Traditional analyses of Early Phanerozoic marine diversity at the genus level show an explosive radiation of marine life until the Late Ordovician, followed by a phase of erratic decline continuing until the end of the Palaeozoic, whereas a more recent analysis extends the duration of this early radiation into the Devonian. This catch-all approach hides an evolutionary and ecological key event long after the Ordovician radiation: the rapid occupation of the free water column by animals during the Devonian. Here, we explore the timing of the occupation of the water column in the Palaeozoic and test the hypothesis that ecological escalation led to fundamental evolutionary changes in the mid-Palaeozoic marine water column. According to our analyses, demersal and nektonic modes of life were probably initially driven by competition in the diversity-saturated benthic habitats together with the availability of abundant planktonic food. Escalatory feedback then promoted the rapid rise of nekton in the Devonian as suggested by the sequence and tempo of water-column occupation. h Devonian, diversity, ecology, food webs, nekton, plankton, radiation.

Chondrichthyans (including living sharks, skates, rays and chimaeras) have a fossil record of scales and dermal denticles perhaps dating back to the Late Ordovician period, about 455 million years ago. Their fossil tooth record extends to the earliest Devonian period, almost 418 million years ago, whereas the oldest known articulated shark remains date from the Early Devonian period, about 394 million years ago. Here we report the discovery of an articulated shark that is almost 409 million years old from the Early Devonian (early Emsian) period of New Brunswick, Canada. The specimen, identified as Doliodus problematicus (Woodward), sheds light on the earliest chondrichthyans and their interrelationships with basal jawed vertebrates. This species has been truly problematic. Previously known only from isolated teeth, it has been identified as an acanthodian and a chondrichthyan. This specimen is the oldest shark showing the tooth families in situ, and preserves one of the oldest chondrichthyan braincases. More notably, it shows the presence of paired pectoral fin-spines, previously unknown in cartilaginous fishes.

MOTS CLÉSChordata, Craniata, animal-conodonte, codage des caractères, analyse cladistique, paléohistologie. KEY WORDSChordata, Craniata, conodont animals/ elements, character coding, cladistic analysis, palaeo-histology. ABSTRACTAn evidence-based reassessment of the phylogenetic relationships of conodonts shows that they are not "stem" gnathostomes, nor vertebrates, and not even craniates. A signifi cant group of conodont workers have proposed or accepted a craniate designation for the conodont animal, an interpretation that is increasingly becoming established as accepted "fact". Against this prevailing trend, our conclusion is based on a revised analysis of traditional morphological features of both discrete conodont elements and apparatuses, histological investigation and a revised cladistic analysis modifying that used in the keystone publication promoted as proof of the hypothesis that conodonts are vertebrates. Our study suggests that conodonts possibly were not even chordates but demonstration of this is beyond the scope of this paper. To summarize, in conodonts there is low cephalization; presence of simple V-shaped trunk musculature and unique large-crystal albid material in the elements; lack of a dermal skeleton including characteristic vertebrate hard tissues of bone, dentine and enamel; lack of odontodes with bone of attachment and a unique pulp system; lack of segmentally-arranged paraxial elements and dermal elements in median fi ns, all of which supports neither a vertebrate nor a craniate relationship for conodonts. RÉSUMÉ Des pseudo-dents : une réévaluation des relations phylogénétiques des conodontes et des vertébrés.Une réévaluation des relations phylogénétiques des conodontes est fondée sur de nouvelles preuves. Elle montre que les conodontes ne sont ni des gnathostomes-souches, ni des vertébrés, ni même des crâniates. Un groupe signifi catif de spécialistes des conodontes a proposé, ou accepté, que ces organismes soient considérés comme des crâniates, une interprétation qui est en train de s'installer en tant que fait avéré. Notre conclusion va à l'encontre de cette tendance ; elle est fondée sur une révision des traits morphologiques traditionnels à la fois des éléments isolés et des assemblages de conodontes, sur les données histologiques et sur une analyse cladistique révisée, ce qui modifi e les conclusions de la publication principale qui a promu l'hypothèse selon laquelle les conodontes seraient des vertébrés. Notre étude suggère même que les conodontes n'aient pas été des chordés, mais la démonstration de cette hypothèse va au-delà de l'objectif de cet article. En résumé, chez les conodontes, le degré de céphalisation est faible ; la musculature du tronc a une forme simple en V ; les éléments isolés montrent un tissu blanc avec des cristaux de grande taille, uniques pour ce tissu ; il n'y a pas de squelette dermique incluant les tissus durs caractéristiques des vertébrés tels que l'os, la dentine et l'émail ; il n'y a pas d'odontodes avec leur os et leur système pulpaire unique ; il ...