One conclusion of this chapter will be that the very "first" galaxies have almost certainly not yet been observed. But in recent years we have undoubtedly witnessed an observational revolution in the study of early galaxies in the young Universe which, for reasons outlined briefly below, I have chosen to define as corresponding to redshifts z > 5 (a good, up-to-date overview of the physical properties of galaxies at z = 2-4 is provided by Shapley 2011).
The discovery and study of galaxies at redshifts z > 5 is really the preserve of the 21st century, and has been one of the most spectacular achievements of astronomy over the last decade. From the ages of stellar populations in galaxies at lower redshifts it was known that galaxies must exist at z > 5 (e.g. Dunlop et al. 1996), but observationally the z = 5 "barrier" wasn't breached until 1998, and then only by accident (Dey et al. 1998). Although this discovery of a Lyman-α emitting galaxy at z = 5.34 was serendipitous, it in effect represented the first successful application at z > 5 of the long-proposed (e.g. Patridge & Peebles 1967a, b) and oft-attempted (e.g. Koo & Kron 1980; Djorgovski et al. 1985; Pritchet & Hartwick 1990; Pritchet 1994) technique of searching for "primeval" galaxies in the young universe on the basis of bright Lyman-α emission. This discovery was important not just for chalking up the next integer value in redshift, but also because this was the first time that the redshift/distance record for any extra-galactic object was held by a "normal" galaxy which had not been discovered on the basis of powerful radio or optical emission from an active galactic nucleus (AGN). Later the same year, two more galaxies selected at z > 5 on the basis of their starlight (Fernandez-Soto et al. 1999) were spectroscopically confirmed at z = 5.34 by Spinrad et al. (1998), and the Lyman-α selection record was advanced to z = 5.64 (Hu et al. 1998).
In this chapter I will explain how these breakthroughs heralded a new era in the study of the high-redshift Universe, in which conceptually simple but technologically challenging techniques have now been successfully applied to discover thousands of galaxies at z > 5, and to extend the redshift record out to z ≃ 9. The key instrumental/observational advances which have facilitated this work are the last two successful refurbishments of the Hubble Space Telescope (HST; first with the ACS optical camera, and most recently with the near-infrared WFC3/IR imager), the provision of wide-field optical and near-infrared imaging on 4-8-m class ground-based telescopes (Suprime-Cam on 8.2-m Subaru telescope, WFCAM on the 3.8-m UK InfraRed Telescope (UKIRT), and ISAAC/Hawk-I on the 8.2-m Very Large Telescope (VLT)), the remarkable performance of the 85-cm Spitzer Space Telescope at mid-infrared wavelengths, and finally the advent of deep red-sensitive optical spectroscopy on the 10-m Keck telescope (with LRIS & DEIMOS), the VLT (with FORS2), and on Subaru (with FOCAS).
I will also endevour to summarize what we have learned about the properties of these early galaxies from this multi-frequency, multi-facility investigation and, as a result, what new information we have gleaned about the evolution of the universe during the first ≃ 1 Gyr of cosmic time. I conclude with a very brief discussion of the prospects for further progress over the next decade; a more detailed description of future facilities is included elsewhere in this volume.
The cosmological parameters of relevance to this work are summarized briefly in the next section. Where required, all magnitudes are reported in the AB system, where mAB = 31.4 - 2.5log(fν / 1 nJy) (Oke & Gunn 1983).