On a theory of amplitude vacillation in baroclinic waves