4 Discussion

We have identified clear signs of tidal interactions and possibly a merger toward HCG 54. The morphology and kinematics of the atomic gas, a sensible tracer of interactions, is strongly perturbed, including a long (20 kpc) HI tidal tail, plus two more HI extensions that emerge from the main body. The stellar component also shows features that are characteristic of interactions or mergers, in particular several shells around the bright central area. We propose here that the formation of the HI tail and the optical ripples involved interactions and a merger of at least two galaxies.

4.1 The isolated environment of HCG 54: only one (dwarf) companion

The proximity of HCG 54 to our Local Universe (v = 1470 km s^-1) makes the study of its large scale environment difficult. We find in the NASA Extragalactic Database (NED) that all galaxies with measured redshifts in a square of 1 Mpc radius centered in HCG 54 have velocities larger than 4000 km s^-1. Within a distance of 500 kpc none of the galaxies without redshift determination have a similar size or magnitude as the HCG 54 members. In fact, the closest and brightest galaxy to HCG 54, A1127+2057, is a background galaxy (see Sect. 3). Similarly, another optically selected potential neighbor A1126+2051 is also shown to be a background galaxy.

The VLA primary beam (30$^\prime$) covers a diameter of 170 kpc at the distance of HCG 54, and we found only one object (A1127+2054, see Sect. 3) in this area at the group redshift in HI emission, within the observed range of 645  km s^-1. The dwarf disk galaxy A1127+2054 is the only identified companion in the environment of HCG 54 and is located at 20 kpc from its center.

   
4.2 Origin of the stellar ripples/shells

Shells are usually thought to form in unequal mass interactions involving a bright elliptical accreting a small companion (see e.g. Quinn 1984; Hernquist & Quinn 1987). The work by Schweizer & Seitzer (1988) showed that a smooth triaxial potential, or even a disk, can also give rise to the formation of shells and ripples in the stellar component of an interacting smaller companion. It does not seem to be the case in HCG 54, where we do not detect a large bulge component nor any other massive system. The two galaxies should have been most probably late-type systems, judging from the post-starburst nature of the spectrum shown by HCG 54a (without clearly detected features typical of early type systems) which could be invoked as the possible nucleus of the system. Furthermore, the low chemical abundance derived from the ionized gas, 12+log(O/H) $\sim$ 8.3, is more typical of irregular galaxies. Few examples of shells around low mass late type galaxies are found in the literature, as NGC 7673 (Homeier & Gallagher 1999). If we relax the low mass criteria there is also NGC 3310 (see Mulder & van Driel, 1996) which is a peculiar Sbc. In both cases a tidal interaction origin has been attributed to the shells. In general, little observational evidence exists for ripples in disk-disk interactions, (see e.g. Homeier & Gallagher 1999; Charmandaris & Appleton 1996; Kemp & Meaburn 1993 or Schweizer & Seitzer 1988). Probably for this reason few models are found focused on reproducing HCG 54-like systems. In this sense, the simulations presented by Hernquist & Spergel (1992) seem to provide with a more suitable model to describe the system. The model represents a merger of two equal disks galaxies in a close collision from a parabolic orbit. At the final stages the remnant has a central bar-like structure embedded in a disk like envelope, together with several shells at different radii. This resembles quite well the optical structure of HCG 54, centered on HCG 54a, with an associated bar-like bluer structure, whose colours are suggestive of star formation, in the middle of a red disk-like component and surrounded by 3 shells. The shells are centered somewhere close to HCG 54a and b, but the data do not allow a determination of the precise location of their focus (Sect. 3.1). The outer shells (t1 & t3, see Fig. 2) have bluer colors, consistent with an origin in the late type progenitor disk. The inner one (t2) is slightly redder, and this might be due to contamination by the population of the disk-like component associated with HCG 54a. Therefore the morphology of HCG 54 can be well explained by the collision of two similar mass disk systems. The model does not allow further test or comparison, since no prediction is given neither for the kinematics nor for the stellar populations/colors of the colliding systems. Here we will try to shed more light in the identification of the involved systems based on the rest of the information provided by our optical and HI data.

We have tried to determine the number of present galaxies using the observed HI mass $(\log[M({\rm ~HI})_{\rm obs}/M_{\odot}] = 8.75$). We have calculated the expected mass as a function of optical luminosity and morphological type via the relationships obtained by Haynes & Giovanelli (1984). Assuming that HCG 54 is composed of 4 irregular galaxies as classified in Hickson catalog, we predict a mass of $\log[M({\rm ~HI})_{\rm pred}/M_{\odot}] = 9.16 \pm 0.23$, but for a single irregular galaxy the expected mass would be $\log[M({\rm ~HI})_{\rm pred}/M_{\odot}] = 9.19 \pm 0.30$. Unfortunately these cases are indistinguishable since they are within one sigma, and all we can conclude is that the HI content of HCG 54 does not deviate significantly from normal.

HCG 54a is the best defined object of the system, with a defined nucleus, embedded in a very elongated blue component and surrounded by a rounder and redder stellar envelope. This ensemble seems to be aligned with a perturbed but still visible  HI velocity gradient. This might indicate that HCG 54a is one of the interacting galaxies, whose disk is a red round component that still keeps memory of the original rotation according to the HI kinematics (Fig. 7b). We find however several arguments against this hypothesis. The atomic gas is decoupled from the stellar component both in morphology and kinematics. The optical emission is displaced to the north by $\sim$1 kpc with respect to the atomic gas, and this shift is too large to be explained only by stripping of the northern HI into the long tail. Furthermore, the HI velocity gradient does not have a counterpart in the optical velocities. The signature of a rotation curve is not found along any of the observed directions, except for a small velocity gradient of 50 km s^-1 in the inner 5 $^{\prime \prime }$ of HCG 54a (see Sect. 3.2). The HI radial velocities along the direction of the optical spectra differ from the optical ones by $\sim$30 km s^-1, also consistent with a decoupling between the stellar and gaseous components. Finally, since the HI velocity gradient is the only trace of a possible disk system in the central part of HCG 54, then the second disk would be a minor body when compared with HCG 54a, in contradiction with the equal masses involved in the simulations. Furthermore, if the observed HI velocity gradient of $\sim$180 km s^-1 had to be attributed to a single galaxy, it would be more characteristic of an Sb type system, which is not observed here.

A more plausible scenario is that we are observing a mix of the two components of the original systems, that have almost coalesced and now present a similar distribution as found in the simulations. The irregular shape of the optical velocity curve supports an advanced merger stage. HCG 54b shows a clear WR feature in the spectrum, indicative of a recent and strong burst of star formation that was possibly triggered by the interaction, while the underlying population, if present, is negligible, as suggested by the very high equivalent widths of the brightest emission lines (see Sect. 3.1.). In fact we found a possible signature of gas infall in our spectra (see Sect. 3.2) that would be feeding the recent violent star formation observed. HCG 54c and d do not have a well defined photometric structure, but consist of an aggregate of knots, and have HII regions features in their spectra. The very blue color of HCG 54d suggests a pure star formation episode younger than several million years old. They plausibly constitute the debris of the interacting systems. The HI component of the kinematics we observe might be the relic of the orbital motions of the interacting systems. However projected radial motions cannot be negelected and detailed numerical simulations including a kinematical study are needed to interpret the data.

In summary, no clear individual galaxies could be identified toward HCG 54 while strong signs of interactions and a merger are found. The overall morphology is in good agreement with the model by Hernquist & Spergel (1992), and we propose that HCG 54 is in the final stages of a merger of two similar mass disk systems.

4.3 Formation of the HI tidal tail

While a chance projection is possible, the location of A1127+2054, a dwarf galaxy at the redshift of the group, closely aligned with the large HI tidal tail is strongly suggestive of its role in shaping the appearance and the extent of the HI tail. A tidal tail pointing back at the responsible intruder is a commonly observed generic feature, particularly soon after the closest approach, as the exchange of momentum between the intruder and the tidally disrupted material tends to bring them close to the common area in the phase space.

A1127+2054 is not likely to be a tidal dwarf since it is not located within the tail and has its own rotating HI disk or cloud. The fact that the associated HI feature is not well centered on the stellar body is consistent with the tidal disruption scenario. No stellar counterpart to the HI tail is detected down to a surface brightness level of 27 mag/( $^{\prime \prime }$)² in R along its length.

This tidal disruption scenario has some potential difficulties however. Assuming an inclination of $30^\circ$, the estimated dynamical mass for A1127+2054 inside 2.1 kpc radius is only about $4\times 10^8~M_\odot$, which is about 4% of the dynamical mass of HCG 54. On the other hand, the estimated HI mass alone for the 20 kpc long tidal tail is about $10^8~M_\odot$ (see Sect. 3.3), and raising such a massive tidal tail from the gravitational potential of HCG 54 may require a far more massive perturber.

An alternative scenario for the formation of the large HI tail is the merger of two equal mass progenitors as we have already proposed in the previous section in order to account for the disrupted stellar features around HCG 54. In numerical simulations of galaxy mergers including gas particles such as by Mihos & Hernquist (1996), formation of one or more massive gaseous tails is commonly seen. For example, the projected appearance of the HI tail with respect to the stellar remnant in HCG 54 is similar to the "$t\sim 70$'' stages shown in Figs. 1 and 2 of Mihos & Hernquist (when mirrored about the vertical axis). This corresponds to a late stage in this particular simulation where the stellar cores finally merge, similar to what we infer in HCG 54. If we are allowed to press this comparison a little further, the extended HI structures seen to the south and southeast of the stellar body may be interpreted as the tidal tail emerging from the second progenitor, as seen in the simulation. Of course, these inferences are qualitative at best since the initial conditions of the Mihos & Hernquist simulation are likely to be different from those of the merger involved in HCG 54.

In this scenario, an active role of A1127+2054 in shaping the appearance of the large HI tail is still allowed. For example, a dwarf companion UGC 957 is found within one of the large HI tails resulting from the merger of two massive disk galaxies in NGC 520 (see Hibbard & van Gorkom 1996) even though UGC 957 may be only a third party to the ongoing or recent merger.