We compare the observed radial velocity of different arm tracers, taken near the tangent to a spiral arm. A slight difference is predicted by the density wave theory, given the shock predicted at the entrance to the inner spiral arm. In many of these spiral arms, the observed velocity offset confirms the prediction of the density wave theory (with a separation between the maser velocity and the CO gas peak velocity, of about 20 km/s)-when the observed offset is bigger than the error estimates.