For the patient reader, it should be clear from the above that this field has had a very exciting first fifty years or so. It must be admitted that this is in large measure due to outside influences. Like all other parts of astronomy, the study of galaxy collisions has ridden the breaking waves of the vast technological advances in detectors, satellite engineering, and computational resources. The subject has received further boosts from the enormous interest in parallel areas of study within the fields of galaxy evolution and star formation, and has contributed back to those areas. Never again will so many new, information-rich wavebands be opened. On the other hand, wide scientific frontiers remain to be explored with the aid of continuing increases in observational sensitivity, resolution, computational power, and synergistic interactions with allied fields.
In the last few sections I have attempted to clarify where we stand on the key questions posed at the end of Section 1. The first group of questions concerned the role of galaxy collisions within the overall picture of galaxy formation and evolution. Toomre's work in the 1970s held out the possibility that collisions and interactions were a dominant process, and that possibility has energized the field for most of the time since. However, there have always been counter-arguments. One of the strongest in the present era is the modest increase in the merger rate with redshift found in deep surveys. On the other hand, the relations between cosmic SFR or ULIRG numbers and redshift tell a different story. Downsizing may be an important part of the resolution between the different stories. Presently, we only see a hazy outline of the full portrait of the relation between galaxy morphology and redshift. Progress has been rapid in these areas, and we can expect a great deal more in the next decade or two. For optimistic theorists the answers are already available (if not yet fully extracted) from large-scale numerical models of structure formation.
The second group of key questions concerned the role of environment on collision dynamics and evolutionary processes. Although the study of galaxy collisions in groups and clusters has been around since Spitzer and Baade's work, it is being reborn in the present era. There is currently a great deal of interest in groups and clusters among observers, with new tools to facilitate that work. The theory and modeling side of this area is more complicated than that of binary collisions and mergers because of the interaction between several strong dynamical processes (e.g., ram pressure stripping, group/cluster direct or tidal effects, etc.). Nonetheless, it is also reasonable to expect significant advances in this area on decadal timescales.
The third group of questions concerned the orchestration of SF and nuclear activity by large-scale interaction dynamics. In the recent past it seemed likely that progress in this area would be hindered by the interplay of a number of complex dynamical processes. It now appears that this view was overly pessimistic. Due to the universal properties of turbulent interstellar gas, it now seems that wherever you compress cool gas you will enhance SF (in quantifiable ways). Thus, large-scale orchestration is mostly about gathering and compressing gas; feedback effects are mostly about heating and dispersing gas. There are more complexities than this, but the big picture does not appear impossibly complicated. Work in the coming decades should provide a much firmer foundation for this scenario, and a much better understanding of the exceptional cases.
So the reviewer's crystal ball conveys a bright outlook for answers to the first three groups of key questions. The glass gets more murky when we ask the last couple of questions. The fourth group of questions concerned secular effects and the fifth was about the archaelogy of individual systems. Of course we can model long-term processes on the computer with ever more precision. However, as discussed in several contexts above, it is hard to compare to observation either statistically or in individual cases.
That said, I would expect more progress from statistical studies, even though that will require the relatively slow accumulation of good datasets on numerous systems, for example acquiring large libraries of faint tidal structures in numerous galaxies. That slow work is not likely to be taken up by professional astronomers, but with the increasing availability of moderate sized telescopes and sensitive CCD detectors, it could become the realm of serious amateurs or robot astronomers.
It seems possible that the majority of key questions discussed above will be resolved within the next 50 years. However, new phenomena will be discovered, and more detailed understandings will be demanded. Recent, and possible near-future, examples support the point. As an example, consider the exotic forms or products of SF, like the ULX sources, and the possibility that some of these X-ray sources are intermediate mass black holes formed in dense, young star clusters. It will take a lot more observational work to explicate this phenomenon, and probably new theoretical insights to explain it. We could use several more Chandra observatories!
There are also still a few wavebands that remain largely unexplored, including low-frequency radio waves, very high-energy gamma rays (new more sensitive Cherenkov telescope arrays are presently coming on line), and gravitational waves of many frequencies. Equally exciting is the prospect that within a few decades our understanding of galaxy disk hydrodynamics may advance to point that we understand both the small scale, relatively short time, weather changes occurring in isolated disks, and the longer term climate changes wrought by various types of collisions and interactions. However, there is a great deal of work to be done before goal is achieved.
Acknowledgements
I am very grateful to my research collaborators for teaching me much about colliding galaxies and related topics. I want to thank Bev Smith, in particular, for making a number of helpful suggestions on this manuscript. I'd also like to acknowledge support from a NASA Spitzer GO Cycle 1 grant.