Annu. Rev. Astron. Astrophys. 1996. 34: 155-206 Copyright © 1996 by Annual Reviews. All rights reserved

6.2. Galactic Encounters

There is strong observational evidence that a number of spiral galaxies are interacting gravitationally with a neighbor. The clearest nearby example is M81, which is believed to have undergone a recent encounter with NGC 3077 (probably less than 10⁹ yr ago). Because the orbit of NGC 3077 is approximately in the disk plane of M81, this system is particularly well suited to simulation, and Thomasson & Donner (1993) predict nonaxisymmetric velocities of order 10 km s^-1 in the disk of M81. With _t ~ 10²⁶ cm² s^-1 and L ~ 1 kpc, the magnetic Reynolds number UL / _t is then about 30, quite large enough to affect significantly the disk fields (Vallée 1986). Interestingly, M81 appears to have a strong bisymmetric field component. M33 also may have some bisymmetric field structure, and it is believed to be interacting with M31. Recently, at least weak evidence has been found for BSS in the interacting galaxy NGC 2276 (Hummel & Beck 1995) and for MSS in M51 (EM Berkhuijsen et al, in preparation).

If we consider a Fourier decomposition of and into parts _m, _m, corresponding to an azimuthal wave number m, then the induction term × ( × ) can give rise to a bisymmetric field component in two ways. If the dynamo basically generates an axisymmetric field ₀, then ₁ can generate a slaved m = 1 component ₁ from the ₁ × ₀ interaction. Moss et al (1993b) investigated this possibility in a nonlinear model with a relatively thick disk, using a velocity field based on the Thomasson & Donner (1993) simulation. They found that a globally modest bisymmetric field component could be generated, concentrated to the outer part of the disk, where it may dominate. More subtly, the ₂ × ₁ interaction (giving rise directly to m = 1 and m = 3 field components) may be such as to increase the linear growth rate of the bisymmetric field component compared to that of the axisymmetric component, so that in the nonlinear case a substantial bisymmetric field could survive. Moss (1995) showed that, in a simple linear model, the m = 1 field could then be excited at lower dynamo number than the m = 0 field, but a nonlinear investigation using a more realistic model is needed to clarify the importance of this mechanism. The remarks concerning the modal interactions apply, of course, whatever the mechanism providing the velocity field. In particular, it may be relevant that a ₂ × ₀ interaction can give rise to a slaved m = 2 field component.