Plagiosuchus pustuliferus

Damiani, Ross, Schoch, Rainer R., Hellrung, Hanna, Werneburg, Ralf & Gastou, Stéphanie, 2009, The plagiosaurid temnospondyl Plagiosuchus pustuliferus (Amphibia: Temnospondyli) from the Middle Triassic of Germany: anatomy and functional morphology of the skull, Zoological Journal of the Linnean Society 155 (2), pp. 348-373 : 368-369

PALAEOBIOLOGY OF PLAGIOSUCHUS PUSTULIFERUS

The cranial anatomy of Plagiosuchus pustuliferus is so extremely derived and unique that it merits a range of functional considerations. The most striking features of the skull include: (1) several bones that plesiomorphically frame the orbit of temnospondyls are absent in P. pustuliferus ; (2) the orbits are large and ‘merge’ posteriorly with an open temporal region, forming enormous orbitotemporal fenestrae; (3) the subtemporal fenestrae of the palate and adductor fossa of the mandible are correspondingly elongated, which correlates with a modification of parts of the cheek; and (4) the dentition consists of minute teeth restricted to very short tooth rows in both the upper and the lower jaws. Two points are particularly important when discussing these features. First, all traits are apparently correlated, which indicates a functionally integrated system, and second, the listed features are unique even when compared with other plagiosaurids.

In temnospondyls, the primitive condition was a completely closed (stegokrotaphic) skull roof. Nearly all temnospondyl taxa share this feature, and with it the possession of a large complement of dermal bones in the skull roof, palate and mandible. In stegokrotaphic skulls, which are also known from most other basal tetrapods, early amniotes and many osteichthyans, the adductor musculature is constrained in its dorsal extension by the temporal, medial and cheek series of skull elements ( Jarvik, 1980). These form the normal and probably plesiomorphic sites for its attachment, as is still seen in the extant actinopterygians Amia and Lepistosteus ( Lauder, 1980). In such crania, evolutionary changes in the length of jawclosing muscles are necessarily restricted to rearrangements within the rigid frame of the dermal skull, such as a re-direction of adductor muscles into an anterior alignment.

Stereospondyls, the most speciose and long-lived clade of temnospondyls, probably took rise from taxa with elongated snouts and correspondingly long gapes, such as Sclerocephalus and Archegosaurus ( Boy, 1990; Milner, 1990; Schoch & Milner, 2000; Yates & Warren, 2000), and on that basis evolved as the most impressive aquatic predators of their time. Despite evident changes in modes of feeding, highlighted by divergent dentition patterns, tooth shapes and muscle attachment sites, these taxa throughout retained the stegokrotaphic skull. During feeding, these large temnospondyls were not able to manipulate their mouth margins. Instead, their skulls were completely akinetic, as evidenced by their complicated, often three-dimensional sutures (Schoch, 1999). This stands in stark contrast to modern teleosts and salamander larvae, which utilize various forms of cranial kinetism to generate suction feeding ( Lauder, 1980). Note, however, that temnospondyl larvae, especially those of branchiosaurids, may have had somewhat kinetic skulls, although branchiosaurids are rather exceptional in many aspects of their larval development ( Schoch, 2002b, 2004).

A feature shared between teleosts, larval salamanders and temnospondyls (among many other vertebrates) is the hyobranchial skeleton. Evidence of this apparatus has been reported in numerous temnospondyls (e.g. Boy, 1974; Boy & Sues, 2000; Witzmann, 2004, 2005), most of which were more or less aquatic; this is indicated by their retention of lateral line sulci into adult stages, among other features ( Boy, 1974; Schoch, 2001). The preserved parts of the temnospondyl hyobranchial apparatus resemble most closely those of larval and neotenic salamanders ( Bystrow, 1938; Boy, 1974; Boy & Sues, 2000), and this has been used as a basis to infer suction feeding by enlargement of the buccal chamber. Thus, among the two components of suction feeding used by teleosts and salamander larvae, temnospondyls probably had one (buccal floor retraction) but lacked the other (manipulation of mouth margin by cranial kinetism).

With regard to plagiosaurids, the anatomy of Gerrothorax spp. provides interesting insight into the hyobranchial apparatus. Both G. rhaeticus and G. pustuloglomeratus possessed a well-ossified hyobranchial apparatus ( Nilsson, 1946; Hellrung, 2003), which includes a full range of ceratobranchials, ceratohyals and hypobranchials. In most other temnospondyl taxa these elements apparently remained cartilaginous, and only parts of the hypobranchium were ossified ( Bystrow, 1938; Olson, 1979; Witzmann, 2005, 2006; Witzmann & Schoch, 2006). The design of this apparatus in Gerrothorax spp. resembles that of the salamander Ambystoma tigrinum , which is a suction feeder during its larval stage ( Lauder & Shaffer, 1985, 1986). As Deban & Wake (2000) reported, aquatic salamanders with heavily ossified hyobranchia are very effective suction feeders. The very broad but short skull and the orientation of the teeth in Gerrothorax spp. are consistent with the hypothesis of buccal-floor-driven suction feeding in this genus: prey would have been easily grasped and secured by the battery of small, densely set and buccally curved teeth. Gerrothorax spp. was evidently a bottom-dweller, and Hellrung (2003) has argued for the existence of a branchial chamber framed chiefly by the massive pectoral girdle, the mandibles and the flank osteoderms, which probably ended in a posterodorsal gill opening. If this conclusion is correct, then Gerrothorax spp. could be envisaged as a suction feeder that used a unidirectional water current to capture and transport prey. It most likely employed a passive, sit-and-wait strategy for prey capture rather than active hunting ( Nilsson, 1946; Hellrung, 2003), given that the postcranial anatomy of the genus is indicative of poor locomotory ability ( Hellrung, 2003).

Despite its unique morphology (flattening of skull and body, extensive bony armour, etc.), Gerrothorax spp. retained a full complement of skull bones, relatively small orbits and a conventional suite of cranial sutures. Plagiosuchus pustuliferus differs substantially from Gerrothorax spp. in those features: the large orbits merge into massive orbitotemporal fenestrae, the prefrontal, postfrontal and (probably) postorbital are absent, and the jugal, quadratojugal and pterygoid have a highly apomorphic morphology. In addition, the marginal teeth are reduced in size and number, fangs are (apparently) absent, and the jaw elements are foreshortened to give an almost trans- verse tooth arcade. The composition and morphology of the hyobranchial skeleton is mostly unknown, but the?ceratobranchial elements indicate that it was at least partially ossified. The effective gape of P. pustuliferus was much shorter than is suggested by the position of the jaw articulation: the extended subtemporal fossa (palate) and adductor fossa (mandible) indicate that a large portion of the cheek margin housed adductor musculature and must have been covered by labial folds. The mouth entrance was therefore restricted to a very narrow opening anteriorly, a unique situation among temnospondyls. Even in the much shorter-skulled Gerrothorax spp. , the gape was substantially longer than in P. pustuliferus .

From this limited set of data, and in the absence of crucial anatomical information (particularly on musculature), only limited conclusions can be drawn with regard to what was probably a highly elaborate functional system. For instance, the relative elongation of the skull in Plagiosuchus pustuliferus , compared with Plagiosternum granulosum or Gerrothorax spp. , is evidently correlated with the elongation of the attachment area for the adductor musculature. By analogy with modern salamanders and frogs, the jawclosing musculature inserted in the adductor fossa of the mandible and along its lingual surface. The origin of these muscles clearly occupied a much larger area, ranging from the internal surface of the cheek (jugal, squamosal, quadratojugal) to the ventral side of the parietal and the flank of the sphenethmoid. In other temnospondyls, muscle scars are also found on the internal surface of the postorbital, postfrontal and supratemporal (e.g. Sclerocephalus , Mastodonsaurus ). In P. pustuliferus these bones are absent (postfrontal,?postorbital) or reduced (supratemporal), thereby reducing the surface area for muscle attachment. Whether the loss of certain circumorbital bones is a paedomorphic condition cannot, at present, be determined because of the absence of earlier growth stages. However, we contend that the loss of these elements to form the large emargination posterior to the orbit was probably related functionally to an increase in the area for expansion of adductor musculature. This situation parallels that of modern salamanders and frogs. By analogy with extant batrachians, the adductors probably originated at the free margins of the parietal, supratemporal and jugal. This situation forms an obvious convergence with that of modern lissamphibians.

In Plagiosuchus pustuliferus , the adaptive requirements for expansion of the jaw-closing musculature is puzzling. On the one hand, the relatively smaller mouth opening, which indicates more specifically directed suction on prey, together with the minute dentition and probable lack of palatal fangs, all suggests that P. pustuliferus preferred small prey items. On the other hand, the voluminous adductor musculature indicates the ability to generate strong force when closing the jaws. Stronger muscle forces than in other plagiosaurids are also indicated for the jaw-opening (depressor mandibulae) muscles, which attached to the hugely elongated PGA of the mandible. Experimental and functional–anatomical studies on the salamander Cryptobranchus have revealed how important behaviour is in this field ( Elwood & Cundall, 1994), which makes functional analysis of extinct systems very difficult. Based on the limited set of data, we hypothesize that P. pustuliferus retained a hyobranchial skeleton that enabled it to generate suction by depressing the buccal floor. The narrow mouth, which opened only anteriorly, and the weak dentition suggest a focus on small prey which was captured by specifically directed suction. Therefore, a powerful dentition with fangs to fix the prey was not essential. The powerful depressor mandibulae, indicated by the elongated PGA, probably played a role in the initial phase of suction, as has been found in actinopterygians and larval salamanders ( Lauder, 1980; Lauder & Shaffer, 1986). Conceivably, strong adductors were also required in such a system.