Introduction

What did galaxy clusters look like ten billion years ago? Simulations (eg. Brainerd & Villumsen 1994, Jenkins et al. 1998, Cen 1998) suggest that the ancestors of present day rich galaxy clusters contained very little mass at redshifts above two: there simply hadn't been time enough to assemble much dark or baryonic matter. Even the richest proto-clusters would only have had about double the average density of the universe. These proto-clusters would typically have been filamentary in shape, and due to their small masses would not have been in virial equilibrium.

Despite their modest overdensities, these protoclusters could have been the sites of the first galaxy formation. If galaxies had formed in these proto-clusters but not in the field, the overdensity of galaxies within the proto-cluster would be much greater than the overdensity of mass (biasing, eg. Fry 1996, Bagla 1998, Baugh et al. 1999, Tegmark & Peebles 1998).

The observational situation is less clear. Galaxy clusters are now routinely being studied out to redshifts . These clusters are remarkably similar to low redshift clusters: they are massive virially bound objects with strong X-ray emission (eg. Rosati et al. 1998, Deltorn et al. 1997). Thus galaxy clusters appear to be well established by redshift one (see also Renzini 1997, Donahue & Voit 1999).

Observations at higher redshifts still are sparse. It is now clear that galaxies at z>2 are at least as strongly clustered as galaxies today (eg. Heisler, Hogan & White 1989, Quashnock, Vanden Berk & York 1996, Malkan, Teplitz & McLean 1996, Steidel et al. 1998, Giavalisco et al. 1998, Campos et al. 1999, Pascarelle, Windhorst & Keel 1998, Adelberger et al. 1998, Djorgovski et al. 1999), as predicted by the biasing model. The structure, mass and galaxy populations of these galaxy concentrations are, however, obscure.

A growing body of evidence suggests that high redshift clusters contain substantial quantities of gas. The strongest evidence comes from high redshift radio galaxies, which are believed to sit in dense cluster environments (eg. Pentericci et al. 1997, Ivison et al. 2000). Many show high rotation measures (Carilli et al. 1997, Pentericci et al. 2000), possibly extended X-ray emission (Carilli et al. 1998), extensive emission-line halos (Bicknell et al. 2000) and associated absorption lines (Binette et al. 2000), suggesting the presence of dense inhomogeneous gas around the galaxies.

Are the clusters around radio galaxies typical? The evidence for gas in other high redshift clusters is sparse. Francis et al. (1996) and Steidel et al. (1998) both found QSO absorption lines seemingly coincident with high-z clusters. Tentative Sunyaev-Zel'dovich effect decrement measurements have been claimed around other possible high-z clusters (eg. Campos et al. 1999). 100 Kpc scale diffuse `blobs' of Ly emission have been found in the centres of two high-z clusters (Francis et al. 1996, Francis, Woodgate & Danks 1997, Steidel et al. 2000).

Most of these observations suggest the presence of hot ionised gas in the central few hundred kpc of the clusters. Francis & Hewett (1993), however, presented tentative evidence for the existence of large concentrations of neutral gas extending over Mpc scales in the high redshift universe. One of their concentrations was associated with a group of galaxies (Francis, Woodgate & Danks 1997). Their evidence was not conclusive, and there is currently no such data available for other clusters, but it is at least possible that similar extended halos of neutral gas surround all high redshift clusters.

In this paper, we concentrate on one of Francis & Hewett's clusters: the 2142-4420 galaxy concentration at redshift 2.38. In Section 2 we present new data on this cluster, strengthening the case for a large halo of neutral gas surrounding it. In Section 3 we investigate ways of reconciling high neutral gas masses with cosmological models. In Section 4 we describe our attempt to observe diffuse Ly emission from the hypothesised gas halo, and place upper limits on its intensity. Except where stated, we assume

, and an open universe with

and either 0 or 0.8 (both values of give the same angular size distance and luminosity distance at redshift 2.38).

© Copyright Astronomical Society of Australia 1997