Search

The water-to-land transition was a major step in vertebrate evolution and eventually gave rise to the tetrapods, including amphibians, reptiles, birds, and mammals. The first land invasion of our fish ancestors is considered to have occurred during the late Devonian period ~370 million years ago (Mya) (Daeschler et al., 2006). Many fossils from important transitional species, such as Tiktaalik, Acanthostega, and Ichthyostega, have helped to identify key morphological and anatomical structures crucial to vertebrate terrestrial adaptation (Coates, 1996; Johanson & Ahlberg, 2001; Shubin et al., 2006). However, homologous analyses of these body forms and structures in more ancient species have suggested that some of the morphologies related to vertebrate land dispersal were already present in early bony fish species. For instance, the presence of shoulder girdles on the articular surface of the endoskeleton in Late Lochkovian Psarolepis indicates that stem sarcopterygians already possessed an endoskeletal fin pattern similar to that of tetrapod stylopods (Zhu & Yu, 2009). In addition, primitive lungs, which originated from the respiratory pharynx and were located on the ventral side of the alimentary tracts, can be observed in several extant basal actinopterygians (bichirs, reedfish) and all extant sarcopterygians, as well as some fossils of coelacanths and salamanders (Cupello et al., 2017; Tissier et al., 2017) (Figure 1). This evidence suggests that, instead of relying on genetic innovations evolving after the first fish left their water habitat, this transition may have been accomplished by adopting physical traits and genetic components that already existed far earlier than when the transition occurred. Whether such an ancestral developmental regulatory network was present or not and how far this ancestral network can be traced in history are challenging questions for paleontologists. Three recent papers published in Cell provide new insights into this hypothesis. Wang et al. (2021) sequenced the giant genome of lungfish, the closest fish species to tetrapods, and Bi et al. (2021) sequenced the genomes of multiple early divergent ray-finned fish. Comparative genomic analyses from these two studies confirmed the presence of ancestral genetic regulatory networks that likely played essential roles in the development and evolution of various biological functions related to vertebrate land invasion. Although certain ancestral features have been lost in teleosts, the most derived fish lineage to evolve after whole-genome duplication (Sato & Nishida, 2010), they have been recreated in zebrafish by modifying their genetic makeup to reactivate the ancestral genetic network (Hawkins et al., 2021).

The known diversity of tetrapods of the Devonian period has increased markedly in recent decades, but their fossil record consists mostly of tantalizing fragments1–15. The framework for interpreting the morphology and palaeobiology of Devonian tetrapods is dominated by the near complete fossils of Ichthyostega and Acanthostega; the less complete, but partly reconstructable, Ventastega and Tulerpeton have supporting roles2,4,16–34. All four of these genera date to the late Famennian age (about 365–359 million years ago)—they are 10 million years younger than the earliest known tetrapod fragments5,10, and nearly 30 million years younger than the oldest known tetrapod footprints35. Here we describe Parmastega aelidae gen. et sp. nov., a tetrapod from Russia dated to the earliest Famennian age (about 372 million years ago), represented by three-dimensional material that enables the reconstruction of the skull and shoulder girdle. The raised orbits, lateral line canals and weakly ossified postcranial skeleton of P. aelidae suggest a largely aquatic, surface-cruising animal. In Bayesian and parsimony-based phylogenetic analyses, the majority of trees place Parmastega as a sister group to all other tetrapods. Three-dimensionally preserved fossils of Parmastega aelidae, a newly described tetrapod from the earliest Famennian (Late Devonian) of Russia, provide detailed insights into the morphology and palaeobiology of the earliest tetrapods.

International audience; The origin of tetrapods is one of the key events in vertebrate history. The oldest tetrapod body fossils are Late Devonian (Frasnian–Famennian) in age, most of them consisting of rare isolated bone elements. Here we describe tetrapod remains from two Famennian localities from Belgium: Strud, in the Province of Namur, and Becco, in the Province of Liège. The newly collected material consists of an isolated complete postorbital, fragments of two maxillae, and one putative partial cleithrum, all from Strud, and an almost complete maxilla from Becco. The two incomplete maxillae and cleithrum from Strud, together with the lower jaw previously recorded from this site, closely resemble the genus Ichthyostega, initially described from East Greenland. The postorbital from Strud and the maxilla from Becco do not resemble the genus Ichthyostega. They show several derived anatomical characters allowing their tentative assignment to a whatcheeriid-grade group. The new tetrapod records show that there are at least two tetrapod taxa in Belgium and almost certainly two different tetrapod taxa at Strud. This locality joins the group of Devonian tetrapod-bearing localities yielding more than one tetrapod taxon, confirming that environments favourable to early tetrapod life were often colonized by several tetrapod taxa.

Genome size spans over a 300-fold range among vertebrates (132 pg for Protopterus aethiopicus, the marbled lungfish, and 0.35 pg for Tetraodon nigroviridis, the green spotted pufferfish). While phylogenetic analysis of genome size has helped clarify how this variation evolved in multiple tetrapod groups, the ancestral tetrapod condition still remains poorly characterized, and this obscures our understanding of character state polarity and macroevolutionary trends in genome size. To address this problem, we used phylogenetic comparative methods to analyze paleohistological data from eight taxa of the Middle and Late Paleozoic to the Early Mesozoic: Eusthenopteron, Ichthyostega, Acheloma, Eryops, Trimerorhachis, Wetlugasaurus, an unidentified dissorophoid, and Chroniosaurus. Five other extinct taxa were included from previous studies to better frame our results, including Marmorerpeton, Cardiocephalus, Diplocaulus, an unidentified basal sauropsid, and Mycterosaurus. We augmented a previously reported histological and genome size data set (including data from 14 lissamphibians, three testudines, Sphenodon, five squamates, two crocodilians, 11 birds, and 22 mammals) with genome size and histological data from extant Latimeria and three extant actinopterygians. Our results suggest that all eight of the newly analyzed extinct taxa had genome sizes ranging between 3.2 and 3.9 pg. These results imply that basal tetrapods had genome sizes (and underlying genomic architectures) similar to extant mammals and lepidosaurs. We find no major shifts in genome size during the tetrapod water-to-land transition. Our analysis suggests that Eusthenopteron and Ichthyostega had genome sizes well within the range of extant actinopterygians and Latimeria, despite several whole-genome duplications in actinopterygians.

The fish-to-tetrapod transition—followed later by terrestrialization—represented a major step in vertebrate evolution that gave rise to a successful clade that today contains more than 30,000 tetrapod species. The early tetrapod Ichthyostega was discovered in 1929 in the Devonian Old Red Sandstone sediments of East Greenland (dated to approximately 365 million years ago). Since then, our understanding of the fish-to-tetrapod transition has increased considerably, owing to the discovery of additional Devonian taxa that represent early tetrapods or groups evolutionarily close to them. However, the aquatic environment of early tetrapods and the vertebrate fauna associated with them has remained elusive and highly debated. Here we use a multi-stable isotope approach (δ13C, δ18O and δ34S) to show that some Devonian vertebrates, including early tetrapods, were euryhaline and inhabited transitional aquatic environments subject to high-magnitude, rapid changes in salinity, such as estuaries or deltas. Euryhalinity may have predisposed the early tetrapod clade to be able to survive Late Devonian biotic crises and then successfully colonize terrestrial environments. An approach using multiple stable isotopes reveals that early tetrapods of the Devonian period were euryhaline animals that inhabited aquatic environments of highly variable salinity.

Available information on the development of primitive urodele (Hynobiidae) limbs and limb structure in Devonian tetrapods provide the basis for formulating the hypothesis on the existence of a special phase, the phase of the biserial limb in tetrapod history. The limb of the Devonian amphibian Ichthyostega with two groups of digits, preaxial and postaxial, corresponds well to this phase. Based on the structure of the Ichthyostega limb, it is suggested that the fin type ancestral to terrestrial limbs was asymmetrical biserial, possessed a short axis, unbranched jointed preaxial radials deviating one by one from each mesomere of the axis, and unbranched jointed postaxial radials deviating from the distal mesomeres of the axis in two or more radials.

Synopsis Over the past century, various modern analogs have been used to infer the evolution of locomotor performance in stem tetrapods and their fish ancestors, with varying success. Here, we conduct a phylogenetic review of these modern analogs, from chondrichthyans to mammals, highlighting the broad spectrum of vertebrate clades and locomotor behaviors. The pros and cons behind utilizing modern analogs for the early stages of the transition from water to land also are discussed. In particular, it is noted that any hypothesis about locomotion not only must be supported by evidence from living animals but must also be consistent with character transformations in the fossil record. A ‘‘totalevidence’’ approach that emphasizes what extinct taxa could not do, rather than focusing on the specifics of how they functioned, is thus recommended. An example of this approach, which investigates mobility of the limb joints in modern semi-aquatic animals and in the Devonian stem tetrapod Ichthyostega, is detailed. We propose that various locomotion behaviors of modern quadrupeds can be ruled out for Ichthyostega, but that forelimb ‘‘crutching’’ motions, as seen in living mudskippers, may have been possible. The potential for movement in other known Devonian stem tetrapods is assessed through an anatomical comparison of limb joint morphology—and associated mobility—with Ichthyostega, and deemed to have been quite similar. Limbs before terrestriality

This article introduces the tetrapods, both modern and extinct, and sets out what we know or can infer about the origin of the group. It introduces fossil members of the group that gave rise to tetrapods and discusses some of the conditions in which tetrapods developed limbs and walking. Terrestrial locomotion followed the evolution of limbs with digits once weight-bearing and flexible joints were developed.

This chapter describes the ear regions of tetrapods from the Paleozoic. In the past couple of decades, understanding of their morphology has been increased by discoveries of fossils from the Late Devonian and early Carboniferous (Mississippian) periods. The primitive condition, as found in several only distantly related taxa, consisted of a bulky stapes with a large footplate and a stout wing-shaped distal portion. In some, the stapes seems to have been a strut supporting the braincase. The hearing capabilities of these taxa were probably poor, confined to low-frequency detection primarily in water but, possibly, also in air. Ichthyostega had a highly modified version probably specialized for aquatic audition. Taxa from the later Carboniferous (Pennsylvanian) and Permian show a great deal of variation, especially among more terrestrially adapted forms. These include the larger stem amniotes, such as seymouriamorphs and diadectomorphs. All these taxa have some kind of a notch or embayment at the back of the skull, which in early forms was probably part of a spiracular mechanism but, in later ones, might have housed a tympanic membrane that closed off an air-filled middle ear cavity. The small recumbirostran microsaurs and the earliest amniotes had no such notch. Microsaur stapes had large footplates and short, stubby shafts. These and the earliest amniotes probably had no middle ear cavity. The earliest amniote stapes were robust, long, and laterally or downwardly projecting. In some cases they contacted the cheek bones or the jaw joint, likely precluding good aerial sound reception.

Three-dimensional reconstruction and modelling of limb joint mobility in the early tetrapod Ichthyostega is used to provide insights into an important step in vertebrate evolution—the transition from swimming to walking. The locomotion of the early tetrapod Ichthyostega is a puzzle because it had an unusual skeleton — a robust ribcage, a spine adapted for up-and-down rather than lateral movement, and paddle-like limbs that seem unsuited to walking. This paper presents a three-dimensional reconstruction and computer modelling of limb-joint mobility based on Ichthyostega fossils. The models suggest that Ichthyostega was unable move its limbs in a walking pattern. Instead it may have shuffled on land using movements similar to those seen in seals. This implies that the transition from swimming to walking was a gradual process. The origin of tetrapods and the transition from swimming to walking was a pivotal step in the evolution and diversification of terrestrial vertebrates. During this time, modifications of the limbs—particularly the specialization of joints and the structures that guide their motions—fundamentally changed the ways in which early tetrapods could move1,2,3,4. Nonetheless, little is known about the functional consequences of limb anatomy in early tetrapods and how that anatomy influenced locomotion capabilities at this very critical stage in vertebrate evolution. Here we present a three-dimensional reconstruction of the iconic Devonian tetrapod Ichthyostega and a quantitative and comparative analysis of limb mobility in this early tetrapod. We show that Ichthyostega could not have employed typical tetrapod locomotory behaviours, such as lateral sequence walking. In particular, it lacked the necessary rotary motions in its limbs to push the body off the ground and move the limbs in an alternating sequence. Given that long-axis rotation was present in the fins of tetrapodomorph fishes5,6,7, it seems that either early tetrapods evolved through an initial stage of restricted shoulder8,9 and hip joint mobility or that Ichthyostega was unique in this respect. We conclude that early tetrapods with the skeletal morphology and limb mobility of Ichthyostega were unlikely to have made some of the recently described Middle Devonian trackways10.

A new genus and species of Devonian tetrapod has been identified from material collected in 1947 from the southern slope of Mt. Celsius, Ymer phi, North-East Greenland. The specimen preserves both ...

The Late Devonian early tetrapods in East Greenland occur in the Celsius Bjerg Group. Key occurrences are located in a detailed stratigraphic section used here to interpret the sedimentary palaeoenvironments. The palaeoenvironment for the Britta Dal Formation (which contains both Ichthyostega and Acanthostega) is reinterpreted. The Britta Dal Formation channels have flat bases, are poorly channelized, are of low sinuosity and are part of a very major distributory system that periodically experienced extreme flooding. The tetrapod fossils were recovered from an ephemeral system that was not permanently habitable in the immediate area. Plant megafossils are poorly preserved casts and impressions dominated by lycopsids and fern-like plants. The overbank siltstones are dominated by arid soil forming processes and comprise a spectacular sequence of vertisols. The 1174 m in situ Ichthyostega locality in Paralleldal was relocated and occurs just below the midpoint of the second megacycle in the Britta Dal Formation.

After a brief historical review of the fin to limb transition and consideration of a theoretical “prototetrapod,” this article considers new ideas generated from recent fossil finds and from developmental biology that bear on the question of how limbs, digits, limb joints, and pentadactyly evolved. Among the first changes to take place were those to the humerus, in concert with those to the breathing apparatus, and these adaptations were acquired while the animals were still basically aquatic with the evolution of digits occurring during this phase. Studies from developmental biology of modern taxa can be integrated with information from fossils to produce a fuller picture. The acquisition of pentadactyly was among the last changes to occur in the modification of a fin into a limb. This vision differs radically from older theoretical ideas which perceived land locomotion as the prime evolutionary force driving the transition.

Ichthyostega and Acanthostega are the earliest tetrapods known from multiple near-complete skeletons, with Acanthostega generally considered the more primitive. New material indicates differing ontogenetic trajectories for their forelimbs: In Ichthyostega , the pattern of muscle attachment processes on small humeri (upper arm bones) resembles that in “fish” members of the tetrapod stem group such as Tiktaalik , whereas large humeri approach (but fail to attain) the tetrapod crown-group condition; in Acanthostega , both small and large humeri exhibit the crown-group pattern. We infer that Ichthyostega underwent greater locomotory terrestrialization during ontogeny. The newly recognized primitive characteristics also suggest that Ichthyostega could be phylogenetically more basal than Acanthostega .

Wrists, ankles and digits distinguish tetrapod limbs from fins, but direct evidence on the origin of these features has been unavailable. Here we describe the pectoral appendage of a member of the sister group of tetrapods, Tiktaalik roseae, which is morphologically and functionally transitional between a fin and a limb. The expanded array of distal endochondral bones and synovial joints in the fin of Tiktaalik is similar to the distal limb pattern of basal tetrapods. The fin of Tiktaalik was capable of a range of postures, including a limb-like substrate-supported stance in which the shoulder and elbow were flexed and the distal skeleton extended. The origin of limbs probably involved the elaboration and proliferation of features already present in the fins of fish such as Tiktaalik.

Study of Devonian tetrapods and their relatives spanning the so-called ‘fish–tetrapod’ transition has expanded almost exponentially in the last 15 years or so. This evolutionary event is now represented by at least nine named genera of Devonian tetrapod, several new ‘near-tetrapods’ and a number of new tetrapods from the Early Carboniferous. The anatomy of Acanthostega has radically changed ideas about how this transition took place, and more recently the anatomy of Ichthyostega is being reassessed with some startling conclusions such as the unique construction of its ear region. The current state of research on this range of animals is summarized, followed by a consideration of the acquisition of limbs and digits among tetrapods including their possible forerunners, the development of digits and their original function, and the onset of pentadactyly. The faunal relations and palaeoecological contexts of the Devonian tetrapods are brought together in an initial though necessarily brief survey, followed by an assessment of Devonian tetrapod diversity, which is seen to be much greater than previously realised. Finally, some hopes and ambitions for the future are set out.

Ichthyostega was the first Devonian tetrapod to be subject to a whole-body reconstruction. It remains, together with Acanthostega, one of only two Devonian tetrapods for which near-complete postcranial material is available. It is thus crucially important for our understanding of the earliest stages of tetrapod evolution and terrestrialization. Here we show a new reconstruction of Ichthyostega based on extensive re-examination of original material and augmented by recently collected specimens. Our reconstruction differs substantially from those previously published and reveals hitherto unrecognized regionalization in the vertebral column. Ichthyostega is the earliest vertebrate to show obvious adaptations for non-swimming locomotion. Uniquely among early tetrapods, the presacral vertebral column shows pronounced regionalization of neural arch morphology, suggesting that it was adapted for dorsoventral rather than lateral flexion.

The presence of a petaloid pattern (previously known as ‘dark dentine’) in crosssections of teeth of the embolomere Pholiderpeton attheyi has been used as a synapomorphy of theembolomeres or of the embolomeres plus the stem tetrapod, Crassigyrinus scoticus . Among the taxastudied, dentine that appears dark results from closely packed dentine tubules and can be found inany part of a tooth section in which such crowding occurs. The petaloid pattern is restricted to toothsections of a particular diameter, and is obliterated in larger sections of teeth that show complexfolding. Petaloid dentine has been found in all tetrapod teeth with plicidentine that were sectioned inthis study, whether from stem tetrapods, the Embolomeri, Temnospondyli, or Stereospondyli, andhas been recognised in some sarcopterygian sh, an extant actinopterygian sh, ichthyosaurs, andVaranus . The presence of petaloid dentine is neither a synapomorphy of the tetrapod node nor of anynode within tetrapods.KEY WORDS: dark dentine, Embolomeri, petaloid dentine, Rhizodontida, stem tetrapods,Stereospondyli, Temnospondyli, tooth histology.The pattern known as ‘dark dentine’ in sections of earlytetrapod teeth with folded dentine (plicidentine) was rstrecognised in 1876 by Thomas Atthey when describing teeth ofthe Carboniferous embolomere, Pholiderpeton attheyi (Clack1987; =‘ Anthracosaurus russelli ’ in error in Atthey 1876 p162,pl. 11; Fig. 1a). Atthey described the appearance of thePholiderpeton attheyi tooth sections as stellate (=petaloid inthe present paper) with a number of fusiform bodies of lightcoloured dentine radiating from the pulp cavity to the circum-ference and embedded in dentine of a dark colour. The rstreference to ‘dark dentine’ was to the interstices between thefusiform bodies, but Panchen (1977, 1985) noted that thecolour arrangement could be reversed. Tooth sections thatshow ‘dark dentine’ may resemble either a light coloured daisyon a dark background when seen in transmitted light (e.g.Fig. 1a) or a dark coloured daisy on a light background whenseen in re ected light (e.g. Fig. 3a).The petaloid pattern is known in the early tetrapodliterature as ‘dark dentine’ (e.g. Panchen 1985). The presenceof ‘dark dentine’ has been used as a synapomorphy ofCrassigyrinus scoticus and embolomeres (Panchen 1985;Panchen & Smithson 1988), and of Ichthyostega stensioei andmost other tetrapods (Coates 1996), although Smithson(1980) implied that the phenomenon might be size-related,having been described in only the larger anthracosauroids(most embolomeres) with the condition unknown in smallerforms. The purpose of this paper is to review the pattern oftooth folding in early tetrapods, and to determine thedistribution of petaloid dentine among them. The Ducabrooklocality from the Lower Carboniferous of Queensland,Australia (Thulborn et al . 1996), yields histologically intactspecimens of both the stem tetrapod Ossinodus pueri(Warren & Turner 2004) and the rhizodontid Strepsodus sp.(Johanson et al . 2000) with which to test the hypothesis ofthe phylogenetic usefulness of tooth dentine characters, inparticular the phenomenon of petaloid dentine among earlytetrapods.

Lepidotrichia are dermal elements located at the distal margin of osteichthyan fins. In sarcopterygians and actinopterygians, the term has been used to denote the most distal bony hemisegments and also the more proximal, scale-covered segments which overlie endochondral bones of the fin. In certain sarcopterygian fishes, including the Rhizodontida, these more proximal, basal segments are very long, extending at least half the length of the fin. The basal segments have a subcircular cross section, rather than the crescentic cross section of the distal lepidotrichial hemisegments, which lack a scale cover and comprise short, generally regular, elements. In rhizodonts and other sarcopterygians, e.g. Eusthenopteron, the basal elements are the first to appear during fin development, followed by the endochondral bones and then the distal lepidotrichia. This sequence contradicts the 'clock-face model' of fin development proposed by Thorogood in which the formation of endochondral bones is followed by development of lepidotrichia. However, if elongate basal 'lepidotrichia' are not homologous with more distal, jointed lepidotrichia and if the latter form within a distal fin-fold and the former outside this fold, then Thorogood's 'clock-face' model remains valid. This interpretation might indicate that the fin-fold has been lost in early digited stem-tetrapods such as Acanthostega and Ichthyostega and elongate basal elements, but not true lepidotrichia, occur in the caudal fins of these taxa.

Two morphologically distinct assemblages of the Late Devonian tetrapod Ichthyostega from East Greenland are described on the basis of a large collection of skulls and postcranial elements. Skull specimens collected on Gauss Halvo show that the assemblage from the lower Aina Dal Formation has proportionally narrower skulls and finer skull roof sculpture than the assemblage from the higher Britta Dal Formation. The assemblages are compared with the stratigraphically unconstrained type material from the north side of Celsius Bjerg on Ymer O, allowing a taxonomic revision. From the original description of seven taxa, three species are recognized and related to the assemblages from Gauss Halvo. I. stensioei is comparable with the Aina Dal Formation assemblage while I. eigili and I. watsoni are valid for specimens found in the Britta Dal Formation. Ichthyostegopsis wimani is not a valid genus or species and may be regarded as a juvenile I. eigili. This example of species-level variability is the earliest known from the fossil record of early tetrapods.

The Late Devonian genus Ichthyostega was for many decades the earliest known tetrapod, and the sole representative of a transitional form between a fish and a land vertebrate. However, despite being known since 1932 (ref. 1) from a large collection of specimens, its morphology remained enigmatic and not what was expected of a very primitive tetrapod2. Its apparent specializations led it to be considered as a “blind offshoot”3 or “sidebranch”4 off the tetrapod family tree, and recent cladistic analyses have disagreed about its exact phylogenetic position5,6,7,8 within the tetrapod stem group. In particular, its braincase and ear region defied interpretation, such that conventional anatomical terms seemed inapplicable4. Using new material collected in 1998 (ref. 9), preparation of earlier-collected material, and high-resolution computed tomography scanning, here we identify and interpret these problematic anatomical structures. They can now be seen to form part of a highly specialized ear, probably a hearing device for use in water. This represents a structurally and functionally unique modification of the tetrapod otic region, unlike anything seen in subsequent tetrapod evolution. The presence of deeply grooved gill bars as in its contemporary Acanthostega10 suggest that Ichthyostega may have been more aquatically adapted than previously believed.

The fossil record of early tetrapods has been increased recently by new finds from the Devonian period1 and mid–late Early Carboniferous period2. Despite this, understanding of tetrapod evolution has been hampered by a 20-million-year gap (‘Romer's Gap’3) that covers the crucial, early period when many key features of terrestrial tetrapods were acquired. Here I describe the only articulated skeleton of a tetrapod, Pederpes, yet found from the Tournaisian epoch (354–344 million years ago (Myr)). The new taxon includes a pes with five robust digits, but a very small, possibly supernumerary digit preserved on the manus suggests the presence of polydactyly. Polydactylous early tetrapods may have survived beyond the end of the Devonian and pentadactyly cannot be assumed for the pes. However, the pes has characteristics that distinguish it from the paddle-like feet of the Devonian forms and resembles the feet of later, more terrestrially adapted Carboniferous forms. Pederpes is the earliest-known tetrapod to show the beginnings of terrestrial locomotion and was at least functionally pentadactyl. With its later American sister-genus, Whatcheeria4,5, it represents the next most primitive tetrapod clade after those of the Late Devonian, bridging the temporal, morphological and phylogenetic gaps that have hitherto separated Late Devonian and mid-Carboniferous tetrapod faunas.

Recent finds of early tetrapods have established that the most primitive form, Acanthostega , retained internal gills and other fish-like features; this has led to the conclusion that it was a primarily aquatic animal. Other Late Devonian tetrapods, such as Ichthyostega and Tulerpeton , provide no evidence of internal gills, but have also been interpreted as inhabiting an aquatic environment. The probable aquatic habits of a diversity of Devonian tetrapods has led to the suggestion that the entire early tetrapod radiation may have been an aquatic one, with terrestriality having evolved in later forms. However, consideration of the physiology of living amphibious vertebrates suggests that this scenario is unlikely. The use of the gills for the excretion of carbon dioxide and ammonia appears to be a fundamental feature of all primarily aquatic vertebrates. No living fish loses its internal gills, even if it excretes a significant portion of its nitrogenous waste as urea via the kidney in the water. Gills are simply too valuable to be lost by an aquatic animal, even in those air-breathing fishes that no longer use the gills for oxygen uptake. We suggest that the apparent loss of the gills in tetrapods more derived than Acanthostega signals their descent from a more terrestrial phase in tetrapod evolution, following the primary assumption by the kidney of the excretion of nitrogenous wastes. Without this new role of the kidney, loss of the gills would have been impossible. With this new kidney role, loss of the gills may have been advantageous in reducing desiccation on land.

The lower jaw of the Devonian tetrapodAcanthostegais described for the first time. Redescriptions are provided for the lower jaws of the elpistostegidPanderichthys, the Devonian tetrapodsElginerpeton, Obruchevichthys, Metaxygnathus, VentastegaandIchthyostega, and the Carboniferous tetrapodsCrassigyrinus, MegalocephalusandGephyrostegus. The character distri- butions thus revealed differ considerably from previous accounts, particularly in the wide distribution of certain primitive characters. Meckelian ossification in the middle part of the jaw is widespread among Devonian tetrapods, being demonstrably absent only inAcanthostega.Among Carboniferous tetrapods, a tooth-bearing parasymphysial plate is shown to be present inCrassigyrinusandMegalocephalus(having already been demonstrated by other authors inWhatcheeriaandGreererpeton).A phylogenetic analysis of 26 early tetrapods including all the aforementioned genera, scored for 51 lower jaw characters, produces at least 2,500 equally parsimonious trees. However, the lack of resolution lies largely in a big top end polychotomy containing anthracosaurs, temnospondyls, seymouriamorphs, microsaurs and a nectridean-amniote clade. Below this polycho- tomy, which may correspond approximately to the tetrapod crown group, there is a well-resolved stem-group containing, in descending order,Megalocephalus, Greererpeton, Crassigyrinus, (jaws associated with)Tulerpeton, Whatcheeria, Acanthostega, Metaxygnathus, Ichthyostega, VentastegaandMetaxygnathus(unresolved), anElginerpeton-Obruchevichthysclade, andPanderichthys.This conflicts with recently published phylogenies by Coates and Lebedev & Coates, which placeTulerpetonand all post-Devonian tetrapods in the amphibian or amniote branches of the tetrapod crown group.

A new reconstruction of the skull ofCrassigyrinus scoticus(Lydekker) is provided, based on newly prepared material. A revised interpretation of the naris shows it in most respects to be like that of other Carboniferous tetrapods, lacking both an anterior tectal and a lateral rostral. However it does incorporate an enlarged septomaxilla, and part of the dorsal surface of the vomer is also visible within the naris of most specimens. Prominent paired buttresses run along the midline of the snout, and a single bowl-shaped interpremaxillary fenestra pierces the snout and palate but is not confluent with the naris or choana. A preopercular is probably absent. A new cladistic analysis placesCrassigyrinuswith the Viséan genusWhatcheeria, and puts those two taxa as a sister-clade to the anthracosaursProterogyrinusandPholiderpeton.Temnospondyls form a clade with loxommatids, and these two arms form a dichotomy crownward toGreererpeton, IchthyostegaandAcanthostegawhich are successive stemward plesions.

Recently, the origin of four-limbed vertebrates (tetrapods) was re-assessed, based on footprints and trackways that were made by early land animals almost 400 million years (my) ago. That is 18 my earlier than the oldest known tetrapod body fossils, such as Acanthostega and Ichthyostega from Greenland, and 10 my older than the earliest ‘tetrapodomorph fishes’ (Panderichthys, Tiktaalik etc.). These and other facts suggest that the first tetrapods may have thrived in the marine-influenced intertidal and/or lagoon zone as well as in brackish and freshwater environments associated with land vegetation, as previously thought. Here, we discuss the controversial question whether or not extant air-breathing fishes, such as mudskippers, amphibious gobioids that inhabit mangrove swamps, can be interpreted as living model organisms, with reference to the earliest land plants. In addition, recent developmental and behavioural studies on lungfishes (Dipnoi) are summarized and evaluated. We conclude that mudskipping ‘walking fishes’ (Periophthalmus sp.) and Dipnoi (Protopterus sp.) shed light on the gradual evolutionary transition of ancient fishes to early tetrapods that occurred during the Devonian in muddy, salty waters. However, they are not the ancestors of tetrapods, because extant organisms cannot be progenitors of other living beings.

X-ray synchrotron microtomography has revealed the three-dimensional vertebral architecture of Ichthyostega, and other crucial and celebrated early tetrapods; a surprising feature is the relationship between the vertebral elements, with the pleurocentra unexpectedly attached to the succeeding intercentrum, suggesting a ‘reverse’ rhachitomous design. A backbone of interlocking vertebrae is a prerequisite for a tetrapod — a vertebrate living on land. Skeletal details in fossils of the very earliest tetrapods, however, are very hard to make out, as they are often swathed in rocky matrix. This inconvenience has now been overcome by the application of the latest synchrotron microtomography technology to the three best-known early tetrapods — Ichthyostega, Acanthostega and Pederpes. Early tetrapods were presumed to have had 'rhachitomous' vertebrae, with a neural arch and spine above, a single intercentrum front and ventral, and paired pleurocentra to the rear and ventral. Surprisingly, the new scans reveal a 'reverse' rhachitomous design, effectively rewriting the 'textbook' view of vertebral column evolution in the first limbed vertebrates. The construction of the vertebral column has been used as a key anatomical character in defining and diagnosing early tetrapod groups1. Rhachitomous vertebrae2—in which there is a dorsally placed neural arch and spine, an anteroventrally placed intercentrum and paired, posterodorsally placed pleurocentra—have long been considered the ancestral morphology for tetrapods1,3,4,5,6. Nonetheless, very little is known about vertebral anatomy in the earliest stem tetrapods, because most specimens remain trapped in surrounding matrix, obscuring important anatomical features7,8,9. Here we describe the three-dimensional vertebral architecture of the Late Devonian stem tetrapod Ichthyostega using propagation phase-contrast X-ray synchrotron microtomography. Our scans reveal a diverse array of new morphological, and associated developmental and functional, characteristics, including a possible posterior-to-anterior vertebral ossification sequence and the first evolutionary appearance of ossified sternal elements. One of the most intriguing features relates to the positional relationships between the vertebral elements, with the pleurocentra being unexpectedly sutured or fused to the intercentra that directly succeed them, indicating a ‘reverse’ rhachitomous design10. Comparison of Ichthyostega with two other stem tetrapods, Acanthostega7 and Pederpes8, shows that reverse rhachitomous vertebrae may be the ancestral condition for limbed vertebrates. This study fundamentally revises our current understanding11 of vertebral column evolution in the earliest tetrapods and raises questions about the presumed vertebral architecture of tetrapodomorph fish12,13 and later, more crownward, tetrapods.

The postcranial skeleton ofAcanthostega gunnarifrom the Famennian of East Greenland displays a unique, transitional, mixture of features conventionally associated with fishand tetrapod-like morphologies. The rhachitomous vertebral column has a primitive, barely differentiated atlas-axis complex, encloses an unconstricted notochordal canal, and the weakly ossified neural arches have poorly developed zygapophyses. More derived axial skeletal features include caudal vertebral proliferation and, transiently, neural radials supporting unbranched and unsegmented lepidotrichia. Sacral and post-sacral ribs reiterate uncinate cervical and anterior thoracic rib morphologies: a simple distal flange supplies a broad surface for iliac attachment. The octodactylous forelimb and hindlimb each articulate with an unsutured, foraminate endoskeletal girdle. A broad-bladed femoral shaft with extreme anterior torsion and associated flattened epipodials indicates a paddle-like hindlimb function. Phylogenetic analysis placesAcanthostegaas the sister-group ofIchthyostegaplus all more advanced tetrapods.Tulerpetonappears to be a basal stemamniote plesion, tying the amphibian-amniote split to the uppermost Devonian.Caerorhachismay represent a more derived stem-amniote plesion. Postcranial evolutionary trends spanning the taxa traditionally associated with the fish-tetrapod transition are discussed in detail. Comparison between axial skeletons of primitive tetrapods suggests that plesiomorphic fish-like morphologies were re-patterned in a cranio-caudal direction with the emergence of tetrapod vertebral regionalisation. The evolution of digited limbs lags behind the initial enlargement of endoskeletal girdles, whereas digit evolution precedes the elaboration of complex carpal and tarsal articulations. Pentadactylous limbs appear to have stabilised independently in amniote and amphibian lineages; the colosteidGreererpetonhas a pentadactylous manus, indicating that basal amphibian forelimbs may not be restricted to patterns of four digits or less.

Postcranial remains of the Russian Late Devonian tetrapod Tulerpeton include the hexadactylous fore limb, hind limb, anocleithral pectoral girdle, squamation, and associated disarticulated postcranial bones. A cladistic analysis indicates that Tulerpeton is a reptiliomorph stem-group amniote and the earliest known crown-group tetrapod: Acanthostega and Ichthyostega are successively more derived plesion stem-group tetrapods and do not consititute a monophyletic ichthyostegalian radiation. Previous analyses suggesting a profound split in tetrapod phylogeny are thereby corroborated, and likewise the interpretation of Westlothiana as a stem-group amniote. The divergence of reptiliomorphs from batrachomorphs occurred before the Devonian-Carboniferous boundary. Tulerpeton originates from an entirely aquatic environment with a diverse fish fauna. The morphologies of its limbs and those of Devonian stem-tetrapods suggest that dactyly predates the elaboration of the carpus and tarsus, and that polydactyly persisted after the evolutionary divergence of the principal lineages of living tetrapods. The apparent absence of a branchial lamina and gill skeleton suggests that Tulerpeton was primarily air-breathing, whereas contemporary stem-group tetrapods and more recent batrachomorphs retained greater emphasis on gill-breathing.