The aim of the present paper is to propose a new interpretation of the zirconium dissolution in liquid aluminum calorimetry experiments performed in the past in order to reconcile some apparently contradictory results and observations. It is supported by the development of a dissolution kinetics model. We show that probably most of the experiments interpreted in terms of dissolution must be considered in terms of (partial or total) transformation of zirconium into zirconium aluminide (Al3Zr). In addition, on the basis of the developed model, we propose some recommendations in terms of experimental conditions to improve the dissolution process. These recommendations are consistent with some empirical rules derived in the past. It also puts in question past standard enthalpy measurements of some compounds in the Cu-Zr, Ni-Zr, Co-Zr and U-Zr systems. (C) 2018 Elsevier Ltd.