1 Introduction

Neutral gas at high redshifts is easiest to detect in absorption against bright background sources. Not surprisingly, most of what we know about the content and evolution of neutral gas in the universe comes from the study of absorption lines seen in the spectra of distant QSOs. The number density (per unit redshift) of these absorption lines is a strong function of the column density, with low column density ( $N_{\rm HI} \sim 10^{13}$ atoms cm^-2) systems being several orders of magnitude more common than high column density ( $N_{\rm HI} \ga 10^{20}$ atoms cm^-2) systems. Nonetheless, the bulk of the neutral gas at high redshift is contained in these rare high HI column density absorbers. It is principally for this reason that these objects (called damped Lyman-$\alpha$  systems or DLAs) are obvious candidates for the precursors of today's galaxies. Further, the gas mass in DLAs at $z \sim 3$ is comparable to the stellar mass in galaxies at z=0 (e.g. Storrie-Lombardi et al. 1996; Storrie-Lombardi & Wolfe 2000), consistent with the idea that the absorbers have converted their gas into stars in the intervening period. Understanding the nature of DLAs at different redshifts is clearly important in tracing the evolution of galaxies; for this reason, the absorbers have been the subject of considerable study over the last two decades.

Unfortunately, since QSOs are point sources, optical absorption studies alone are unable to constrain the transverse size of the absorbers. The typical size and mass of DLAs have hence long been controversial issues. Traditionally, DLAs have been modelled as large proto-spirals (Wolfe et al. 1986). Some support for this model comes from the shapes and widths of the absorption profiles produced by ions such as SiII (which are associated with neutral HI). The large velocity widths ( $\Delta V \sim 300$ km s^-1) and the asymmetric shapes of these lines have been successfully modelled as arising from gas in a thick spinning disk (Prochaska & Wolfe 1997, 1998). However, models involving infall or random motions of smaller gas clouds have also been found to succesfully reproduce the observed velocity profiles (Haehnelt et al. 1998; McDonald & Miralda-Escude 1999).

At low redshifts, the galaxies in which the damped absorption arises can be directly studied by ground-based or HST observations. Contrary to expectations, low redshift DLAs appear to be associated with a wide variety of galaxy types, including dwarf and low surface brightness (LSB) galaxies (e.g. Le Brun et al. 1997; Nestor et al. 2001) and not exclusively (or even predominantly) with spiral galaxies. Besides this, the majority of DLAs tend to have low metallicities ($\sim$0.1 solar) at all redshifts, with very little evolution in their metallicity with redshift (Pettini et al. 1999); further, they also do not show the expected $\alpha$/Fe enrichment pattern expected for spiral galaxies (Centurión et al. 2000), suggesting a different star formation history than that of spirals. Finally, 21 cm absorption studies (Chengalur & Kanekar 2000; Kanekar & Chengalur 2001) have shown that the majority of DLAs have high spin temperatures ( $T_{\rm s} \sim 1000$ K), far higher than those observed in the Milky Way or local spirals ( $T_{\rm s} \sim 200$ K). All these results indicate that damped absorption is likely to originate in all types of galaxies and not merely in luminous disk systems.

Of course, the above results do not preclude the possibility that the galaxies responsible for the damped absorption are indeed massive gas-rich disks, but have not undergone much star formation and hence have both low luminosities as well as low metallicities. Such low surface brightness systems could well have a larger fraction of the warm phase of neutral HI and hence, a high spin temperature (Chengalur & Kanekar 2000). However, for systems at low redshift, 21 cm emission studies can be used to get direct estimates of the HI mass of the absorbers; one can thus directly test the above hypothesis, that DLAs are massive, low luminosity galaxies. Deep searches for HI emission from two nearby DLAs have resulted in non-detections (Kanekar et al. 2001; Lane 2000). In both cases, the 3 $\sigma$ upper limit to the HI mass is $\sim$1/3 the HI mass of an L_* spiral. Thus, in these two cases at least, the absorption does not arise in an optically faint, but extremely gas rich galaxy.

In this paper, we discuss Giant Metrewave Radio Telescope (GMRT) 21 cm emission observations of a third low redshift DLA, the z=0.009 absorber towards the QSO HS 1543+5921. The damped absorption has been identified as arising in a low surface brightness galaxy SBS 1543+593 (Reimers & Hagen 1998; Bowen et al. 2001a). HI emission has also been detected from this galaxy using the Bonn Telescope (Bowen et al. 2001b). The extremely low redshift of the absorber allows us to make, for the first time, spatially resolved images of the 21 cm emission; besides the HI mass, this also enables us to determine the velocity field of the galaxy and, hence, to estimate its dynamical mass.

The GMRT observations are detailed in Sect. 2, while the results are presented in Sect. 3.1 and discussed in Sect. 3.2. Throughout the paper, we use a Hubble constant of $H_{\rm0}=75$ km s^-1 Mpc^-1, i.e. a distance to SBS 1543+593 of 38 Mpc.