GALAXIES, HIGH REDSHIFT PATRICK J. McCARTHY The search for distant galaxies was first motivated by the desire to use them in the classical cosmological tests for determining the value of the deceleration parameter (qo). It quickly became clear that the evolution of galaxies must be taken into account before cosmological questions could be addressed. Thus the emphasis of high-redshift galaxy research shifted to empirically quantifying the evolution of galaxies and identifying the epoch of galaxy formation. If the latter can be achieved, it will provide an important constraint on competing models of galaxy formation and the growth of structure in the Universe. Nearly all galaxies known to have redshifts greater than unity were recognized on the basis of their unusually powerful radio emission. The optical properties of these galaxies are closely linked to their radio properties, making them less useful for cosmological and evolutionary studies. Some important inferences regarding the formation of massive galaxies can, however, be drawn from this highly unusual population of galaxies. FINDING GALAXIES AT HIGH REDSHIFT The night sky, when viewed at the faintest levels in optical light, is crowded with galaxies. The most difficult aspect of high-redshift galaxy research is to distinguish distant, intrinsically luminous galaxies from inherently faint foreground galaxies. Our lack of an a priori knowledge of how distant, and hence young, galaxies should appear compounds the problem. Redshift surveys of complete samples of faint galaxies have unveiled information concerning the properties of galaxies out to redshifts of z=0.75 or so. The redshifts of these galaxies are determined both from stellar absorption lines (primarily ionized calcium) and from strong emission lines from ionized gas (usually doubly ionized oxygen). The shape of the luminosity function of galaxies ensures that most faint galaxies will be objects of more or less average luminosity at modest distances, making it unlikely that galaxies with Z>1 will be found in random surveys. Rich clusters of galaxies, readily identified on photographic plates because they stand out from the more uniform distribution of field galaxies, have been identified out to redshifts of 0.9. These observations have shown significant evolution in the properties of cluster galaxies in the last 4-8x10* years, but foreground confusion problems have prevented the detection of clusters beyond z=1, if any exist. The most productive method of isolating distant galaxies has been identifying the optical counterparts of bright radio sources. High-luminosity radio sources can be readily distinguished from low-luminosity sources on the basis of their double-lobed morphology. Bright radio sources are less crowded than faint optical galaxies, and thus confusion is not as likely. The culmination of high-redshift radio galaxy identification in the era of photographic astronomy was Rudolph L. Minkowski's 1960 measurement of z=0.46 for the galaxy associated with the radio source designated 3C 295. The advent of digital detectors allowed other observers to carry on the approach pioneered by Minkowski to much fainter limits. The first galaxy with a redshift greater than unity, 3C 368, was observed by Hyron Spinrad at Lick Observatory in 1982. In the past five years galaxies identified with relatively strong radio sources have been found out to redshifts of 2, providing the first glimpse of galaxies when the Universe was less than * of its current age. The redshifts of these faint galaxies are determined from their strong emission lines of ionized oxygen, carbon, and hydrogen. Recently, astronomers at the University of Hawaii and Johns Hopkins University have extended the search to somewhat fainter radio fluxes and have located galaxies with redshifts approaching 4, a distance at which the Universe was between 15 and 30% of its current age. PROPERTIES OF HIGH-REDSHIFT RADIO GALAXIES Only recently has comparison between the optical and radio properties of distant galaxies become possible. Multiwavelength investigations have shown that strong correlations exist between the two. High-luminosity radio galaxies are often associated with large emission-line nebulae, some extending >100 kpc (300,000 ly). For redshifts greater than -0.5, these giant emission-line regions are found exclusively along the axis defined by the double-lobed structure of the radio source (see Fig. 1). These emission-line regions are extremely luminous and thus require a large input of energy to maintain their highly ionized state. One potential source of energy is the active nucleus itself. A number of researchers have recently suggested that high-redshift radio galaxies are closely linked to radio-loud quasars and may only differ in their angle to the line of sight to the observer, quasars being the pole-on subset of radio galaxies. A more striking correlation between radio and optical properties at high redshifts concerns the rest-frame ultraviolet continuum, which is believed, but not conclusively proven, to be the light of massive young stars. Powerful radio galaxies with z>1 often have ultraviolet continuum emission distributed in highly lumpy linear structures extending 30 kpc or more (see Fig. 2). These elongated continuum structures are in turn closely aligned with their radio lobes. While the origin of this radio/optical alignment is unclear, it is strong evidence that the high-redshift radio galaxies are qualitatively different from present-day massive galaxies (often identified as the descendants of powerful radio galaxies) in a fundamental manner. Thus, by selecting galaxies on the basis of their radio emission, we have found a population of objects that are not likely to be representative of most galaxies, either now or in the distant past. All hope of learning about galaxy evolution and formation from these unusual objects is not lost. The most natural part of the electromagnetic spectrum in which to study high-redshift galaxies is the near infrared (wavelengths, ***1-3 **). As we observe galaxies at higher and higher redshifts from our fixed visual window we observe them deeper into the rest-frame ultraviolet, where the light is dominated by the contribution from a minority population of massive stars. We must shift our observing window to the infrared to examine the bulk of the stars, which emit their light in the rest-frame visual and near infrared. Observers on Mauna Kea in Hawaii have found that most distant radio galaxies have well-developed old stellar populations, even at redshifts corresponding to ages of only a few x 10* years. This is strong evidence that these galaxies formed at very high redshifts, greater than 5 or 10, depending on the value of the deceleration parameter qo. Thus these highly unusual objects tell us that there were massive stellar systems very early in the history of the Universe. If this same conclusion could be shown to hold for the majority of massive galaxies, then certain cosmological models, the cold dark matter model in particular, would be in serious difficulty. The matter is not so simple. The latest generation of infrared arrays have allowed us to obtain images in the rest-frame visual and near infrared. The images show that the light at long wavelengths, believed to be the light of old stars, is also strongly aligned with the radio axis. This alignment is not seen in present-day radio galaxies, and its implications are not yet clear. THE FUTURE OF THE DISTANT PAST The future of high-redshift galaxy research is likely to lie with objects that are less extreme than the powerful radio galaxies. The latter have provided us with our first glimpse of galaxies in the early universe and have given us important clues concerning the formation and evolution of galaxies, but they are not well suited to addressing the development of ordinary galaxies. The near infrared, both from the ground and, ultimately, from space, will likely be our most productive window on galaxies and their stellar populations in the distant past. Additional Reading Bergeron, J. and Kunth, D., eds.(1988). High Redshift and Primeval Galaxies. Editions Frontieres, Paris. Chambers, K.C.(1989). Radio galaxies at z>2. In The Hubble Symposium on the Evolution of Galaxies, R.G. Kron, ed. Publ. Astron. Soc. Pacific Conference Series 10 373. Frenk, C. et al., eds.(1989). The Epoch of Galaxy Formation. Kluwer, Dordrecht. McCarthy, P. and van Breugel, W.(1989). Emission line properties of high redshift radio galaxies. In Extranuclear Activity in Galaxies, E.J.A. Meuers and R.A.E. Fosbury, eds. European Southern Observatory, Garching p. 55. Spinrad, H.(1986). Faint galaxies and cosmology. Publ. Astron. Soc. Pac. 98 269. van Breugel, W. and McCarthy, P.(1989). Relations between radio and optical properties in redshift radio galaxies. In Extranuclear Activity in Galaxies, E.J.A. Meuers and R.A.E. Fosbury, eds. European Southern Observatory, Garching p. 227.