Traditionally, two types of rib are distinguished in gnathostomes: dorsal (upper) and ventral (lower, pleural) ribs. They are defined according to their position in the connective tissue system of the body: dorsal ribs develop at the intersection of the serially arranged myosepta with the horizontal septum that separates epaxial from hypaxial musculature, whereas ventral ribs develop at the intersection of myosepta with the peritoneum and usually encircle the body cavity. Distribution of rib types among gnathostomes has traditionally been reported as follows: elasmobranchs have dorsal ribs; all Actinopterygii have only ventral ribs with the exception of polypterids, and two subgroups of teleosts, which supposedly also have dorsal ribs; within Sarcopterygii tetrapods have dorsal ribs, whereas dipnoans have ventral ribs. Here, we report the development of ribs in polypterids, a taxon playing a crucial role in discussions on rib homology. We demonstrate that putative dorsal ribs of polypterids have a unique ontogeny and represent an autapomorphy of this taxon. We discuss previous hypotheses of rib homology and offer a more plausible (i.e. more parsimonious) alternative to the conventional interpretation. We conclude that dorsal ribs do not exist and that ribs of gnathostomes are ventral ribs.