§3.1 Detailed Notes on Classifications and Ordering

The first class of objects in the Gallery is Galaxies with Extended HI Envelopes, where the extent is compared to the underlying optical light rather than to a fixed physical scale. The rest of the Gallery is dominated by interacting systems, many of which also have very extended distributions. In this first class, we include objects which do not apparently owe their extended to an interaction . By this we mean that there is not an obvious partner with which the system is interacting, nor are there tell-tale signs of a prior minor or major merger event (see below), such as stellar shells and loops or multiple nuclei.

This classification includes many Blue Compact Dwarfs (BCDs), although BCDs also appear elsewhere in the Gallery. For example, UGC00521 (Fig. 14) appears with warps, and there are several BCDs filed under Miscellaneous, which may not have been mapped with enough sensitivity to determine whether or not they have extended gas envelopes. BCDs are some of the most unevolved systems in the nearby universe, in terms of their stellar content and metallicity, and as such this Extended HI class may include examples of galaxies actually forming out of the gaseous reservoirs in which they are embedded. For this reason, we have arranged galaxies within this class based on the regularity of their HI morphology: galaxies with very irregular outer HI are at the beginning, and galaxies with disk-like HI distributions are at the end.

The next class is Galaxies with HI Extensions. The first subclass is Galaxies with Two-Sided Warps. This is the perhaps the most under-represented class included in the Gallery, as warps are rather common (Bosma 1991). Here we have simply collected some representative examples. This includes the very peculiar system NGC3718 ( Fig. 13). We have placed it here because, like the two-sided warps, it has a 180 deg sense of symmetry, and we could not think of a more suitable place for it.

The two-sided warps are followed by Galaxies with One-Sided HI Extensions. Within this subclass are galaxies whose HI morphology is believed to be due to Ram Pressure Stripping (Fig. 21 - Fig. 26). The next system, the BCD galaxy NGC4861 (Fig. 27), has both an extended HI envelope and an HI cloud with no optical counterpart, making it the perfect connector to the next subclass, Galaxies with Detached HI Clouds . This is another group of galaxies which is poorly represented in the Gallery, and many more examples appear in the literature. The qualifier ``no optical counterpart'' obviously depends on the quality of the existing optical data, and to hedge our bets, we have placed this classification between ``HI extensions'' and the next class of systems, the Minor Mergers. This way, if an optical counterpart to the HI cloud is subsequently identified, the Gallery needs not be re-ordered. Similarly, the first member of the Minor Mergers, M108 (Fig. 30), has an HI extension toward an optical companion, but the optical companion does not have a known redshift. If it is subsequently shown to be a background object, then this system naturally falls into the prior subclass.

As just mentioned, the next class of systems is the Minor Mergers. These are two (or more) galaxies which are physically close to each other and show signs of interacting, and in which one of the galaxies is much larger or much brighter than the other(s). While we have by no means quantified the relative size or luminosity ratio, we estimate that the companion is of order 1/4 or less the size or luminosity of the primary. These systems are arranged in order of decreasing separation between the component galaxies. With the exception of M108, all the pairs of Minor Mergers have known redshifts which imply a physical associaton.

Within the class of Minor Mergers, we include several subclasses. The first is M51 Types: large grand-design spirals with small companions at the end of one of the spiral arms. The next subclass is 3-body Encounters, of which M81 (Fig. 53) is the prototype. These are different from the class of Triples later in the Gallery, as one of the participants appears much larger or brighter than the others. The Mrk348 system (Fig. 52) is a member of both of these subclasses: it has a small companion at the end of a spiral arm, and an HI distribution very similar to M51 (although on a much larger physical scale); but there is also a large neighbor (NGC266) to the northwest, which may have played some role in shaping the outer HI morphology.

After the 3-body encounters comes NGC1097 (Fig. 57), which is placed here because its companion (NGC1097A) appears about ready to merge with it. This is followed by the Minor Merger Remnants. These have a single identifiable nucleus, but optical morphological peculiarities typically ascribed to strong gravitational disturbances, such as shells, ripples, tails, and plumes. They appear here, as opposed to the later class of Merger Remnants, because of the continued survival (or re-formation?) of a large disk. It is widely believed that major mergers destroy disks. While we are not convinced of this fact, the literature on these objects discusses them almost exclusively in terms of a minor merger origin, so we adopt those results here.

The next class of galaxies is the dominant class represented in the Gallery: Interacting Doubles or Major Mergers. These are galaxies of apparently similar mass which are physically associated. This classification has taken precedence over other classifications except Triples -- Groups. For example, if the outer HI morphology appears clearly tidal in origin, we have placed the system in this class rather than under the category of HI Extensions.

We have subdivided this category into five separate subclasses. In doing so we draw heavily on the lessons learned in the pioneering study of interacting galaxies by Toomre & Toomre (1972; see also Barnes 1998). Specifically, from that work we learn the following: (1) tidal features first form shortly after first orbital periapses; (2) tails are formed at the rate of one per prograde disk; (3) high-inclination or retrograde encounters lead to large epicyclic motions within the disk, but do not form well-defined tails; and (4) bridges form from a wide range of encounter geometries. Further, we know that later-type spirals tend to be rich in HI, whereas lenticulars and ellipticals tend to be gas poor (Roberts & Haynes 1994).

With these considerations in mind, we have morphologically defined several subclasses which we think are dynamically related to the Hubble types of the participants and the spin geometry of the encounter. We emphasize that this classification is purely morphological: we have made no attempt to check the proposed spin geometries against the velocity fields of the galaxies. The subclasses are defined as follows:

Two HI Systems; Two Tails: from the above dynamical considerations, we suspect that these are interactions between two prograde disk galaxies, which we indicate in Table 1 by the notation (), where Sp denotes a spiral progenitor, and the + superscript denotes the suggested prograde spin geometry. Objects within this subclass are arranged by decreasing nuclear separation, with well-separated objects at the beginning and single objects with two tails at the end. II Zw 40 (Fig. 6) may reasonably be placed midway through this sequence, but we left it with the BCDs in the class of Galaxies with Extended HI Envelopes.

Two HI Systems, One HI Tail: from the above dynamical considerations, we suspect that these involve interactions between two spiral disk galaxies, only one of which has a prograde geometry. The second disk might have a highly inclined or a retrograde spin geometry, indicated by a superscript 0 in the () notation in Table 1. Objects within this subclass are again arranged by decreasing nuclear separation, with well-separated objects at the beginning and single objects with a single tail at the end. Arp 295 ( Fig. 159) belongs near the beginning of this sequence, but has been placed with Groups... due to its large number of HI companions. The first member of this subclass, the LMC/SMC/MS system ( Fig. 73), may more properly belong to the class of 3-body encounters under Minor Mergers, but we decided to place it according to the morphology exhibited in the figure. Arp 215 (Fig. 60) might reasonably be placed at the end of this sequence, rather than with the Minor Merger Remnants.

Two HI Systems, Bridge, No Tails: we suspect that these involve interactions between two highly inclined or retrograde disk galaxies, denoted by the () notation in Table 1. The first members of this subclass may well have different spin geometries and be caught prior to first orbital periapse, but they fit the above morphological definition so are placed here. The galaxies in this subclass are arranged with nuclear separation first decreasing as morphological distortion increases, then with nuclear separation increasing as more bridge material appears between the two systems. Notably, there are no Merger Remnants included at the end of this sequence. Since the tidal signatures of such encounters are not as well defined, it is much harder to make a confident classification of such objects after the progenitors have merged. VV 114 (Fig. 107 under Merger Remnants) may be an example of such an object. The reader will notice that this sequence ends with Ring Galaxies, which also starts the next subclass.

Two Systems, Only One with HI: These are presumably encounters between one gas-rich and one gas-poor progenitor (E - Sp notation). The first few systems in the subclass exhibit no tails, while the remaining systems all show one tail. The first three continue the theme of the last five systems of the previous subclass by showing Ring Galaxies, but in this case the smaller penetrating galaxy has no gas. It is possible that the gas was stripped as this system passed through the gas-rich target system, or that the progenitor was always gas-poor. In the Arp 104 system (Fig. 101) it looks like the southern system has a gas-rich tail, but the morphology of this feature resembles numerical simulations in which bridge material passes through the companion and emerges on the opposite side.

The final subclass of Interacting Doubles is that of Merger Remnants of Indeterminate Origin. These systems are clearly the result of the coalescence of separate stellar systems, but it is really not possible to say what has merged or how.

All of the previous classes have been dominated by late-type or spiral galaxies. The next major class, Peculiar Early Types or Early Types with Peculiar HI, is centered around early types and ellipitcals. Many of these may be Merger Remnants, others may be Minor Merger Remnants, and yet others may owe their gas and/or optical morphological peculiarities to their dense local environments. Since these origins are very difficult to distinguish, we have made subclasses based on both their optical and HI morphology.

The first subclass is Peculiar Ellipticals with HI Outside the Optical Body, and is ordered by the amount of HI in the outer regions (from lots of HI to no HI). This sequence may be considered a possible extension of the Toomre Sequence of Major Mergers (Toomre 1977), demonstrating how gas-rich disk galaxies might fall together, merge, and leave a gas-poor bulge-dominated galaxy in their place. The amount of optical peculiarity decreases along this sequence (but not uniformly), with more subtle optical peculiarities in the later than in the earlier stages. At the end of this sequence we have included a montage of optically peculiar early types mapped in HI, but in which the HI is not associated with the early type galaxy. These provide an interesting counterpart to optically peculiar systems in which HI has been detected. It is possible that more sensitive HI observations might uncover some HI in these systems, but clearly it will be less than in those that have already been detected. These systems are ordered by decreasing optical peculiarity, as quantified by the Fine Structure Index (FSI) of Schweizer & Seitzer (1992). This index quantifies the number of shells, jets, plumes, ripples, and ``X''-structures, as well as the boxiness of the galaxy. We note that many of these galaxies exhibit quite striking peculiarities when imaged with modern CCDs, and the reproductions here, taken from the Digital Sky Survey, really do not do them full justice.

The next subclass is Peculiar Early Types with HI Within the Optical Body. These are arranged by the degree of regularity of the HI. The early systems have a very irregular HI distribution, and the distribution becomes more symmetric and disk-like as the sequence progresses. This sequence demonstrates the intriguing possibility that in some cases enough cold gas is accreted onto a bulge-dominated galaxy to form (or re-form) a disk.

This is followed by Normal Early Types with Peculiar HI . These galaxies have no obvious optical peculiarities (certainly at a much lower level than the preceding two classes), but have some very interesting HI distributions. This emphasizes the point that it is very difficult, if not impossible, to guess the HI morphology based on a system's optical appearance. The first system of this subclass, the polar ring galaxy UGC7576 (Fig. 130), could equally well have been the last system of the previous subclass. After this, the HI distribution becomes increasingly irregular along the sequence. Unlike the previous two subclasses, this ordering is not meant to suggest an evolutionary sequence. Rather, this ordering forms a natural transition from the disk-like HI distributions of the previous subclass, to the irregular distribution of the next class of objects.

Given the normal optical appearance of the hosts, we are not sure what to make of this subclass of objects. It is possible that future optical observations will reveal as-yet-undiscovered optical peculiarities in these systems, and that they may fit naturally into one of the previous two categories. It is also feasible that the HI has a tidal origin, and for some reason the encounter geometry left the outer tidal gas in an irregular distribution long after the inner regions have relaxed. In this case, these systems would be an extension of the previous two categories. Yet another possibility is suggested by the fact that most, if not all, of the members of this subclass live in group environments; in this case the gas may have been stripped from the outer regions of other members and accreted onto the early type, which is usually one of the largest members of the group. These objects may therefore belong to the class of Interacting Triples -- Groups -- Clusters. In the latter class however the effects of interaction are manifested optically, so we have elected to keep these classes separate. The most interesting possibility (in our opinion) is that the gas in these systems was never in a galaxy, and represents accretion from a primordial reservoir. At present it is not possible to discriminate between these and other possible scenarios, but this is in any case one of the most intriguing categories of objects in the Gallery.

The next class of objects is just as interesting, comprising galaxy systems with Intergalactic Debris with No Optical Counterpart. These three objects could easily have been categorized into other classes (the Leo Ring [Fig. 141] with the previous class; NGC5291 [Fig. 142] with peculiar early types with HI within the optical body; and NGC4532 [Fig. 143] with galaxies with one-sided extensions, or with the minor mergers). We have placed them into a separate category since in these cases the relation between the HI and the neighboring galaxies is less clear.

Interacting Triples -- Groups -- Clusters form the next major class. As mentioned above, there are a lot of similarities between the HI distribution seen in triples and groups and those shown in the previous three subclasses of objects, but here there are clear optical distortions suggesting more directly an interaction origin for the intergalactic gas. This class is ordered (for the most part) by the increasing number of members. The exception is at the end, where we have put three early-type-dominated groups. In each of these last three systems there is a significant extended X-ray component, which must have some effect on the presence or absence of cold gas. [For the on-line version of the catalog, we have separated these classes into three separate webpages].

Finally, there is the unavoidable Miscellaneous class. These objects do not obviously belong to any of the preceding classes, nor are there enough similar characteristics to warrant the creation of additional classes. The first examples of this class (the low-redshift QSOs and the E+A galaxy) have been called interacting galaxies, but this conclusion was based on the HI distribution. If there is anything we have learned from the compilation of the Gallery, it is that weird HI distributions need not always arise due to interactions.