During the last few years a variety of observational campaigns, in particular by ISO from space in the far-IR and by large mm telescopes from ground, have started to provide a complementary view of the distant universe at long wavelengths with respect to that offered by standard optical-UV-NIR deep explorations. Also of crucial importance in this context was the discovery of an intense diffuse background radiation in the far-IR/sub-mm of extragalactic origin, the CIRB. These results are challenging those obtained from optical-UV observations only, by revealing luminous to very luminous phases in galaxy evolution at substantial redshifts, likely corresponding to violent events of star-formation in massive systems. In the most extreme of these sources, however, a quasar contribution cannot be excluded, and sometimes has indeed been proven.
Whereas the process of optical identification and spectroscopic characterization of the long-wavelength selected high-redshift sources is only at the beginning (and will keep being a challenging task for the next several years because of the faintness of the optical counterparts), some interesting constraints on the cosmic evolution can already been inferred from observations of the CIRB spectral intensity and the multi-wavelength source counts. The most robust conclusions at the moment appear to be those of a very rapid increase of galaxy long-wavelength emissivity with redshift, paralleled by an increased incidence in high-redshift sources of dust extinction and thermal dust reprocessing with respect to locally observed sources.
A way to interprete these results is to consider as a crucial cosmogonic ingredient the role of galaxy interactions and merging. The strong increases with redshift of the probability of interactions (as partly due to a plain geometrical effect in the expanding universe) and of the effects of interactions (due to the more abundant fuel avaliable in the past), likely explain the observed rapid evolution.
Altogether, the large energy content of the CIRB is not easily explained, unless the powerful infrared starburst phase is characterized by a stellar IMF somewhat deprived in low-mass stars.
Although the subject is presently subject to some controversies, we think we have provided enough evidence, based on pioneering efforts of deep sky surveys in the IR and mm, that only such long wavelengths contain the clue to an exhaustive description of the star formation phenomenon, now and in the past. It seems clear that there are no alternatives, neither in X-rays, optical nor radio, to the IR/mm flux measurement for a reliable determinantion of the rate of SF in galaxies, simply because it is there that a dominant fraction of photons from young very luminous stars emerges, and no ways are available to determine "a priory" what precisely this fraction is. Fundamental aspects of galaxy formation and evolution (e.g. the origin of galaxy spheroids, and the onset of quasar activity) can effectively be observed at long wavelengths. In this sense the variety of ground-based and space projects in this field planned for the present decade promises extremely rewarding benefits for observational cosmology.
This paper has benefited by a large collaboration, in particular concerning items discussed in the last chapters, including some yet unpublished results. I want to mention the people who have particularly contributed: H. Aussel, S. Bressan, C. Cesarsky, D. Clements, FX. Desert, D. Elbaz, D. Fadda, R. Genzel, G.L. Granato, M. Harwit, S. Oliver, B. Poggianti, J.L. Puget, D. Rigopoulou, M. Rowan-Robinson, L. Silva. I am also glad to thank L. Danese, G. De Zotti for a long-standing collaboration in this field, A. Cavaliere and C. Chiosi for many fruitful discussions. Finally, I want to warmly thank the organizers of the Canary Islands Winter School on "High-Redshift Galaxies" for their kind invitation.