There was once a prissy British civil servant who, when he came across a passage in a memo that displeased him, wrote “round objects” in the margin as a synonym for something ruder. This arch circumlocution was lost on the bluff minister he served, who fired back a query as to who this Round fellow was, and why he objected so much.

We can expect there to be plenty of members of the International Astronomical Union (IAU) who, reading the proposed new definition of a planet offered to them by their executive committee, will want to scrawl something equally rude and rather blunter in the margin — and will want to make their objections heard, possibly quite vociferously, at their general assembly in Prague this week (see page 724).

We understand and, to some extent, sympathize. But we would suggest that, instead, they acquiesce in the new definition, which will have the effect of increasing the number of planets in the Solar System to 12, and open the doors to more. They should do this for two reasons: it is not a bad definition; and it will at least stop the rumbling debate over the status of Pluto.

In the 1990s, it became clear that Pluto, the most newly discovered planet, was the most conspicuous of a crowd of icy 'trans-neptunian objects' (TNOs), some of which might well be larger. There was an obvious historical parallel to this situation with asteroids in the nineteenth century. When it was found that there were dozens of asteroids, Ceres, the largest and first discovered, was demoted from its position as a proper planet; it is now a 'minor planet' along with all the other asteroids. Pluto, it was argued by analogy, should be a minor planet with the rest of the TNOs on similar grounds.

This proposal sparked a degree of public debate that irritated many astronomers, who felt that the question of whether a particular body gets called a planet or not is of no scientific interest whatsoever. Still, the IAU decided that it should try and resolve the matter: planets loom large in the public imagination, and it seemed only reasonable for astronomers to be able to say whether a new discovery (or for that matter an old friend) was a planet or not.

The IAU's proposal is that the term 'planet' should apply to an object that has a sufficiently strong gravitational field to have pulled itself into a spherical shape, that is in orbit around a star, but that is not a star itself. This lets in Pluto and 2003 UB313, a TNO that is a touch bigger and not yet equipped with an IAU-approved name. It also readmits Ceres. And in the most peculiar aspect of the whole business, Charon, previously considered to be a moon of Pluto, will become a planet in its own right. Moons, however spherical, will remain satellites, not planets, in the IAU's eyes. But because the centre of mass of the Pluto–Charon system lies outside the body of Pluto, Charon, although tiny compared with, say, Neptune's moon Triton, qualifies as a planet.

All this will doubtless lead to ructions, but it is at least a coherent approach, and it has a fairly clear basis in physical properties.

Nine more TNOs, and three more asteroids, will become candidate planets, pending further investigation of how spherical they are. More planetary TNOs may follow, when discovered. To tidy things up, the minor planets will get renamed: those that don't have enough of a gravitational grip on themselves to be proper planets will now be 'small Solar System bodies'.

All this will doubtless lead to ructions. But it is at least a coherent approach, and it has a fairly clear basis in physical properties. It has been convenient to have a small and easily memorized number of planets in the Solar System, but convenience is not the only thing that counts. The effects of mass define (unofficially) the upper limits of the planetary realm; anything big enough for fusion is a star. It is fitting, then, that mass should define the lower limit too. This, we think, adds up to a case for IAU members to accept the proposal.