1. INTRODUCTION

Wolf-Rayet (WR) galaxies are extragalactic objects whose integrated spectra show direct signatures from WR stars, most commonly a broad HeII 4686 feature originating in the stellar winds of these stars. Since the first detection of such a feature in He 2-10 (Allen et al. 1976), a large number of WR galaxies have been reported, some in systematic searches (e.g. Kunth & Joubert 1985), but mostly serendipitously. For example many objects with WR features have been found in samples of high S/N spectra of low metallicity extra-galactic HII regions aimed at deriving the primordial He abundance (cf. Izotov & Thuan 1998). It must be reminded to use the term WR "galaxy" with caution. Depending e.g. on the distance of the object and the spatial extension of the observation, the region of concern may be "just" a single extra-galactic HII region with a few WR stars in a galaxy or the nucleus of a powerful starburst galaxy harbouring numerous massive stars.

Since the compilation of Conti (1991) listing 37 objects, the number of known WR galaxies has grown rapidly to more than 130 in the present catalogue. Interestingly many objects are now found showing additional features from WR stars in their spectra. E.g. the broad emission lines of NIII 4640 and/or CIII 4650 as well as CIV 5808, among the strongest optical lines in WN and WC stars, are increasingly often being detected. Lines originating from WC stars (representing more evolved phases than WN stars) provide useful independent and complementary information on the massive star content in these regions (e.g. Schaerer et al. 1999).

By definition it is not surprising that WR galaxies do not form a homogeneous class. Indeed, WR galaxies are found among a large variety of morphological types, from low-mass blue compact dwarf (BCDs) and irregular galaxies, to massive spirals and luminous merging IRAS galaxies. Recent studies also quite convincingly show the evidence of signatures from WR stars in Seyfert 2 and LINERs (Osterbrock & Cohen 1982, Ho et al 1995, Heckman et al. 1997, Storchi-Bergmann et al. 1998, Kunth & Contini 1998). Allen (1995) claims even the possible detection of WR stars in central cluster galaxies of two cooling flows out to a redshift of z ~ 0.25.

Empirically all WR galaxies show nebular emission lines. The absolute scales (absolute magnitudes, ionizing fluxes etc.) of the investigated objects vary greatly; generally speaking the properties of WR galaxies overlap with those of other emission line galaxies and form a continuous extension of giant HII regions (Conti 1991).

For most "traditional" WR galaxies (e.g. HII galaxies, BCDs etc.) the nebular spectrum is likely due to photoionization of stellar origin. However, this statement does obviously not hold in general, e.g. for Seyfert 2 and LINER revealing the presence of WR stars. Among the former "class" a considerable fraction (~ 1/3 in the present compilation) of objects also show nebular HeII 4686 emission in addition to the broad WR feature. This line is also present in some giant HII regions where no WR features have been detected. Except in planetary nebulae, nebular HeII 4686 emission is very rarely found in Galactic HII regions (cf. Garnett et al. 1991, Schaerer 1997). Its origin, requiring sources with sufficient photons of energy > 54 eV, has remained puzzling (see Garnett et al. 1991 and references therein). Supported by quantitative modeling, Schaerer (1996) has suggested that the origin of nebular HeII 4686 emission is intimately linked with the appearance of hot WR stars. To facilitate systematic analysis on the origin of nebular HeII 4686 emission we therefore also include the relevant information on objects showing this line. Such studies have a bearing on our understanding of physical processes in HII regions and related nebulae, the ionizing fluxes of starbursts and their contribution to the ionization of the intergalactic medium etc. (cf. Garnett et al. 1991, Schaerer et al. 1998, Stasinska 1998).

The minimum common property of all WR galaxies is (provided our the origin of the considered line and our understanding of stellar evolution is correct) ongoing or recent star formation which has produced stars massive enough to evolve to the WR stage. This indicates typically ages of 10 Myr and stars with initial masses M_ini 20 M (Maeder & Conti 1994).

WR galaxies are therefore ideal objects to study the early phases of starbursts, determine burst properties (age, duration, SFR), and to constrain parameters (i.e. slope and upper mass cut-off) of the upper part of the initial mass function (see e.g. Arnault et al. 1989, Mas-Hesse & Kunth 1991, 1998, Krüger et al. 1992, Vacca & Conti 1992, Meynet 1995, Contini et al. 1995, Schaerer 1996, Schaerer et al. 1999). Conversely studies of the stellar populations in super star clusters frequently formed in starbursts and WR galaxies (Conti & Vacca 1994, Meurer et al. 1995) can also place constraints on stellar evolution models for massive stars, e.g. at extremely low metallicities, which are inaccessible in the Local Group (cf. I Zw 18: Izotov et al. 1997b, Legrand et al. 1997, de Mello et al. 1998).

As galaxies exhibiting intense star formation are being discovered in large numbers at progressively larger distances, "template" systems become increasingly important for our understanding of distant objects which cannot be studied to the same depth. As such WR galaxies represent useful templates of young starbursts which show close resemblance to recently discovered high redshift galaxies (cf. Leitherer et al. 1996, Ebbels et al. 1996, Lowenthal et al. 1997).

The present compilation should facilitate future systematic studies on some of the issues discussed above. The structure of the paper is as follows. In Sect. 2 we review the searches undertaken up to date for WR galaxies. In Sect. 3 the compilation of all known WR galaxies is presented. Brief remarks on each individual object are given in Sect. 4. The list of extra-galactic HII regions showing only nebular HeII 4686 is given in Sect. 5. Suspected WR galaxies are discussed in Sect. 6. A brief discussion and our main conclusions are found in Sect. 7.