The new genus and species Corveolepis elgae is described from the Lochkovian, Severnaya Zemlya formation of October Revolution Island, Severnaya Zemlya. It is ascribed to the family Corvaspididae whose diagnosis is corrected. Corvaspis arctica Loeffler and Dineley is designated as the type species of Corveolepis. Previously described taxa such as “Corvaspis kingi” auct. and Corvaspis graticulata Dineley from Spitsbergen are included with doubt in Corveolepis. Other new material includes Corveolepis? sp. cf. C.? graticulata from the Lochkovian, Pod'emnaya formation of October Revolution Island. This material is used as supporting evidence for a correlation of Lower Devonian vertebrate-bearing strata between Severnaya Zemlya and Spitsbergen.

Two new species of thelodont are described. A new thelodontiform is based on articulated skeletons and isolated scales from late Telychian (Llandovery) to earliest Sheinwoodian (Wenlock) shales of the Avalanche Lake sections, Mackenzie Mountains, N.W.T., Canada. The head is blunt; it and the anterior trunk are broad and dorsoventrally flattened, while the posterior trunk and tail are more slender and laterally compressed. Unlike most other articulated thelodonts, there is no evidence of fins apart from the caudal fin. The hypocercal tail displays no evidence of rays or intermediate secondary lobes. The scales are robust, with large bases, large pulp cavities, and strong crown ridges in a stellate pattern; scale crowns are almost unornamented along the presumed ventral margin of the trunk and tail. An articulated specimen from the uppermost Telychian of Baumann Fiord, Ellesmere Island, Canadian Arctic Archipelago, is considered conspecific because of similarities in scale size and ornament, and extends the stratigraphic range of the species from the late Telychian griestoniensis-sakmaricus Graptolite Zone to the early Sheinwoodian centrifugus-insectus Graptolite Zone/Pterospathodus procerus Conodont Superzone, and the geographic range from the Selwyn Basin of the Yukon and N.W.T. to the Franklinian Basin of the Canadian Arctic Archipelago, Nunavut. A new phlebolepidiform is based on scales only. It also occurs both in the Mackenzie Mountains and the Canadian Arctic Archipelago. The two new species contribute significantly to a preliminary vertebrate faunal succession scheme for both areas and are therefore considered to be biostratigraphically useful.

Acanthodians, osteostracans, and putative chondrichthyans from the Lower Devonian (Lochkovian) ‘MOTH’ locality are found to have granular labyrinth infillings located immediately posterior to the orbits and composed of fine, sand-sized particles. They were examined under light and scanning electron microscopes and an elemental analysis (EDX) was performed. The infillings contain large proportions of silica, calcium carbonate, and dolomite, as well as minor proportions of elements such as aluminum, iron, and potassium. These results were similar to elemental analysis of the rock matrix. The inability of the fish to manufacture some of the minerals confirms that the labyrinth infilling consists at least partly of extrinsic grains. Endolymphatic pores, the openings through which this material probably entered the labyrinth, are seen in osteostracans, but have seldom been reported in acanthodians and never in basal (putative) chondrichthyans. No undoubted otoliths were seen in any of the specimens. The presence of similar statoconia, including extrinsic grains, in all three groups indicates a common inner ear physiology and presence of open endolymphatic ducts. The phylogenetic distribution of this feature suggests that it is primitive for osteostracans gnathostomes, and that placoderm ancestors also had open endolymphatic ducts that functioned to allow entry of extrinsic grains into the labyrinth.

A unique ‘buchanosteid’ arthrodire specimen from Emsian limestones at Burrinjuck represents the complete articulated remains of the head and cheek, and most of the trunk armor bones from a single fish, together with well-ossified braincase and jaw cartilages. The structure of the toothplates and jaws are described, and compared with primitive brachythoracid material from Saudi Arabia of similar age. Both the palatoquadrate and meckelian cartilages are well preserved, and perichondrally ossified as single elements. As previously described in some phlyctaeniids, the dermal gnathal elements carry crowded denticles similar to the normal tubercular ornament, enlarged along anterior and lateral margins. Six characters concerning gnathal plates of brachythoracids, which have been used in phylogenetic analysis, are discussed and reformulated.

Based on new material collected from the Upper Devonian, “Catskill Formation” of Pennsylvania, and a comparison with variation in Bothriolepis canadensis and B. yeungae, it is suggested that the species Bothriolepis minor, B. virginiensis, B. coloradensis, and B. darbiensis, pending the discovery of more complete materials, cannot reliably be distinguished from Bothriolepis nitida (Leidy, 1856). Although some of the present named species may be invalid, Bothriolepis as a whole (some 70 species world-wide, over 20 million years) is not abnormally speciose when compared with modern bottom-living species. A better understanding of the species of Bothriolepis is essential if the potential of this genus for palaeobiogeography is to be realized.

The structure of the orbito-nasal cavity in the antiarchan placoderm fish Asterolepis ornata has been studied, providing a detailed description of the premedian, rostral, and pineal plates, and bones of the sclerotic ring. Both eyes and nostrils were directed antero-latero-dorsally in Asterolepis. Eyes were enclosed within the sclerotic capsule. Nasal cavities were not opened into a so called “prenasal slit,” but directly above the head of fish. A new reconstruction of the nasal openings and position of the nasal sacs shows broader similarities between various antiarchs than was previously claimed, and allows the evaluation of the differences in the structure of the orbito-nasal cavities in antiarchs and osteostracans to disclaim the hypothesis about the presence of nasohypophyseal opening in antiarchs.

Kannathalepis milleri, gen. et sp. nov. is described from the Upper Sheinwoodian to Lower Homerian interval (Wenlock) of the Baillie-Hamilton Island section, Canadian Arctic. Based on several distinct features, such as the occurrence of several canals in the neck region, much smaller length of the wide and short base compared to the crown, presence of compound scales, character of dentine canals and tubules, and specific ultrasculpture of the crown, the new taxon Kannathalepis milleri, gen. et sp. nov. is assigned to Kannathalepididae fam. nov. of uncertain order of the subclass Elasmobranchii, class Chondrichthyes.

The earliest-known chondrichthyan braincase, from the Early Devonian (Emsian) Gydo Formation of South Africa, is described along with parts of the visceral skeleton. Only the ventral surface of the braincase is exposed, but CT-scanning permits the investigation of its unprepared regions and internal features. There is a persistent cranial fissure (as in osteichthyans) separating the trabecular and parachordal regions. The parachordal cartilage is identical to isolated endoskeletal elements referred to Pucapampella from the Middle Devonian of Bolivia. The semicircular canals are arranged as in osteichthyans and chimaeroids, with a crus commune connecting the anterior and posterior canals dorsally. The elongate palatoquadrate has an ethmoidal and palatobasal articulation (it is unknown if a postorbital articulation was present). The mandibular joint is positioned lateral to the widest part of the parachordal region, and the hyoid arch probably helped support the jaw. The South African specimen demonstrates unequivocally that some features previously known only in osteichthyans (e.g., ventral otic fissure, posterior dorsal fontanelle, palatobasal articulation) were also present in primitive chondrichthyans and are actually plesiomorphic hold-overs whose distribution was primitively more universal among gnathostomes. Fossils rarely overturn the phylogenetic status of morphological characters at such deep phylogenetic levels, perhaps because their wider original distribution pattern was short-lived and is rarely recovered from the fossil record.

Late Famennian assemblages of chondrichthyan microremains, especially teeth, from Nevada and Utah representing two zones of different water depth are analysed and compared to formerly described pelagic chondrichthyan biofacies from the areas between S Euramerica and NW Gondwana. The assemblage from the deeper (deep to moderately deep subtidal) zone is comparable to the Phoebodus-Thrinacodus biofacies, but it lacks such typical forms as Ph. gothicus and Th. tranquillus. The assemblage from the shallower (shallow subtidal) zone might be an equivalent of the Protacrodus biofacies from which it differs in containing Th. ferox and some other shallow water taxa, thus far unknown or very rarely found from the Famennian of central Europe and Africa. New definitions of chondrichthyan biofacies are proposed.

An associated specimen of the large fossil lamnid shark Carcharodon angustidens from the Late Oligocene of New Zealand's South Island preserves approximately 165 teeth, and 32 vertebral centra, making it one of the most complete Tertiary lamnids recovered to date, and the most complete fossil shark known from New Zealand. The well-preserved dentition allows for a more thorough description and revised interpretation of the dental morphology of this relatively poorly known species, and the partial vertebral column permits the unequivocal relating of teeth and centra for this taxon. Based on dental and vertebral morphology, C. angustidens is here considered to be properly assigned to the genus Carcharodon, which also includes several other “great-toothed” Tertiary shark species and C. carcharias, the extant Great White Shark. According to this interpretation, Carcharodon has a record extending back to the early Tertiary; this is in sharp contrast to an opposing view, which holds that the genus evolved much more recently, at the Miocene–Pliocene boundary, and that C. angustidens and the other great-toothed forms should be placed in a separate genus (“Carcharocles”).

Previously, known specimens of the Lochkovian (Lower Devonian) diplacanthid acanthodian genus Tetanopsyrus (all from the MOTH locality in the Northwest Territories, Canada) were thought to belong to a single species, T. lindoei. New specimens from the same locality have shown that two species are present. Tetanopsyrus lindoei is revised and T. breviacanthias, sp. nov. is described, leading to a revised diagnosis of the genus. Both species lack rostral tesserae; instead they have irregularly-shaped, monodontode, rostral scales without basal tissue and with an open pulp cavity. Both species also have enlarged, tuberculated, anterior circumorbital plates. Dorsal spines are shallowly inserted and each is supported by a basal plate that ossified early in ontogeny. The two species are distinguished from each other by the structure of the dorsal spines, the length of the pectoral spines, and the shape of the procoracoids.

The revision of Tetanopsyrus has implications also for relationships among acanthodian families. Both Tetanopsyrus (Tetanopsyridae) and Gladiobranchus (Gladiobranchidae) are seen as diplacanthiforms rather than as ischnacanthiforms. Similarities are found in the jaws, circumorbital plates, anterior dorsal spine, dorsal spine basal plates, prepelvic spines, and scapulocoracoids. Other potential relatives of diplacanthids include Uraniacanthus, which has similarities in prepelvic and dorsal spines, and Culmacanthus, whose relationships cannot be settled without further study.

A well-preserved neurocranium with attached partial skull roof is described from the Early Devonian Taemas Formation at Wee Jasper, southern New South Wales, Australia. The dermal ornament, skull roof pattern, and general proportion and structure of the endocranium are typically actinopterygian and the specimen is tentatively assigned to the actinopterygian genus Ligulalepsis. Other features more closely resemble some other groups, such as sarcopterygians (widely-spaced orbital walls, and short, broad telencephalon cavity), acanthodians (position of hyomandibular facet), and placoderms (many similarities including position of foramina for oculomotor, profundus, and trigeminal nerves, pituitary vein, and ophthalmic and orbital arteries in and around the orbit). This specimen is the first early osteichthyan to demonstrate the presence of an eyestalk, previously known only in placoderms and chondrichthyans. The unusual mix of characters and presence of an eyestalk provide new insights into primitive osteichthyan anatomy.

Several hypotheses supporting monophyly of the Teleostei on the basis of synapomorphies have been produced over the last 30 years. The concept of Teleostei sensu Patterson (1977) and Patterson and Rosen (1977), with halecomorphs as the sister-group and †Pachycormiformes and †Aspidorhynchiformes at the base, has been questioned recently. A new hypothesis has been suggested (Arratia, 1999, 2000a) with †Pholidophorus as the basal taxon. Whereas the monophyly of Teleostei has been supported by numerous investigations based on morphological evidence of fossil and living forms and on molecular data, the sister-group of Teleostei is still unresolved. Possible sister-groups are the Amiiformes, Lepisosteiformes, †Dapedium, †Pycnodontiformes, †Pachycormiformes, and †Aspidorhynchiformes. Their relative positions in the cladogram changes when different outgroups are used.

The large actinopterygian clade comprising the stem-groups of teleosts and the Teleostei (including fossil and extant members) and excluding the Halecomorphi (Amia and relatives) and the Ginglymodi (Lepisosteus and relatives) is formally named Teleosteomorpha. The name Teleostei is reserved here for the apomorphy-based taxon Teleostei that includes the fossil basal teleosts and the Teleocephala (crown-group). The monophyly of Teleostei is supported by one uniquely derived character and numerous homoplasious derived characters.

The available information reveals that there is no correlation between the age of the probable sister-groups of Teleostei and their primitiveness. All of them appeared at the same time as the Teleostei (e.g., †Pycnodontiformes) or are younger (e.g., Amiiformes, Lepisosteiformes, †Dapedium, †Pachycormiformes, and †Aspidorhynchiformes).

The closest living (Amia or Lepisosteus) and the closest fossil sister-group of teleosts remains unknown.

A new genus and species of acanthomorph teleost, Spinocaudichthys oumtkoutensis, gen. et sp. nov., is described from the Cretaceous (Cenomanian) of Morocco. This taxon is the first Cretaceous acanthomorph ever recorded from a freshwater deposit. Its caudal fin displays dorsal and ventral procurrent spines, a feature only present in some acanthopterygian fishes. It also shares features with paracanthopterygians. However, this fish cannot be a holacanthopterygian (sensu Johnson and Patterson, 1993) because it lacks a pelvic spine and has a plesiomorphic caudal skeleton. Here, we consider it as an acanthomorph incertae sedis and we discuss relationships of the Acanthomorpha.

The skull roof, palate, and lingual side of the lower jaw of the primitive dipnoan Melanognathus canadensis are described. Melanognathus is a denticulated dipnoan with: a sarcopterygian-like composition of bones arranged in sequence on the external side of the lower jaw; “splenial” and “postsplenial” fused; plow-shaped parasphenoid; and a primitive snout region formed by a mosaic of bones. It is the sister taxon of Uranolophus, both being the most primitive dipnoans in a resolved cladogram of Devonian dipnoans. The first resolved cladogram of Devonian dipnoans shows that denticulation is the primitive feature of dipnoans, it reoccurs independently as a neotenic feature in the Late Devonian (Barwickia to Jarvikia) and also in Carboniferous/Permian (Conchopoma) dipnoans. True dipnoan tooth plates appear with Stomiahykus in the dipnoan phylogeny above Dipnorhynchus. ‘Dipnorhynchus’ lehmanni does not cluster with ‘Speonesydrion’ iani, which is part of the genus Dipnorhynchus. A new genus, Westollrhynchus, has been erected for ‘Dipnorhynchus’ lehmanni. Dipterus appears just above the node with the highest number of changes in Devonian dipnoan phylogeny.

New material from the Emsian (Lower Devonian) of Spitsbergen, consisting of lower jaws and the ethmosphenoid portion of skulls, allows the elucidation of the relations between the dermopalatine, the vomer and the ethmoid region in porolepiformes. It shows that the fenestra endochoanalis was certainly occluded by the ethmoidal process of the palatoquadrate and that a fenestra exochoanalis is not present between the vomer, dermopalatine, maxilla and premaxilla.

We use probability plots to compare metric data for a growth series of the Permian nectridean Diplocaulus magnicornis Cope to a normal (Gaussian) probability distribution function and demonstrate that the ontogeny of Diplocaulus is not a single continuous process. The plots are bimodal, revealing a population of small, ontogenetically young animals whose growth is governed by a distinct process. This population, which comprises ∼20% of the growth series, is not evident in allometric studies. These young animals have a mean mid-line skull length of ∼21 mm, and their orbits are relatively larger, and rounder, than the orbits of the older animals, whose mean skull length is ∼94 mm. Although there is no evidence of lissamphibian-type metamorphosis in the Nectridea, the probability plots clearly demonstrate a two-stage ontogeny, best described as an ontogenetic morphological saltation. This saltation may have been an alternative strategy to the metamorphic heterochrony found in lissamphibians and their relatives.