The late Devonian lungfish Soederberghia (Sarcopterygii, Dipnoi) from Australia and North America, and its biogeographical implications

A new species of rhynchodipterid lungfish, Soedeiberghia simpsoni, sp. nov. , is described on the basis of a complete skull roof, and an incomplete but articulated head plus body, from the Upper Devonian Mandagery Sandstone ne ar Canowindra, New South Wales (NSW. Fig. 1A), Australia. It is compared wi th Soederberghia material from the upper part of the Catskill Formation, Pe nnsylvania (Fig. 1B), and the Cloghnan Shale at Jemalong near Forbas, NSW, Australia. The Catskill and Cloghnan shale material consisting Of two incom plete skull roofs, appears to be conspecific with the type species, Soederb erghia groenlandica from the Famennian Remigolepis Series of Greenland. Thi s is congruent with a recent suggestion that the Mandagery Sandstone is lat e Frasnian in age, whereas the upper Catskill Formation and Cloghnan Shale are Famennian. At the Famennian localities, Soederberghia is associated wit h tetrapods. Articulated postcranial material of Soeder-berghia including t he pectoral girdle, cranial and pleural ribs, unpaired fins and fin support s, is described for the first time. The presence of cranial ribs and a para sphenoid with a posterior stalk suggest that Soederberghia gulped air and p robably inhabited a non-marine or shallow near-shore marine environment. Th e occurrence of Soederberghia groenlandica in the Famennian Old Red Sandsto ne of North America, Greenland and Australia thus furnishes evidence for co ntact or close proximity between Gondwana and Laurussia at this time, in co nflict with some recent paleomagnetic data. The sister group of Soederbergh ia is Rhynchodipterus elginensis from the Famennian of Rosebrae near Elgin, Scotland; together they form the Family Rhynchodipteridae. Griphognathus, previously included in the Rhynchodipteridae, lacks a stalked parasphenoid and cranial ribs, and has a quite different dermal bone pattern from Soeder berghia and Rhynchodipterus. We remove it from the group on these grounds, and interpret it as less crownward than the Rhynchodipteridae.