Annu. Rev. Astron. Astrophys. 1996. 34: 749-792 Copyright © 1996 by Annual Reviews. All rights reserved

8. SUMMARY

Our knowledge of luminous galaxies has increased dramatically as a result of the IRAS all-sky survey. During the past decade, ground-based redshift surveys have succeeded in identifying thousands of LIGs, including several hundred ULIGs in the local Universe (z 0.3), and detailed multiwavelength studies, particularly of the brightest objects, have provided substantial information on what types of galaxies emit these large infrared luminosities. It seems clear that strong interactions/mergers of gas-rich spirals are responsible for the great majority of LIGs and that enormous starbursts must be involved in generating a substantial fraction of the infrared luminosity, at least in the early phases of the interaction.

Theoretical models have added substantially to our understanding of the merger process, in particular by showing how it is possible to build up enormous concentrations of gas in the centers of the merger remnant. Interferometer measurements at millimeter wavelengths show that the most luminous infrared objects often contain as much as 10¹⁰ M of gas within 0.5-kpc radius of the merger nucleus! Such enormous gas concentrations are an ideal breeding ground for a variety of powerful phenomena, including powerful starbursts with their accompanying superwinds that may add substantially to the metal enrichment of the intergalactic medium, the formation of massive star clusters, and most likely the building and/or fueling of an AGN. LIGs very likely represent an important link between starburst galaxies and the AGN phenomena exhibited by QSOs and PRGs. They also likely represent transition objects spanning the gap between merging spirals and emerging ellipticals.

Future infrared space missions and more sensitive ground-based submillimeter surveys will succeed in identifying more distant LIGs, thus allowing a test of whether the infrared luminosity function evolves as steeply as that of QSOs and whether LIGs were more numerous in clusters during the epoch when it is presumed that most of the ellipticals were formed from mergers of spirals.

However, detailed studies of objects already identified from the IRAS survey will probably play the largest role in our ability to determine the nature of the dominant energy sources in LIGs and to understand more precisely what parameters lead to the origin and ultimate fate of these objects. The following research areas could prove particularly productive:

1.	Detailed multiwavelength studies of those LIGs already identified in the IRAS BGS that do not show obvious evidence of current or past interactions (e.g. IRAS 10173+0828, VII Zw 031). Do these ``exceptions'' to the merger hypothesis simply represent the end-stage of the merger after most of the obvious tidal features have disappeared, or do they represent an alternative way of producing LIGs ?
2.	Identification of more ``transition objects'' - those galaxies that are currently LIGs but that also show broad Sy 1 emission lines or powerful radio cores/jets that are characteristics of QSOs and PRGs respectively. These objects currently represent ~ 15% of LIGs in the IRAS BGS and many more should be discovered in follow-up studies of sources in existing catalogs of fainter objects.
3.	A systematic search for possible fossil remnants of the ULIG merger phase (e.g. molecular gas, faint tidal features, etc.) in the host galaxies of QSOs and PRGs.
4.	Mid- and far-infrared spectroscopy with the Infrared Space Observatory (ISO) to better study the nature of the deeply embedded energy sources in LIGs.
5.	More elaborate theoretical models that include a better treatment of star formation, stellar winds, and supernovae explosions to better follow the evolution and fate of the central gas concentration.
6.	VLBI observations of extragalactic OH megamasers - ULIGs host a class of megamasers that may prove to be a powerful dynamical probe that can be used to test for the presence of supermassive black holes. Finally, it should eventually be possible to better discriminate between AGN versus starbursts in ULIGs by direct measurement of the size of the emitting region at mid- and far-infrared wavelengths. This could happen in the next few years as ground-based submillimeter interferometers come into operation and later with proposed airborne and space-based platforms such as SOFIA and FIRST.

ACKNOWLEDGMENTS

It is a pleasure to thank the many people who sent us preprints and reprints of their work on luminous infrared galaxies. We also thank L. Armus, M. Arnaud, W. Baan, J. Ballet, J. Barnes, E. Brinks, R. Cohen, L. Cowie, P. Duc, A. Evans, T. Heckman, E. Hu, R. Joseph, J. Kormendy, J. Mazzarella, R. Norris, T. Soifer, A. Stockton, N. Trentham, W. Vacca, and S. Veilleux for helpful discussions and comments while writing the manuscript. We are grateful to those people who generously made their H I data available to us for use in constructing overlays for the figures: E. Brinks (Figure 9a), C. Simpson and S. Gottesman (Figure 9b), J. Hibbard (Figures 8b and 8d), J. van der Hulst (Figure 8a); to P Bryant for permission to use his new millimeter-wave interferometer CO data in Figure 6d; to J. Wink for permission to use the CO map in Figure 9a; to A. Stockton for permission to reproduce optical images obtained with the CFHT and the UH 2.2-m telescope (Figure 10); and to P. Eisenhart for permission to use the HST image reproduced in Figure 11. We would particularly like to express our appreciation to Pierre-Alain Duc for reducing most of the grey-scale images displayed in Figures 8 and 9, to Karen Teramura for preparing final versions of all of the figures, and to John Kormendy for his TeX macro, which was used to prepare preprints of this article. DBS was supported in part by NASA grants NAG5-1741 and NAGW-3938 and would like to acknowledge the hospitality of the Service d'Astrophysique, Saclay, during extended visits while this article was being written.