Pelvic and femoral anatomy of the Allodaposuchidae (Crocodyliformes, Eusuchia) from the Late Cretaceous of Lo Hueco (Cuenca, Spain)
- 150 Downloads
The fossil record of postcranial remains assigned to Allodaposuchidae is currently sparse. However, the Late Cretaceous paleontological site of Lo Hueco (Cuenca, Spain), from where two new taxa of allodaposuchid have been described, has yielded numerous postcranial remains assignable to this clade. Among them, the large amount of pelvic and femoral material is notable, providing the opportunity to study these allodaposuchid elements and assess their morphological similarity with other eusuchian remains.
The comparison with extant crocodylians was accomplished using traditional morphometric techniques, whereas the comparison with other fossils and establishment of morphotypes was done using morphological criteria.
The results of the cluster and principal components analyses show morphological differences between extant crocodylians and allodaposuchids from Lo Hueco, allowing the segregation of these lineages. The similarities found between the pelvic and femoral remains from Lo Hueco, and those referred to Allodaposuchus precedens and other putative allodaposuchids from the Iberian Peninsula, allows referral of these remains to allodaposuchids. The differences found among the femoral and pelvic remains of Lo Hueco enables us to recognize two morphotypes per each element.
This study allows a better understanding of allodaposuchid postcranial elements that were previously poorly known. The ilia, ischia and femora from Lo Hueco allodaposuchids are distinct from those of other crocodylian lineages. Finally, the fact that there are two morphotypes per each element at Lo Hueco is congruent with the presence of two different allodaposuchids at the site.
KeywordsAllodaposuchidae Iberian Peninsula Postcranial Morphometrics PCA
El registro fósil contiene escasos restos postcraneales asignados a Allodaposuchidae. Sin embargo, en el yacimiento paleontológico del Cretácico Superior de Lo Hueco (Cuenca, España), del cual se han descrito dos nuevas especies de alodaposúquido, se han recuperado numerosos restos fósiles postcraneales asignables a este clado. Entre ellos destaca la cantidad de restos pélvicos y fémorales recuperados. Esto ofrece la oportunidad de estudiarlos y evaluar su semejanza morfológica con restos de otros eusuquios.
La comparación con cocodrilos actuales se realizó empleando técnicas morfométricas tradicionales, mientras que la comparación con otros restos fósiles y el establecimiento de morfotipos se efectuó utilizando criterios morfológicos.
Los resultados de los análisis de clúster y de componentes principales muestran que existen diferencias morfológicas entre los restos de cocodrilos actuales y los restos de los alodaposúquidos de Lo Hueco, permitiendo la segregación de estos linajes. La similitud encontrada entre los restos pélvicos y femorales de Lo Hueco, y aquellos asignados a Allodaposuchus precedens y otros supuestos alodaposúquidos de la península ibérica, permite asignar estos restos a alodaposúquidos. Las diferencias encontradas entre el material femoral y pélvico de Lo Hueco permite reconocer dos morfotipos para cada elemento postcraneal estudiado.
El presente estudio posibilita un mejor entendimiento de estos restos postcraneales en alodaposúquidos, previamente poco conocidos. Los iliones, isquiones y fémures de los alodaposúquidos de Lo Hueco son distintos de los de miembros de los linajes de Crocodylia. Por último, el hecho de que existan dos morfotipos de cada elemento estudiado de Lo Hueco es congruente con la presencia de dos especies distintas de alodaposúquido en el yacimiento.
Palabras claveAllodaposuchidae Península ibérica postcraneal morfometría PCA
Allodaposuchidae is a clade of basal eusuchian crocodyliforms defined by Narváez et al. in 2015. This clade includes several European taxa from the Late Cretaceous (Campanian–Maastrichtian) that, in most analyses, forms the sister group of Crocodylia (Narváez et al. 2016, 2017). The taxa in this clade include the type species Allodaposuchus precedens Nopcsa, 1928; Agaresuchus fontisensis Narváez et al. 2016; Agaresuchus subjuniperus (Puértolas et al. 2013); ‘Allodaposuchus hulki’ Blanco et al. 2015; ‘Allodaposuchus palustris’ Blanco et al. 2014; Arenysuchus gascabadiolorum Puértolas et al. 2011; Lohuecosuchus mechinorum Narváez et al. 2015; Lohuecosuchus megadontos Narváez et al. 2015; and Musturzabalsuchus buffetauti Buscalioni, Ortega, Vasse, 1997.
Allodaposuchid postcranial remains are scarce, even more so in the case of pelvic and femoral elements. In fact, from the abovementioned taxa, only the type series of Allodaposuchus precedens from Vălioara (Romania) includes some associated iliac and femoral remains belonging to the same individual (Nopcsa 1928) and described by Buscalioni et al. (2001). ‘Allodaposuchus palustris’, from Fumanya (Spain) has directly associated pelvic and femoral remains (Blanco et al. 2014). Martin et al. (2015) assigned a left femur from Velaux-La Bastide Neuve (France) to Allodaposuchus precedens. All of these postcranial remains have been briefly described in the literature and thus are poorly known.
The study of several pelvic and femoral remains from Lo Hueco allows us to describe and improve our knowledge about the allodaposuchid postcranial skeleton. As mentioned before, two different allodaposuchid taxa have been described in Lo Hueco and, on this basis, two pelvic and femoral morphotypes are expected to be distinguished.
The idea that the appendicular skeleton of crocodylians is conservative is widespread, but as stated by Chamero et al. (2013) is not accurate. There are few studies considering morphological variation in these elements, and usually they have been conventionally based on qualitative observations (Romer 1956; Erickson 1976). Therefore, a preliminary morphometric study to compare the morphology between allodaposuchids and extant crocodylians has been performed. In this regard, it is hypothesized that there are enough morphological differences to segregate the allodaposuchid pelvic and femoral remains from that of extant crocodylians.
1.1 Institutional abbreviations
HUE, specimens housed in the MUPA, Museo de Paleontología de Castilla-La Mancha (Cuenca, Spain), IGM, Mongolian Institute of Geology (Ulaan Bataar, Mongolia), MACV, Museo de Anatomía Comparada de Vertebrados - Universidad Complutense de Madrid (Madrid, Spain), MAFI, Magyar Állami Földtani Intézet - Hungarian Geological Institute (Budapest, Hungary), MMC, Museu de les Mines de Cercs (Sant Corneli, Spain), MMS/VBN, Musée du Moulin seigneurial (Velaux-La Bastide Neuve, France) MNCN, Museo Nacional de Ciencias Naturales (Madrid, Spain), SMM, The Science Museum of Minnesota (Minneapolis, United States), SMU, Shuler Museum of Paleontology (Dallas, United States), UPUAM, Unidad de Paleontología de la Universidad Autónoma de Madrid (Madrid, Spain).
1.2 Anatomical abbreviations
4t, fourth trochanter; act, acetabulum; ailf, anterior iliac facet; aspe, articular surface of the proximal epiphysis; cf, condylar fold; fc, fibular condyle; ilap, iliac anterior process; ilb, iliac blade; ins, intercondylar sulcus; isap, ischial anterior process; isat, ischial antitrochanter; isb, ischial blade; ispp, ischial posterior process; isva, ischial ventral apex; lc, lateral hemicondyle; mc, medial hemicondyle; pat, proximal anterior tuberosity; pf, articular facet for the pubis; pilf, posterior iliac facet; ppt, proximal posterior tuberosity; prap, preacetabular process; psap, postacetabular process; sactc, supraacetabular crest.
2 Materials and methods
2.1.1 Postcranial eusuchian fossil remains
Locality, horizon and age: Lo Hueco fossil site (Fuentes, Cuenca, Spain) (Fig. 1). Arcas-Fuentes Syncline, southwestern branch of the Iberian Range, Margas, Arcillas y Yesos de Villalba de la Sierra Formation. Late Cretaceous (late Campanian-early Maastrichtian) in age (Barroso-Barcenilla et al. 2009; Ortega et al. 2015).
2.1.2 Postcranial remains belonging to extant eusuchian species
Material: both ilia, ischia and femora from Alligator mississippiensis (adult, MACV-2697), Caiman crocodilus (adult, MACV-6612), Crocodylus niloticus (young adult, MACV-6138), Crocodylus rhombifer (adult, MNCN-45749), Osteolaemus tetraspis (adult, MNCN-42089) and Paleosuchus palpebrosus (adult, MACV-6139).
2.2 Traditional morphometrics
2.2.1 Missing values
Removal of specimens or variables with missing values from the data matrix (listwise deletion of missing data) is a standard method in multivariate analysis, as this type of analysis usually needs a complete set of data (Strauss et al. 2003). However, in this case over ninety percent of the fossil specimens had at least one missing value, so it was desirable to reach a compromise between the exclusion or retention of variables/specimens. The pattern of missing values was analysed with SPSS v20 (IBM Corporation 2011). The consensus was to exclude the most incomplete variables (measurements with over 10% missing values, not included in Fig. 4), and the most fragmentary specimens (over 35% missing values).
After removal, the missing values were 3.03, 0 and 4.71% for the ilium, ischium and femur data matrix, respectively. All these missing values belonged to Lo Hueco fossils. These values can be estimated employing the adequate methods, allowing the inclusion of more material while at the same time minimizing impact on the statistical inference. The method employed was expectation–maximization (EM), which uses the covariance matrix and converges it iteratively to the maximum probability results (Strauss and Atanassov 2006). This method has been reported to have better results than substitution by the mean or regression methods, since them cause loss of variance and bias (Strauss et al. 2003). Due to the use of the covariance matrix by the EM algorithm, the data was log-transformed a priori to standardize variances and linearize the allometric relations between the variables. Because the only missing values belong to some Lo Hueco fossils, and similar cases are needed, only other fossils from Lo Hueco were employed to estimate the missing values. The EM was performed in SPSS v.20 (IBM Corporation 2011).
2.2.2 Multivariate analysis—cluster analysis and principal components analysis
Due to the size of the available sample, only exploration methods were performed in PAST v3.14 (Hammer et al. 2001). Three CA (Cluster Analysis) were performed in order to identify groups of similar specimens. These analyses were done using the log-transformed data matrices, UPGMA algorithm (Unweighted Pair Group Method with Arithmetic mean), Bray–Curtis similarity index and 10.000 permutations. Also, three PCA (Principal Components Analysis) were performed to identify the orthogonal axes of maximal variation in each dataset. For the PCA the non-log-transformed data matrices and their respective variance–covariance matrices were used.
3.1 Pelvic girdle—Ilium
Eight previously unpublished ilia from Lo Hueco are described here (Fig. 2a–h). All bear a substantial blunt-ended posterior process that extends beyond the posterior region of the iliac body. In addition, all ilia bear a reduced anterior process that does not extend anteriorly to the preacetabular process. Another common feature of these ilia is an anteroposteriorly wide acetabulum. The iliac blade lies dorsal to the acetabulum and it has a remarkable height and a convex outline, bearing numerous muscle insertion marks at the margin.
Two different morphotypes can be distinguished in this set of specimens. The first of these morphotypes includes HUE-00993, HUE-01451, HUE-01473, HUE-02282, HUE-03689 and HUE-04677 (Fig. 2a–e, h). HUE-00993 is an almost complete right ilium with the ventral margin of the acetabulum damaged (Fig. 2a). HUE-01451 is a right ilium that lacks the postacetabular process (Fig. 2b) found in association with femur HUE-01451 (Fig. 3c, d), among other elements. HUE-01473 is a left ilium that lacks the posterior process of the iliac blade (Fig. 2c). HUE-02282 is a right ilium that lacks almost the entire iliac blade (Fig. 2d). HUE-03689 is the anterior part of a right ilium (Fig. 2e). HUE-04677 is a left ilium with the posterior process damaged (Fig. 2h). The anterior process in the ilium HUE-00993 is mediolaterally thick, dorsoventrally extended and cranially projected, a morphology shared with HUE-04677 and HUE-01451 (Fig. 2a, b, h). The anterior processes of HUE-01473, HUE-02282 and HUE-03689 are mediolaterally thick, dorsoventrally extended and blunt-ended (Fig. 2c–e). This process is almost connected to the supraacetabular crest, giving a robust appearance to the ilia. The supraacetabular crest of HUE-00993 is rough, dorsoventrally wide and mediolaterally thick (Fig. 2a). Consequently, the acetabulum is markedly concave on its dorsal region. Although this character can be observed in all specimens of this morphotype, it is very notable in HUE-01451 (Fig. 2b). The pre- and postacetabular processes are located anterior and posterior to the acetabulum, and they articulate with the ischium via the anterior and posterior ischial facets, respectively. The postacetabular process has a flat ventral ischiatic facet. This process is mediolaterally wider than the preacetabular process and reaches its maximum mediolateral width at the base of HUE-00993 (Fig. 2a). These characters are shared with HUE-01473, HUE-02282 and HUE-04677 (Fig. 2c, d, h), but cannot be observed in HUE-01451 and HUE-03689 (Fig. 2b, e) because both lack the postacetabular process. The preacetabular process is bilobated, rounded and anteriorly directed in HUE-00993 (Fig. 2a). This morphology of the preacetabular process is shared with all the ilia of this morphotype, and it is very notable in HUE-02282 (Fig. 2d). The iliac blade is blunt-ended and its end is slightly caudolaterally directed in HUE-00993, but also in HUE-01451 (Fig. 2a, b). This character cannot be observed in the other ilia of this morphotype because the iliac blade is damaged or absent in these specimens.
The second morphotype includes HUE-04504 and HUE-04515 (Fig. 2f, g). HUE-04504 is a right ilium that lacks the pre- and postacetabular processes, and has a damaged iliac blade (Fig. 2f). HUE-04515 is a left ilium that lacks the preacetabular process and part of the acetabulum (Fig. 2g). The anterior process of HUE-04515 is anteroposteriorly elongated, sharp-ended and it retains the lateromedial width of the iliac blade (Fig. 2g). Although it is also observable in HUE-04504 (Fig. 2f), this anterior process is more evident in HUE-04515. The supraacetabular crest of HUE-04515 is dorsoventrally narrow and not mediolaterally thick, giving a shallower appearance to the acetabulum (Fig. 2g). Although HUE-04504 lacks the ventral part of the acetabulum and its associated processes, it is still possible to observe a shallow acetabulum (Fig. 2f). The morphology of the anterior process and the supraacetabular crest gives the ilium a slender appearance. In this second morphotype, the preacetabular process remains unknown because both ilia lack it. Only HUE-04515 preserves the postacetabular process, but it is lateromedially crushed and therefore the width cannot be objectively interpreted (Fig. 2g). Even so, in general appearance, the postacetabular process is not massive, and the ischiatic facet is longer than wide. In HUE-04515 the iliac blade is almost vertical, blunt-ended and the posterior margin of the blade is straight, not angled (Fig. 2g). HUE-04505 shares this morphology of the iliac blade (Fig. 2f). HUE-04515 is the only one with a stout tubercle towards the end of the iliac blade among the sample of ilia from Lo Hueco (Fig. 1g).
3.2 Pelvic girdle—Ischium
The ischium is represented by two preserved elements, HUE-00920 and HUE-02524 (Fig. 2i, j). Each ischium belongs to a different morphotype, but they share some common features, indicated below. The proximal end of either ischium has two processes, the anterior and posterior processes, separated by a notch that forms the ventral margin of the perforated acetabulum. These ischial processes, also known as pubic and iliac peduncles respectively, connect the ischium with the ilium through the dorsal iliac facets (Fig. 2i, j). The anterior process of the ischium, which is anterodorsally directed, also includes the articular facet for the pubis. The articular surface for the pubis is located anteroventrally and is separated from the articular surface for the ilium of the anterior process. Therefore, the pubis does not contact the ilium directly and is excluded from the acetabulum. The posterior process, dorsally directed, has a triangular shape in lateromedial view. The iliac peduncle has two separate articular facets: the anterior or ischial antitrochanter, and the posterior or articular facet for the ilium (Fig. 2i, j). The ischial antitrochanter is slightly laterally exposed in both ischia. The diaphysis is short and wide in anteroposterior direction. The lateromedial width of the diaphysis increases from the anterior to the posterior surface, but decreases proximodistally towards the ischial blade, which is anteroposteriorly wide. The ischial blade is angled in relation to the posteroproximal axis of the ischium, with the anterior end of the blade directed medially (Fig. 2i, j). In both ischia, the blade bears a ventral apex, located towards the caudal end.
The first morphotype is represented by HUE-02524 (Fig. 2j), which is a complete left ischium. The anterior process is narrow and proximodistally elongate. The articular facet for the pubis in the anterior process of HUE-02524 is also narrow and elongated, with a ridge around the pubic facet (Fig. 2j). The articular surface for the ilium in the anterior process is surrounded by a ridge, which separates that flat surface from the surroundings. The ischial antitrochanter, located at the posterior process, is lateromedially wide and anteroposteriorly extended (Fig. 2j). In contrast, the articular facet for the ilium is anteroposteriorly shorter than the ischial antitrochanter. The ischial blade of HUE-02524 is anteroposteriorly expanded, projecting caudally (Fig. 2j).
The second morphotype includes HUE-00920 (Fig. 2i), a left ischium lacking the anterior margin of the diaphysis. The anterior process of this specimen is bulbous, rounded and expanded lateromedially (Fig. 2i). The pubic facet, located on the anterior process, is rounded and proximodistally short in HUE-00920. The articular surface for the ilium on the anterior process is barely notable (Fig. 2i). In the posterior process, the ischial antitrochanter is mediolaterally expanded, but it is shorter than the articular facet for the ilium of HUE-00920 (Fig. 2i). The ischial blade in this ischium is anteroposteriorly expanded, but it does not project prominently to the posterior area.
There are six femora from Lo Hueco (Fig. 3). The most complete is HUE-02267 (Fig. 3e-f). There are two different femur morphotypes, but they share some common characters. All have a characteristic sigmoidal profile due to the torsion angle that exists between the proximal and distal epiphyses. The proximal and distal epiphyses have a similar width. The proximal end is slightly flat and wide in anterior and posterior view, with a concave medial margin and convex lateral margin. In dorsal view, the rounded articular surface of the proximal epiphysis reaches its maximum height at the lateral side. The anterior surface of the proximal epiphysis of all femora is slightly concave and bears numerous muscle insertion marks, especially on the lateral side, but also on the surface located over the proximal anterior tuberosity (sensu Brochu 1992). The diaphysis of the femora starts below the condylar fold (sensu Brochu 1992) and has an oval-shaped outline up to the end of the fourth trochanter, from where it changes to a subcircular outline. In all femora, the fourth trochanter, which is symmetric and oval-shaped, is located on the proximal diaphysis, placed obliquely to the longitudinal axis of the femur. The distal epiphysis of the femora has two distinct condyles, the lateral and medial hemicondyles. The lateral hemicondyle is mediolaterally wider than the medial hemicondyle. Between the posteriorly-projected lateral and medial hemicondyles there is a well-developed intercondylar fossa, notable in posterior view. The femora also have a depression between these hemicondyles that is visible in dorsal view. A lateral facet on the lateral hemicondyle, clearly visible in lateral view, is the fibular condyle (Fig. 3c).
The first morphotype of femur includes HUE-01451, HUE-02267, HUE-04111 and HUE-04902 (Fig. 3c–j). HUE-01451 is a left femur with the proximal epiphysis and medial hemicondyle damaged (Fig. 3c, d). HUE-02267 is a complete right femur (Fig. 3e, f). HUE-04111 is a left femur with the diaphysis damaged (Fig. 3g, h). HUE-04902 is a left femur without the distal epiphysis (Fig. 3i, j). In posterior view, below the dorsal articular surface of the femur, is located the proximal posterior tuberosity, which fairly protrudes in HUE-02267 and the other femora of the first morphotype. The medial margin of the proximal epiphysis of HUE-02267 is strongly marked and protrudes medially in relation to the diaphysis, forming the condylar fold (Fig. 3e, f). Although this character is most visible on HUE-02267, all femora of this morphotype share it. The fourth trochanter of HUE-02267 is large and thick, with numerous muscle insertion marks, and protrudes not only medially but also laterally (in anterior and posterior view respectively; Fig. 3e, f). The other femora from this morphotype are very similar, but their fourth trochanter does not project laterally in posterior view. The distal epiphysis of HUE-02267 starts abruptly and gets wider towards the distal end in anterior and posterior views, character shared with HUE-01451 and HUE-04111. This feature, and those that follow, cannot be observed on HUE-04902 as it lacks the distal epiphysis, but it is notable that the diaphysis gets wider towards its end. The lateral hemicondyle of HUE-02267 projects distally more than the medial hemicondyle in anterior view (Fig. 3e): this morphology is also present in HUE-04111 (Fig. 3g). HUE-01451 lacks part of the medial hemicondyle and, thus, the abovementioned character cannot be properly observed. In ventral view, the medial hemicondyle is oval-shaped in HUE-02267 and HUE-04111, whereas the lateral hemicondyle has a rounded shape in HUE-01451, HUE-02267 and HUE-04111. The fibular condyle is rounded and anteroposteriorly enlarged in HUE-01451, HUE-02267 and HUE-04111 (Fig. 3c, f, h).
The second morphotype comprises HUE-00744 and HUE-06070 (Fig. 3a, b, k, l). HUE-00744 is a left femur lacking part of the fourth trochanter and medial hemicondyle (Fig. 3a, b). This femur is covered by an iron crust that cannot be removed because it penetrates into the periostium. HUE-06070 is a left femur lacking part of the diaphysis with a damaged fourth trochanter and medial hemicondyle (Fig. 3k, l). The proximal posterior tuberosity is shallow in both HUE-00744 and HUE-06070. The medial margin of the proximal epiphysis of HUE-06070 shows a smooth transition between the proximal epiphysis and the diaphysis; consequently the condylar fold is not strongly defined in anterior view (Fig. 3k). This morphology is also present in HUE-00744, but the medial margin of the proximal epiphysis protrudes more laterally than in HUE-06070. Although the fourth trochanter is damaged in both femora, it is still possible to observe some characters. Both femora display a fourth trochanter that is narrow, long and which protrudes laterally rather than medially (Fig. 3b, l). The distal epiphyses of HUE-06070 and HUE-00744 start progressively below the diaphysis and do not expand abruptly. Despite the fact that both femora have damaged medial hemicondyles, it is still possible to observe that the lateral hemicondyle projects more distally than the medial hemicondyle, and that the lateral hemicondyle is lateromedially wider (Fig. 3a, k). In ventral and lateral views, the lateral hemicondyle and fibular condyle are subtriangular shaped on HUE-06070. In HUE-00744 the lateral hemicondyle and fibular condyle are narrow and sharp-ended.
Before proceeding with the results of the CA and PCA, it is important to underline that these analyses were performed over a limited sample. Therefore, the obtained results are only exploratory.
4.1 Cluster analysis
4.2 Principal components analysis
In the case of the ilium (Fig. 6a), the first two PC accounted for 98.6% of the total variance. The first PC (PC1; 96.4%) is correlated with ilium blade maximum height point (ILbmxH, r = 0.61) and ilium maximum acetabular width (ILmxaW, r = 0.57). Thus, toward positive values, the length from the anterior process to the point of maximum height of the iliac blade and the acetabular width increase. The second PC (PC2; 2.2%) is correlated with ilium blade height (ILbH, r = 0.81) and ilium blade maximum height point (ILbmxH, r = − 0.47). Hence, the ilia located towards the positive end values of this second PC (Lo Hueco ilia and Crocodylus niloticus, MACV-6612), are those with the highest iliac blades, and also those with least distance from the anterior process to the maximum height point of the iliac blade. The results from PC1 and PC2 are incongruent regarding variable ilium maximum height point, but the first PC is highly correlated with size. Thus, the premise that Lo Hueco ilia are the specimens with the widest acetabula and longest distance from the anterior process to the maximum height point on the blade is a matter of scaling. Actually, Lo Hueco ilia and MACV-6612 have the least distance between the anterior process and the maximum height point of the blade (PC2, Fig. 6a). Moreover, Alligatoridae and Crocodylidae overlap their areas in PC1 and PC2.
For the PCA performed over the ischium matrix (Fig. 6b), the two first PC collected 98% of the total variance. The first PC (PC1; 95.6%) relates to ischium blade maximum width (ISbmxW, r = 0.54) and ischium acetabular facet length (ISafL, r = 0.30). As mentioned above, this first PC is also related to size and, because of it, the second PC should be used to explain morphological variation rather than the first PC. Over the second PC (PC2; 2.4%) the variables with greatest influence were ischium anterior process length (ISapL, r = 0.51) and ischium blade maximum width (ISbmxW, r = − 0.47). On the second PC the specimens are ordered, and those with the narrowest ischial blades and longest anterior processes are at the positive extreme end. These characteristics correspond to that of the ischium from Lo Hueco, located at the positive end of PC2 and segregated from the rest of the sample (Fig. 6b). Again, the alligatorid and crocodylid ranges of variation overlap in the PCA.
For the PCA over the femur matrix (Fig. 6c), the two first PC gathered 99.3% of the total variance. The results show that the first PC (PC1; 98.2%) also appears to be highly related to size, and the variable that contributed the most to PC1 was femur total length (FTL, r = 0.80), followed far behind by femur proximal end to fourth trochanter length (FpIVL, r = 0.30). Concerning the second PC (PC2; 1.1%), the variable with the highest contribution was femur fourth trochanter length (FIVL, r = 0.60), followed distantly by fibular condyle dimensions (FfcL, r = 0.37; FfcH, r = 0.31). Towards the positive end of the second PC are located the femora with the longest fourth trochanter and greatest length and height of the fibular condyle (Fig. 6c). In this case, on the second PC, the superposition between Lo Hueco specimens and Crocodylia occurs due to the position of HUE-02267 and HUE-04111. Once again, alligatorids and crocodylids share a similar space at the PCA.
5.1 Multivariate analysis
The results of the CA and PCA analyses show that the ilia, ischia and femora of alligatorids and crocodylids are more similar to each other than to the allodaposuchids, since they appear mixed in the CA and occupy approximately the same area in the PCA (Figs. 5, 6). Therefore, with the information provided in these analyses, it is difficult to distinguish whether any of these elements belong to one lineage or another in absence of cranial remains or other diagnostic elements. However, the results suggest that although the morphological differences are small, it is possible to distinguish the pelvic and femoral remains between these basal eusuchians and extant crocodylians.
List of taxa used for comparisons in the text
Blanco et al. (2014)
MAFI Ob 3136
MAFI Ob 3140
Buscalioni et al. (2011)
Pol et al. (2009)
The anterior process of the ilium is more reduced in the sample from Lo Hueco (Fig. 2a–h) than in Pachycheilosuchus trinquei and Bernissartia fagesii, but not as reduced as in Borealosuchus formidabilis, Alligator mississippiensis or Crocodylus niloticus. These ilia have a moderately reduced anterior process, shared with ‘Allodaposuchus palustris’ and the ilium from Vălioara attributed to Allodaposuchus precedens by Nopcsa (1928). The rounded and thick anterior process of ‘Allodaposuchus palustris’ is similar to the morphology of HUE-01473 or HUE-02282 (Fig. 2c, d), the first ilium morphotype of Lo Hueco, whereas the sharp-ended anterior process of Allodaposuchus precedens from Vălioara (Nopcsa 1928) is similar to the morphology of HUE-04515 (Fig. 2g), the second ilium morphotype. The wide acetabulum of Lo Hueco ilia is also present in the ilium of ‘Allodaposuchus palustris’ and the Vălioara ilium. The supraacetabular crests on the ilia of Bernissartia fagesii, Pachycheilosuchus trinquei, Borealosuchus formidabilis, Alligator mississippiensis and Crocodylus niloticus are less pronounced, being similar to that of HUE-04515 (Fig. 2g). However, the supraacetabular crest of ‘Allodaposuchus palustris’ and the ilium from Vălioara is thicker than the one present on the second morphotype, resembling that of HUE-01451 or HUE-04677 (Fig. 2b, h). The outline morphology of the iliac blade is different to those of Pachycheilosuchus trinquei or Bernissartia fagesii, which have almost linear iliac blades. As stated by Brochu (1999) the morphology of the iliac blade in mature individuals of Gavialis, most caimans and several fossil crocodylians is rounded and does not bear a constriction. In contrast, the adult morphology of the iliac blade in Osteolaemus, Tomistoma, most Alligator and Crocodylus is modestly indented. This is much more pronounced in Crocodylus except for Crocodylus rhombifer (Brochu 1999). The iliac blades from Lo Hueco (Fig. 2a–h) are rounded, with a robust end of the iliac blade and no clear indentation towards the end of the blade, thus differing from these described taxa. However, they are similar to the iliac blade of ‘Allodaposuchus palustris’ or the Vălioara ilium, being markedly rounded and tall anteriorly and blunt at the posterior end of the iliac blade.
The ilia from Lo Hueco share with those of other eusuchians (Buscalioni and Sanz 1990) a reduced anterior process and a posterior part of the blade that is higher than the anterior part. Furthermore, the morphology of ilia from Lo Hueco shares more similarities with the ilium of ‘Allodaposuchus palustris’ and the ilium from Vălioara than with the rest of taxa and, thus, these ilia are referred to Allodaposuchidae.
Blanco et al. (2014) noted the particular morphology of the iliac blade in ‘Allodaposuchus palustris’ and stated it as an ambiguous autapomorphy for the genus. The presence of these features in the ilia from Lo Hueco suggests these characters could be synapomorphies for Allodaposuchidae.
All analysed ischia (Table 1) have two developed processes at the proximal end. However, the ischia of Pachycheilosuchus trinquei and Shamosuchus djadochtaensis have a posterior process that is proximodistally shorter than the anterior process. The anterior and posterior processes of Lo Hueco ischia (Fig. 2i–j), Borealosuchus formidabilis, Alligator mississippiensis and Crocodylus niloticus have a similar proximodistal development. The anterior process in all ischia have separated anterior iliac and pubic facets, a character shared from basal crocodyliforms to extant crocodylians (Claessens and Vickaryous 2012). The anterior process of the ischium of Pachicheylosuchus trinquei is anteroposteriorly constricted, similar to that of HUE-02524 (Fig. 2j). However, Pachycheilosuchus trinquei does not have the pubic facet strongly marked and surrounded by a ridge like HUE-02524 (Fig. 2j). Borealosuchus formidabilis, Alligator mississippiensis and Crocodylus niloticus have a rounded anterior process, similar to HUE-00920 (Fig. 2i) but smaller. The posterior process is triangular in lateral view in all ischia, and all of them have separated posterior iliac facets and ischial antitrochanters with the exception of Pachycheilosuchus trinquei. In lateral view the ischial antitrochanter of Shamosuchus djadochtaensis, HUE-00920 and HUE-02524 is slightly visible, and it is entirely visible in Borealosuchus formidabilis, Alligator mississippiensis and Crocodylus niloticus. Ischial blade morphology is highly variable but all the comparative sample has in common a caudal region of the blade that projects more than the cranial region of the blade with regard to the dorsoventral axis of the ischium.
Ischia from Lo Hueco, which have well-developed processes with the posterior process being triangular in lateral view and with two well-differentiated articular facets, can be referred to Eusuchia. Unfortunately, there are no ischia assigned to allodaposuchids from other localities to compare with those from Lo Hueco, but HUE-02524 (Fig. 2j) was found associated with HUE-01451 (ilium and femur among others, Figs. 2b, 3c, d). That material is referred to Allodaposuchidae due to the similarity between them and the known femora and ilia of ‘Allodaposuchus palustris’ and the specimens of Allodaposuchus precedens from Vălioara. By extension, these ischia are also assigned to allodaposuchids.
All femora have a twisted shaft, although the degree of torsion is not uniform among the specimens. Femora of Bernissartia fagesii, Pachycheilosuchus trinquei and Pietraroiasuchus ormezzanoi have a longer and more slender appearance, with the proximal epiphysis less lateromedially expanded and a shallow proximal posterior tuberosity. In contrast, the proximal epiphysis of ‘Allodaposuchus palustris’, the classic material of Allodaposuchus precedens from Vălioara (MAFI Ob 3140), Borealosuchus formidabilis, Alligator mississippiensis and Crocodylus niloticus have a similar width and form, resembling that of the femora from Lo Hueco; but the proximal posterior tuberosity is also shallow, like the tuberosity present on HUE-06070 or HUE-00744 (Fig. 3a, b, k, l). The proximal epiphysis of the femur assigned to Allodaposuchus precedens by Martin et al. (2015) is anteroposteriorly compressed and, as a result, the proximal posterior tuberosity seems to stand out more than on the femora from Vălioara. The fourth trochanter of the femora of Bernissartia fagesii, Pachycheilosuchus trinquei and Pietraroiasuchus ormezzanoi is placed nearer to the proximal epiphysis, is less oblique and narrower, and protrudes less than the fourth trochanter of Lo Hueco femora (Fig. 3). The fourth trochanters of the remaining taxa in the sample resemble those from Lo Hueco; they are thicker and project more than in the before-mentioned specimens. The distal epiphysis of Bernissartia fagesii, Pachycheilosuchus trinquei and Shamosuchus djadochtaensis has two distinct hemicondyles, but the size difference between hemicondyles is less pronounced than in the case of Lo Hueco femora and the rest of the comparative sample. In addition, their hemicondyles seem to be more separated and less posteriorly projected than in Lo Hueco femora and the rest of the sample. The femora from Lo Hueco share with all taxa in the comparative sample a lateral hemicondyle that projects distally more than the medial hemicondyle. The fibular condyle of Bernissartia fagesii, Pachycheilosuchus trinquei and Shamosuchus djadochtaensis is rounded and proximodistally enlarged, morphology similar to that of HUE-04111 or HUE-02267 (Fig. 3e, f, g, h). The classic material of Allodaposuchus precedens from Vălioara (MAFI Ob 3140) has a more subtriangular shape at the fibular condyle, similar to that of HUE-06070 (second morphotype), whereas it is rounded at the femur assigned to Allodaposuchus precedens by Martin et al. (2015). The fibular condyle of Borealosuchus formidabilis, Alligator mississippiensis and Crocodylus niloticus is subtriangular and proximodistally shorter than that of Lo Hueco femora.
The strongly twisted shaft of the femora along with the two well differentiated distal hemicondyles allow the referral of Lo Hueco femora to Eusuchia. Their morphology resembles that of Allodaposuchus precedens from Vălioara more than that of the crocodylians or other basal eusuchians, particularly in their condylar morphology. Therefore, these femoral remains could be referred to Allodaposuchidae.
Two different allodaposuchid taxa have been described from Lo Hueco by Narváez et al. (2015, 2016), Lohuecosuchus megadontos and Agaresuchus fontisensis. The postcranial remains presented here could not be referred to either of these taxa, but they belong to two different morphotypes. Hereunder are described the differences found at the ilia, ischia and femora of Lo Hueco that allows the establishment of two morphotypes.
The first morphotype of ilium, or robust, includes HUE-00993, HUE-01451, HUE-01473, HUE-02282, HUE-03689 and HUE-04677 (Fig. 2a–e, h), whereas the second morphotype of ilium, or gracile, includes HUE-4504 and HUE-4515 (Fig. 2f, g). The robust morphotype shows great thickness of the anterior process, iliac blade, supraacetabular crest, pre- and postacetabular processes. The second morphotype presents a slender appearance in these areas. The caudal end of the iliac blade at the robust morphotype is angled, whereas the gracile morphotype has a straight iliac blade.
The first morphotype of ischium, or elongated, includes HUE-02524 (Fig. 2j) whereas the second morphotype of ischium, or narrowed, includes HUE-00920 (Fig. 2i). The anterior process of the elongated morphotype is slim, whereas the narrowed morphotype has a rounded anterior process. The ischial antitrochanter of the elongated morphotype is anteroposteriorly longer and the ischial blade projects much more in caudal direction with regard to the narrowed morphotype. Although these ischia have visible differences, there is only one specimen per morphotype and more material would be desirable to strengthen these findings.
The first morphotype of femur, or robust, includes HUE-01451, HUE-02267, HUE-04111 and HUE-04902 (Fig. 3c–j), whereas the second morphotype of femur, or gracile, includes HUE-00744 and HUE-06070 (Fig. 3a, b, k, l). Again, the greatest difference among these elements is the robustness, reflected in the large proximal posterior tuberosity, as well as the short and thick projecting fourth trochanter of the robust morphotype. The depression between the distal hemicondyles in dorsal view is more pronounced at the robust morphotype. The difference among the distal projection of the hemicondyles is minor at the robust morphotype with regard to the gracile morphotype. The torsion angle between the proximal and distal epiphysis exhibited by the robust morphotype is about forty-five degrees, whereas the torsion angle of the gracile morphotype is about sixty degrees.
HUE-01451 (ilium and femur) is associated with the ischium HUE-02524 (Fig. 2j). This made possible to link together the different morphotypes of ilia, ischia and femora. As result, the robust morphotype of ilium is related to the elongated morphotype of ischium and robust femur. Conversely, the gracile morphotype of ilium is related to the narrowed morphotype of ischium and gracile femur.
Pelvic and femoral remains of allodaposuchids from Lo Hueco have been described in detail, allowing a better understanding and knowledge of the postcranial skeleton in allodaposuchids. Traditional morphometric methods show that, although the morphological differences in these postcranial elements are in many cases subtle, it is possible to establish differences between pelvic and femoral remains of basal eusuchians and those of crocodylians. The major morphological variation found is concentrated in the iliac blade, ischial blade and anterior process of the ischium, and fourth trochanter and fibular condyle of the femur. Pelvic and femoral remains from Lo Hueco show a great similarity with those of other allodaposuchids, in particular ‘Allodaposuchus palustris’ from Fumanya (Spain) and Allodaposuchus precedens from Vălioara (Romania), thus allowing the inclusion of Lo Hueco specimens as allodaposuchid postcranial remains. Moreover, differences found among the material from Lo Hueco enables the establishment of two different morphotypes of pelvic and femoral remains at the site. These findings are congruent with the known presence of two different allodaposuchid taxa at the site. At this stage it is not possible to relate each morphotype to a specific allodaposuchid taxon.
The authors thanks M. Padilla, MACV, and M. Calvo, MNCN, for specimen access. Special thanks to members of Grupo de Biología Evolutiva (UNED) and F. Marcos (UCM) for the preparation of the specimens. Thanks also to C. Brochu and M. Delfino for their useful comments, suggestions and language corrections.
- Barroso-Barcenilla, F., Cambra-Moo, O., Escaso, F., Ortega, F., Pascual, A., Pérez-García, A., et al. (2009). New and exceptional discovery in the Upper Cretaceous of the Iberian Peninsula: The palaeontological site of “Lo Hueco”, Cuenca, Spain. Cretaceous Research, 30, 1268–1278.CrossRefGoogle Scholar
- Blanco, A., Puértolas-Pascual, E., Marmi, J., Vila, B., & Sellés, A. G. (2014). Allodaposuchus palustris sp. nov. from the Upper Cretaceous of Fumanya (South-Eastern Pyrenees, Iberian Peninsula): Systematics, Palaeoecology and Palaeobiogeography of the Enigmatic Allodaposuchian Crocodylians. PLoS One, 9(12), e115837.CrossRefGoogle Scholar
- Brochu, C. A. (1992). Ontogeny of the postcranium in crocodylomorph archosaurs. Unpublished Master of Arts Thesis, The University of Texas at Austin, Austin, Texas, p. 340.Google Scholar
- Brochu, C. A. (1999). Phylogenetics, taxonomy, and historical biogeography of alligatoroidea. Society of Vertebrate Paleontology, Memoir 6, 19(2), 9–100.Google Scholar
- Buffetaut, E. (1975). Sur l’anatomie et la position systématique de Bernissartia fagesii Dollo, L., 1883, crocodilien du Wealdien de Bernissart, Belgique. Bulletin de l’Institut Royal des Sciences Naturelles de Belgique Sciences de la Terre, 51(2), 1–20.Google Scholar
- Buscalioni, A. D., Ortega, F., Weishampel, D. B., & Jianu, C. M. (2001). A revision of the crocodyliform Allodaposuchus precedens from the Upper Cretaceous of the Haţeg Basin, Romania. Its relevance in the phylogeny of Eusuchia. Journal of Vertebrate Paleontology, 21(1), 74–86.CrossRefGoogle Scholar
- Buscalioni, A. D., & Sanz, J. L. (1990). The small crocodile Bernissartia fagesii from the Lower Cretaceous of Galve (Teruel, Spain). Bulletin de L’institut Royal des Sciences Naturelles de Belgique, Sciences de la Terre, 60, 129–150.Google Scholar
- Dollo, L. (1883). Première note sur les crocodiliens de Bernissart. Bulletin du Musée Royal d’Histoire Naturelle de Belgique, 2, 309–338.Google Scholar
- Erickson, B. R. (1976). Osteology of the early eusuchian crocodile Leidyosuchus formidabilis, sp. nov. Monographs of the Science Museum of Minnesota, 2, 1–61.Google Scholar
- Hammer, Ø., Harper, D. A. T., & Ryan, P. D. (2001). PAST: Paleontological statistics software package for education and data analysis. Palaeontologia Electronica, 4(1), 1–19.Google Scholar
- IBM Corporation (2011). IBM SPSS Statistics for Macintosh, v 20.0. Armonk, New York.Google Scholar
- Martin, J. E., Delfino, M., Garcia, G., Godefroit, P., Berton, S., & Valentin, X. (2015). New specimens of Allodaposuchus precedens from France: Intraspecific variability and the diversity of European Late Cretaceous eusuchians. Zoological Journal of the Linnean Society, 176(3), 607–631.CrossRefGoogle Scholar
- Nopcsa, F. B. (1928). Palaeontological notes on Reptilia. 7. Classification of the Crocodilia. Geologica Hungarica, Series Palaeontologica, 1, 75–84.Google Scholar
- Pol, D., Turner, A. H., & Norell, M. A. (2009). The Late Cretaceous crocodylomorp Shamosuchus djadochtaensis and a discussion of neosuchian phylogeny as related to the origin of Eusuchia. Bulletin of the American Museum of Natural History, 324, 1–103.Google Scholar
- Romer, A. S. (1956). Osteology of the reptiles. Chicago: University of Chicago Press.Google Scholar