun's path. Around 1500 the Kerala astronomer Nilakantha Somayaji fixed it. This post explains what he changed, and checks two of his numbers with Class 11 level maths.
Who was Nilakantha Somayaji?
The MacTutor biography from the University of St Andrews gives his birth as 14 June 1444 at Trikkantiyur, near Tirur in Kerala. He studied astronomy under Ravi and under Damodara, the son of the astronomer Paramesvara, and wrote about ten treatises on astronomy. His major work, the Tantrasangraha, has 432 Sanskrit verses in 8 chapters. MacTutor dates it to 1501. K. Ramasubramanian, M. D. Srinivas and M. S. Sriram, in a 1994 paper in Current Science, date it to 1500 on the word of his pupil Sankara Variar. The date of his death is uncertain.
He belonged to the Kerala school that also produced the infinite series covered in our post on Madhava of Sangamagrama.
What was wrong with the older model?
Indian astronomy from at least Aryabhata (499 CE) found a planet's position in two steps, the Current Science paper explains. A manda correction allowed for the eccentricity of the orbit, much like the modern equation of centre. A sighra correction then turned the result into the position seen from Earth.
For Mars, Jupiter and Saturn this worked well. For Mercury and Venus, the older texts took the mean Sun as the "mean planet" and applied the manda correction to the Sun instead of to the planet. Positions came out reasonably, but the texts needed one rule for the latitudes of the outer planets and an entirely different one for the inner planets. Astronomers from Bhaskara I onwards noticed the oddity; Bhaskara II quoted an earlier writer who said it was justified only because it matched observation.
What did Nilakantha change?
In the first chapter of the Tantrasangraha, the paper reports, Nilakantha gives the usual periods of 87.966 days and 224.702 days for Mercury and Venus, but says these are the planets' own revolutions. He then applies the manda correction to the planet itself. With that change, one rule for latitudes works for all five planets.
In later works he describes the geometry this implies. The authors translate a passage from his Aryabhatiya commentary: "The earth is not circumscribed by their orbits." In his model the five planets move in eccentric orbits around the mean Sun, which itself goes around the Earth. The paper notes that many of the ideas were already present in Paramesvara's work, and the journal's editor added that whether this amounts to a heliocentric system hinges on how the texts are read. The authors also argue that the correct rule for applying the equation of centre to an inner planet first appears in Europe only with Kepler.
Worked example 1: from orbital period to synodic period
Venus returns to the same position relative to the Sun, as seen from Earth, every synodic period S. In one day Venus gains 1/224.702 of a revolution and the Earth 1/365.256, so
1/S = 1/224.702 − 1/365.256 = 0.0044503 − 0.0027378 = 0.0017125,
which gives S = 583.9 days. The NASA Venus fact sheet lists 583.92 days. Nilakantha's period, read as Venus's own orbit, gives the right cycle of appearances.
Worked example 2: why Venus stays near the Sun
Indian texts specified a ratio r/R for the sighra correction. The Current Science paper tabulates the Aryabhatiya values: 0.712 to 0.737 for Venus and 0.361 to 0.387 for Mercury, against modern values of 0.723 and 0.387. If Venus moves on a circle of radius r around the Sun and Earth is at distance R, the largest angle between Venus and the Sun, seen from Earth, comes when the line of sight is tangent to Venus's orbit:
sin θ = r/R = 0.723, so θ = 46.3°.
For Mercury, sin θ = 0.387 gives θ = 22.8°. That is why Venus is an evening or morning star and never rises at midnight. Real orbits are slightly elliptical, so the actual greatest angle varies a little from one appearance to the next.
Which mistakes does this help you avoid?
Subtracting periods instead of rates. 365.256 − 224.702 is not the synodic period. Subtract the rates 1/T, then invert.
Mixing up sidereal and synodic periods. The 224.7-day period is relative to the stars; 583.9 days is relative to the Sun as seen from Earth.
Overstating the history. The model keeps the Earth at the centre of the Sun's motion. Call it a planets-around-the-Sun picture within a geocentric frame, as the authors describe it.
What can a student learn from Nilakantha?
He took a rule that gave acceptable answers and asked why it needed a special case. Then he checked that his new version reproduced the same positions while removing the special case. In physics problems the same habit pays: if your method needs a different trick for one case, look for the version that handles every case at once.
Orbits and relative motion are easier to grasp when you can watch them. The Fermi Simulation Lab has interactive simulations for ideas that are easier to see than read.
Try it now: use 87.966 days for Mercury to find its synodic period, and check your answer against the NASA Mercury fact sheet.
