Introduction

The Permian-Triassic mass extinction (PTME) was the largest biodiversity crash of the Phanerozoic, witnessing the death of almost all life on earth1,2. The timing of biotic recovery in the sea from this mass extinction has been puzzling. It is known that some taxa such as ammonites3,4, conodonts4,5 and benthic foraminifera6 had a higher recovery rate than other marine taxa and species numbers in some transient biotas recovered to the pre-extinction level within a very short interval7,8,9. Furthermore, some transient metazoan reefs even appeared shortly after the catastrophe10. More recently, a survey of marine vertebrate predators has shown that multi-level food webs had already become established very shortly after the PTME11. On the other hand, the traditional view is that complete recovery of life in the sea was much delayed until the rebuilding of stable and complex ecosystems some 5–10 million years after PTME, likely the result of a combination of intrinsic causes, primarily the severity of the extinction and destruction of key life modes such as reefs, as well as extrinsic causes, primarily repeated periods of global warming, ocean acidification and anoxia linked to volcanic activity throughout the Early Triassic12,13,14,15.

Although devastating to the pre-existing biota, recovery from the PTME facilitated emergence of several entirely new groups, most strikingly, the appearance of Mesozoic marine reptiles13. The earliest Mesozoic marine reptiles–nothosaurs, pachypleurosaurs, ichthyosaurs, hupehsuchids and probably thalattosaurs–appeared in the late Early Triassic and further new groups, including placodonts, saurosphargids and probably some archosauromorphs, emerged soon after, in the early Middle Triassic11,16,17. These predators roamed in the shallow epicontinental seas and intraplatform basins around the supercontinent Pangea and explored many different diets, ranging from cephalopods and other invertebrates to fishes, while the placodonts specialized in crushing hard-shelled invertebrates and the larger marine reptiles, such as some ichthyosaurs and nothosaurs, may have expanded their diet to include small marine reptiles11,18.

The first diversity peak of marine reptiles occurred in the Anisian16,19. Interestingly, the first appearance of gigantic apex predators (defined here as those having body size > 5 m long) in the sea from the eastern Panthalassic and western Tethyan provinces (Fig. 1) coincides with this diversity peak. The appearance of such gigantic apex predators, together with associated complex food webs, has been taken as an indicator of the full recovery of marine ecosystems from the PTME20. Surprisingly, Middle Triassic gigantic apex marine predators have not yet been recorded from the eastern Tethyan/western Panthalassic province, where several Lagerstätten with beautifully preserved fossils have been intensively sampled in recent decades21. The lack of gigantic apex predators in the Middle Triassic of this region could imply significant diachronic timing of recovery in different parts of the Triassic ocean, as often suggested by research on invertebrates6,7,8,9,10. Here we report a gigantic nothosaur and associated diverse marine reptiles from the Middle Triassic Luoping biota22,23 in the eastern Tethyan province, which indicates that the appearance of such apex predators and complex food webs in the sea was globally synchronous after the devastating PTME, rejecting the hypothesis of diachronous ecosystem recovery across different parts of the Triassic ocean.

Figure 1
figure 1

Palaeogeographic reconstruction (244 Ma) showing the global distribution of Anisian gigantic apex predators in the sea (Generated from http://fossilworks.org/).

The underlying source of the data is the Paleobiology Database.

Results

Systematic Palaeontology

Sauropterygia Owen, 1860

Nothosauria Baur, 1889

Nothosauridae Baur, 1889

Nothosaurus Münster, 1834

Nothosaurus zhangi sp. nov.

Holotype

LPV 20167, a complete lower jaw associated with partial postcranial skeleton (Fig. 2), catalogued at the Chengdu Center, China Geological Survey.

Figure 2
figure 2

Holotype of Nothosaurus zhangi (LPV 20167).

(a) Photo of the skeleton. (b) Line drawing. (c) Mandibular symphysis and a dislocated tooth; the black arrow indicates the rough position of crown/root boundary. (d) Cross section of the left dentary and an associated small functional dentary tooth with tooth wear. (e) Postcranial skeleton. (f) Anteroposteriorly exposed centrum. Scale bar equals 10 cm in b and e, 5 cm in c and 5 mm in d. Abbreviations: an, angular; ar, articular; c, centrum; d, dentary; ns, neural spine; par, prearticular; r, rib; sa, surangular; sp, splenial; t, tooth.

Type Locality and horizon

The holotype and only known specimen was collected from Bed 165 of the Dawazi section24 in Luoping County, Yunnan Province, by Qiyue Zhang and his mapping team in 2008. Dawazi is the nominal location of the Luoping biota22,23. The fossiliferous horizons of the Luoping biota are contained within the upper member of the Guanling Formation and conodont study suggests dating to the Pelsonian substage of the Anisian24.

Etymology

The species name is in honour of Qiyue Zhang who discovered the Luoping biota.

Diagnosis

A gigantic species of Nothosauria whose size approached or slightly exceeded that of Nothosaurus giganteus, the largest known Triassic sauropterygian. The new species differs from N. giganteus by the presence of a short mandibular symphysis, a relatively higher neural spine, a distinct medial expansion of the prearticular (autapomorphy among Sauropterygia) and a distinctly short retroarticular process (autapomorphy among Nothosauria).

Description and comparisons

The lower jaw is prepared in ventral view. Although broken into two pieces, the right mandibular ramus is nearly complete, only lacking some fragments along the inner broken edge (Fig. 2a,b). Its length along the midline is about 65 cm, while the width is about 45 cm, comparable to the largest known specimens of Nothosaurus giganteus25. The lower jaw is strongly constricted at a level just posterior to the mandibular symphysis (Fig. 2c).

The mandibular symphysis is very short, a primitive character among Nothosauria. Dividing the length of the symphysis by the width of the lower jaw in the constriction yields a ratio of 0.64, much smaller than all known mandibular symphyses of Nothosaurus baruthicus (synonymized with N. giganteus) which yield ratios between 1.0 and 1.3 (ref. 25). Many small pits are present on the surface of the symphysis. The mandibular symphysis is almost completely fused, leaving only the trace of a suture line in the posterior half of the mandible. The splenial is excluded from the formation of the symphysis.

Both the splenial and the dentary extend posteriorly for two-thirds of the length of the mandible. In the posteriormost part of the left mandible, the surangular and prearticular form the lateral and medial margins respectively. The lateral ridge of the surangular is present. The prearticular is expanded medially to a degree not seen in other Nothosauria. The posterior margin of the prearticular is smoothly concave. Part of the articular is exposed posteromedial to the prearticular. The retroarticular process is distinctly short, a morphology that is known among Nothosauria only from a specimen reported as Nothosaurus cf. giganteus from the Middle Triassic of Makhtesh Ramon in Israel26.

Four anteriormost dentary teeth are exposed. Judging from the shape of the symphysis, there were at least four dentary fangs present on either side. These fang-like teeth are procumbent and implanted in deep sockets in the dentary. The crown surface is striated. Three dislocated fangs are scattered on the surface of the blocks. All three fangs are strongly curved. One of them is completely preserved, with a striated crown surface and a smooth root. The crown of this tooth is about 34 mm long while the root measures 56 mm. There is no constriction between root and crown. The crown has a conical shape, while the root seems more compressed.

The broken surface of the left dentary exposes the sagittal section of a small functional dentary tooth just posterior to the mandibular symphysis, which indicates that the bottom of the tooth root is firmly attached to the dentary bone (Fig. 2d). The crown is represented only by an impression and is slightly curved inward. The lingual (inner) side of the crown has an apparent constriction slightly below the level where the pulp cavity extends (Fig. 2d), a character unknown in all other nothosaurs, but likely caused by wear.

There are five articulated centra preserved and exposed in ventrolateral view, as well as an anteroposteriorly exposed centrum (Fig. 2e). All preserved centra are platycoelous. The anteroposteriorly exposed centra show that the vertebrae are non-notochordal (Fig. 2f). The main part of this centrum is round. The dorsal part that connects with the neural arches develops an articular facet for the rib. The dorsal margin of this centrum is concave, forming the floor of the neural canal. All of the centra are slightly constricted, but still retain parallel edges. Small nutritive foramina are found on the ventral surface of the centra. However, they are by no means comparable to those found in pistosauroid sauropterygians, where the subcentral foramina are relatively large and symmetrically paired.

One isolated neural arch is also preserved (Fig. 2e). The neural spine is relatively low compared with those found in Nothosaurus mirabilis, N. haasi or N. tchernovi, but relatively higher than a typical N. giganteus neural spine25.

The articular head of the dorsal ribs is unicipital and expanded. It becomes thinner along the curved shaft, but expanded again at the distal end. The cross sections of the dorsal ribs are oval in shape. No groove is found on any of the dorsal ribs.

There is one partly preserved rib with a rectangular articular head, which is different from all other preserved rib heads that have a more rounded articular head. This rectangular-headed rib may belong to the posterior dorsal region.

Phylogenetic relationships

Nothosaurus zhangi shares a set of derived characters with other Nothosauria, including the well developed lateral ridge of the surangular, the elongated and “scoop”-like mandibular symphysis, strongly procumbent and fang-like teeth in the anterior region of jaws and platycoelous centra with parallel lateral edges. Phylogenetic analysis based on a new species-level data matrix (Supplementary Note) recovered N. zhangi as one of the most primitive taxa within Nothosauria. This is relatively well supported by the Decay index (also known as Bremer support), although the relationship with several other basal taxa remains unresolved (Fig. 3). Bootstrap values (1000 replications) across the tree are generally low, mainly because of the fragmentary nature of Germanosaurus and many European Nothosaurus species and the new taxon introduced here, which can be only scored with ca. 12% of all characters. Only the clades consisting of (N. jagisteus (N. mirabilis, N. tchernovi)) have bootstrap values more than 50%.

Figure 3
figure 3

Phylogenetic relationships and stratigraphic occurrences of Nothosauria.

Consensus of the seven most parsimonious trees (TL = 234, CI = 0.3761, HI = 0.6239, RI = 0.4931, RC = 0.1854). Time scale was generated using TSCreator 6.1 (Available free at https://engineering.purdue.edu/Stratigraphy/tscreator/index/index.php). Schematic body outlines are scaled on the maximum size of individual species. Taxa with gigantism are highlighted with blue. Numbers beside nodes indicate Decay indices.

One of the important results of this phylogenetic analysis is the collapse of the monophyly of traditionally recognized Nothosaurus and Lariosaurus. To force the monophyly of Nothosaurus and Lariosaurus as traditionally recognized clades would require 16 additional steps, clear evidence that the two genera, as generally constituted, are not monophyletic. This is, however, not a surprise as some recent Chinese taxa have shown many mixed morphologies27,28 intermediate between Nothosaurus and Lariosaurus. Systematic revision of these two genera will be presented elsewhere. We adopt the conservative approach of referring this new species to Nothosaurus, with the acknowledgement that this genus requires further clarification in light of the phylogenetic results presented here.

Discussion

Among Triassic sauropterygians, N. giganteus from Central Europe achieved the largest skull size. The largest complete skull of N. giganteus reached a length of ca. 61 cm, while the largest complete lower jaw is 59 cm25. In the type of N. zhangi, the lower jaw measures 65 cm long, representing the largest jaw ever reported in a Triassic sauropterygian. Although N. zhangi is similar to N. giganteus in size, it differs significantly from the latter in morphological details, as recovered in the phylogenetic analysis (Fig. 3). This result demonstrates parallel evolution of gigantism in Triassic sauropterygians.

Estimating the body size of an extinct organism from incomplete remains is fraught with difficulty and it is easy to exaggerate. In this case, however, the close relationship of Nothosaurus zhangi to other nothosauroid sauropterygians and close similarity of skull and vertebral proportions among them suggests that a direct comparison might be possible. Among the more complete specimens from Europe, the ratio of skull: body length ranges from about 0.12 in N. giganteus to about 0.09 in Lariosaurus29. The length from the tip of the snout to the jaw articulation of N. zhangi measures about 60 cm, which gives a rough estimate of a 5–7 m total body length in N. zhangi.

Biotic recovery after the PTME in the sea was a protracted process lasting 5–10 million years, probably slowed down by the harsh environment in the Early Triassic, with repeated global warming and environmental crises and destruction of key habitats such as reefs12,13,14,15. Immediately after the catastrophe, most marine communities13, if not all11, were characterized by primary producers and opportunistic consumers, followed by the addition of meso-consumers. Predatory invertebrates and vertebrates appeared later. Gigantic apex predators, however, did not evolve until the establishment of a complex and stable ecosystem, as evidenced by the occurrence of the ichthyosaurs Thalattoarchon and Cymbospondylus in western USA, eastern Panthalassic province and Cymbospondylus and Nothosaurus giganteus in central Europe, western Tethyan province, both in the mid-late Anisian20,25 (Fig. 1).

Recently, Scheyer et al.11 reported a huge ichthyopterygian humerus supposedly from the Lower Triassic of western USA. The body size of this animal, by extrapolation, is around 11 m long11. However, this new specimen was collected from the surface of private land11, so its stratigraphic provenance needs to be confirmed by further scientific excavation. In addition, gigantic size alone does not necessarily indicate the existence of a long food chain. For example, the largest marine mammals, the baleen whales, mainly feed on zooplankton and small schooling fish30. The whale shark and the basking shark, the only two living shark species with body length over 10 m, are both planktivorous31. Such gigantic marine vertebrates all have very low trophic levels31,32,33. Therefore, if a gigantic ichthyosaur with body length of around 11 m was indeed present in the Early Triassic, it is possible that this ichthyosaur fed on small organisms by batch feeding34 like the gigantic Late Triassic shastasaurid ichthyosaurs35,36, baleen whales30 and whale and basking sharks31. Whereas such planktivory has been suggested in giant ichthyosaurs, there is no evidence that sauropterygians, including nothosaurs, ever adopted such a diet. Therefore, a large ichthyosaur bone might indicate an apex predator or planktivore, whereas a giant nothosaur was almost certainly an apex predator. This is confirmed in the case of our specimen and the other giant Middle Triassic nothosaurs, by their elongate, predatory teeth25.

In the eastern Tethyan province, gigantic apex predators were previously unknown in the Middle Triassic, not appearing until their occurrence in the Carnian Guanling biota in South China21,37. Considering the intensive sampling in several beautifully preserved fossil Lagerstätten in South China in recent decades21, this might have implied that ecosystem recovery in the eastern Tethyan region was somehow slower than in the western Tethyan and the eastern Panthalassic region. Now, the discovery of Nothosaurus zhangi in the Luoping biota fills this spatio-temporal gap and indicates a globally synchronous complete biotic recovery of shallow marine ecosystems after the PTME. The diversity of marine reptiles (Fig. 4 and Table 1) and the reconstructed complex community structure in the Luoping biota (Fig. 5) also supports the ecological recovery evidenced by the occurrence of the gigantic nothosaur.

Table 1 Ecological guilds of marine reptiles from Luoping biota
Figure 4
figure 4

Marine reptiles from Luoping biota.

(a) Phalarodon atavus (LPV 30872)38. (b) Mixosaurus cf. panxianensis (LPV 30986)40. (c) cf. Atopodentatus (LPV 30172)41. (d) Sinosaurosphargis yunguiensis (LPV uncatalogued)42. (e) Diandongosaurus acutidentatus (IVPP V 17761)45. (f) Dianopachysaurus dingi (LPV 31365)44. (g) Lariosaurus sp. (LPV 301881). (h) Nothosaurus zhangi (LPV 20167). (i) cf. Qianosuchus (LPV 31411). (j) Dinocephalosaurus cf. orientalis (LPV 30174). Scale bar equals 1 cm in f and i and 10 cm in all others.

Figure 5
figure 5

Hypothesized food web of Luoping biota, updated from ref.

23. Silhouettes 6, 10, 14 and 17 were adapted and revised from ref. 17, which were the original artwork of B. Scheffold (Zurich).

Among the marine reptiles in the Luoping community, ichthyosaurs dominate the fauna. They are exclusively composed of medium- to large-sized (see ref. 21 for definition of size classes in Triassic marine reptiles) mixosaurid ichthyosaurs. Mixosaurs are inferred to have been widely foraging predators that actively searched for their prey38. At least two forms could be differentiated among the Luoping mixosaurs. One is Phalarodon atavus38, characterized by a largely homodont dentition adapted for externally soft prey (Fig. 4a). The other (Fig. 4b) is a taxon that resembles Mixosaurus panxianensis, for which the heterodont dentition suggests a broad spectrum of prey39,40.

A recently described large-sized marine reptile, Atopodentatus unicus, a possible relative of the sauropterygians41, has been interpreted as a bottom-living filter feeder that consumed microorganisms or benthic invertebrates such as sea worms (Fig. 4c). Two medium-sized saurosphargids42,43 (Fig. 4d) might have been omnivorous41 but given their body plan likely fed on slow-moving or benthic prey. Among sauropterygians from Luoping, eosauropterygians include two small-sized pachypleurosaurs44,45,46 (Fig. 4e,f), a yet undescribed medium-sized Lariosaurus (Fig. 4g) and the gigantic Nothosaurus zhangi (Fig. 4h), all characterized by fang-like and conical teeth in the anterior region of the jaws and these all specializing in pincering their prey.

Archosauromorphs from Luoping comprise archosaurs and protorosaurs. Archosaurs are represented by scattered teeth (Fig. 4i), in which the serrated margin is a typical feature of the clade. From the time-equivalent Panxian fauna, Qianosuchus has been reported47, a large-sized marine predator with dagger-like teeth, which could catch any prey available with a forceful strike. It is probable that archosaurian teeth from Luoping may come from the same species or a closely related taxon. Considering the fragmentary nature of the material from Luoping, however, discovery of more complete specimens is needed to confirm the presence of Qianosuchus-like predators in the Luoping community. Another large-sized archosauromorph from Luoping is the protorosaur Dinocephalosaurus (Fig. 4j), characterized by an elongated neck that presumably helped it to catch its prey with a rapid strike before the prey had even detected it48.

With its gigantic skull and the presence of large and conical canine teeth, Nothosaurus zhangi must have occupied the top level of the food web in Luoping (Fig. 5). Although the postcranial skeleton of N. zhangi is incomplete, its phylogenetic placement and the generally conservative postcranial anatomy of nothosaurs suggests that this species may have had a similar body shape to other nothosauroids, which were probably ambush predators49 using their forelimbs as the primary propulsive organs50,51. Nothosauroid sauropterygians in general were well adapted to prey on fish and cephalopods52 and occasionally, the preservation of rare gastric contents53 and coprolites54 provides direct evidence that small marine reptiles were part of their diets. Thus, N. zhangi could prey on large fish and other marine reptiles from Luoping and attack their prey with a quick strike and pierce them with its large fang-like teeth. The discovery of N. zhangi at Luoping thus provides direct evidence that a complex and stable ecosystem in the ocean had been established globally by the early Middle Triassic, after the devastating PTME.

Although biotic recovery after the PTME has generally been thought to have been slow in the sea13, some recent discoveries have challenged this hypothesis. For example, some marine clades such as ammonites3 and conodonts5 diversified rapidly within a short interval after the catastrophe, leading to the suggestion that post-extinction recovery may not have been as slow as previously envisioned. However, the diversity of most other groups remained relatively low until the Middle Triassic13. In addition to the generally low diversity, the Early Triassic ocean was also characterized by a harsh and unstable environment, as evidenced by large fluctuations of carbon12 and oxygen isotopes14,15, sharp warming episodes, episodic widespread oceanic anoxia, the coral gap13 and the lack of rich and diverse marine faunas across the globe hosting gigantic apex predators. Thus, it is unlikely that healthy and stable marine ecosystems could have become globally re-established before the environment had stabilized. The appearance of gigantic marine reptile predators across the Panthalassic and Tethyan regions in the Middle Triassic represents an important indicator of climax recovery.

Methods

To clarify the phylogenetic relationships of Nothosaurus zhangi, we constructed a new species-level data matrix (Supplementary Note), which is a combination of phylogenetically informative characters for resolving the phylogenetic interrelationships of Nothosaurus55 and Lariosaurus27, as well as 20 new morphological characters that have not been used previously. These new characters are Characters 7, 9, 14, 16, 19, 23, 25, 32, 34, 36, 38, 39, 40, 41, 42, 43, 44, 46, 59 and 66 listed in the Supplementary Note. Thus, there are 74 informative morphological characters in total. Informative proportional characters from previous analyses27,55 were retained. However, many states were redefined for those proportional characters based on gap coding.

Based on several recent global phylogenetic analyses of sauropterygians17,44, Pachypleurosauria and Simosaurus were selected as successive outgroups for the analysis of interrelationships of Nothosauria. All currently recognized species of Nothosauria were included in the data matrix, except for Micronothosaurus stensioi, Nothosaurus cymatosauroides and Ceresiosaurus lanzi, which are less well known. In addition, coding of N. winterswijkensis was combined with N. marchicus based on recent revisionary work56,57.

PAUP Version 4.0 Beta 10 for Windows58 was used to analyze the data matrix. Heuristic search (ADDSEQ = RANDOM, NREPS = 1000, HOLD = 100, with other settings default) was performed to search the most parsimonious trees. The character list and data matrix were constructed using NDE Version 0.5.0 (available free at http://taxonomy.zoology.gla.ac.uk/rod/NDE/nde.html).