Introduction

Although the neutral (see van Zee et al. 1998 and references therein) and molecular hydrogen (Young & Knezek 1989; Israel, Tacconi & Bass 1998) are quite abundant in Blue Compact Dwarfs and dwarf irregulars, they are not enough to reproduce the flattening of the rotation curve. Like in spiral galaxies, the existence of this flattening in the rotation curve of dwarf galaxies has been related with the presence of large amounts of dark matter in galaxy outer regions (Carignan & Freeman 1988; Carignan & Beaulieu 1989; Broeils 1992). The dark matter content derived indicates that dark matter is even more abundant in dwarfs that in more massive galaxies (see Moore 1994 and references therein). In fact, standard cold dark matter (CDM hereafter) models predict that low-mass halos are denser than more massive systems, because their higher formation redshift (Navarro, Frenk & White 1997, NFW hereafter). The density profiles of the simulated CDM halos fall with radius as r^-2. This is density profile expected for a flat rotation curve body. The competition between the dark matter and the stellar mass components within the optical radius difficults the analysis of solid-body portion of the rotation curve. Several works have argued that dark matter in dwarf galaxies dominates the total mass density profile even within their optical radius (Carignan & Beaulieu 1989; Broeils 1992), showing a constant density dark matter core (Moore 1994; Flores & Primack 1994; Salucci & Persic 1997). On the other hand, Lo, Sargent & Young lo and Staveley-Smith, Davies & Kinman ss deduced reasonable virial mass to blue light ratios, M_V/L_B<7M $_{\sun}$ /L $_{\sun}$ , for a large fraction of their samples. Loose & Thuan loose86 found that the virial mass of Haro 2 can be reproduced just adding the stellar and HI mass components. Also the study of Swaters swaters of the rotation curves of 44 dwarf galaxies indicates that the mass of a large fraction of these galaxies could be dominated by the stellar component, even at distances larger than three disk scale lengths.

One of the main sources of uncertainty in all these studies is the mass-to-light ratio adopted for the stellar component (Meurer, Staveley-Smith & Killeen 1998; Swarters 1998). Therefore, high quality optical and near-infrared imaging and spectroscopy in order to obtain the physical parameters of the stellar populations and derive reasonable mass-to-light ratios is mandatory to prevent this inconvenient which is inherent to this kind of kinematical studies.

**Figure:** Long-slit positions superimposed on the Mrk86 B-band image. Close to the region B, size and position of the Mrk86-B bubble H $\alpha$ lobes are shown (see Figures a & b). Relative coordinates (in arcsec) are referred to the r-band outer isophotes centre.
$\begin{figure}\epsfxsize=80mm \epsfysize=80mm \epsffile{fig1.ps} \par\end{figure}$

Superimposed on the regular solid-body portion of the velocity field, peculiar motions of the ionized gas have been observed in many star forming dwarf galaxies (Tomita et al. 1997; Petrosian et al. 1997). They have been commonly explained as infalling motions of HII regions (Saito et al. 1992), multiple clouds merging (Skillman & Kennicutt 1993) and local peculiar gas motions induced by violent star formation events (Petrosian et al. 1997). Very high star formation rates associated with these intense star forming events have been demonstrated to be able to produce a cavity of shock-heated gas due to the energy input provided by supernovae and stellar winds (Chevalier & Clegg 1985; Vader 1986, 1987). This hot gas will accelerate the ambient interstellar medium resulting in a collective supernova-driven wind. In fact, several galactic supernova-driven winds phenomenae have been found to be associated with violent star formation places in dwarf galaxies (Roy et al. 1991; Izotov et al. 1996; Martin 1996, 1998, CM98 hereafter). They have been generally detected as holes in the neutral hydrogen distribution (Puche et al. 1992; Brinks 1994), bubbles or shells in H $\alpha$ emission (Marlowe et al. 1995, MHW hereafter) or from their hot gas X-ray emission (Bomans et al. 1997). The existence of these phenomenae could produce the loss of a significant fraction of the galaxy interstellar medium. Depending on the final destination of the accelerated gas, these structures could produce no mass loss, blow-out, only affecting the galaxy chemical evolution, or blow-away processes, with a significant loss of interstellar mass (Young & Gallagher 1990; CM98; Mac Low & Ferrara 1998). Consequently, these supernova-driven galactic winds are accepted to be a key parameter in the dwarf galaxy formation (Silk et al. 1987; Mori et al. 1997) and evolution (MHW; Mac Low & Ferrara 1998).

**Figure:** Interpolated 2D velocity field. Only regions with $\Sigma_{\mathrm{H}\alpha}>$ 1.5 $\times10^{-16}$ ergs^-1cm^-2arcsec^-2 are shown.
$\begin{figure}\epsfxsize=80mm \epsfysize=80mm \epsffile{fig2.ps} \par\end{figure}$

Blue Compact Dwarf galaxies, with intense recent or ongoing star forming activity, are those systems where the interplay between star formation and the interstellar medium is more feasible to be studied. However, although the majority of the BCD galaxies are iE type BCDs ( $\simeq$ 70 per cent; Thuan 1991), with star formation spreading over the whole galaxy, the effects of the supernova-driven winds have been mainly studied in dwarf amorphous galaxies (see, e.g. MHW), which show nuclear star forming activity.

The galaxy Mrk86=NGC 2537 (Shapley & Ames 1932; Markarian 1969), also known as Arp 6 (Arp 1966), constitutes an excellent laboratory to test the properties and effects of the supernova-driven winds on the interstellar medium of dwarf galaxies, as a nearby prototype of the iE BCD galaxies class.