Another possibility was demonstrated by Wilson in 1986: new magnetic
flux may be generated in the form of loops by the action of oscillatory
velocity fields.
Later, Spruit, Title and Ballegooijen (1987) assumed
that a large loop rises from the convection zone and brakes into many
small loops as it crosses the surface.
There are at least two obstacles with these models.
As noted by Lee (1992), the models will
generate MMFs everywhere in the moat, which does not agree with the
earlier studies (Vrabec 1971; Sheely 1972; Harvey and Harvey 1973).
The models will produce randomly oriented pairs of MMFs, which
also contradicts the result on non-random (radial) orientation of MMFs.
However, Lee's conclusion favors the Wilson model.