Relationships of tristichopterids (osteolepiform lobe-finned fishes) from the Middle–Late Devonian of East Gondwana

Abstract

Two well-preserved articulated tristichopterids are known from the Canowindra fossil fish locality (Frasnian) in central New South Wales: Mandageria Johanson & Ahlberg, 1997 and Cabonnichthys Ahlberg & Johanson, 1997. Two other large tristichopterids are Notorhizodon Young et al., 1992 from the Aztec Siltstone fauna of Antarctica (Givetian), and the Famennian Eusthenodon Jarvik, 1952, first described from East Greenland and recently identified in eastern Australia. Relationships of the East Gondwana tristichopterids were previously assessed as Notorhizodon [Platycephalichthys [Cabonnichthys [Mandageria [Eusthenodon]]]], with no endemic Gondwanan clade, and more basal tristichopterids (Tristichopterus, Eusthenopteron, Jarvikina) being Northern Hemisphere forms. This suggested that the group may have originated in Laurussia, with advanced tristichopterids later reaching Gondwana (Middle–Late Devonian). Reconsideration of evidence from structure of the palate and scale morphology suggests an alternative phylogeny, with an endemic Gondwanan clade (subfamily Mandageriidae nov.) comprising at least three genera: Cabonnichthys and Mandageria from Canowindra, an undescribed tristichopterid from Eden, NSW, and possibly Notorhizodon from Antarctica. Distinctive isolated scales of ‘mandageriid’ type are illustrated from the Antarctic Aztec Siltstone assemblage and contrasted with typical tristichopterid scales, including new examples illustrated from the Givetian Bunga beds, NSW. Biogeographic implications include a possible Gondwanan origin for tristichopterids.

Key words:

• tristichopterid osteolepiform fish
• Devonian
• Canowindra
• Australia
• Antarctica
• morphology
• phylogeny
• biogeography

Acknowledgements

Alex Ritchie (Australian Museum) and Craig Dickman (National Parks, Merimbula, NSW south coast) are thanked for providing casts of the Canowindra and Eden tristichopterids respectively. Ben Young is thanked for preparation and photography, and with Dan Evans for assisting in preliminary assessment of the Eden material, including one field visit (NSW National Parks is acknowledged for a collection permit, and Monica Yeung and Bruce Loomes for assistance at the locality). John Long (Museum Victoria, Melbourne) and Keith Crook (ANU) discussed sarcopterygian phylogeny and south coast stratigraphy respectively. Per Ahlberg (Uppsala), Gäel Clément (Paris), Ted Daeschler (Philadelphia), and Zerina Johanson (London) gave advice on relevant sarcopterygian material. I thank Ella Hoch and Gilles Cluny (Copenhagen), and Gäel Clément (Paris) for assistance in examining material from Greenland and Belgium, respectively (August, 2007). Brian Harrold (ANU) is thanked for computer assistance. Provision of facilities in the ANU Department of Earth & Marine Sciences is acknowledged. I thank reviewers Gäel Clément and John Long and editor Steve McLoughin for many suggestions that greatly improved the manuscript. This research is a contribution to IGCP Project 491 and is supported by ARC Discovery Grants DP0558499 and DP0772138.