## 4. Concluding remarksWe recognize that the model investigated in this paper is very idealized, and so the solutions discussed can only be considered to be illustrative. Nevertheless we believe that we have captured something of the essence of the physical situation, and that the solutions presented do have some relevance to the richness and wide variety of behaviour that can be expected in a close binary comprising two late-type stars. The solutions are naturally nonaxisymmetric. However this is more than just the obvious asymmetry obtained, eg, by viewing two individual, almost axisymmetric, stars from a distance. The `natural' parity of a spherical dynamo, or of two such fields viewed together, is odd (). We see from Tables 1 - 3 that many of the solutions are even with respect to the orbital plane (). We emphasize that we have taken the simplest form of nonlinear dynamo, without inclusion of any of the effects that have previously been shown to favour the growth of nonaxisymmetric fields, such as nonuniform distribution of dynamo source terms (Rädler et al 1990; Moss et al 1991), anisotropy of (Rüdiger & Elstner 1994) or dynamically driven large-scale motions in meridional planes, perhaps with a latitudinally dependent angular velocity (Moss et al 1995). The sensitivity of the solutions to the imposed parameters in certain ranges is hinted at by the last entry of Table 2, where a small change in the parameters alters quite radically the solution. There are other indications of the presence of uninvestigated bifurcations as the physical parameters are altered: the computations are sufficiently time-consuming, and the model too preliminary, for it to be yet worthwhile to explore this issue systematically We can deduce that close late-type binaries can be expected to
exhibit large-scale nonaxisymmetric fields. These may be of even
parity, especially if the components are detached (Tables 1 and
2), which is not the parity of the stable solutions for the individual
stars. The visible fields then have maxima at the longitudes
corresponding to the intersection of the line of centres with the
stellar surfaces. We believe that these results may have relevance to
the preferred longitudes of active regions on the surfaces of RS CVn
systems, as referenced in the Introduction. © European Southern Observatory (ESO) 1997 Online publication: June 30, 1998 |