Form and function of the pelvic girdle of Thalattosuchia and Dyrosauridae (Crocodyliformes) Author Scavezzoni, Isaure Universite de Liège, Evolution and Diversity Dynamics Lab, All. du Six Août 14, 4000 Liège (Belgique) isaure. scavezzoni @ gmail. com v. fischer @ uliege. be isaure.scavezzoni@gmail.com Author Fischer, Valentin Universite de Liège, Evolution and Diversity Dynamics Lab, All. du Six Août 14, 4000 Liège (Belgique) isaure. scavezzoni @ gmail. com v. fischer @ uliege. be v.fischer@uliege.be Author Johnson, Michela M. Department of Palaeontology, Staatliches Museum für Naturkunde Stuttgart, Museum am LÖwentor, Rosenstein 1, 70191 Stuttgart (Germany) michela. johnson @ smns-bw. de michela.johnson@smns-bw.de Author Jouve, Stéphane Sorbonne Universite, BUPMC - Pôle Collections, Tour Zamansky, 15 étage, bureau 1513, 4 place Jussieu, 75252 Paris Cedex 05 (France) stephane. jouve @ sorbonne-universite. fr stephane.jouve@sorbonne-universite.fr text Geodiversitas 2024 2024-05-02 46 6 135 326 https://sciencepress.mnhn.fr/sites/default/files/articles/pdf/g2024v46a6.pdf journal article 10.5252/geodiversitas2024v46a6 1638-9395 11106598 urn:lsid:zoobank.org:pub:6ACF6A79-9149-4781-808D-478668673EB6 CONGOSAURUS BEQUAERTI DOLLO , 1914 For measurements, see Tables 7-9 . The pelvic girdle of Congosaurus bequaerti MRAC 1806 ( Fig. 72 ) is limited to the ilium, as the ischium and pubis have not been recovered for this taxon. The ilium of Congosaurus bequaerti sharply contrasts with that of other thalattosuchians in possessing a well-developed postacetabular process (which specifically differs from that of metriorhynchoids), along with a short (anteroposteriorly) but thick (lateromedially) preacetabular process, and deeply carved medial attachments sites for the sacral ribs.Yet, while teleosauroid thalattosuchians also display a postacetabular process, its relative size proportionally to the total anteroposterior length of the ilium is still inferior to what is observed in Congosaurus bequaerti , Hyposaurus natator or Acherontisuchus guajiraensis . Compared to Hyposaurus natator and Acherontisuchus guajiraensis , the ilium of Congosaurus bequaerti shows a greatly reduced acetabular perforation and, in parallel, a taller bony acetabulum. The ilium of Congosaurus bequaerti also displays a shorter preacetabular process in relation to Hyposaurus natator and Acherontisuchus guajiraensis , but a more massive and higher postacetabular process compared to Hyposaurus natator (as that of Hyposaurus natator is concave ventrally, and that of Congosaurus bequaerti is not). However, the postacetabular process of Acherontisuchus guajiraensis is dorsoventrally taller than that of Congosaurus bequaerti due to a more convex iliac crest. Other great differences between Hyposaurus natator and Congosaurus bequaerti include the difference of inclination between the pubic and ischial peduncles, the number of attachments sites for the sacrals medially (in Congosaurus bequaerti the posterior-most imprints are almost fused), and the roughness of the iliac blade. The depth and number of attachments sites is a distinctive feature of dyrosaurids versus thalattosuchians and extant crocodylians. The preacetabular process of Congosaurus bequaerti is short and thick: anteriorly, the preacetabular process does not protrude much from the main mass of the ilium similar to Dyrosaurus maghribensis , but unlike Hyposaurus natator and Acherontisuchus guajiraensis . In addition, the dorsoventral thickness of the preacetabular process of Congosaurus bequaerti reaches both its anteroposterior length and mediolateral width. The peak of the preacetabular process is truncated, and thus is not positioned midway but rather ventrally. Still, the preacetabular process points anteriorly, but with a small dorsal component. There is a rugged area covering the lateral side of the preacetabular process which stretches out both ventrally and posteriorly, but also laterally up until it meets with the supraacetabular crest. Posteriorly, this rugged area stops just under the start of the convex iliac blade dorsally. The region directly bordering the supraacetabular crest shows a subtle change in coarseness. The preacetabular process reaches its maximal mediolateral thickness at its junction with the supraacetabular crest. In Congosaurus bequaerti ( Fig. 72 ), the supraacetabular crest appears to be made of two distinct portion: a relatively wide anterior rugged and laterally protruding part, and a more slender posterior rim (like in Acherontisuchus guajiraensis and Hyposaurus natator ). This rim may actually not be part of the supraacetabular crest as it appears to be a simple byproduct of two adjacent convex areas, even if it is delimiting the acetabulum and the postacetabular process. Comparatively, Dyrosaurus maghribensis also displays an anterior portion laterally prominent but its posterior rim is wider than in Congosaurus bequaerti . In Mecistops cataphractus the supraacetabular crest is identified thanks to its relief and rugged texture, but it is slightly more difficult in Caiman crocodilus as only the depth is present. For Congosaurus bequaerti , the supraacetabular crest will be limited to the coarse and prominent ridge, similar to extant crocodylians and thalattosuchians. Acherontisuchus guajiraensis and Hyposaurus natator possess a similar supraacetabular crest, but Hyposaurus natator YPM VP. 000753 shows a smoother dorsal area overhanging the process. The supraacetabular crest borders the anterior half of the acetabulum dorsally, and was presumably the attachment site for a soft tissue structure equivalent to the acetabular labrum of extant crocodylians. The dorsal margin of the ilium of Congosaurus bequaerti ( Fig. 72 ) is almost exclusively convex, with a very localized shallow recess at the dorsal base of the preacetabular process. Throughout its length, the iliac blade is scarred perpendicularly to its extension; this coarseness indicates the presence of a cartilage cap in vivo . The extremity of the postacetabular process points posteriorly, with a small dorsal component. As both dorsal and ventral borders of the postacetabular process are convex, the peak takes the shape of a ribbed vault similar to other dyrosaurids (e.g. Hyposaurus natator , Dyrosaurus maghribensis , Acherontisuchus guajiraensis ). Yet, the postacetabular process of Dyrosaurus maghribensis appears relatively less convex. Comparatively, the postacetabular process in extant crocodylians strongly differs from dyrosaurids and teleosauroids in being slender (i.e. more elongated anteroposteriorly and thinner dorsoventrally) and in possessing an enlarged rugged area in the place of its peak. In Congosaurus bequaerti , there is seemingly no transition between the convex ventral margin of the postacetabular process and the ischial peduncle, whereas in Hyposaurus natator the transition is marked by an inversion of concavity. The absence of a recessed area posteriorly to the ischial peduncle in Congosaurus bequaerti accounts for the thickness its postacetabular process ( Fig. 72 ). Anteriorly, the pubic peduncle of the ilium forms a thick rounded area which breaks the straight monotony of the anterior margin of the ilium. The junction between the anterior and dorsal margins of the pubic peduncle is achieved through a re-entrant angle giving the impression of an inverted triangle. Laterally, the facet of the pubic peduncle bears two triangular shapes as it is the case in other dyrosaurids (i.e. Hyposaurus natator , Dyrosaurus maghribensis , Acherontisuchus guajiraensis ): the apex of the anterior triangle meets with the anterior margin of the bone whereas the posterior shape appears like an isosceles triangle dorsally. The maximal dorsoventral height of the pubic peduncle appears to reach that of the ischial peduncle, as in Hyposaurus natator and Acherontisuchus guajiraensis . FIG . 72. — Left ilium of Congosaurus bequaerti Dollo,1914 , MRAC 1806 (holotype): A , lateral view; B , medial view. Arrow points anteriorly.Scale bar:1 cm. In Congosaurus bequaerti , the ischial peduncle is a large process resembling an isosceles triangle whose vertex angle is dorsally facing. The ischial peduncle borders the acetabulum posteriorly as it markedly protrudes laterally. Hence, it was presumably the attachment site for a structure equivalent to the crocodylian antitrochanter in vivo ( Tsai & Holliday 2015 ). At about 1/4 of its height starting from its base (ventrally), the lateral surface of the ischial peduncle is truncated to form the articular facet it shares with the ischium. The ischial and pubic peduncles are clearly separated by a gap, the acetabular perforation (similar to teleosauroids but contrary to metriorhynchoids), whose anteroposterior length is greater than its dorsoventral height, similar to Dyrosaurus maghribensis contra Hyposaurus natator and Acherontisuchus guajiraensis . In this way, the acetabular perforation of Congosaurus bequaerti appears relatively reduced. It is possible that the ischium (not preserved) bore a greater acetabular perforation to counter this structure on the ilium. In parallel, the inclination of the pubic peduncle could also help increase the size of the acetabular perforation, notably by necessitating a longer peduncle bridge on the ischium ( Fig. 73 ). The relative shortness of the acetabular perforation of Congosaurus bequaerti brings the ilium of teleosauroids like Lemmysuchus obtusidens to mind. Still, the acetabular perforation of Congosaurus bequaerti is more pronounced than in most teleosauroids (e.g. Lemmysuchus obtusidens , Charitomenosuchus leedsi , Neosteneosaurus edwardsi ). The ventral margin of the ischial peduncle is parallel to the tangent to the ventral margin of the postacetabular process: as a consequence, the distal margin of the ischial peduncle points both ventrally and posteriorly (similar to what is observed in Acherontisuchus guajiraensis , Mecistops cataphractus or Caiman crocodilus ) and forms an angle of approximately 120-125° with the ventral margin of the pubic peduncle (seeTable 10). Conversely, in Hyposaurus natator and Dyrosaurus maghribensis , the ventral margins of the ischial and pubic peduncles appear almost parallel and are mostly ventrally oriented. The different orientation and shape of the peduncles between the ilia of Congosaurus bequaerti and Hyposaurus natator plus Dyrosaurus maghribensis would presumably imply dissimilarity in the way the ischium connects to the ilium as well. The ‘open’ orientation of the iliac peduncles of Congosaurus bequaerti ( Fig. 72 ) resembles the configuration of both Acherontisuchus and Dyrosaurus maghribensis , but also that of extant crocodylians (e.g. Palaeosuchus palpebrosus RVC-JRH-PP1 [ Fig. 7 ], Mecistops cataphractus [ Fig. 8 ], Caiman crocodilus [ Fig. 9 ]) for which the anterior peduncle of the ischium and the pubic peduncle of the ilium do not contact each other, at least not entirely, and are covered (presumably for extinct taxa) with hyaline cartilage in vivo . It is possible that the configuration of Congosaurus bequaerti approximated that of extant crocodylians, in which the ischium either partly contacted the ilium anteriorly (see below), or was set further ventrally avoiding contact between the ischium and pubic peduncle of the ilium ( Fig. 73 ). In the case of Dyrosaurus maghribensis , however, the pubic peduncle of the ilium appears to have been in contact in its entirety with the dorsal articular surface of the ischium. The ventral margin of the pubic and ischial peduncles are not parallel which conveys the idea that the ischium of Congosaurus bequaerti presented an anterior peduncle similar to those of extant crocodylians (e.g. Palaeosuchus palpebrosus RVC-JRH-PP1 [ Fig. 7 ], Mecistops cataphractus RBINS 18374 [ Fig. 8 ], Caiman crocodilus NHMW 30900 [ Fig. 9 ]). Hence, for Congosaurus bequaerti , it is possible that the anterior peduncle of the ischium presented a short anteroposterior articular surface with a long dorsoventral articular surface like Mecistops cataphractus ( Fig. 8 ), with only a fraction of the anterior peduncle of the ischium meeting with the pubic peduncle of the ilium. Contrastively, it is also possible that there was a gap between the pubic peduncle of the ilium and the anterior peduncle of the ischium similar to Hyposaurus natator . The vertex angle of the pubic peduncle points dorsally, and the peduncle borders the acetabulum ventrally. The short dorsal extension of the pubic peduncle was presumably intended to leave more room to the acetabulum. Anteriorly, the margin ilium of Congosaurus bequaerti is concave, and is bordered ventrally by the pubic peduncle, and dorsally by the preacetabular process. The bony acetabulum of Congosaurus bequaerti , which can be viewed as a 3D parabola, is mediolaterally deep (i.e. along the coronal plane), similar to Hyposaurus natator . The deepest point of the bony acetabulum of Congosaurus bequaerti is located near the dorsal peak of the ischial peduncle.In contrast, the bony acetabulum of extant crocodylians appears greatly shallower (e.g. Palaeosuchus palpebrosus [ Fig. 7 ], Alligator mississippiensis [ Fig. 74 ], Crocodylus niloticus , Mecistops cataphractus [ Fig. 8 ], Caiman crocodilus [ Fig. 9 ]). This difference in depth, which is better observed in ventral view ( Fig. 74 ), is not given by the distance between the anterior-most peak of the pubic peduncle and the posterior-most peak of the ischial peduncle, but rather by their relative inclination: in extant crocodylians, both peduncles are mainly oriented laterally, and in Congosaurus bequaerti and Hyposaurus natator the peduncles appear to be predominantly facing each other. This difference gives extant crocodylian a relatively more open but shallow bony acetabulum, whereas the acetabulum of Congosaurus bequaerti , Hyposaurus natator and Acherontisuchus guajiraensis appears more narrow and deep. This fact is reinforced by the presence of an even more pronounced supraacetabular crest in Congosaurus bequaerti and Hyposaurus natator than in extant crocodylians ( Fig. 74 ). The significant mediolateral depth and relative anteroposterior narrow appearance of the bony acetabulum in Congosaurus bequaerti , Hyposaurus natator and Acherontisuchus guajiraensis presumably conveys a better bony congruence between the femoral head and the ilium than what is observed in extant crocodylians ( Tsai & Holliday 2015 ; Tsai et al. 2019 ). Similarly, the more prominent attachment sites for the capsular soft tissues on the ilium of Congosaurus bequaerti and other dyrosaurids (namely the supraaacetabular crest and ischial peduncle) hypothetically helped better border the femoral head (again better bony congruence), so that the articular capsule in Congosaurus bequaerti and other dyrosaurids was presumably formed by slightly more calcified elements than what is observed in extant crocodylians. Still, the majority of the caspular articulation was seemingly composed of soft tissues, as the dyrosaurid ilium does not display an actual ball and socket articulation in the way of extant birds ( Kuznetsov & Sennikov 2000 ; Tsai & Holliday 2015 ) and appears close to extant crocodylians. In parallel, the shape of the crocodylian and Congosaurus bequaerti and Hyposaurus natator femoral head is slightly different, with the femur of Congosaurus bequaerti and Hyposaurus natator displaying a rounder outline in dorsal view and a globally thicker head in the dorsoventral direction (visible in anteroposterior views). This larger femoral head could potentially account for the deeper bony acetabulum on the ilium. Besides the depth, the acetabulum of Congosaurus bequaerti also covers an extensive area both dorsoventrally and anteroposteriorly (i.e. within the sagittal plane), like the acetabulum of Hyposaurus natator and Lemmysuchus obtusidens . Comparatively, the acetabulum of Dyrosaurus maghribensis appears proportionally larger both dorsoventrally and anteroposteriorly. In comparison, metriorhynchoids differ from Congosaurus bequaerti and other dyrosaurids as they display a more limited acetabulum along both the sagittal and coronal planes (e.g. Tyrannoneustes lythrodectikos , Thalattosuchus superciliosus , Suchodus durobrivensis , etc.). FIG . 73. — Pelvic reconstruction of Congosaurus bequaerti Dollo, 1914 , MRAC 1806 (holotype); ischium and pubis have been recreated based on those of Hyposaurus natator ( Troxell, 1925 ) , NJSM 23368; A , lateral view; B , anterior view; C , ventral view; D , dorsal view. Arrow points anteriorly. Target indicates anterior. Original 3D models of NJSM 23368 courtesy of Candice Stefanic. Reconstructed bones only serve as a qualitative representation of the pelvic girdle of Congosaurus bequaerti . Scale bar: 5 cm. The relative extension (especially the dorsoventral height) of the bony acetabulum and acetabular perforation (see Table 11 ) differs between Congosaurus bequaerti and other dyrosaurids ( Fig. 72 ). Proportionally, the acetabular perforation of Congosaurus bequaerti appears limited facing the large bony acetabulum (which reaches about 13 times the height of the acetabular perforation), whereas Hyposaurus natator and Dyrosaurus maghribensis possesses a more developed acetabular perforation for a shorter bony acetabulum (about 4 times the height of the acetabular perforation). Comparatively, Acherontisuchus guajiraensis also displays a short acetabular perforation anteroposteriorly, but the latter forms a greater dorsal indentation than in Congosaurus bequaerti . The small size of the acetabular perforation in Congosaurus bequaerti could possibly be linked to the relative inclination of the pubic and ischial peduncles ( Figs 72 ; 73 ), as a similar relation is observed in extant crocodylians (e.g. Mecistops cataphractus , Caiman crocodilus ). Since the hip joint capsule presumably extended as far ventrally as the ischium as in extant crocodylians ( Tsai & Holliday 2015 ; Tsai et al. 2019 ), the relative dorsoventral height of the bony acetabulum and acetabular perforation probably did not impact the size of the hip joint capsule in vivo individually.Nevertheless, the dimensions of the acetabular perforation has a direct influence over the potential excursion of the femur, and its sole presence informs on the existence of intrinsic capsular ligaments ( Tsai & Holliday 2015 ). In extant crocodylians, the soft inner wall covering the acetabular perforation acts like a buffer during the femoral excursion, preventing the articular capsule to be sucked in ( Kuznetsov & Sennikov 2000 ). A smaller acetabular perforation could presumably mean that the need for a buffer is lower during a hypothetical high walk posture. It could also potentially imply that the femoral excursion and/or its constraints are somehow less significant than what is observed in extant crocodylians ( Kuznetsov & Sennikov 2000 ; Tsai et al. 2019 ). Yet, the actual acetabular perforation of Congosaurus bequaerti was probably greater than what is observed on the ilium solely. Indeed, the ischium of Congosaurus bequaerti has not been recovered and presumably largely contributed to the acetabular perforation. In extant crocodylians, the acetabular perforation is concurrently composed by the ilium and the ischium, but only the portion formed by the ilium bears one of the insertion of the ligamentum capitis femoris along its margin ( Kuznetsov & Sennikov 2000 ; Tsai & Holliday 2015 ; Tsai et al. 2019 ). Hence, a difference in height of the acetabular perforation may presumably change the location of the ligamentum capitis femoris insertion. This could have an influence on either the length of the ligament, or its insertion on the femoral head, or the relative position of the femur within the joint cavity. In extant crocodylians, this ligament holds the femur during its excursion provoked by the high walk posture (and the initial lack of congruence between the femur and the acetabulum) ( Kuznetsov & Sennikov 2000 ; Tsai et al. 2019 ). It is relatively safe to infer that Congosaurus bequaerti and other dyrosaurids possessed intrinsic ligaments as well, which accommodated the movements of the femur during elevated postures. The femoral excursion is an ability inherent to the shape of the femoral head and the shallowness of the bony acetabulum ( Kuznetsov & Sennikov 2000 ; Tsai & Holliday 2015 ); consistent shapes betraying a lack of congruence between the femur and acetabulum are encountered in the pelvic girdles of Congosaurus bequaerti and other dyrosaurids. Medially, the ilium of Congosaurus bequaerti ( Fig. 72 ) bears three distinct indentations indicating the sacral rib attachment sites for the sacral ribs, as in Acherontisuchus guajiraensis . These markings are borne directly medially to the pubic and ischial peduncles, and are thus separated by the acetabular perforation. The anterior attachments sites, corresponding to the first sacral, are composed of two elliptic imprints which are fused ventrally. The anterior-most imprint of the first sacral is wide, and extends from the tip of the preacetabular process to the mid-length of the ventral margin of the pubic peduncle. Its concavity is strictly anteriorly oriented so that its peak points posteriorly. The second indentation of the first sacral is directly annexed to the anterior-most imprint, with which it only shares a thin separating wall dorsally. Its elliptic shape is more squeezed, its peak points dorsoposteriorly, and its base is entirely comprised within the remaining half of the ventral margin of the pubic peduncle as it is bordered by the acetabular perforation posteriorly. Posterodorsally to the second attachment site is an oval rugged portion, which presumably molded the shape of the receding posterior part of the second sacral rib (which presumably ensures extra support) like similar shapes in extant crocodylians (e.g. see Alligator mississippiensis UF Herp 21461 on Fig. 74 ). As this part is convex, it probably did not serve as an anchor point. The attachment site for the second sacral appears like the mirrored version of the posterior-most indentation of the first sacral, yet slightly wider. Indeed, the greater axis of both ellipse seems to share about the same length, as well as a similar inclination angle with a 90° difference, so that the peak of the elliptic attachment site belonging to the second sacral points dorsoanteriorly rather than dorsoposteriorly. Ventrally, the attachment site for the second sacral extends from the posterior-most margin of the ischial peduncle to about the 2/3 of its length anteriorly. Around the center of the surface corresponding to the attachment site for the second sacral, there is a shallow isolated ridge, which is not connected to either the ventral margin of the ilium nor the dorsal margin of the attachment site. Looking at this area from a ventral view of the ilium reveals that the ridge actually defines the junction between two slightly different portions of the second attachment site: indeed, the posterior half is slightly more medially driven (i.e. deeper) than the anterior half making them appearance like a pair of steps. This difference in depths within the posterior attachment site reflects the existence of a minor subdivision in the ending of the second sacral rib. Nevertheless, the ridge does not define partially nor completely distinct elliptic indentations as only one summit is present for this attachment site, which contrasts with Hyposaurus natator and Dyrosaurus maghribensis . Despite this, the depth of the sacral rib attachment sites is relatively similar throughout the ilium. Undoubtedly, the depth of the sacral rib attachment sites forms one of the typical dyrosaurid features; in this way, dyrosaurid ilia differ from those of extant crocodylians and thalattosuchians. In Congosaurus bequaerti , the relatively small size of the posterior attachment site compared to the anterior one implies that the contribution of the first sacral in holding the pelvic girdle exceeded that of the second one ( Fig. 72 ), similar to what is observed in Dyrosaurus maghribensis Jouve et al. (2006) . Conversely, the contribution of each sacral rib appears to have been slightly more balanced for Hyposaurus natator and Acherontisuchus guajiraensis . In extant crocodylians the relation is inverted as the posterior attachment site (for the second sacral) is greater than the anterior one (e.g. Mecistops cataphractus [ Fig. 8 ] and Caiman crocodilus [ Fig. 9 ]). This dissimilarity between both Congosaurus bequaerti and Hyposaurus natator could be explained by the distinct orientation of their peduncles: the inclination angle between the peduncles of Congosaurus bequaerti possibly conveys a slightly different orientation of the whole ilium compared to that of Hyposaurus natator . A different adjustment of the ilium potentially impacted the transmission of load from the limbs, and thereby required differing anchor sites. The dissimilarity could also be caused by the general position of the ilium relatively to the axial skeleton: in Alligator mississippiensis ( Fig. 74 ) the ilium is shifted posteriorly compared to the centre of the sacral region, so that it is the second sacral that supports most of the ilium; in Palaeosuchus palpebrosus ( Fig. 7 ) the ilium is shifted anteriorly and thereby the opposite relation is observed where the first sacral bears most of the weight. In Palaeosuchus palpebrosus ( Fig. 7 ), two conditions are observed with the left ilium being supported by three processes (the two sacrals plus the first caudal), whereas in the right ilium the sacral ribs cover almost the entirety of the sacral rib attachment sites, leaving very little to no room for the lateral process of the first caudal.