1 Introduction

Blue Compact Dwarf galaxies present striking properties, a UV-excess in the continuum produced by a very young integrated stellar population, as well as a strong nebular emission spectrum similar to H II regions. (Terlevich et al. 1991; [Storchi-Bergmann et al.1995]). The generally high surface brightness of their star-forming regions and their small intrinsic sizes have led observers to call them "compact'' objects. Low luminosity (dwarf) systems have absolute magnitudes fainter than -18 ([Thuan & Martin1981]) or -16.5 ([Zwicky1971]), depending on the luminosity cutoff arbitrarily defined by the authors, and extending to -12. BCDGs appear to have little dust and low metallicity compared to giant starburst galaxies (from 1/6 to 1/50 $Z_{\odot}$,[Kunth & Sargent1986]; [Thuan et al.1997]). They are experiencing a strong star formation episode that will exhaust their gas content in a few 10⁸ years if no replenishment occurs.

Hypothetical evolutionary schemes ([Burkert1989]; Kruger et al. 1994) place them between the slowly evolving gas rich Magellanic dwarf irregulars (dIs) and evolved gas-free dwarf ellipticals (dEs), implying that dIs could be the quiescent phase of the BCDGs, and dEs the evolved results. More recently, [Sung et al.(1998)] have argued that BCDGs and dEs have similar intrinsic shapes, i.e. their axis ratios show a similar distribution. Although this assumption presents the advantage of drawing an unified scheme of the evolution of low luminosity galaxies, it is still confronted with serious problems related to the differences in light distribution and dynamics between the three classes of dwarfs ([Van Zee et al.1998]).

Early studies by Loose & Thuan (1986a and b), and Kunth et al. (1988) have shown that the morphology of BCDGs varies greatly from galaxy to galaxy. From morphology considerations, Loose & Thuan (1986b) have proposed the following classification where a large majority is composed of analogs to the quiescent dwarf ellipticals and dwarf irregulars, a- with regular isophotes like elliptical dwarf galaxies (dE), b- with a bright central star forming region like the nucleated dE (n, dE), c- with very irregular isophotes like the dwarf irregulars (iI, with cometary (C), and merger (M) as peculiar subdivisions). Finally, a small fraction of the BCDGs have been classified as i0-BCDGs because they are unresolved without detected low surface brightness underlying compoment.

This morphological classification has been proposed regardless of a quantitative determination of the surface brightness distribution whereas the dynamics of the starburst and its host galaxy need detailed investigations.

Recent photometric studies have shown that the BCDGs appear to follow a somewhat simpler classification: in our previous paper, (Paper I) we report that among the 23 BCDGs observed, 1/3 are classified as spheroidal analogs (showing r^1/4 law dominated light profiles), 1/3 are classified as exponential law dominated galaxies, and 1/3 have composite light profiles. The fact that these galaxies have similar components as normal giant galaxies suggests that they may have a similar dynamical history, related to their star formation history. Dynamically, BCDGs seem to be complicated systems, as suggested by detailed surface photometry ([Papaderos et al.1996a]; [Telles et al.1997]; [Telles & Terlevich1997]). Some BCDGs present signatures of interactions and merging, although very few are known to have companions or to be found in close groups ([Pustil'niketal.1995]).

Our goals are to study the intrinsic properties of BCDGs derived from their light distribution and their structure; we have searched for the areas of star formation, for the presence of disks, spheroids, or non-axisymmetric components, for traces of possible interactions, and for color gradients. Also, we have searched for relationships between the photometric parameters and the spectroscopic indicators of galactic evolution (e.g. the metallicity of the ionized gas, the ages of the starburst and its possible host galaxy); finally, we have studied the environment of the BCDGs by searching for interacting neighbours (galaxies or clouds). This paper presents the results of the second part of the imaging and photometric observing program at visible wavelengths, whose first part was reported in Paper I. This program is coupled to another one at high spectral resolution, using multi-pupil spectroscopy, of the continuum and emission lines of BCDGs. Near-infrared images of the BCDGs presented in this paper have also been obtained ([Doublier et al.1999], in preparation).