Three cases have been documented in which abdominal heterotaxy with absent spleen was accompanied by a discordant anatomy between the tracheobronchial tree and the atria. In 1 instance the tracheobronchial anatomy was as in situs inversus but the venous connections and the atrial appendages suggested situs solitus. The second case exhibited a situs solitus anatomy of the tracheobronchial tree, but the morphological features of the atria were mixed up in the sense that the venous connections and atrial septum suggested situs inversus, whereas the atrial appendages were of a right isomeric type. The third case was interesting in showing a symmetrical tracheobronchial tree not accompanied by isomeric atrial appendages. The latter and the atrial septum exhibited situs solitus characteristics, though the venous connections suggested 'situs ambiguus'. Such observations have led us to reconsider the criteria that determine the classification of atrial situs. It is our contention that in cases characterised by major abnormalities in atrial anatomy, particularly with respect to the venous connections, the prime feature for identification of atrial situs would be the morphology of the atrial appendages, which represent the 'true' atrial Anlagen. Positive identification of atrial appendages would also enable a more precise definition of the term 'atrial situs ambiguus'. This term can then be defined as the situs characterised by either bilateral right-or bilateral left-sided morphological characteristics of the atrial appendages, irrespective of the venous connections or atrial septal anatomy.