The Baryonic Tully-Fisher relation revisited

The Baryonic Tully-Fisher relation (BTF) can be substantially improved when considering that the galactic baryonic mass is likely to consist not only of the detected baryons, stars and gas, but also of a dark baryonic component proportional to the HI gas. The BTF relation is optimally improved when the HI mass is multiplied by a factor of about 3, but larger factors up to $11{-}16$ still improve the fit over the original one using only the detected baryons. The strength of this improved relation is quantified with up-to-date statistical tests such as the Akaike Information Criterion or the Bayesian Information Criterion. In particular they allow us to show that supposing a variable $M_\star/L$ ratio instead is much less significant. This result reinforces the suggestion made in several recent works that mass within galactic disks must be a multiple of the HI mass, and that galactic disks are substantially, but not necessarily fully, self-gravitating.