In the Magellanic Clouds (MCs) giant loops of H II regions can be recognized in deep pictures taken in H light. These H II structures were divided by Goudis & Meaburn (1978) into two distinct groups: shell structures called giant shells (GSs) with diameters of 20- pc and the huge supergiant shells (SGSs) with diameters of 600- pc. Employing unsharp masking techniques and high contrast copying of the long exposed H images of Davies et al. (1976, hereafter DEM), Meaburn and collaborators (see Meaburn 1980) subsequently identified 85 GSs and 9 SGSs in the Large Magellanic Cloud (LMC) and 1 SGS in the Small Magellanic Cloud (SMC).
Unlike GSs whose stucture can be explained by the combined activity of supernova explosions, stellar winds and radiation pressure of the central star association(s), SGSs (i.e. structures about kpc in diameter) need very effective, gigantic mechanisms for their creation. Such large scale features, which rival in size only with spiral structures, might be created, according to the appraisal of mechanisms by Tenorio-Tagle & Bodenheimer (1988), by the collision of high velocity clouds (HVCs) with the disk of the galaxy or by stochastic self-propagating star formation (SSPSF).
The infall of an HVC would cause a well defined velocity deviating from the velocity of the original disk gas. Also, because of the short time scale of such a collision, an almost identical age is expected for all stars whose formation was triggered by that event.
In the case of SSPSF (see Feitzinger et al. 1981) star formation propagates from one point (e.g. the centre of a SGS) in all directions (e.g. towards the rim of a SGS). Thus one should see in projection two velocity components in the H I layer (a receeding and an approaching component) as well as a clear age gradient of the star populations inside a SGS.
However, the observation of a central area of a SGS with little gas (i.e. a 'hole' in the H I layer) and of an approximately pc thick shell of neutral material, which is ionized at the inner edge (visible as H filaments) by the radiation from the associations containing numerous early type stars (see Lucke & Hodge 1970; Lucke 1974) in the centre of the SGS, would be consistent with both scenarios.
The importance of understanding the formation and the structure of SGSs is evident if one realizes how big they are. They may dominate large portions of a galaxy, in particular of the smaller irregular galaxies, such as the Magellanic Clouds are. Knowledge about the creation of SGSs leads to a better knowledge of the more recent development of the MCs and of their youngest star populations.
The shape and size of LMC 4 is such, that SSPSF has been considered as the most likely explanation for the formation of this SGS (see Dopita et al. 1985). However, observations of star groups along the edge of LMC 4 showed that the ages of these groups were of the same order as the time needed to create the SGS in the case of SSPSF (see Vallenari et al. 1993; Petr et al. 1994).
Reid et al. (1987) found from V, I photometry of Shapley Constellation III (the southern half of the SGS LMC 4, McKibben Nail & Shapley 1953) no clear age gradient, also inconsistent with a global SSPSF model.
Furthermore, Domgörgen et al. (1995) presented an investigation of LMC 4 based on H I data and IUE spectra. One result is that there is only one distinct velocity component towards us with and just a diffuse rear component. This is not consistent with an undisturbed expanding shell, but it indicates a break-out of the SGS at the back side of the LMC.
So LMC 4 resembles a cylinder rather than a sphere which should be expected in a galaxy with an H I scale height well below pc. We note, however, that it is notoriously difficult to derive depth structure in a reliable way.
All these points show that we still do not well understand the history of LMC 4. To improve this situation, photometry of star groups inside LMC 4 was taken in 1993 with the goal to derive ages. This paper presents that data and the results. Additionally the star formation history is further investigated in Sect. 4by looking for overlapping age groups in the derived mass functions.
© European Southern Observatory (ESO) 1997
Online publication: March 24, 1998