Numerical models predict that TDGs, or at least the more massive of them, should be supported by rotation. Optical slit spectroscopy of numerous TDG candidates has been carried out [29, 18], showing strong velocity gradients, consistent with rotation, but also with artificial slit effects . Integral field spectroscopy data , as well as high resolution HI and CO datacubes [15, 5] are available for several TDGs and confirm that TDGs rotate, though at velocities much lower than measured with slit spectroscopy. For the TDGs located in the collisional ring of NGC 5291, the data quality was sufficient to allow a determination of their total mass, and comparing it with the luminous mass (HI, H2, stars), of their dark matter content. As predicted by numerical simulations and early estimates of the dynamical mass based on the millimeter CO line width , the inferred M/L is much lower than in regular, dark matter dominated dwarf galaxies. However an unanticipated mismatch by a factor of 2-3 between the dynamical and luminous mass has been noticed . A similar discrepancy was found for the TDG candidate VCC 2062, in the Virgo cluster , and more recently in one of the TDGs hosted by the prototypical advanced merger NGC 7252. The results of this latter study are illustrated in Fig. 3.
Figure 3. Prototypical merger NGC 7252. The moment maps of the HI towards the TDG candidate (top), as well as its position-velocity diagrams are consistent with rotation (Belles et al. in prep.)
Various hypotheses have been put forward to account for the missing mass in TDGs. Cosmological dark matter accreted from disrupted satellite galaxies  might be present in the disk of spirals and thus in TDGs. Theory of modified gravity, such as MOND, predicts rotation curves for TDGs similar to the observed ones [13, 20]. An alternative idea is that spirals disks contain dark baryons, for instance in the form of very cold molecular gas not accounted for by CO observations . The presence of dark gas clouds in the Milky Way had been inferred using gamma rays . In the far-infrared domain, Planck is also making a census of the molecular component not mapped by standard tracers. Whether it may explain entirely the missing mass observed in TDGs is still an open question.