Do maser stars trace a rotating Galactic bar?

I have analyzed the radial velocities in two large samples of maser stars that cover the Galactic plane from l = +40deg to l = −40deg through the Galactic Centre. I compare a diagram of longitude versus radial velocity as observed with a diagram obtained from the calculation of orbits in a simple, but realistic potential containing a weak, rotating bar. Almost all stars belong to the Galactic disk but there is a small sample with high velocities close to the Galactic Centre and in the so-called forbidden quadrants. These velocities cannot be explained by a combination of pure Galactic rotation and a modest dispersion in velocity. An acceptable explanation is one in which the stars of this small sample move up and down a rotating Galactic bar.