A Comment on the Stability Property of Baroclinic Models