3.3.1. The Cartwheel Galaxy and its Companions

An analysis of the velocity field of the Cartwheel is consistent with a slowly rising rotation curve (Struck-Marcell and Higdon 1993; Higdon 1993) similar to that found in NGC 2793. The ratio of ring expansion to ring rotation in the HI gas is significantly lower than that reported in the ionized component by Fosbury and Hawarden (1977), being around 52 km s^-1 (Higdon 1993). Fosbury and Hawarden obtained a radial expansion velocity of 89 km s^-1 based on optical spectroscopy of 5 HII regions. Unpublished work by Taylor and Atherton (1991) using Fabry-Perot imaging of the H line emission provide a value of 61 km s^-1, in closer agreement with that of the HI observations. Based on Higdon's analysis, the time taken for the ring to reach its current radius (i.e., the time since the impact of the intruder) is approximately 300 million years. (assuming a value for H₀ of 100 km s^-1 Mpc^-2.) Another interesting result of the Cartwheel HI observations is that neutral hydrogen surface density is significantly lower in the region of the bright star-forming knots in the southern quadrant of the ring. This may be due to the disruption of the HI clouds by the intense winds from the O-stars known to be present in the ring (Fosbury and Hawarden 1977).

Since the work of Davies and Morton (1982) there has been considerable debate not only about the identity of the intruder, but also whether the intruder candidates are massive enough to drive the observed expansion of the ring, and also create such a strong ring. The two galaxies nearest the Cartwheel (see Figure 9) were labeled G1, G2 by Higdon (1993) and a third more distant companion 3 arcminutes northeast of the Cartwheel G3. The question of the strength of the ring is addressed below. All three potential intruders have velocities close to the Cartwheel and are probable members of the group. G1 and G3 contain significant quantities of HI. The SO companion G2 was originally though to be a high surface brightness elliptical by Davies and Morton (called Galaxy 3 by them). Higdon (1993) has analyzed the times needed for the three galaxies to reach their present projected separations if each was postulated to have passed through the center of the Cartwheel, in the hope that this might rule out one or other of the companions. However, taking into account the projected distances and observed radial velocities, it appears that all three galaxies could have reached their current positions in the time needed to create the ring. There is no unambiguous "smoking gun".

Davies and Morton (1982) explored the idea that the early-type galaxy was the likely intruder. Based on a measurement of the central velocity dispersion, they calculated that the mass of G2 was approximately 4 × 10¹⁰ M (this would correspond to a mass of 1.8 × 10¹⁰ M for H₀ = 75 km s^-1 Mpc^-2, a value we will use in the subsequent discussion here). Higdon's HI observations (Higdon 1993) provide a mass (similarly adjusted to common value for H₀) for the Cartwheel based on the last measured point of the rising rotation curve of M(total) = 3.46 × 10¹¹ M, for an assumed inclination of 44 degrees. Davies and Morton (1982) pointed out that this very small implied mass ratio ( 1:20) was a potential problem for the classical stellar model for ring formation (e.g., Toomre 1978). How could such a small mass intruder create such a dramatic ring? Indeed the ratio of the ring expansion to ring rotation velocity in the Cartwheel is approximately 20%, suggesting a larger perturbation than could be delivered by such a small companion. Do more recent observations shed further light on the problem? Davies and Morton speculated that the high contrast in the Cartwheel ring might result from the triggering of stellar birth in the ring, rather than merely a classical bunching of old stars in the Lynds and Toomre picture. We believe that this is essentially correct. However, the question of relatively high ring expansion velocity is not so easily disregarded. We therefore reexamine the question of the mass ratio of the possible intruder galaxies.

The near-IR photometry of Marcum, Appleton and Higdon (1992) provided 2.2 µm magnitudes for all three potential companions as well as the Cartwheel itself. A surprising result of the K-band observations is that the early-type galaxy, G2 (in Higdon's terminology) and the more distant G3 are only one magnitude fainter than the Cartwheel at this wavelength. The third late-type companion (G1) is significantly less luminous at IR wavelengths although it exhibits evidence for new stars in its irregular disk (Figure 1). If we make the naive assumption that all the K-band light comes from old stars, we can use the relation of Thronson and Greenhouse (1988) to estimate the mass in old stars of the three companions and the Cartwheel. The result of this approach is companions G1 through G3 have masses of 0.3, 2.5 and 1.9 × 10¹⁰ M, respectively, compared with the Cartwheel which, using the same argument, would have 6 × 10¹⁰ M of old stars. Notice that based on this simplistic argument, the mass ratio of old stars for G2 and G3 is 41% and 31% respectively of that found in the Cartwheel. Hence we find that, in the absence of dark matter, the luminous mass of the companions is a substantial fraction of the luminous mass of the Cartwheel. The assumption that all the K-band light originates from old stars is a dubious approximation, but seems in reasonable agreement with the dynamical mass derived for G2 by Davies and Morton (1982). In the Cartwheel, supergiants may contribute a fraction of the light (see MAH) but this will serve to further increase the relative (old stellar) mass of the companion to the target galaxy.

We cannot, of course, ignore the dark matter. Based on the above argument, the Cartwheel contains significant quantities of dark matter, since M(total) / M(old-star) = 5.8. (Here M(total) is taken from the HI work of Higdon (1993)). The crucial question, therefore, becomes one of the dark matter component of the companion galaxies. It is clear that if the companion galaxies (G2 or G3) contained dark matter fractions of the same order as the Cartwheel, the perturbations needed to produce a substantial ring expansion velocity would be easily achieved. How likely is it that the companions have massive halos? Higdon (1993) has measured the masses of G1 and G3 using his VLA HI observations and finds total masses similar to those quoted above based on the old stellar population. However, it is not clear that the HI measures all the mass since the spatial resolution of the observations was insufficient to determine if the rotation curves were falling. Also, the fact that the mass derived for G2 based on the optical stellar velocity dispersion by Davies and Morton is comparable with the approximate mass derived from old starlight might suggest that this galaxy has little dark matter. However, it must be borne in mind that this was a mass derived for the central bulge of the galaxy. Recent observations with the HST confirm that G2 has two very extensive spiral arms and most likely has a larger mass than Davies and Morton inferred from what we now know to be the rather bar-like bulge of an SO. In conclusion, it seems likely that with the addition of modest dark halos, companion masses of the order of 20% of the Cartwheel are not out of the question for both G2 (the SO galaxy) and G3 (the more distant edge-on galaxy). Davies and Morton (1982) were probably correct to suspect that the contrast in the ring is a consequence of the triggering of star birth in the ring. Models suggest that such an effect could be produced in a companion with a mass as low as 15-20% of the mass of the Cartwheel (Struck-Marcell and Higdon 1992; Hernquist and Weil 1993). More discussion of this interesting point can he found in Section 6.3.