Strong starbursts typically have low ages. Therefore we expect to be able to trace the history of this dramatic event. What triggered the burst? Hunter and Elmegreen (2004) found no support for tidally triggered starbursts in their sample of BCGs. Several strong starburst dwarfs live in isolation but show signs of multiple nuclei, unordered velocity fields, morphological large-scale distortions etc which supports mergers as the most common triggering mechanism (e.g., Östlin et al. 2001). Understanding the triggering mechanism in low mass starbursts is more of a challenge since the turbulent velocities caused by stellar winds are comparable to the gravitational effects and the possible merging components have no prominent central concentrations.
A burning question is how the different types of dwarf galaxies are related and the possible transition from one type to another. This is an ongoing discussion since the 1970s (Searle et al. 1973, Thuan 1983, Loose and Thuan 1986, Davies and Phillipps 1988, Hoffman et al. 1989, Hoffman, Helou, and et al., Drinkwater and Hardy 1991, James 1994, Papaderos et al. 1996, Papaderos et al. 1996, van Zee et al. 2001). It is clear of course that gas rich galaxies in isolation turn into gas poor. But the reverse should not be excluded. There are various possibilities of how this could be achieved except for pure gas consumption in star formation. In starbursts, gas may be expelled as consequence of winds from massive stars and supernovae. Ram pressure stripping as the galaxy falls in towards a cluster with a higher density of the intergalactic medium is another option. Harassment and tidal stirring have also been proposed as important mechanisms. The first step in these studies is to make a comparative investigation of the stellar populations and their distributions in gas-rich vs. gas-poor galaxies.
Papaderos et al. (1996) compare the structural properties of different dwarf galaxy types. The BCDs typically has a central surface brightness 1.5 mag. brighter and a scale length a factor of 2 smaller than dEs and dIs. They conclude that there may be evolutionary connections between BCDs, dEs provided the BCD modifies its structural properties during the transition by a change in the potential field as a result of e.g. a global in- or outflows of gas (Larson 1974). Outflows of gas in galactic superwinds (Heckman et al. 1993) is found in many starburst galaxies (e.g., Israel and van Driel 1990, Papaderos et al. 1994, Marlowe et al. 1995, Lequeux et al. 1995, Izotov et al. 1996, Yang et al. 1996, Kunth et al. 1998, Méndez et al. 1999, Summers et al. 2001, Heckman et al. 2001) and the energetics cope with those needed to expel the gas. In some SFDGs it has been argued that outflows have been efficient mechanisms to remove the gas (e.g., Efstathiou 2000, Grimes et al. 2007). This has been supported by some models of SN driven gas outflows (Dekel and Silk 1986, Mac Low and Ferrara 1999, Efstathiou 2000, Heckman et al. 2001). From a cosmological point of view, dwarf galaxies are remarkably devoid of baryons (McGaugh et al. 2010) and it has been argued that outflows could explain this condition (Silk 2003). E.g. Garnett (2002), Pilyugin et al. (2004), Tremonti et al. (2004) and Dalcanton (2007) all find that the effective yields are lower than expected in a closed-box scenario, supporting an ejection of metals. Dalcanton argues however that the amount of mass lost need not be more than 15%. In the dwarf starburst Mrk 33 Summers et al. (2001) find support for ejection of most of the produced metals and a small amount, 1%, of the ISM. These findings are in conflict with other results (Larsen et al. 2001, Lee et al. 2006) which find support for the closed-box model. A potential problem when investigating the validity of the closed box model is to derive reliable gas masses at low metallicities, in particular molecular hydrogen. The galaxies producing the most metals are also supposed to have the largest amounts of molecular hydrogen and therefore the total gas mass, and effective yield, may be underestimated.
While the question about selective metal outflows is still unsettled there is little support of SN blowouts of large fractions of the ISM both from observations and recent modelling (Recchi and Hensler 2006). No large sinks of gas, either neutral, molecular or ionized are found in the intergalactic medium in groups (e.g., Dahlem et al. 2001, Hoffman et al. 2003) and when outflows are observed, it appears difficult to transfer the gas to the IGM environment (e.g., Strickland and Stevens 1999, Summers et al. 2003). Rather, to explain the remarkably low baryon content in dwarf galaxies, X-ray observations (Dai et al. 2010) seem to favour a baryon loss due to preheating. A second possibility is that large quantities of baryons never was involved in galaxy formation but reside as warm-hot gas in the large scale cosmic walls (Fang et al. 2010). The scatter in the fractional gas/star content however is increasing with decreasing mass so at the lowest masses, SN winds or other mechanisms may play a fundamental role.
So, what are then the dominating mechanisms that can explain a relationship between more massive dwarf galaxies? What will happen to a BCG after a bursting phase? Meurer et al. (1998) find that the DM halos in two BCGs is ~ 10 times higher than in dIs. A high central mass density is found also in Haro 11 (Cumming et al. 2009) which makes it unlikely that BCGs can transform into dIs. It seems more probable that BCGs could develop into dEs (e.g., Papaderos et al. 1996, Östlin et al. 2004, Gil de Paz and Madore 2005, Marquart et al. 2007). An attractive mechanism to reduce the gas content seems to be the scenario of tidal stirring, tested by Mayer et al. (2001) on the local group galaxies. But, while it appears attractive in an group environment it will not work on isolated BCGs so the problem remains.