Ardeosaurus digitatellus, Hoffstetter, 1964

OSTEOLOGY OF ARDEOSAURUS DIGITATELLUS

Cranium

The holotype of A. digitatellus is represented by a slab containing a single individual in dorsal view ( Fig. 5A View Figure 5 ). Osteological material is preserved in the skull and postcranium, along with impressions of some of the missing elements in the surrounding matrix.

The snout elements are not preserved in the holotype of A. digitatellus , but some impressions indicate their outline. The anterior-most elements preserved are the prefrontals, which are quite large compared with the prefrontals of most lizards known to us, also suggesting the absence of lacrimals ( Fig. 5B, C View Figure 5 ). They do not bear surface ornamentation or a prefrontal boss. They are connected posteromedially to the frontals. The frontals are fused anteriorly, but still preserve a sutural line in their posterior half. It is not possible to determine if they became completely fused later in ontogeny. The frontals are expanded at the frontoparietal contact, and become constricted between the orbits. As in A. brevipes , the frontals do not expand anteriorly ( Mateer, 1982) and are less constricted between the orbits than they are in E. schroederi ( Table 1).

The frontoparietal suture is straight in A. digitatellus , and the parietals are completely fused ( Fig. 5A, B View Figure 5 ), bearing no remnants of the sutural line between them. The pineal foramen is present and located in the centre of the parietal table. The parietal is weakly emarginated on both lateral margins. This weak emargination, along with the orientation of the right squamosal (which is less displaced than its left counterpart), indicates that the upper temporal fenestra is narrower than the one observed in E. schroederi . The supratemporal processes are present, elongate, and also possess a medial excavation ( Fig. 6A View Figure 6 ). The posterior margin of the parietal between the supratemporal processes is reduced, forming a V-shaped posterior margin. A small protuberance observed posteromedially to the parietal table may represent a posteromedial process, or part of the processus ascendens of the supraoccipital, but we could not determine either because most of this element has eroded away. The supratemporal bone is short and is located between the posterior end of the supratemporal process of the parietal medially, and the squamosal laterally.

The jugals are not entirely preserved, but their postorbital process is evident on both sides of the skull, the left element still contacting the ventral margin of the posterior process of the postorbitofrontal. The postorbital and postfrontal on the right side of the skull are fused into a postorbitofrontal ( Fig. 6B View Figure 6 ). Medially, the postorbitofrontal has one parietal and one comparatively longer frontal process, both processes contributing to an extensive contact with the frontal and parietal. The posterior process of the postorbitofrontal extends far posteriorly, beyond the midpoint of the upper temporal fenestra.

The right squamosal is preserved lateral to the postorbitofrontal, indicating that the jugal extended far posteriorly, intervening between the postorbitofrontal and the squamosal. The posterior end of the left jugal is placed more anteriorly and did not reach beyond the posterior end of the postorbitofrontal. Whether the left jugal was displaced more anteriorly or the right one more posteriorly cannot be determined, but the impression of the anterior end of the left squamosal is located lateral to the posterior end of the left jugal, as observed on the right side. Therefore, it is concluded that the jugal prevented the postorbitofrontal from contacting the squamosal.

The squamosal is a moderately stout element, bearing a posteroventral process contacting the cephalic condyle of the quadrate. The posterodorsal process contacting the supratemporal process of the parietal is absent. Parts of both quadrates are preserved in articulation with the squamosal, supratemporal, and probably the supratemporal process of the parietal. The ventral margin of both quadrates has been displaced anteriorly, such that they lie with their anterior surfaces facing dorsally. The right element is better preserved and still preserves a quadrate conch posterolaterally, as well as a distinct tympanic crest.

Postcranium

The holotype of A. digitatellus has the postcranium preserved mostly as impressions in the calcareous matrix, with a few cervical and dorsal vertebrae, as well as parts of the pelvic girdle preserved as osteological material. Yet, the quality of preservation allows many details of the vertebral and especially the fore- and hindlimb anatomy to be discerned.

There are five or six cervical vertebrae ( Fig. 5B View Figure 5 ), a relatively low cervical vertebral count amongst lizards, and only the posterior-most one or two cervicals bear ribs, as also observed in the holotype of E. schroederi . The total presacral vertebral count is estimated to be 27, followed by two sacrals and six caudals. The anterior cervicals are preserved in dorsal aspect, and have very low neural spines. The cervical pleurocentra are very fragmented, preventing a detailed description, but they are similar in length to the anterior dorsal pleurocentra. The dorsal vertebrae are preserved in lateral aspect, revealing some information about the morphology of their centra, despite some limitations owing to deformation and some degree of shattering. The dorsal centra show no indication of bearing a posterior condyle. A structure resembling a condyle is observed anteriorly on the dorsal vertebrae however, as seen in lateral aspect ( Fig. 7A View Figure 7 ). In the caudal region, impressions of the anterior caudals also show the anterior border of the vertebrae to be convex, and the posterior border to be concave ( Fig. 7B – D View Figure 7 ), indicative of opisthocoely, a condition currently unknown for any extant or fossil squamate. However, it could also represent amphicoely with a sediment mould forming in the space between them. Therefore, the form of articulation cannot be determined with confidence, but it certainly is not procoelic as in most squamates. The neural arches are partially preserved and are as tall as the centra.

The ribs are single headed and articulated on circular synapophyses of the centra. The dorsal ribs are present up to the level of the penultimate presacral, but the condition in the last presacral cannot be determined owing to poor preservation. The sacral ribs, as well as the anterior caudal ribs, are preserved only as impressions, being laterally orientated and not forked.

No elements of the pectoral girdle are preserved, but parts of both forelimbs are preserved as impressions ( Fig. 5A View Figure 5 ). The humeri, radii, and ulnae do not reveal finer details of their anatomy, but it is possible to detect that the forelimbs were relatively small in relation to SVL ( Table 5). The left manus ( Fig. 7E View Figure 7 ) has digits 3, 4, and 5 preserved, indicating a phalangeal formula of?-?-4-5-3. The penultimate phalanges of those digits are elongate compared with the intermediate phalanges ( Tables 2 and 4), and are followed by unguals that are curved and tall at their bases.

Both ilia are preserved in dorsal view, revealing elongate posterior blades. The impression of the right pubis indicates that this element was narrow and strongly angled anteromedially. No impressions of the pubic tubercle are evident, suggesting that it was very small or absent. The ilia are not preserved either as bony elements or impressions.

The hindlimbs are also proportionally short compared with SVL, and similar in length to the forelimbs ( Table 5). The femora are relatively slender and not sigmoidal, as was also observed for E. schroederi . The tibia and fibula on both sides are of similar width to each other and diverge distally. The left pes is better preserved than the right, with parts of all digits in articulation, indicating a phalangeal formula of?-3-4-5-4. As in the forelimbs, the penultimate phalanges are elongate compared with the intermediate ones, being c. 40% longer than the immediately preceding phalanx in the third digit ( Fig. 8C; Tables 3 and 4). Most articulatory surfaces are not well preserved, but in the left pes, articulation between phalanges I and II in digit II is concave-convex, as reported for E. schroederi above.

The unguals are recurved and relatively tall at their bases owing to a ventral expansion beyond the level of contact with the penultimate phalanx, as in G. gecko ( Fig. 9 View Figure 9 ) and E. schroederi (see above): pedal claw height equals 0.65 mm. The ratio between claw height/area of contact between the claw and the penultimate phalanx equals 1.625.

FUNCTIONAL MORPHOLOGY

The functional morphologies of Eichstaettisaurus and Ardeosaurus have never been given extensive attention, despite some particularly interesting aspects of their limb morphology. Russell, Bauer & Laroiya (1997) briefly mentioned some gecko-like aspects of the feet of E. schroederi , suggesting a reduction in length of the fourth metatarsal and a slight divergence between metatarsals III and IV, but did not go further.

Relatively short fore- and hindlimbs

One of the most apparent aspects of the body form of both E. schroederi and A. digitatellus is the relatively short fore- and hindlimbs ( Figs 1A, B View Figure 1 , 5A View Figure 5 ), which are also similar in length to each other (see Tables 5 and 6). Short fore- and hindlimbs have been FL, forelimb length; F+T, femoro-tibial length; HL, hindlimb length; H+R, humero-radial length; MT, metatarsal; SVL, estimated snout – vent length. Absolute values represent means calculated from the right and left sides of the holotype.

proposed to be functionally advantageous for scansoriality (the capacity to climb) as shorter limbs bring the centre of gravity closer to the substrate and reduce the rotatory moment of the body in relation to the inclined plane ( Cartmill, 1985). This is further enhanced by the overall body depression observed in E. schroederi and A. digitatellus , a feature also observed amongst gekkotans, and previously suggested to enhance climbing performance in G. gecko ( Russell, 1975) - see also Table 6. However, recent studies have contested the correlation of short fore- and hindlimbs with scansoriality for lacertids and geckoes ( Vanhooydonck & Van Damme, 1999; Zaaf & Van Damme, 2001) and this particular feature may instead be related to phylogeny (see below).

Fore- and hindlimbs of similar lengths

Fore- and hindlimbs of similar lengths to each other may also contribute to stable climbing, as limbs of very distinct lengths would result in different stride lengths, and in the tendency to have fewer limbs maintaining contact with the substrate during fast locomotion. It is important to maintain grip in some climbing lizards, as documented for Lacerta oxycephala , which maintains three or four limbs in contact with the substrate most of the time during fast locomotion ( Arnold, 1973). Yet, the latter may not be applicable to geckoes with highly developed adhesive toepads, such as G. gecko , as the latter keeps only two limbs (and occasionally only one) in contact with the substrate most of the time ( Russell, 1975). This latter attribute, as displayed by pad-bearing geckoes, may be limited to those climbing lizards with adhesive toepads, owing to their greater clinging capacity in relation to frictional grip relative to other lizards.

The correlation between fore- to hindlimb ratio and habitat preference has been tested for lacertid lizards, and no significant correlation was found ( Vanhooydonck & Van Damme, 1999). However, as acknowledged by these authors, lacertids may use several other habitats, thus not usually being specialized in that regard. Therefore, the relevance of fore- to hindlimb ratios is dependent on particular clades owing to phylogenetic signal and the level of specialization for a particular habitat, as well as to the influence of functionally related structures (e.g. adhesive toepads). Therefore, fore- to hindlimb ratios, when taken alone, have to be interpreted with caution when trying to differentiate between scansorial and ground-dwelling habits for lizards.

Claws tall at their bases

Despite some ambiguity regarding the relevance of fore- and hindlimb lengths as adaptations to scansoriality, both the manus and pedes of E. schroederi and A. digitatellus bear claws that are relatively tall at their bases ( Figs 4E, 7E View Figure 7 , 8A – D, and descriptions with ratios above). This feature has been positively correlated with climbing in lizards, and advocated to be a functional adaptation, even when phylogenetic history is taken into consideration ( Zani, 2000; Tulli et al., 2009; Crandell et al., 2014). A possible explanation for this correlation is that higher claws exhibit a ventral expansion relative to the level of contact with the penultimate phalanx, when compared with terrestrial nonscansorial lizards ( Fig. 9 View Figure 9 ). The flexor tendon, which runs ventrally and inserts proximally in each phalanx (including the claw), would have a greater lever arm and increase holding strength against the substrate, thus aiding lizards in climbing ( Russell, 1975). In addition to height, the claws of A. digitatellus are more elongate and curved relative to those of E. digitatellus . However, the relevance of claw length to clinging performance is debatable, owing to contrasting conclusions ( Zani, 2000; Tulli et al., 2009; Crandell et al., 2014), especially when gekkotans (typically with short claws) are considered.

Elongate penultimate phalanges

Eichstaettisaurus schroederi and A. digitatellus also bear elongate penultimate phalanges in both their fore- and hindlimbs ( Fig. 8A – D). This feature has been strongly correlated with climbing habits in a variety of reptiles, including lizards ( Kavanagh et al., 2013). It has also been observed in other fossil lizards that have been interpreted as having had a scansorial lifestyle, such as Scandensia from the Early Cretaceous of Spain ( Evans & Barbadillo, 1998; Bolet & Evans, 2011) and Calanguban from the Early Cretaceous of Brazil (Simoes et al., 2015a).

Foot symmetry

Geckoes bear radiating digits, instead of the subparallel digits evident in most lizards, creating a symmetrical foot in relation to the typical lizard condition ( Fig. 8E, F). This feature is associated with the adhesive toepad climbing mechanism of geckoes, as this facilitates the spreading of the seta-bearing surfaces about a broad arc ( Russell, 1975). The latter allows for various combinations of digital orientation that maximize their potential in passive loading, thus aiding in the maintenance of grip in a variety of body orientations ( Russell, 1986; Russell et al., 1997; Russell & Oetelaar, 2016). Additionally, a symmetrical foot with digits that radiate distally helps in providing grip, as it allows the first and fifth digits to develop opposability ( Robinson, 1975; Rewcastle, 1983).

Foot symmetry is achieved by a variety of factors. Whereas most lizards have the first pedal metatarsal (MT I) of about half the length of MT II and III (MT III/MT I length ratio of 2:1) geckoes have a proportionally longer MT I in relation to MT III, with a MT III/MT I ratio between 1.3 and 1.5, with the greatest average amongst geckoes (1.47) being found in padless diplodactylines ( Russell et al., 1997). Eichstaettisaurus schroederi and A. digitatellus display a MT III/MT I ratio of 1.41 and 1.48, respectively ( Table 5), falling within the range reported by Russell et al. (1997) for extant geckoes.

Another factor contributing to foot symmetry is the reduction in length of MT IV. Whereas most lizards have a MT IV that is longer than MT III, geckoes have a MT IV shorter than MT III ( Russell et al., 1997; Russell & Bauer, 2008). Some varanids have similar lengths for MT III and IV ( Russell et al., 1997). This is also seen in Heloderma , but in both cases they are usually never shorter than MT III (T. R. Simoes, pers. observ.). Reduction in length of MT IV is observed in E. schroederi ( Fig. 8A – C), and may be the case for A. digitatellus ( Fig. 8C) as well, although this cannot be confirmed.

Other features related to foot symmetry are the broadened proximal head of MT IV, which greatly increases the angle between the shafts of MT III and MT V, as well as the reduction of imbrication amongst the metatarsals proximally. Both features help to create an expanded digital arc, and contribute to foot symmetry ( Russell et al., 1997). These features, however, are not observable for either E. schroederi or A. digitatellus . Eichstaettisaurus schroederi has a somewhat broadened proximal head of MT IV, but it does not seem to be proportionally larger than the condition seen in Iguana iguana and most other lizards. Therefore, both species display partial development of foot symmetry, which is more developed in E. schroederi .

A few important considerations regarding the foot symmetry of E. schroederi and A. digitatellus are relevant to understanding their consequence for the locomotion of these lizards, even though they are not necessarily linked to scansorial habits.

Metatarsals of quite distinct lengths contribute to the highly asymmetrical feet of most lizards. The distal tips of the metatarsals form a straight metatarsophalangeal line ( Fig. 8E), which is directed perpendicular to the parasagittal plane at rest. During limb retraction in sprawling locomotion this aids in maintaining the first three digits in contact with the substrate and provides even support amongst these digits for bearing the animal’s body weight ( Brinkman, 1980; Rewcastle, 1983). If they were of equal length, most of the weight would be concentrated on the first digit only, as a consequence of the lateral orientation of the femur. Metatarsals of similar length are usually observed only in lizards with more anteriorly orientated feet ( Brinkman, 1980). Therefore, the highly symmetrical metatarsals of E. schroederi and A. digitatellus indicate that their feet were probably more anteriorly, rather than laterally, orientated.

Another character of laterally orientated feet is the development of the ginglymoid bicondylar articulation observed with greater development on the first three pedal digits of most lizards. These joints, along with tendinous bands that lie along the lateral sides of the digits contribute to resisting lateral displacement while enabling dorsoventral flexion of the FL, forelimb length; HL, hindlimb length; SVL, estimated snout – vent length. Source references indicate studies where measurements were obtained, or calculated from.

phalanges ( Russell, 1975; Landsmeer, 1981). This is important considering that the posteriorly directed thrust during limb retraction has a perpendicular orientation relative to the laterally orientated phalanges ( Rewcastle, 1983) – see Fig. 8E and F – thus an interlocking mechanism represented by the bicondylar articulations provides greater stability. In lizards with more anteriorly orientated feet, this morphology would have a reduced adaptive significance.

Bicondylar articulations are not observed in E. schroederi or A. digitatellus , which have convexconcave joints in the intermediate phalanges where articulatory surfaces are well preserved. This is another indication that E. schroederi and A. digitatellus had more anteriorly orientated feet than most lizards, which bear bicondylar articulations. Finally, in E. schroederi and A. digitatellus the long axis of the first metatarsal makes a right angle with that of the tibia, a condition observed in G. gecko . This is compatible with anteriorly orientated feet. Although it is possible that this latter anatomy is an artefact of post-mortem changes in orientation, the metatarsal proportions and shape of interphalangeal articulations do support interpretation of the pedes of E. schroederi and A. digitatellus as having an anteromedial orientation similar to that found in geckoes and some platynotans.

In conclusion, there are limitations in differentiating ground-dwelling from scansorial lizard species based on limb to SVL ratios, as well as fore- to hindlimb length ratios, as discussed above. However, a dorsoventrally expanded claw and elongate penultimate phalanges have a significant correlation to habitat usage in lizards and other reptiles ( Zani, 2000; Tulli et al., 2009; Kavanagh et al., 2013; Crandell et al., 2014). Therefore, the combination of the latter, along with a depressed body and skull in E. schroederi and A. digitatellus , provides a suite of features that are usually found together only in species specialized to climbing. Additionally, metatarsal proportions indicate the feet of E. schroederi and A. digitatellus were oriented more anteriorly than in most lizards.