Chersina langebaanwegi Delfino, Cohen & Macaluso

Chersina langebaanwegi Delfino, Cohen & Macaluso sp. nov.

( Figs 1–10 View Figure 1 View Figure 2 View Figure 3 View Figure 4 View Figure 5 View Figure 6 View Figure 7 View Figure 8 View Figure 9 View Figure 10 )

1973 Chersina sp. - Hendey, p. 13.

1974 Chersina sp. - Hendey, p. 35.

1981 Chersina sp. - Hendey, p. 31.

1986 Chersina sp. - Meylan and Auffenberg, p. 288.

1997 Chersina sp. - Broadley, p. 111.

2000 Chersina sp. - Lapparent de Broin, p. 64.

2007 Chersina sp. - Daniels et al., p. 916.

2021 Chersina sp. - Georgalis et al., p. 49.

2023 Chersina angulata (Duméril in Schweigger. 1812) - Bhat et al., p. 2.

Derivation of name: The species name,meaning ‘of Langebaanweg’, derives from the site where it was collected, and is currently the only site at which it has been identified. Worth noting is that turtles represent the most common vertebrate taxon at Langebaanweg, and that this species is by far the most common taxon in the turtle assemblage. This species is dedicated to Prof. Margaretha D. Hofmeyr (1950–2020) for her commitment to the study of African extant tortoises and for having promoted the study of the fossil tortoises from Langebaanweg.

Zoobank LSID: urn:lsid:zoobank.org:pub:A8B93307-034E - 4D7F -96BC -B17F8F32B122.

Holotype: SAM-PQL-20734 : nearly complete female shell ( Figs 1–3 View Figure 1 View Figure 2 View Figure 3 ) GoogleMaps .

Paratypes: SAM-PQL-20925 : partial male shell with a nearly complete plastron, several peripherals and pygal ( Fig. 4 View Figure 4 ); GoogleMaps SAM-PQL-21109 : partially preserved shell of a male spe - cimen missing most of the costals and all the neurals ( Fig. 5 View Figure 5 ); GoogleMaps SAM-PQL-21174 : highly fragmentary, incomplete shell whose best -preserved elements are the nuchal, fifth neural, pygal, and, above all, both epiplastra ( Fig. 6 View Figure 6 ); GoogleMaps SAM-PQL-21749 : crushed but nearly complete shell with partially articulated elements, including well -preserved and well -exposed nuchal, pygal, epiplastra, and xiphiplastra ( Fig. 7 View Figure 7 ); GoogleMaps SAM-PQL-22076 : a com - plete male plastron ( Fig. 8 View Figure 8 ); GoogleMaps SAM-PQL-35023d : partially pre - served shell of a very large specimen represented mostly by an epiplastra, entoplastron, left hyoplastron, three neurals, pygal with the surrounding fragmentary peripherals, and three peri - pherals ( Fig. 9 View Figure 9 ) GoogleMaps .

Referred specimens: SAM-PQL-12740a : right epiplastron; SAM-PQL-12750c : left epiplastron; SAM-PQL-12754a : two epiplastra sutured to the entoplastron; SAM-PQL-12754b : left epiplastron; SAM-PQL-12754c : left epiplastron; SAM-PQL-24826b : right epiplastron; SAM-PQL-24827a : left epiplastron; SAM-PQL-24862a : right and left epiplastrons; SAM-PQL-24862b : right epiplastron; SAM-PQL-52640 : right epiplastron; SAM-PQL-73745 : right epiplastron; SAM-PQL-73747 : nuchal; SAM-PQL-73749 : right epiplastron; SAM-PQL-73750 , left epiplastron; SAM-PQL-73751 : nu - chal; SAM-PQL-73754 : nuchal; SAM-PQL-73757 : nuchal; SAM-PQL-73761 : right epiplastron; SAM-PQL-73763 : right epiplastron; SAM-PQL-73765 : nuchal; SAM-PQL-73769 : entoplastron; SAM-PQL-73777 : nuchal; SAM-PQL-73781 : left epiplastron; SAM-PQL-73789 : right epiplastron; SAM-PQL-73792 : left epiplastron; SAM-PQL-73795 : entoplastron; SAM-PQL-73799 : left epiplastron; SAM-PQL-73803 : left epiplastron; SAM-PQL-73807 : pygal; SAM-PQL-73808 : nuchal; SAM-PQL-73809 : pygal; SAM-PQL-73811 : right epiplastron; SAM-PQL-73819 : entoplastron; SAM-PQL-73825 : right epiplastron; SAM-PQL-73828 : right epiplastron; SAM-PQL-73829 : entoplastron; SAM-PQL-73831 : nuchal; SAM-PQL-73832 : pygal; SAM-PQL-73834 : right epiplastron; SAM-PQL-73836 : left epiplastron; SAM-PQL-73837 : left epiplastron; SAM-PQL-73845 : left epiplastron; SAM-PQL-73860 : right epiplastron; SAM-PQL-73862 : entoplastron; SAM-PQL-73863 : pygal; SAM-PQL-73876 : entoplastron; SAM-PQL-73877 : right epiplastron; SAM-PQL-73878 : right epiplastron .

Type locality and age: ‘E’ Quarry, Langebaanweg   GoogleMaps (18°9’ E, 32°58’ S), approximately 110 km NNW of Cape Town, South Africa. Upper Varswater Formation; Early Pliocene, about 5.1 Mya ( Roberts et al. 2011).

The fossil vertebrates of this locality have been studied since the late 1950s ( Singer and Hooijer 1958, Simpson 1971, Hendey 1973, 1974a, b, 1976) and have been still actively in - vestigated in recent years (among others see Groenewald et al. 2020, Valenciano and Govender 2020a, b, Govender 2021, Nacarino -Meneses and Chinsamy 2022, Rabe et al. 2022, Valenciano et al. 2022, Doughty et al. 2023, Govender and Marx 2023, Jannello and Chinsamy 2023, Jiangzuo et al. 2023, Matthews and Steininger 2023; and literature therein).

Differential diagnosis: Member of testudinid genus Chersina be - cause of the following combination of characters: wide vertebral scutes almost equal to the vertebral scutes; single trapezoidal suprapygal, notched pygal, and distinct gular protrusion covered by fused gulars scutes.

Differing from Chersina angulata regarding the following characters (see Supporting information for the comparison with C. angulata ): the shell is distinctly domed and not relatively low and elongated as in C. angulata ( Fig. 1E, F View Figure 1 ; Supporting in - formation, Fig. S1 View Figure 1 ); the nuchal is not as significantly arched (in anterior view) as that of C. angulata ( Fig. 1G View Figure 1 ; Supporting in - formation, Fig. S2 View Figure 2 ) and does not develop a ventral transversal ridge with medial notch and a dorsal shelf as in C. angulata ( Fig. 6B View Figure 6 ; Supporting information, Fig. S3A, B View Figure 3 ); the cervical is, on the dorsal surface, much broader and proportionally longer than that of C. angulata ( Fig. 2A, B View Figure 2 ; Supporting information, Fig. S3C, D View Figure 3 ); the first neural is hexagonal and not rectangular as in C. angulata ( Figure 1A, B View Figure 1 ; Supporting information, Fig. S4 View Figure 4 ); most of the neurals host the ribheads ( Figs 3A View Figure 3 , 9B View Figure 9 ; Supporting infor - mation, Fig. S5 View Figure 5 ) and not regularly the costals as in C. angulata ; the sixth marginal scute does not contact the third pleural scute as in C. angulata ( Fig. 1E, F View Figure 1 ; Supporting information, Fig. S6 View Figure 6 ); the anterior protrusion of the plastron (the gular protrusion of the epiplastra) is comparatively more developed than in C. angulata (especially in males; Figs 4A, B View Figure 4 , 5C, D View Figure 5 , 6D, E View Figure 6 , 8A, B View Figure 8 , 9C, D View Figure 9 , 10 View Figure 10 ; Supporting information, Fig. S7 View Figure 7 ); the dorsal surface of the epiplastral lips is nearly flat or very slightly convex and not vari - ably concave as in C. angulata ( Fig. 10 View Figure 10 ; Supporting information, Fig. S8 View Figure 8 ); the epiplastral lips in most of the cases end posteriorly with a nearly straight or moderately convex profile (in most cases not producing any significant pocket) and not with a very strong convexity as in C. angulata (which regularly has a very deep gular pocket) ( Figs 3B View Figure 3 , 5D View Figure 5 , 6D View Figure 6 , 8B View Figure 8 , 9D View Figure 9 , 10A, D View Figure 10 ; Supporting informa - tion, Fig. S9 View Figure 9 ); the entoplastron is wider than long instead of being longer than wide, or circular, as in C. angulata ( Figs 1D View Figure 1 , 3B View Figure 3 , 4A, B View Figure 4 , 5C, D View Figure 5 , 8A, B View Figure 8 , 9C, D View Figure 9 ; Supporting information, Fig. S10 View Figure 10 ).

Description: If not otherwise stated, the following descrip - tion is mostly based on the holotype SAM -PQL -20734 ( Figs 1–3 View Figure 1 View Figure 2 View Figure 3 ) as it is the only nearly complete and mounted shell so far available. There are no cranial, axial, or appendicular elements associated with the holotype and the referred specimens.

Carapace

The nearly complete female shell SAM -PQL -20734 (holotype) indicates that the carapace of this species was highly domed (length 23.6 cm, width 15.7, height 14.7 mm), with a nearly smooth dorsal profile both in lateral and anterior/posterior views. The posterior border is relatively curved inwards and not significantly flared. Very weak bosses are visible in lateral view ( Fig. 1C, D View Figure 1 ) that correspond to five vertebral scutes. In dorsal view, the anterior edge of the shell ( Fig. 2A, B View Figure 2 ) has a modest notch laterally delimited by the small protrusion of the first peripherals (corresponding to the sulcus between the first and second marginal) and, moreover, the central area of the carapace, corresponding to the bridge, is somewhat swollen if compared to the areas anterior and posterior to it. The anterior opening is kidney shaped because the anterior lobe of the plastron is ele - vated in the dorsal direction ( Fig. 3A View Figure 3 ; see description of the plas - tron). The posterior opening is rather broad ( Fig. 2C, D View Figure 2 ) and markedly much broader than in extant Chersina angulata . The carapace has no hinge and moreover there is no hinge between the carapace and plastron [as also indicated, among other, by the partial plastron SAM -PQL -20925 ( Fig. 4 View Figure 4 ) that is still sutured with some bridge peripherals] or within the plastron (see also SAM -PQL -22076— Fig. 8 View Figure 8 ).

Bony plates: The nuchal is a hexagonal element, about 1/3 broader than long (65 mm wide and 47 mm long in the holo - type SAM -PQL -20734), with borders that can be nearly straight (SAM -PQL -21174) or weakly concave (SAM - PQL -20734, SAM -PQL -21174) (the character has been scored as straight because of the weak concavity). The anterior edge can be straight (SAM -PQL -21174) or shallowly concave with, at least in some cases (SAM -PQL -20734), a small, acute notch at the centre. In anterior view, the anterior edge is only modestly arched (but not deeply arched as in C. angulata ). On the ven - tral surface, the transversal ridge marking the line of maximum thickness of the element (slightly posterior to the end of the ventral fold of the scutes) is straight and medially notched as in C. angulata that develops a sort of dorsal shelf (Supporting information, Fig. S3A, B View Figure 3 ).

There are eight neural elements and the neural formula is: 6,8,4,8,4,6A,4,6A. The first neural of SAM -PQL -20734 is anteroposteriorly elongated and hexagonal with anterolateral sides approximately equal to the posterolateral ones. Neurals from two to seven are wider than long (in absence of the proper state, the third has been scored as square but it is in fact rect - angular). The eighth is longer than wide and, at least on the better -preserved right side, it seems to be hexagonal (with a short anterolateral side). On the visceral surface of the neurals from two to five there is clear evidence for the emergence of very thin costal processes from the neural and not from the corres - ponding costals. In addition, the sixth neural may show the same condition as indicated by a small irregularity located close to the posterior border (and by the absence of a costal process on the corresponding costal). The costovertebral tunnel is, therefore, very narrow (much narrower than the width of the neurals). The two joined neurals of SAM -PQL -35023d ( Fig. 9 View Figure 9 ) have a sort of tubercle on the visceral surface close to the lateral edge. The preservation of the ventral surface of the isolated neural of SAM - PQL -21174 does not allow confirmation of this morphology in another specimen but its irregular surface could indicate the presence of costal processes ( Fig. 6 View Figure 6 ). Noteworthy is that the three preserved costals of SAM -PQL -21109 do not have a costal process on the medial area of their visceral surface. There is only one suprapygal that is trapezoidal in shape; its posterior edge is approximately straight. The female pygal of SAM -PQL -20734 is quadrangular and shows a small notch centrally located on the free, ventral edge; it is only moderately convex. The male pygal SAM -PQL -20925 is also quadrangular. The male pygal of SAM - PQL -21109 and likely also the partly preserved large male pygal SAM -PQL -35023d are hexagonal, with short anterolateral sides and long, concave posterolateral sides; the free, ventral edge is broadly concave in both cases.

There are eight pairs of costals. Those from two to six are markedly trapezoidal with alternate broad and narrow medial and lateral ends. The sixth and, above all, the seventh and eighth costals are posteriorly bent. Only the first, seventh, and eighth costals bear costal processes on their visceral surfaces (see de - scription of neurals above). The first costals have a very stout costal process, and the rib is raised in an evident, anteriorly con - cave arched ridge. The external surface of costals from two to five shows in the medial sector (close to the neurals) marked growth grooves and ridges.

There are 11 peripherals. Those from three to seven are involved in the bridge; their external surface is smooth, devoid of any signifi - cant keel (that is conversely present in Chersina angulata ). In the holotype SAM -PQL -20734 the first and second peripherals de - velop a small, distal projection corresponding to the intermarginal sulcus, but there are no intermarginal projections or interperipheral serrations in the posterior area. Conversely, in SAM -PQL -21109 there is a marked intermarginal projection on the ninth periph - eral (best preserved on the right side). The third peripheral hosts a small axillary scute that develops on the hyoplastron buttress as well. The seventh peripheral hosts part of the large inguinal scute (that is mostly developed on the hypoplastron).

Horny scutes: The cervical scute is rather large on both the dorsal and ventral surface of the nuchal. On the dorsal surface, it can be only slightly longer than broad (in SAM -PQL -20734, respect - ively 15.0 and 14.3 mm; Fig. 2A, B View Figure 2 ), and it is slightly longer than 25% of the nuchal. The cervical scute of SAM -PQL -21174 is distinctly longer than broad but its length corresponds to about 25% of the nuchal. On the ventral surface of the nuchal, the cer - vical is very long and has nearly parallel margins. The first pleural scutes slightly overlap the nuchal in SAM -PQL -20734 ( Fig. 2A, B View Figure 2 ) even less in SAM -PQL -21174 (where only the left side is well preserved) but not in SAM -PQL -21109 (that also preserves only the left side).

The first vertebral scute is as long as it is broad and has nearly parallel sides; its posterior border (the sulcus between the first and second vertebral) crosses the first neural close to its pos - terior edge.

Intervertebral sulci are located on the first, third, fifth, and sev - enth costals, whereas the interpleural sulci on the second, fourth, sixth, and eighth costals. Since the posterior edge of the fifth ver - tebral is located only slightly posteriorly to the suprapygal -pygal suture (and is, therefore, visible 1–2 mm posterior to the suture on the pygal of SAM -PQL -20734), this character was scored as if the sulcus coincides with the suprapygal -pygal suture (char - acter 112, state 1). With the exception of the nuchal and pygal areas, the pleuro -marginal sulcus overlaps the costo -peripheral suture. The five vertebral scutes are rather large, nearly equalling the width of the pleurals. The intervertebral sulci are located on the suture between the first and the second neurals, on the third, fifth, and eight neurals (very close to the anterior suture of the latter, not much posteriorly as in C. angulata ). The second marginal scute does not contact the nuchal or the first vertebral. The sixth marginal corresponds entirely to the second pleural (does not contact the third pleural; unlike in C. angulata ). The tenth marginal scute does not contact the fifth vertebral scute. The posterior sulcus of the fifth vertebral scute is placed slightly posterior to the suprapygal -pygal suture in SAM -PQL -20734 ( Fig. 2C, D View Figure 2 ) and even more posteriorly (therefore, clearly on the pygal) in SAM -PQL -20925, SAM -PQL -21109, and, above all, in the very large male SAM -PQL -35023d. The twelfth marginal scutes are fused in a single supracaudal scute.

Plastron

The large (293 mm long) and complete male plastron SAM - PQL -22076 ( Fig. 8 View Figure 8 ) provides a unique opportunity to assess the general morphology of this shell portion that otherwise is not entirely preserved in a single specimen. The anterior lobe is somewhat trilobed, longer than the posterior lobe (93 and 68 mm, respectively) and does not host any hinge. In dorsal (or ventral) view, the gularo -humeral sulcus corresponds to a dis - tinct constriction in the lateral profile of the element. Because of the anterodorsal orientation of the hyoplastra, the epiplastra are raised from the horizontal plane and their anterior protru - sion is anteriorly directed and approximately parallel to the hori - zontal plane but about 2 cm above it. Such a general morphology is also present in all the preserved plastra. The ventral plastral surface is distinctly medially concave in correspondence of the posterior region of the hyoplastra, the whole hypoplastra and at least the anterior region of the xiphyplastra, indicating that this specimen was likely a male (as SAM -PQL -20925 and the very large SAM -PQL -21109). The holotype SAM -PQL -20734 pre - serves a nearly complete plastron (only the left hypoplastron is missing) that is fully congruent with the one previously de - scribed but slightly smaller, less protruding anteriorly and flat, at least the latter character indicating that this specimen could be a female. The plastron extends anteriorly up to the third peripheral and posteriorly up to the seventh one.

Bony plates: The epiplastra are characterized by anterior projec - tions that in males can be very long, accounting in most cases (SAM -PQL -21174; Fig. 6 View Figure 6 ) for about half of the anteropos - terior length of this element, but in SAM -PQL -35023d they are even longer and represent nearly 80% of the total length (total length 95 mm; projection length 75 mm; Fig. 9 View Figure 9 ). In females, the projection is much less developed as shown by the holotype SAM -PQL -20734 and, among the referred materials, by SAM - PQL -73878 (compare male epiplastron SAM -PQL -12750c, Fig. 10A–C View Figure 10 , with female epiplastron SAM -PQL -73878, Fig. 10D–F View Figure 10 ).

In most cases, the projections are quite asymmetric (one of the two is much more developed than the other one), and, therefore, the interepiplastral suture is not located sagittally but deviates laterally. The long and broad dorsal surface of the lip is generally rather flat (SAM -PQL -21174) or, very rarely, modestly convex (SAM -PLQ -24826a). In dorsal view, the pos - terior edge of the gular lip is nearly straight or modestly convex but in some cases it can be moderately overhanging. There are however some exceptions such as the female right epiplastra SAM -PQL -73749 and SAM -PQL -73878 ( Fig. 10D–F View Figure 10 ) (both fitting with Chersina langebaanwegi because the golarohumeral sulcus indicates that the gular area reaches but does not enter the entoplastron), and the two epiplastra SAM -PQL -12754a (fitting with C. langebaanwegi also because of the shape of the associated entoplastron) showing a gular lip moderately convex and the male left epiplastron SAM -PQL -73837 or the female right epiplastron SAM -PQL -24826b that both have a distinctly overhanging posterior edge of the gular lip, forming a rather deep pocket (the edge of the lip reaches in the first and even sur - passes in the second the anterior margin of the epi -entoplastron suture). The visceral surface of the epiplastra, posterior to the pad, is raised in a variably developed tubercle along the suture with the entoplastron (approximately at mid -length of the an - terolateral suture of the entoplastron). The ventral surface of the epiplastra is relatively flat in the area covered by the single gular. The epiplastra thin out gradually in a lateral direction along the suture with the hyoplastron, in correspondence with the dorsal area covered by the humeral shield. The entoplastron is regu - larly wider than long but only slightly in the holotype SAM - PQL -20734 where it is 24.5 mm wide and 22.0 mm long, and much more in SAM -PQL -22076 where it is 8.1 mm wide and 36.3 mm long. The visceral surface of the epiplastron hosts a Y -shaped ridge whose anterior rami originates from the above described tubercles on the epiplastra. The central depression between the anterior rami is in all cases very modest. The ven - tral surface of the element is weakly concave. The hyoplastra are slightly longer than wide. They develop a rather robust and long buttress that reaches the first costal plate. Their medial sector is sexually dimorphic, being concave in males and flat in females. The hypoplastra are approximately rectangular. They develop a very robust posterior buttress that contacts the fifth and sixth costal plate. These elements show the same sexual dimorphism of the hyoplastra. The xipiplastra are not significantly dimorphic, with the exception of the presence of a medial concavity devel - oped in hyo -, hypo -, and xiphyplastra of males (the plastron is rather flat in females). In both sexes, the lateral edge is slightly convex with nearly no (SAM -PQL -22076; Fig. 8 View Figure 8 ) or a very modest (SAM -PQL -20734; Fig. 1 View Figure 1 ) notch corresponding to the femoroanal sulcus. The anal notch is in both sexes rather shallow and delimiting an open angle of about 125–140°. However, the male xipiplastra of SAM -PQL -21109 ( Fig. 5 View Figure 5 ) are different from those of the equally large male SAM -PQL -22076 ( Fig. 8 View Figure 8 ) be - cause their tip is bent upwards and the anal notch is much deeper and narrower (115°).

Horny scutes: The gular scutes are fused in a single unit that, in all cases ( Fig. 1C, D View Figure 1 ), except one, reaches posteriorly to the anterior edge of the entoplastron (without entering it). In SAM -PQL -21109 ( Fig. 5 View Figure 5 ), however, it stops anteriorly to the entoplastron as shown by the incomplete, but informative, left epiplastron. The gularohumeral sulcus is laterally concave on the dorsal surface and weakly anteromedially convex on the ven - tral surface. The angle at the junction of the two gularohumeral sulci (the right and the left) varies in relation to the form of the gular projection: from slightly less than 90° (SAM -PQL -21174), to approximately 90° (SAM -PQL -20734, 20925, 35023d), to slightly more than 90° (SAM -PQL -22076). The ratio between the length of the gulars and the humerals (the median antero - posterior length of the humerals) seems to vary according to the development of the gular projections and, therefore, the sex of the specimen: the female SAM -PQL -20734 has a gular as long as the humerals, whereas all the male plastral have humerals that are at least slightly longer than the gular. The humeropectoral su - ture regularly passes posterior to the entoplastron (that, there - fore, is crossed only by the interhumeral suture), and is regularly concave in the anterior direction but has a rather variable trend: it is much more open in the males SAM -PQL -20925 ( Fig. 4 View Figure 4 ) and SAM -PQL -22076 ( Fig. 8 View Figure 8 ), where the medial sector of the sulcus is nearly straight and perpendicular to the sagittal axis of the plastron, than in the holotypic female shell SAM -PQL -20734 ( Fig. 1C, D View Figure 1 ). The pectoroabdominal sulci are modestly convex in the anterior direction and reach medially the mid of the interhyoplastral suture. The abdominofemoral sulci are located close to the hypoxiphyplastral suture but do not reach it medi - ally. They are laterally slightly convex in the anterior direction and medially nearly straight. The femoroanal sulci are mod - erately sigmoid and reach, with an acute posterior angle, the interxiphyplastral suture slightly anterior to its midlength.

Axillary and inguinal scutes are present. The first ones (one per side) are relatively small and develop mostly on the third peripheral but also on the hyoplastron. The large inguinal scutes develop mainly at the posterolateral corner of each hypoplastron as well as partly on the seventh peripheral. The inguinal scute contacts the femoral scute both on the ventral and (especially) on the lateral surface of the hypoplastron. In visceral view, there is a significant covering of the plastral scutes on the dorsal sur - face of the bony elements delimiting the anterior and posterior aperture of the shell.

Critical specimens referred to Chersina langebaanwegi

The tentative referral of the female right epiplastron SAM - PQL -24826b to C. langebaanwegi , despite the dorsal convexity of the gular lip and its exceptional posterior development (somehow similar to the extant C. angulata ), is based on the fact that the same degree of lip convexity is shared by the anterior portion of the female anterior plastral lobe SAM -PQL - 12754a that includes the two epiplastra and the entoplastron: the latter has an entoplastron wider that long, and the posterior edge of the gular lip in the epiplastron is straight, as typically shown by most of the specimens of C. langebaanwegi . A convex epiplastral lip also characterizes the male left epiplastron SAM - PQL -24827a which shows a straight posterior edge of the gular lip. The same condition (convex epiplastral lip poster - iorly straight) is shown by the right female epiplastron SAM - PQL -73749. Both SAM -PQL -24826b and SAM -PQL -24827a have the tip of the gular protrusion slightly pointing in the anteroventral direction. Among the referred material, a convex epiplastral lip associated with a variably shaped posterior edge of the lip (and, therefore, a variably developed pocket) is pre - sent in the male epiplastra SAM -PQL -73750 and SAM -PQL - 12740c. The female right epiplastron SAM -PQL -73878 shows an intermediate condition in having a rather flat gular lip that ends posteriorly with a marked convexity that gives origin to a moderate pocket ( Fig. 10F View Figure 10 ).

Phylogenetic relationships of the new extinct species from Langebaanweg

The phylogenetic analysis resulted in the consensus trees re - ported in Figure 11 View Figure 11 . The first round of the analysis without a molecular scaffold ( Fig. 11A View Figure 11 ; Consistency Index (CI) = 0.293; Retention Index (RI) = 0.620; Most Parsimonious Trees (MPT) = 47.997) found two most parsimonious trees, and the second round of tree bisection and reconnection retained one most parsimonious tree.

In the analysis with the molecular scaffold ( Fig. 11B View Figure 11 ; CI = 0.268, RI = 0.572; MPT 53.1423), the New Technology Search found two most parsimonious trees, and the second round of tree bisection and reconnection retained two most par - simonious trees. Both the analyses confirm that the new extinct species from Langebaanweg is the sister taxon of C. angulata . Remarkably, despite sharing with Astrochelys spp. a highly domed shell, and with A. yniphora a long gular protrusion and a single gular shield, it is not retrieved in the clade of Astrochelys .