1. INTRODUCTION

The simplest definition of a primeval galaxy (PG) is ``a galaxy viewed during its formation''. The precise meaning of ``formation'' depends, however, on one's vantage point. Perhaps the most exact meaning of ``formation'' is ``assembly of mass'', as Peebles (1989) has emphasized; but mass assembly may be extraordinarily difficult to observe if the only release of energy during this phase is radiation of gravitational binding energy. A more pragmatic definition of formation epoch is that time at which a PG is at or near maximum luminosity. By the time this phase has been reached, a great deal of information about the earliest history of protogalaxies may have already been erased. But, at the present stage of our understanding of PG's, this may be the only period of PG evolution that we have a chance of observing.

Why search for primeval galaxies? There are many gaps in our understanding of galaxy formation (e.g., review by Silk and Wyse 1993); an empiricist's approach to improving our understanding of galaxy formation theory would be simply to search for PG's, and then compare the observed properties of these PG's with theoretical expectations. In other words, studying PG's can provide important constraints on galaxy formation theory. Furthermore, studying forming galaxies in situ provides information on the formation of stars under very different conditions than found locally, and hence may ultimately provide important insight into star formation theory. Finally, the search for PG's may eventually lead to the discovery of objects at z > 5, and studying these objects should provide important insight into the physical processes that obtained at such redshifts.

As will be discussed in Section 2.6, there are many objects now known to possess properties that resemble the predicted characteristics of PG's. However, the problem is not one of locating such objects, which have been recognized in various surveys because of one or more extreme characteristics. Rather, the problem is to find the expected widespread population of galaxies undergoing their first starburst; these objects are expected to be numerous (~ 10^4-5 deg^-2) and relatively bright (apparent magnitude +24 in the AB system ⁽¹⁾), yet detection of this pervasive population has so far eluded detection. This is the basic problem facing observers: where are these objects?

The first part of this review will concentrate on the predicted (and, to a surprising degree, model independent) properties of PG's (Section 2), with special attention to the problem of the redshift of galaxy formation. This paper will then turn to known isolated PG candidates (Section 2.6) and PG models (Section 3). Finally a discussion of PG searches (Section 4), complications (Section 5), and future search strategies (Section 6) will be presented.

Several excellent discussions on primeval galaxies and related topics have appeared over the past few years (Koo 1986; Djorgovski and Thompson 1992; Djorgovski 1992). The reader is also referred to excellent reviews of galaxy formation and related subjects by Cowie (1988), White (1989), Larson (1990, 1992) and Silk and Wyse (1993).