5 Do shell galaxies have the kinematic structure of merger remnants?

The shell structure is widely indicated as a signature of a merging event, (see Barnes & Hernquist 1992 and reference therein). On the contrary, Thomson (1991) suggests that a "weak intercation'' may produce long lasting shells. Both hypotheses explain several properties of the shells and of the host galaxy.

As discussed above, in E 2400100 two nuclei are visible. Borne (1990 and references therein) and Combes et al. (1995) have simulated interpenetrating encounters of two galaxies showing that a "bona fide'' indication of the ongoing interaction is U-shape (and the correlated inverse U-shape) of the stellar velocity profile. In our view, the velocity profile of the two components in E 2400100 is an example of this behaviour. The velocity dispersion curve is asymmetric, showing a rapid rising (although with large uncertainties) in the NW part of the galaxy. The U-shape profile is developed both during bound and during un-bound encounters (see Borne 1990); so it is difficult on this basis to know the dynamical evolution of E 2400100. If this object represents the final phase of an encounter, when the two galaxies are going to coalesce, and if the formation of shells in the galaxy is the results of such an encounter, these structures have to be very recent, since we still distinguish the two nuclei of the progenitors.

Recently Heyl et al. (1996) performed numerical experiments in order to identify kinematical signatures of "old'' merging events in elliptical galaxies. The basic idea is that during a merging process there is a significant outward of angular momentum transfer. This will end in a merger remnant which will be supported by random motion in the centre but characterized by rapid rotation in the outer portions. This translates directly in a $V/\sigma$ increase with respect to the radius, up to 0.5 in the outskirts. This evidence adds to the kinematic misalignments noticed in the simulated merger remnants. Both these characteristics are not present when the galaxy formation takes place through hierarchical collapse (Warren et al. 1992). We have calculated the $V/\sigma$ ratio for some of the sample galaxies, and their trends indicate different conditions in the galaxies nuclei. E 3420390 is dominated by anisotropy just like the nuclear parts of NGC 1549 and NGC 2865. Velocity curves available in the literature for NGC 1549 (Rampazzo 1988; Franx et al. 1989), which extend in radius up to $\approx 1$$r_{\rm e}$, indicate that $V/\sigma$ grows at $\approx 0.5$. In this galaxy, kinematic misalignments (rotation along apparent minor axes, strong twist of isopothes $\gt 120^\circ$) are further noticed. For NGC 2865 too, stellar kinematics data of Bettoni (1992) suggest an increase of $V/\sigma$ up to 1.3- 1.4 in the outer parts. The HI gas detection extend further out the kinematics, showing an increase in the rotation velocity. The lack of gas in the central part of this galaxy has led Schiminovich et al. (1995) to consider the hypothesis that it has been converted to stars through a starburst. In Paper III we suggest, on the ground of spectrophotometric models, that a burst can be happened in the centre of this galaxy about 2 Gyr ago (depending on the selected models).

NGC 6776, NGC 813, NGC 1316 and NGC 1553 are in the range of oblate rotators. The kinematical and structural differences between the two companion galaxies NGC 1549, NGC 1553 are particularly significant. Differently from NGC 1549, NGC 1553 does not rotate along the minor axis (Kormendy 1984; Rampazzo 1988) and the isophotal twisting is $\leq 2^\circ$ (Rampazzo 1987). For both these shell galaxies spectro-photometric models suggest that the last episode of star formation is quite old (see Paper III).