Interaction of galaxies is a way to reveal extended gas around galaxies, that could stay cold and dark in quiescent phases. The hypothesis that the baryonic dark matter around galaxies could be cold H2 clouds has been developped in recent years, in different models. The cold clouds could be rotationally supported, orbiting in the outer parts of the galaxies in a flaring plane, associated to the HI gas (Pfenniger et al 1994, Pfenniger & Combes 1994). They are distributed in a fractal structure, with the smallest fragments (the clumpuscules) having a mass M ~ 10-3 M, a molecular density nH2 ~ 1010 cm-3, and a radius R ~ 10 AU. Alternatively, the cold H2 gas could be associated to brown dwarfs in clusters, distributed in the galactic halo, or spheroidal component (de Paolis et al 1995; Gerhard & Silk 1996). Their temperature would then be higher (T = 20K).
These clumps are seen in ESE (extreme scattering events) in front of quasars and in HI - VLBI in absorption against extended radio-sources (Fiedler et al 1987, Walker & Wardle 1998, Faison et al 1998). The statistics of observations are compatible with a large number of these clumps in the Galaxy, so that the total mass could be comparable to that of the visible Milky Way.
The existence of this cold H2 gas has also been invoked to explain the extra- rays detected by EGRET, on board GRO (de Paolis et al. 1999, Sciama 1999) and some observational consequences have been studied (Combes & Pfenniger 1997, Shchekinov 1999). Cooling flows in clusters could also form these clumpuscules (Ferland et al 1994).
Let us recall that the visible matter in the Universe corresponds to only vis ~ 0.003, while the Big Bang nucleosynthesis implies that the baryonic density is b ~ 0.01 h-2, or that 0.01 < b < 0.04. Most of the baryons are therefore dark, and the MACHOs cannot be a significant part of them, according to the microlensing experiments.