Though bright-rimmed shells seem to be uniquely associated with elliptical galaxies, other types of galaxies show peculiar structures that can only be explained by invoking some external influence, ranging from the accretion of gas from a companion to a full-scale merger. However, not all anomalous optical features can be explained in this way (eg. NGC 1313, Ryder et al. 1995) and not all galaxies that show such structures have detectable HI or obvious dust. Equally, not all elliptical galaxies with detected HI are optically abnormal. Here we will look some examples of HI ellipticals, with and without shells.
One of the best examples of a shell galaxy with neutral hydrogen is NGC 5128, Centaurus A. In addition to its distinctive dust lanes, it has an extensive low surface brightness optical structure which extends well beyond the shell structure discovered by Malin, Quinn & Graham (1983). The location and spacing of the shells, together with a model of the shell-forming process, permitted some estimate of the age of the merging event that produced them. In the case of NGC 5128, using Quinn's (1984) model, it was shown that this was a few times 109 years ago. However, it seems probable that the faint outer material noted by Johnson (1963) and shown in Malin (1981) was displaced at some earlier period, perhaps by the current merger victim on a previous close passage. This faint outer envelope extends well over 80 arc minutes along the NE-SW axis of the galaxy, and no distinct shell structure is visible, though it is not completely featureless.
Schiminovich et al. (1994) mapped the distribution of HI around the central 30 arcmin (about 30 kpc) of the galaxy, detecting 1.5 × 108 M of HI, which seems loosely associated with the outermost shells, noting that it appears beyond the optical boundary of the shell in each case where it is found. From the spatial distribution and velocity structure, they suggest that it forms a partial ring, and question the 109 year timescale for the merger derived by Malin, Quinn & Graham (1983).
NGC 2685 is another shell galaxy that was included in the original Malin-Carter (Malin & Carter 1983) catalogue. Schiminovich et al. (1995) have shown that the mass of neutral hydrogen it contains is similar to that found in NGC 5128, and it is likewise distributed in a large loop or arc, loosely associated with, but just beyond, the low contrast optical features. A deeper image of this galaxy is included here as Fig.1. If the HI found in these galaxies is associated with the shells the velocities are difficult to reconcile with any of the current shell-producing models. Clearly, HI observations are vital new evidence that can be used to refine models of shell formation.
Figure 1. The main image of NGC 2685 was made by combining photographically amplified derivatives from three UKST IIIa-J plates and shows a series of low surface brightness loops extending from the galaxy. Inset is a normal contrast photograph of NGC 2865 made to the same scale. The inset print is 5 arc min on a side.
NGC 5266 is a polar ring galaxy where the polar ring is hidden on deep exposures. However, the low surface brightness envelope is unusually large, and subtle evidence of the dust lane can be seen in the `pinched waist' of the galaxy, which is noticeable on the original of Fig. 2. This illustration also reveals an even fainter, smooth extension to the SE of the galaxy. Observations by Morganti and Sadler (this volume) show that the optical extensions coincide quite well with the HI detection.
Figure 2. This image shows the enormous extent of the faint outer envelope of NGC 5266 compared with the superimposed positive, which is to the same scale. Note the large, faint, structureless SE extension. The scale bar is 5 arc min.
NGC 5018 (mentioned in Schiminovich et al. 1994) is an E3-E4 giant elliptical galaxy that was included in the Malin-Carter (1983) shell galaxy catalogue. Deep images show that it is obviously distorted and interacting with a gas-rich spiral companion NGC 5022, with a faint, probably stellar `bridge' between the two bright galaxies, clearly shown for the first time in Fig. 3.
Figure 3. The interacting system NGC 5018 / 22 is clearly linked by a low surface brightness `bridge' of luminous material. The arrow indicates the optical detection of a companion dwarf galaxy detected in HI by Kim et al. (1988). Scale bar is 10 arc min.
Though NGC 5018 itself has only a marginal HI detection coincident with the optical image, NGC 5022 has an extension of its HI profile in the direction of the giant elliptical (Kim et al. 1988). This coincides with the brightest part of the stellar bridge between the galaxies. These authors also show that there are three blobs of HI with masses of a few times 108 M (and velocities which are similar to both galaxies) in the field around NGC 5018, though they were only able to identify two with extended optical counterparts. Our deep image reveals that Kim's source S3 is also a galaxy (arrowed in Fig. 3) which is featureless and of low surface brightness. Other low surface brightness galaxies which correspond to marginal HI detections are also indicated with arrows and all are probably members of a small group associated with NGC 5018 / 22.
Kim et al. describe NGC 5018 as perhaps being `caught in the act' of acquiring its HI from a companion galaxy; however, the galaxy has a dust lane (Fort et al. 1986) which seems to be deeply embedded, and an unusually low metallicity for a giant elliptical of this intrinsic luminosity. Bertola et al. (1993) consider the galaxy to be prototypical of a class of low metallicity giant ellipticals. In view of the unusual juxtaposition of low metallicity and evident dust, it seems probable that these galaxies have interacted more than once, perhaps resulting in the shells and the material which produced the arc of HI-rich companions, but they are now in the early stages of a closer encounter.
Other optically interesting elliptical galaxies with detected HI are IC 1459, which appears to be an S0 but shows low contrast spiral arms and a counter-rotating core (illustrated in Malin 1985) and the interacting pair NGC 1549 - 1553 (in Malin & Carter 1983). All these galaxies seem to be examples of `galaxy harassment' of the kind described by Moore et al. (1996), where multiple, high speed encounters drive star formation and leave detectable arcs of debris. These interactions will also leave disturbed spirals.