SYSTEMATIC PALEONTOLOGY
Order ANSERIFORMES
Family PARANYROCIDAE Miller and Compton, 1939
Genus
CYGNOPTERUS Lambrecht, 1931
Type Species—
Sula affinis van Beneden, 1883.
Included Species—Type species from the lower Oligocene of Belgium and England,
C. lambrechti Kurochkin, 1968from the lower Oligocene of Kazakhstan, and (tentatively)
C. alphonsi Cheneval, 1984from the Lower Miocene of France.
Remarks—
Cygnopterus, as a fossil genus of swans (Cygnini), was erected for Sula affinis van Beneden, 1883from the lower Oligocene of Belgium (Lambrecht, 1931, 1933). Later, two more species,
Cygnopterus lambrechti from the lower Oligocene of Kazakhstan (Kurochkin, 1968) and
C. alphonsi from the Lower Miocene of France (Cheneval, 1984), were described within this genus. A probable larger representative of
Cygnopterus was described from the upper Eocene of Xinjiang (North-Western China; Stidham & Ni, 2014). Most recently,
C. neogradensis Kessler and Hír, 2009, was described from the Upper Middle Miocene of Hungary (Kessler & Hír, 2009). The assignment of
Cygnopterus to Cygnini has been questioned by Louchart et al. (2005) based on the structure of the humerus, and Mayr (2005, 2022) noted that phylogenetic affinities of this genus require a revision. Mayr and Smith (2017) showed that the tarsometatarsus of the youngest species
C. alphonsi is very similar to that of species of
Paranyroca, swan-sized anseriform birds known from the Early Miocene of Europe and North America. However, close affinities between the Early Miocene
C. alphonsi and the early Oligocene species of
Cygnopterus have remained unconfirmed so far (Mayr & Smith, 2017).
Cygnopterus alphonsi was synonymized with an Early Miocene taxon
C. senckenbergi Lambrecht, 1931, a swan- like anseriform bird from France (Mlíkovský, 2002), but this action was not accepted by Louchart et al. (2005).
The genus
Cygnopterus has been repeatedly compared with flamingos (Phoenicopteridae) and the flamingo-like birds Palaelodidae.
Cygnopterus lambrechti was synonymized with the fossil phoenicopteriform bird “
Agnopterus”
turgaiensis Tugarinov, 1940 from the same deposits (Mlíkovský & Švec, 1986) and I previously also considered
C. lambrechti to represent a separate form of flamingo, distinct from
A. turgaiensis (see Zelenkov, 2013). Mayr (2008, 2022; Mayr & Smith, 2002) noted that the humerus, scapula and coracoid of
C. affinis are very similar to the corresponding bones of the alleged phoenicopteriform
Agnopterus (
Headonornis)
hantoniensis (Lydekker, 1891) from the upper Eocene and (presumably) lower Oligocene of England (Harrison & Walker, 1977, 1979; but see below). In addition, the coracoid of
C. alphonsi is similar to that of the coeval Early Miocene palaelodid
Megapalaelodus goliath Milne-Edwards, 1868 (see Mayr, 2008). However, in contrast, the tarsometatarsus of
C. alphonsi has a typical anseriform morphology (Mayr, 2008), and the femur of
C. affinis differs from that of flamingos and
Headonornis (Mayr, 2022)
This apparent morphological conflict has led to confusion regarding the previous identification of
C. alphonsi and
Agnopterus (
Headonornis)
hantoniensis bones. First of all, it should be noted that the morphology of
C. affinis, the type species of
Cygnopterus, agrees better with anseriforms than with phoenicopteriforms. This species was originally established based on five bones (syntypes) said to be from one individual, which is possible but not certain (Mlíkovský, 2002). The coracoid from the syntype series of
C. affinis is similar to those of anseriforms in that the acrocoracoid part does not expand in depth cranially: i.e., the ventral margin of the shaft at the level of the facies articularis humeralis runs subparallel to the dorsal margin of the bone (see Lambrecht, 1933:fig. 127J, K), while in phoenicopteriforms and palaelodids, the coracoid expands markedly cranially, and the ventral margin of the shaft is significantly inclined relative to the dorsal one. The latter morphology is clearly visible in the coracoid of
Headonornis (see Harrison & Walker, 1976
l. 3, N, O), which is similar to that of
C. alphonsi and indeed strongly resembles those of phoenicopteriforms and palaelodids, as noted earlier (Mayr, 2008). The scapula of all phoenicopteriforms and palaelodids is characterized (Fig. 2H) by a straight or even slightly convex dorsal margin of the bone at the level of the facies articularis humeralis (Ericson, 1999; Woolfenden, 1961; Zelenkov, 2021b). The scapula, which was attributed to
C. affinis by Lambrecht (1931), has a different morphology, with a characteristic concavity in the outline of the dorsal margin of the bone at the level of the facies articularis humeralis. This morphology is more consistent with the condition in modern anseriforms (with the exception of the aberrant anhimids and anhimid-like
Anachronornis ahnimops Houde, Dickson, and Camarena, 2023, from the Paleocene of North America (Houde et al., 2023), and some geese have a similarly concave, albeit less pronounced, dorsal bone outline. The femur morphology of
C. affinis also does not support assignment to flamingos, because of the well-developed collum femoris (reduced in phoenicoteriforms), poorly protruding proximally trochanter femoris, lacking an extensive pneumatization (strongly proximally protruding and pneumatized in phoenicopteriforms), gracile shaft (robust in phoenicopteriforms), and medially expanded distal end (practically unexpanded in phoenicopteriforms).
The humerus of
C. affinis is definitively sigmoid in dorsal view (see Lambrecht, 1931
l. 2, figs. 7, 8; 1933:fig. 127B), as in anseriforms (in modern members of the order, the curvature is less expressed, although it is still quite pronounced in anhimids). Species of Phoenicopteriformes, in contrast, have a straight humerus in dorsal view. The overall shape of the distal end in
C. affinis is quite similar in shape to that of swans, being distinctly expanded dorsoventrally (especially due to the enlarged epicondylus dorsalis; see Lambrecht, 1933:fig. 127A). The distal end of the humerus in phoenicopteriforms and palaelodids is much less expanded. The incisura between the condyles is deep in
C. affinis as in non-anhimid Anseriformes (so that the condylus dorsalis protrudes conspicuously distally in cranial view), while in phoenicopteriforms and palaelodids this incisura is distinctly smaller, and the dorsal condyle does not protrude distally (Fig. 2A, B). The humerus assigned by Cheneval (1984) to
C. alphonsi (the specimen FSL 91923, which I was able to examine first hand), most likely belongs to a species in Palaelodidae. It is not expanded distally and has a poorly developed incisura intercondylaris and a well-developed fossa brachialis located in the center, as in palaelodids.
The holotype and only known specimen until now of
C. lambrechti, a distal humerus (PIN, 1399/123; Zelenkov & Kurochkin, 2015
l. XXII, figs. 5, 6) shares with
C. affinis a distal expansion of the humerus and a deep incisura intercondylaris with a distally protruding condylus dorsalis (Fig. 2). The epicondylus dorsalis is less wide in
C. lambrechti than in
C. affinis but still it is more prominent than that of flamingos and palaelodids. The area positioned proximal to the “tuberculum” supracondylare ventrale is rather flat as in anseriforms, while there is a distinct blunt ridge (Fig. 2, r, fa) in this position in flamingos. The fossa brachialis is offset from the tuberculum supracondylare ventrale, as in
Cygnus but unlike phoenicopteriforms. The attachment site of the ligamentum collaterale ventrale is not restricted to a tuberculum (as in anatids) but rather forms a flat, broad, proximally expanded and proximally pointed area, as in
Anseranas semipalmata of modern Anseranatidae. A similar morphology is present in
C. affinis. The condylus dorsalis is smaller in
C. lambrechti than in anatids, but is similar to that of anseranatids. These features support anseriform affinities of C. lambrechti and do not contradict the assignment of this species to the genus
Cygnopterus (Kurochkin, 1968; contra Mlíkovský & Švec, 1986; Zelenkov, 2013; Zelenkov & Kurochkin, 2015). Additional bones of large anseriforms from the lower Oligocene of Kazakhstan are described below, and can be assigned to
C. lambrechti based on the same relative size and similarity to other species of
Cygnopterus (see below)
A humerus from the Bembridge Marls (England; late Priabonian; terminal Eocene), specimen NHMUK 5105, assigned to
Agnopterus (
Headonornis)
hantoniensis and illustrated by Mayr (2022:fig. 5.1f), shows a well pronounced ridge running proximally from the tuberculum supracondylare ventrale, and a fossa brachialis positioned close to this ridge. These are features of Phoenicopteridae, while in
C. lambrechti and Anatidae this area is flat and the fossa brachialis is located more dorsally. Thus, the morphology of NHMUK 5105 is consistent with the flamingo-like holotype coracoid of
Headonornis hantoniensis and supports the assignment of this species to the flamingo evolutionary lineage (Mayr, 2022). However, the scapula from the lower Oligocene of Hamstead Member of the Bouldnor Formation, assigned to the same species based on size, bears a clear resemblance to that of
C. affinis (see Mayr, 2008) and strongly differs from the scapula of flamingos and palaelodids. Its distinct morphology and younger age indicate that it represents a distinct taxon from
Headonornis, and may indeed be referred to the coeval
Cygnopterus affinis.
Summarizing the above data, it is concluded here that
C. affinis is an anseriform bird characterized, among other features, by a distinctive morphology of the scapula. A similar scapula from the lower Oligocene of England (described within
Headonornis hantoniensis by Harrison & Walker, 1979) may be assigned to this taxon. The youngest described species,
C. neogradensis, has a more derived morphology of the scapula (see Kessler & Hír, 2009:fig. 1) and thus represents a species within Anatidae. Additional material of
Headonornis hantoniensis from older (upper Eocene) deposits of England most likely belongs to a flamingo-like bird (Mayr, 2022). The hypodigm of
C. alphonsi from the Lower Miocene of France also contains flamingo bones, but the holotype tarsometatarsus of this species (FSL 91883) does represent Anseriformes. It has been shown to be similar to
Paranyroca magna Miller and Compton, 1939 (Paranyrocidae), but the generic attribution of
C. alphonsi was considered uncertain due to the lack of overlapping material (Mayr & Smith, 2017). The new material for
C. lambrechti described below includes a scapula very similar to that attributed to
C. affinis and tarsometatarsi that are morphologically similar to those of
Paranyroca magna and
C. alphonsi. The similarity between the early Oligocene members of the genus
Cygnopterus and Early Miocene species of
Paranyroca allows these genera to be considered within the same taxon at the family level, although the exact generic assignment of
C. alphonsi remains tentative.
The position of Paranyrocidae within Anseriformes has been discussed by Mayr and Smith (2017), who considered them to represent a stem-Anatidae lineage based on the structure of the tarsometatarsus. The moderate degree of development of the epicondylus dorsalis and non-elevated elongate scar of the “tuberculum” supracondylare ventrale in the humerus of
C. lambrechti are morphologically closer to species of Anseranatidae. However, the better development of the incisura intercondylaris of the humerus and the fact that the foramen vasculare distale opens plantarly into a recessed notch (see below) makes
C. lambrechti evolutionarily more advanced towards crown- group Anatidae—these observations are consistent with the conclusions by Mayr and Smith (2017). Additional similarities with the anatid taxon
Cereopsis (see below) further support this view.
CYGNOPTERUS LAMBRECHTI Kurochkin, 1968
(Fig. 2C, F, J, L, N, O, P, R, S, T, U, W, Y, Z, AA, CC)
Cygnopterus lambrechti Kurochkin, 1968:92, fig. 1 (original description).
Cygnopterus lambrechti Kurochkin, 1968: Brodkorb, 1971:175; Zelenkov, 2013:1327, fig. 2f, g; Zelenkov and Kurochkin, 2015:195, pl. XXII, figs. 5, 6.
Agnopterus turgaiensis Tugarinov, 1940: Mlíkovský and Švec, 1986:266.
Holotype—PIN 1399/123, distal fragment of the left humerus.
Type Locality and Horizon—Chelkar-Tengiz, Kur-Sai comb; Turgai Depression, Kustanai Region, some 190 km to south- south-east of Turgai city; Central Kazakhstan. Chliktinskaya (Chelkarnurinskaya) Svita; late Rupelian, lower Oligocene
Newly Referred Material—PIN 2974/19, symphysial fragment of furcula (Fig. 2T, U); PIN 2974/29, fragmentary left scapula of a subadult individual; PIN 2974/30, fragmentary right scapula (Fig. 2F); PIN 2974/17, proximal fragment of left tarsometatarsus (Fig. 2J, L, N, O); PIN 2974/36, fragmentary distal end of a right tarsometatarsus (Fig. 2P–S); PIN 2974/13, proximal phalanx of the right pedal digit IV (Fig. 2W, Y, Z, AA, CC); all from Donguz- Tau (Donyztau) locality; Chelkarnurinskaya Svita, late Rupelian, Lower Oligocene. Collected by E.N. Kurochkin in 1968.
Measurements (mm)—Humerus (holotype PIN 1399/123): distal width, 28.1; depth across condylus dorsalis, 15.1; shaft width at the level of the fossa brachialis, 19.9. Furcula: craniocaudal width in medial part, 8.8; dorsoventral width in medial part, 7.6. Scapula (PIN 2974/30): length as preserved, 38.4; depth at the level of the facies articularis humеralis, 12.7. Tarsometatarsus, proximal part (specimen PIN 2974/17): shaft width, 7.6; shaft depth, 7.1; distal part (specimen PIN 2974/36): width 6.6; depth of trochlea metatarsi III, ∼11.3. Proximal phalanx of pedal digit IV: length, 38.2; proximal width, 8.6; proximal depth, 9.8; minimal shaft width, 4.8; distal width, 5.7; distal depth, 6.9.
Description and Comparisons—The remains of a goose-sized bird from Donguz-Tau locality can be assigned to the genus
Cygnopterus due to the characteristic scapular morphology shared with that of the type species
C. affinis from the similar- aged sediments of Belgium. The scapula of both
C. affinis and specimen PIN 2974/30 is characterized by an elongate cranial part (only partly preserved in new specimens) and a strongly dorsally protruding acromion. Further in both forms, the dorsal margin of the bone is distinctly concave above the glenoid area (Fig. 2, con), and there is a projection on the ventral margin of the shaft. Thus, the scapula looks sigmoid in lateral and medial views. The form from Kazakhstan is somewhat smaller than
C. affinis and corresponds in size and proportions of all known skeletal elements to smaller specimens of modern
Cygnus melancoryphus.
The partial proximal tarsometatarsus (specimen PIN 2974/17) resembles anatids in the morphology of a partly preserved proximally positioned and strongly asymmetrical hypotarsus that shows at least two incomplete sulci and a distally elongated medial ridge. In phoenicopteriforms, the hypotarsus is distally displaced (relative to its position in anseriforms), unisulcular and symmetrical, with the lateral crest being subequal to or even longer than the medial one (Mayr, 2017). Palaelodids also have a rather symmetrical proximal tarsometatarsus and a distally displaced and mediolaterally constricted (narrow) hypotarsus with a symmetrical distal margin in plantar view unlike the condition in the described fossil. In both phoenicopteriforms and palaelodids, the plantar surface of the tarsometatarsus distal to the hypotarsus forms a centrally located ridge, while in PIN 2974/17 and anseriforms, the corresponding plantar surface is inclined to various degrees (Fig. 2, ss), and the most prominent ridge is located closer to the medial margin of the bone (continuous with the medial hypotarsal ridge).
Specimen PIN 2974/17 preserves a well-developed dorsomedial ridge (Fig. 2, dmr), bordering the fossa infracotylaris and moderately developed sulcus extensorius medially, as in
C. alphonsi. Such a ridge is not prominent in anatids, but is present in Anseranas semipalmata,
C. alphonsi, and
Cousteauvia kustovia (see Mayr & Smith, 2017; Zelenkov, 2020b). The dorsolateral ridge is still notably thicker than the dorsomedial one as in all anseriforms. The tuberculi musculi tibialis cranialis are rather distally located as in anatids and C. alphonsi. The sulcus extensorius is wide and extends distally as in
C. alphonsi. The plantar openings of the foramina vascularia proximalia are large as in members of
Paranyroca and
C. alphonsi (see Mayr & Smith, 2017). Most importantly, the proximal half of the shaft is not flattened dorsoplantarly, as is characteristic of most anseriforms, but forms a sloping surface (Fig. 2, ss). The plantar surface is prominent medially, forming a blunt ridge-like continuation of the crista medialis hypotarsi. Similar sloping plantar surface of the proximal tarsometatarsus is present in anhimids, members of
Paranyroca and
C. alphonsi. Among modern anatids, it is only characteristic of
Cereopsis novaehollandiae (personal observation), but other anatids have a flattened proximal part of the shaft. Only a part of the cotyla lateralis is present, which shows that the position of the cotyla relative to the fossa infracotylaris was like that of anatids. A tuberosity formed by the lateral wall of the tarsometatarsus at the level of the fossa infracotylaris in most anseriforms, is not distinct as in members of
Paranyroca and
C. alphonsi (see Mayr & Smith, 2017). The walls of the tarsometatarsus are thin in the cross section, unlike
Cousteauvia kustovia and diving anatids (see Zelenkov, 2020b).
The fragmentary distal tarsometatarsus (specimen PIN 2974/ 36) has a more derived anatid-like morphology compared with the proximal part of the bone, a condition already noted for Paranyrocidae by Mayr and Smith (2017). Specimen PIN 2974/36 shows a very large dorsal opening of the foramen vasculare distale, as in Paranyroca magna, and the plantar opening is a recessed notch connected to the intertrochlear incision as in all species of Paranyrocidae (Mayr & Smith, 2017). The plantar articular surface of trochlea metatarsi III is pointed proximally and extends proximally to the level of the foramen vasculare proximale, resembling the condition in Paranyroca magna. In this character, specimen PIN 2974/36 differs from the large-sized romainvilliid Saintandrea chenoides Mayr and De Pietri, 2013, from the upper Oligocene of France (Mayr & De Pietri, 2013). Trochlea metatarsi III is laterally concave in distal view in S. chenoides, while in members of Paranyroca and specimen PIN 2974/36 it is dorsoventrally oriented and straight. In an unnamed putative romainvilliid bird from the upper Eocene of Xinjiang (North-Western China; Stidham & Ni, 2014), the trochlea metatarsi III is dorsally asymmetrical in distal view and shorter than in specimen PIN 2974/36 in distal and plantar views. The medial semi-trochlea of trochlea metatarsi III in specimen PIN 2974/36 is more robust and protruding than the lateral one, like in members of Paranyroca, in contrast to those of
Cygnus and
Cygnopterus alphonsi.
Mayr and Smith (2017) mentioned several characters of the tarsometatarsus distinguishing members of
Paranyroca and
C. alphonsi, of which only one is reliably traceable in the fossil from Kazakhstan. Trochlea metatarsi III in specimen PIN 2974/ 36 is the same length as in
Paranyroca, while in
C. alphonsi it is proportionally shorter (with a comparable distal width). This difference alone is not sufficient to draw conclusions about the relationships between the discussed forms, but it may support a separate generic status of
C. alphonsi within Paranyrocidae
The proximal phalanx of the right fourth pedal digit (specimen PIN 2977/13) is relatively larger than other elements (agrees with modern C. columbianus) and, therefore, is provisionally assigned to C. lambrechti here. If correctly attributed to the species, it may indicate a sexual dimorphism in
C. lambrechti. Otherwise, this specimen may indicate the presence of another large anseriform in the fauna, which is now considered less likely. In general proportions, it is quite similar to
C. melancoryphus, but is somewhat more robust and differs in some morphological details. The proximolateral tubercle is rudimentary, whereas in swans it is well developed, the collateral concavities of the distal extremity are well developed and located plantarly (practically absent in swans). The distal articular part is distinctly constricted (“waisted”) and more asymmetrical than in swans, i.e., its medial rim is enlarged in the distal and plantar views.
The phalanx (specimen PIN 2977/13) differs from the corresponding bone of flamingos in its notably more elongate proportions, more symmetrical proximal extremity, less developed plantar ridges, oval (rather than sub-triangular) cross section, and much wider distal articular surface in distal view.
The rami of the fragmentary furcula (specimen PIN 2974/19) are similar in size to those in the furcula of
C. columbianus in the cranial and caudal views, but are significantly more robust in internal and external (“hypocleidum”) views. There are no signs of a hypocleidium.
Remarks—
Cygnopterus lambrechti, previously known only from a partial distal humerus, has been earlier synonymized with the fossil flamingo
Agnopterus turgainensis from the same strata of Central Kazakhstan (Nessov, 1992; Mlíkovský & Švec, 1986) or considered as a separate flamingo taxon (Zelenkov, 2013). Nevertheless, the restudy of the holotype and evidence from new material testify to its anseriform affinities (see above).
Earlier, Cheneval (1984) noted similarities between
C. alphonsi and modern
Cereopsis novaehollandiae in the hindlimb structure, and this is also evident in the new fossils from Central Kazakhstan. The lateral orientation of the sulcus musculi hallucis longus, the more plantar position of sulcus m. flexor perforans digiti II and a well-developed sulcus musculi fibularis longus in
C. novaehollandiae are all plesiomorphic features shared with anseranatids. Mayr and Smith (2017) have suggested that the lateral sulci of the hypotarsus of
Anseranas semipalmata may have evolved independently from anatids, but the intermediate morphology of
C. novaehollandiae still indicates that the anatid hypotarsus may have evolved from the anseranatid-like one. Thus, the hypotarsus of
C. novaehollandiae may be close to that of the common ancestor of Anatidae, and the similarity between
C. novaehollandiae and Paranyrocidae in the morphology of the hypotarsus (and the proximal tarsometatarsus in general) is remarkable. However, the absence of a closed canal for the tendon of musculus flexor digitorum longus of the hypotarsus is obviously a plesiomorphic character of paranyrocids (Mayr & Smith, 2017), which distinguishes them from those of all Anatidae and
A. semipalmata. The distinct scapular morphology further supports the position of paranyrocids being outside the clade Anseres (Anseranatidae + Anatidae)
Fred
FIGURE 2. Cygnopterus lambrechti from the lower Oligocene of Kazakhstan compared with selected fossil and recent taxa. A, H,
Phoenicopterus chilensis (PIN 29-2-1), left humerus in cranial view. B,
Palaelodus ambiguus from the lower Miocene of Saint-Gérand-le-Puy, France (MB Av-188), left humerus in cranial view. C, F, J, L, N–P, R–U, W, Y–AA, CC,
Cygnopterus lambrechti from the lower Oligocene of Chelkarnura Formation, Kazakhstan: C, left distal humerus (holotype PIN 1399/123) in cranial view; F, right scapula (PIN 2974/30) in lateral view; J, L, N, O, proximal left tarsometatarsus (PIN 2974/17, reversed to facilitate comparisons) in plantar (J), dorsal (L), lateral (N), and medial (O) views; P–S, distal right tarsometatarsus (PIN 2976/36) in distal (P), plantar (R; coated with magnesium oxide) and dorsal (S) views; T, U, furcula (PIN 2974/19) in caudal (T) and ventral (U) views; W, Y–AA, CC, proximal phalanx of the right pedal digit IV (2974/13) in medial (W), plantar (Y), dorsal (Z), proximal (AA, BB), and distal (CC, DD) views. D, G, V, X, BB, DD,
Cygnus melancoryphus (PIN 38-3-2): D, left humerus in cranial view; G, right scapula in lateral view; V, X, BB, DD, proximal phalanx of the right pedal digit IV in medial (V), plantar (X), proximal (BB), and distal (DD) views. E,
Cygnopteus affinis from the lower Oligocene of Belgium (syntype part of the holotype partial skeleton if all bones are from one individual; after Lambrecht, 1933; see Mlíkovský, 2002:112), left humerus in cranial view. I, K, M, Q,
Cygnopterus alphonsi from the Lower Miocene of Saint-Gérand-le- Puy, France (holotype FSL 91.883), right tarsometatarsus in plantar (I), dorsal (K), lateral (M), and distal (Q) views. Abbreviations: cd, condylus dorsalis; cmh, crista medialis hypotarsi; con, concave dorsal margin of the scapula; di, deep incisura between condyles; dlr, dorsal lateral ridge of tarsometatarsus; dmr, dorsal medial ridge of tarsometatarsus; ed, epicondylus dorsalis; fa, flat and wide area proximal to tuberculum supracondylare ventrale; fb, fossa musculi brachialis; fdl, sulcus for tendon of musculus flexor digitorum longus; fic, fossa infracotylaris dorsalis; fp2, sulcus for tendon of musculus perforans digiti II; fv, ventral openings of the proximal vascular foramina; plt, proximolateral tubercle; pro, projection on the ventral margin of the scapular shaft; r, blunt ridge proximal to tuberculum supracondylare ventrale; se, sulcus extensorius; ss, sloping plantar surface of the tarsometatarsus just distal to hypotarsus; tsv, tuberculum supracondylare ventrale; tmIII, trochlea metatarsi III. Scale bars equal 1 cm. E, P, Q, AA–DD, not to scale.
