Early vertebrate evolution: pharyngeal structure and the origin of gnathostomes

Some authors have claimed that pharyngeal morphology supports the classical division of vertebrates into Gnathostomata and Agnatha. Assuming the branchial arches of jawed and jawless fish to be homologous, and noting that the arches lie on opposite sides of the gills (internal vs. external), these authors necessarily concluded that the gills evolved independently in the two groups (ectobranchiates, endobranchiates). A gill‐less common vertebrate ancestor must have been very simple and ancient (early Cambrian or older). The present paper challenges this scheme by arguing that the gills of lampreys and gnathostome fish are too structurally similar to have been independently derived. Rather, the branchial arches in the two groups are non‐homologous, for they differ in location, not only with respect to the gills, but also with respect to other major organs (ventral aorta, thyroid, visceral musculature). It is concluded that the common ancestors of gnathostomes and lampreys were not simple animals, but fish with fully developed gills that could have existed well after the Cambrian period. Constructing a gnathostome‐like pharynx requires only the addition of internal branchial arches to an agnathan pharynx (larval lamprey). It is widely held that vertebrates initially fed on microscopic particles; if so, the mere addition of larger, zooplankton‐sized items to the pre‐gnathostome diet would have provided selection for protectively re‐enforcing the medial margins of the gills with internal arches.SummaryContrary to one widely held view, pharyngeal morphology does not separate jawless vertebrates (lampreys) from gnathostome fish. Despite differences in branchial arch location, gills of lampreys and gnathostomes are fundamentally alike. Phylogenetic considerations might take a fruitful new direction by asking how the branchial arches came to differ in the two groups.Thanks are extended to Quentin Bone, Philippe Janvier, R. P. S. Jefferies, Malcolm Jollie, Ken Kardong, Colin Patterson, Barbara Stahl, Keith Thomson and Richard Wassersug for discussion of the ideas in this paper, and to Barbara Comstock and Sue Rose for typing the manuscript.