2007A&A...471..103B

CO observations have been the best way so far to trace molecular gas in external galaxies, but in low metallicity environments the gas mass deduced could be largely underestimated due to enhanced photodissociation of the CO molecule. Large envelopes of H₂ could therefore be missed by CO observations. At present, the kinematic information of CO data cubes are used to estimate virial masses and trace the total mass of the molecular clouds. Millimeter dust emission can also be used as a dense gas tracer and could unveil H₂ envelopes lacking CO. These different tracers must be compared in different environments. This study compares virial masses to masses deduced from millimeter emission, in two GMC samples: the local molecular clouds in our Galaxy (10⁴-10⁵M_☉), and their equivalents in the Small Magellanic Cloud (SMC), one of the nearest low metallicity dwarf galaxies. In our Galaxy, mass estimates deduced from millimeter (FIRAS) emission are consistent with masses deduced from gamma ray analysis and therefore trace the total mass of the clouds. Virial masses are systematically larger (twice on average) than mass estimates from millimeter dust emission. This difference decreases toward high masses and has been reported in previous studies. This is not the case for SMC giant molecular clouds: molecular cloud masses deduced from SIMBA millimeter observations are systematically higher (twice on average for conservative values of the dust to gas ratio and dust emissivity) than the virial masses from SEST CO observations. The observed excess cannot be accounted for by any plausible change of dust properties. Taking a general form for the virial theorem, we show that a magnetic field strength of ∼15µG in SMC clouds could provide additional support for the clouds and explain the difference observed. We conclude that masses of SMC molecular clouds have so far been underestimated. Magnetic pressure may contribute significantly to their support.