Astron. Astrophys. 346, 359-368 (1999)

1. Introduction

The identification and study of very high-redshift galaxies () is probably one of the most direct methods to constrain the scenarios of galaxy formation and evolution. The observation of such galaxies presents a major challenge complementing the statistical studies of lower redshift galaxies (CFRS: Lilly et al. 1996; Hawaii Deep Fields: Cowie et al. 1996; HDF: Sawicki et al. 1997; Lowenthal et al. 1997; Steidel et al. 1996b). The key information to retrieve is the star formation history in the universe and the evolution of the different morphological types of galaxies. The main problems are the identification of high-z galaxies and the construction of a sample as bias-free as possible. Steidel & Hamilton (1992, 1993) and Steidel et al. (1996a, 1998) first used the Lyman dropout technique to photometrically identify galaxies in empty fields, and they succeeded in confirming spectroscopically a large sample of them with the Keck telescope. A similar technique has been used on the HDF to identify galaxies (Steidel et al. 1996b; Lowenthal et al. 1997) also confirmed with the Keck telescope. The number of massive star-forming galaxies at high redshift can be used to constrain the cosmological models (see Baugh et al. 1998). All these galaxies are in general very compact in their restframe UV, with the bulk of their star formation located on high surface-brightness regions. Their morphologies might give some hints on the physical processes involved in galaxy evolution. In particular, they might provide useful constraints on the formation of the different systems (spheroids, bulges, disks), and its close relation with metal enrichment timescales (see Trager et al. 1997 for a detailed discussion). At 3 to 4, the corresponding lookback time is 13.7 to 14.4 Gyr, 15 Gyr being the present age of galaxies ( km s^-1 Mpc^-1, and ). Thus, such high redshift systems provide a direct measure of the early star-formation processes.

The photometric technique based on the identification of Lyman dropouts in the U-band has shown to be a useful tool to select galaxies, and it has been successfully extended to the B-band to locate candidates (Steidel et al. 1998). Nevertheless, two major biases will appear in all such selection techniques: one towards intrinsically luminous galaxies because the sample is limited in apparent magnitude, and another towards galaxies with strong star-formation activity, because the rest-frame UV will be seen at the visible wavelengths. When high-redshift galaxies are selected through photometric redshift techniques based on a large wavelength domain, including near-IR, and close to critical-lines of cluster-lenses, the resulting sample is less sensitive to these biases. This paper presents the first results on this original method to build-up and study an independent sample of galaxies, by combining photometric redshifts (visible and near-IR) and lens modelling. One of the first examples of such a technique on lensed galaxies is the star-forming object in A2218 (Ebbels et al. 1996), and a number of recent discoveries of lensed-galaxies in clusters, sometimes serendipituous, strongly encourages this approach (Yee et al. 1996a; Trager et al. 1997; Franx et al. 1997; Soifer et al. 1998; Seitz et al. 1998; Frye & Broadhurst 1998; Bunker et al. 1998).

The two lensed galaxies discussed in this paper were identified in the cluster Abell 2390 (), a gravitational lens which has been extensively studied. The first detailed photometric and spectroscopic surveys on this target were performed by Le Borgne et al. (1991), and recently enlarged by the CNOC group (Yee et al. 1996b; Abraham et al. 1996; Carlberg et al. 1996). This cluster, which is known as a strong X-ray emitter (Ulmer et al. 1986; Pierre et al. 1996), has an elongated shape (, based on the luminosity map of the photometrically selected cluster-galaxies, Pelló et al. 1991) and a high velocity-dispersion (1090 km s^-1, Abraham et al. 1996). A near-IR photometric survey is also available (Miralles et al. 1999, in preparation). We will focus here on two multiple images at , namely H3 and H5, according to the nomenclature by Kneib et al. (1999, see also Fig. 1). The redshift of the former had already been confirmed by Frye & Broadhurst (1998) at the Keck telescope.

Fig. 1. Image of Abell 2390 in logarithmic scale, obtained by coadding the final HST images in and . A zoom on H3 and H5 is also shown, with the identification of the different components. The critical lines at 1, 2.5 and 4 (from the inner to the outer part, respectively) are also displayed, according to the model by Kneib et al. (1999).

A plan of the paper follows. First, we give a summary of the spectroscopic and photometric data in Sect. 2. Sect. 3 is devoted to the study of the main properties of H3 and H5, namely the spectroscopic and photometric redshifts, the lensing properties and the morphology of these objects. The stellar population of these objects is characterized in Sect. 4. In Sect. 5 these results are discussed, with a special attention to the implications for galaxy formation and evolution. Throughout this paper, we assume km s^-1 Mpc^-1 and .

© European Southern Observatory (ESO) 1999

Online publication: May 21, 1999
helpdesk.link@springer.de