Report

A tetrapod fauna from within the Devonian Antarctic Circle

See allHide authors and affiliations

Science  08 Jun 2018:
Vol. 360, Issue 6393, pp. 1120-1124
DOI: 10.1126/science.aaq1645

Out of Antarctica

When we think of Devonian tetrapods, the ancestors of all modern vertebrates, we tend to picture amphibian-like creatures emerging from the water into a wet tropical forest or swamp. Indeed, all previously described specimens of this group have been recovered from the tropics. Gess and Ahlberg now describe two fossil tetrapods from Devonian Antarctica. Thus, the distribution of tetrapods may have been global, which encourages us to rethink the environments in which this important group was shaped.

Science, this issue p. 1120

Abstract

Until now, all known fossils of tetrapods (limbed vertebrates with digits) and near-tetrapods (such as Elpistostege, Tiktaalik, and Panderichthys) from the Devonian period have come from localities in tropical to subtropical paleolatitudes. Most are from Laurussia, a continent incorporating Europe, Greenland, and North America, with only one body fossil and one footprint locality from Australia representing the southern supercontinent Gondwana. Here we describe two previously unknown tetrapods from the Late Devonian (late Famennian) Gondwana locality of Waterloo Farm in South Africa, then located within the Antarctic Circle, which demonstrate that Devonian tetrapods were not restricted to warm environments and suggest that they may have been global in distribution.

The fossil locality at Waterloo Farm, near Grahamstown, South Africa (Fig. 1A), features an exceptionally preserved biota, including examples of soft-tissue preservation (14), deposited in the south polar region close to paleolatitude 70°S (Fig. 1B). In contrast, all previously known Devonian tetrapod and elpistostegid localities lie within about 30° of the palaeoequator (5). The Waterloo Farm fossils are metamorphosed and strongly flattened, with the bone tissue replaced by secondary metamorphic mica, partially altered to chlorite. Two tetrapods—Tutusius umlambo gen. et sp. nov. and Umzantsia amazana gen. et sp. nov., both represented by disarticulated material (Figs. 2 and 3 and figs. S1 and S2)—are present in the assemblage and are described here (formal taxonomic descriptions are in the supplementary materials). Tutusius is represented by a single cleithrum (Fig. 2, A and B) with a broad, flat, unornamented blade, resembling that of the early Famennian Russian genus Jakubsonia more than the slender cleithra of the late Famennian Ichthyostega and Ventastega (68) (Fig. 4).

Fig. 1 Maps of the fossil locality.

(A) Map of South Africa showing the Waterloo Farm fossil locality (black asterisk). (B) South-polar projection of Gondwana, modified from (20), showing Waterloo Farm (black asterisk) in relation to the reconstructed position of the South Pole 360 million years ago (21). Blue asterisks indicate the other two known Devonian tetrapod localities in Gondwana—Genoa River [left; footprints (22)] and Jemalong [right; Metaxygnathus, a single lower jaw ramus (23)]—both in Australia. The landmass with a dashed outline below Waterloo Farm is an emergent part of the Falklands Plateau, forming the outer margin of the semi-enclosed Agulhas Sea (24). Brown denotes land; pale blue, shallow shelf; blue, deep shelf.

Fig. 2 Material of Tutusius and Umzantsia.

(A and B) Photograph and line drawing of AM7527, a left cleithrum, the holotype and only known specimen of Tutusius umlambo. (C to P) AM7528a to -f, the bones of the holotype of Umzantsia amazana, believed to represent one individual. (C and D) AM7528a, right cleithrum (line drawing incorporates information from the counterpart); (E to G) AM7528b, left jugal, showing part, counterpart, and line drawing; (H and I) AM7528c, right preopercular; (J and K) AM7528d, incomplete left frontal; (L and M) probable left supratemporal; (N to P) AM7528e, a bone assemblage comprising a chain of two partial infradentaries, one near-complete infradentary, a probable premaxilla, and an unidentified tooth-bearing ossicle (see also Fig. 3). In all drawings, thick outlines denote true margins, and thin outlines denote broken or covered margins. In (D), gray shading indicates the dermal ornament. Anterior is to the left in (A), (B), (E) to (G), (J), and (K) and to the right in (C), (D), (H), and (I). All scale bars, 10 mm. (C) to (P) are shown to the same scale.

Fig. 3 The lower jaw of Umzantsia.

(A to C) AM7529, left mandibular ramus and infradentaries of the right mandibular ramus of Umzantsia amazana. (A) Photograph of the specimen. The splenial of the left ramus partly overlies an infradentary of the right ramus; the area within the white box is shown on the left with the splenial in place and on the right as an excerpt box with the splenial removed. (B) Photograph overlaid with interpretative line drawing. (C) Interpretative line drawing. Light gray shading indicates infradentaries of the right mandibular ramus; dark gray, the sensory canals on these bones. Fine parallel gray lines on the left jaw ramus represent dermal ornament. (D) Sketch reconstruction of the left mandibular ramus (AM7529). Scale bars, 10 mm.

Fig. 4 Comparison of cleithra.

A series of cleithra from the tetrapod stem group, spanning the fin-to-limb transition, placed on a simplified phylogeny that reflects recent analyses (8, 9), illustrating the morphological transformation of the shoulder girdle and the tentative phylogenetic positions of Tutusius and Umzantsia. Not to scale. For each taxon, the upper image shows the cleithrum in external view, and the lower image shows it in internal view. Anterior is to the left in all cases. Only for Eusthenopteron (25) is the entire scapulocoracoid shown; in other taxa, the projecting ventral part has been cut off. For comparison, a complete scapulocoracoid plus cleithrum of Ichthyostega is shown (bottom right). The arrowhead branch of the phylogeny leads to the tetrapod crown group. Small gray arrows indicate the posteroventral buttress of the cleithrum. Character states at nodes: 1, the scapulocoracoid is small and concealed by the cleithrum in lateral view, and the cleithrum has a broad ventral lamina and is entirely covered with ornament (primitive condition, widely shared among Osteichthyes); 2, the cleithrum tapers to a point anteroventrally and attaches along the anterodorsal margin of the scapulocoracoid; 3, the dorsal end of the scapulocoracoid forms a v-shaped peak, and the cleithrum carries a posterodorsal buttress; 4, the cleithrum lacks ornament.

Dermal bones of Umzantsia carry a distinctive ornament consisting of fine parallel ridges reminiscent of water ripples. This allows identification of a number of cranial bones and a cleithrum from one bedding plane as probably derived from a single individual, designated the holotype (Fig. 2, C to P). A lower jaw ramus from another bedding plane (Fig. 3) is also assigned to Umzantsia. Scaling the bones of the Umzantsia holotype to the skull reconstruction of Ventastega (8) suggests a head length of ~13 cm. The lower jaw is 17.9 cm long. The dermal ornament also covers much of the cleithrum of Umzantsia; this fishlike characteristic contrasts with the unornamented cleithra of other Devonian tetrapods, suggesting a phylogenetic position between those tetrapods and Tiktaalik (9) (Fig. 4).

The largest skull bone is a jugal (Fig. 2, E to G), identifiable from the presence of an orbital margin and characteristic set of sutural margins (5, 7, 8). The orbital margin is short [suggesting a triangular orbit with a ventral apex, similar in shape to that of Anthracosaurus (10), unless the eye was extremely small], the lacrimal is excluded from the orbital margin by a jugal-prefrontal contact, there is no distinct dorsal postorbital process, and the notch for the quadratojugal is deep. The preopercular (Fig. 2, H and I) is similar to that of Ventastega, with a rounded posterior margin that projects as a short process beyond the quadratojugal contact. The frontal resembles those of previously described Devonian tetrapods. A probable supratemporal is the only recovered skull table element.

The lower jaw is slender and gently curved. The splenial is the longest of the infradentaries, occupying about half the jaw length. The left lower jaw ramus has five infradentaries, instead of the normal four. This may be an autapomorphy of Umzantsia, but the associated infradentaries from the right ramus appear proportionately longer, raising the possibility that there were only four infradentaries on the right side and that this individual was asymmetrical. The infradentaries carry the typical tetrapod “starburst” ornament, grading dorsally into a ripple-like ornament. A series of short tooth-bearing ossifications appears to represent the coronoid series (Fig. 3C), implying that, of the three coronoids normally seen in tetrapods, at least the posterior one has been replaced by a chain of smaller elements in this taxon. An isolated element of this kind is also associated with the holotype (Fig. 2, N to P). The lateral line canals of the skull and lower jaw appear as a combination of continuous and discontinuous grooves, similar to the condition in other Devonian tetrapods (5, 8), though poorly preserved.

Waterloo Farm demonstrates that the early evolution of tetrapods did not play out exclusively in tropical and subtropical environments. The late Famennian to Tournaisian witnessed a gradual transition from greenhouse to icehouse conditions, punctuated by an end-Famennian glaciation (11). Exact timing of this glaciation relative to Waterloo Farm deposition is uncertain, but late Famennian diamictites in South Africa are probably glaciogenic, and carbonates are entirely absent from the region (11). Thus, even though Waterloo Farm yields a rich terrestrial flora that rules out a truly polar climate (12), it cannot have been very warm, and proximity to the pole implies several months of complete winter darkness.

The presence of tetrapods in such an environment raises the question of whether high-latitude environments played a distinctive role in the fish-tetrapod transition—for example, as drivers of innovation or as refuges for archaic taxa. The combination of autapomorphic and primitive characters in Umzantsia has bearing on this problem. All Devonian tetrapod cleithra described to date (68, 1315), including fragmentary late Frasnian material associated with Elginerpeton (16), lack dermal ornament (Fig. 4). This suggests that Umzantsia represents a deep but specialized branch of the tetrapod lineage that had been in existence since at least the Frasnian, a time interval of some 12 million years.

The Waterloo Farm tetrapod fossils and the Middle Devonian tetrapod trackways from Poland and Ireland (1719) challenge the popular scenario of a tropical origin of tetrapods during the Late Devonian (5). Tetrapods originated no later than the Eifelian (early Middle Devonian), when they were present in southern Laurussia; by the late Famennian (latest Devonian), they ranged from the tropics to the south polar regions. This geographic pattern could still point to a tropical origin but may simply be a sampling artifact. Against this background, the continued investigation of nontropical localities such as Waterloo Farm must be a priority. Waterloo Farm is also the only known Devonian tetrapod locality to feature soft-tissue preservation, as exemplified by the earliest known lamprey, Priscomyzon (1). The locality thus has the potential not only to cast new light on early tetrapod biogeography and evolution, but also to illuminate unknown aspects of their morphology.

Supplementary Materials

www.sciencemag.org/content/360/6393/1120/suppl/DC1

Materials and Methods

Systematic Paleontology

Figs. S1 and S2

Reference (26)

References and Notes

Acknowledgments: The South African National Roads Agency Limited supported rescue of shale during roadworks. R.G. acknowledges useful early discussions with M. Coates and P. Janvier regarding the morphology and confirming the identity of AM7527. B. Nosilela (Department of African Languages, Rhodes University) advised on taxonomic names. Funding: R.G. acknowledges funding from the South African Millennium Trust and the South African DST-NRF Centre of Excellence in Palaeosciences (CoE-Pal). P.E.A. acknowledges a Wallenberg Scholarship from the Knut and Alice Wallenberg Foundation. Author contributions: Fieldwork, collection and preparation of material, initial identification of tetrapod specimens (cleithra), and project conceptualization and design, R.G.; identification of additional specimens, R.G. and P.E.A.; manuscript writing and illustrations, P.E.A. and R.G. Competing interests: None declared. Data and materials availability: Formal taxonomy is presented in the supplementary materials. Specimens are accessioned at the Albany Museum, Grahamstown, South Africa, as AM7511 to AM7513.
View Abstract

Navigate This Article