Numerical results of the "ao-" and the "o,'V~-formulation" of the atmospheric energy conversion integral, as derived from ECMWF-data, are presented and discussed. I t is shown that even global integrals of either formulation are considerably different. If the integration area is restricted to hemispheres, the differences --though growing somewhat largerare of similar magnitude. The results suggest that numerical evaluations of the baroclinic energy conversion process are still assochted with a considerable degree of uncertainty.