This beautiful picture of Mercury is compiled from images of the planet sent back by the Mariner 10 spacecraft in 1974 and 1975. The Mariner 10 data also presented several puzzles. Writing in Earth and Planetary Science Letters (218, 261–268; 2004), Oded Aharonson and colleagues revisit one of them — why, against expectations, does Mercury have a global magnetic field?

The planet's diminutive size means that it should have cooled quickly after it formed. Any molten core would have become solid, or almost completely so. A magnetic-field-generating dynamo in an outer core (as is the case for Earth) should have long since seized up. But the news from Mariner 10 meant a rethink was needed, and since then attempts to account for the magnetic field have centred on how the core might have remained largely molten.

Credit: MARINER 10/ASTROGEOLOGY TEAM/US GEOL. SURV.

Aharonson et al. examine another possibility — that the field originates from magnetization of Mercury's crust, the remnant of a powerful core dynamo early in the planet's history. After considering some of the arguments against this possibility, the authors tackle the main objection, which comes in the form of Runcorn's theorem. This holds that a uniform shell, magnetized by an internal source, cannot have an external field if the source is removed.

Aha! thought Aharonson et al., what if the shell — the crust — is not uniform, and the magnetized layer varies in thickness? They discuss why this could be so, work through the relevant mathematics, and conclude that Mercury's magnetic field could be crustal in origin under certain conditions.

Core or crust? The question should be answered by the MESSENGER mission, due for launch in May. The spacecraft will make two passes by Mercury in 2007 and 2008, before going into orbit around the planet in 2009. On board are a laser altimeter and a magnetometer, which can respectively provide evidence of the planet's structure and any asymmetries in the magnetic field. The resulting data should give a definitive answer.