Phase coupling between movement trajectories has been proposed as the basic mechanism of hand coordination in the production of bimanual rhythmic movements with a 1:2 frequency ratio. Here a central temporal coupling view is proposed as an alternative. Extending previous models of two-handed synchronic and alternate-hand tapping, we hypothesized that 1:2 tapping is performed under the control of a single internal timekeeper set at the frequency required for the fast hand. The fast hand is assumed to use every signal and the slow hand every other signal of the timekeeper, to produce actions coordinated in time. The model's predictions for the variance-covariance pattern of tap timing within and across hands were tested in an experiment that required tapping with both hands with 1:1 or 1:2 frequency ratio. The finger contact on the response plate was to be short or long, according to instruction. Prolonged finger contact entailed profound modifications in the movement trajectories but failed to modify the variance-covariance pattern of the tap timing. This pattern proved to conform to predictions under both the short and the long contact conditions, thus supporting the central temporal coupling hypothesis.