The origin of magnetism on the upper main sequence

We consider the incidence of magnetism in main-sequence stars with mainly radiative envelopes. We propose that the small fraction, which increases with mass, of stars which are magnetic can be explained if towards the end of the formation process, after the stars have developed a substantial radiative envelope, a correspondingly small fraction of stars merge. Such late mergers would produce a brief period of strong differential rotation and give rise to large-scale fields in the radiative envelopes. Such late mergers can also account for the lack of close binaries among these stars