Conodont affinity and chordate phylogeny

Current information on the conodonts Clydagnathus windsorensis (Globensky) and Promissum pulchrum Kovács‐ Endrödy, together with the latest interpretations of conodont hard tissues, are reviewed and it is concluded that sufficient evidence exists to justify interpretation of the conodonts on a chordate model. A new phylogenetic analysis is undertaken, consisting of 17 chordate taxa and 103 morphological, physiological and biochemical characters; conodonts are included as a primary taxon. Various experiments with character coding, taxon deletion and the use of constraint trees are carried out. We conclude that conodonts are cladistically more derived than either hagfishes or lampreys because they possess a mineralised dermal skeleton and that they are the most plesiomorphic member of the total group Gnathostomata. We discuss the evolution of the nervous and sensory systems and the skeleton in the context of our optimal phylogenetic tree. There appears to be no simple evolution of free to canal‐enclosed neuromasts; organised neuromasts within canals appear to have arisen at least three times from free neuromasts or neuromasts arranged within grooves. The mineralised vertebrate skeleton first appeared as odontodes of dentine or dentine plus enamel in the paraconodont/euconodont feeding apparatus. Bone appeared later, co‐ordinate with the development of a dermal skeleton, and it appears to have been primitively acellular. Atubular dentine is more primitive than tubular dentine. However, the subsequent distribution of the different types of dentine (e.g. mesodentine, orthodentine), suggests that these tissue types are homoplastic. The topology of relationships and known stratigraphic ranges of taxa in our phylogeny predict the existence of myxinoids and petromyzontids in the Cambrian.