Aquilolamna milarcae gen. et sp. nov Etymology Generic name combines ʿ Aquila ʾ (Latin, eagle), and the shark genus name ʿ Lamna ʾ, in reference to the wing-shaped pectoral fins and aquilopelagic ecomorphotype of this probable lamniform. Specific epithet refers to the Museo La Milarca (San Pedro Garza García, Nuevo León State, Mexico), where the holotype is housed and exhibited. Proposed vernacular name: eagle shark. Holotype and only known specimen Instituto Nacional de Antropología e Historia (INAH) 2544 P. F.17, an articulated and nearly complete skeleton with associated soft tissue (Figs. 1 and 2 and Figs. S3 to S6). Locality and horizon Vallecillo quarry area, Nuevo León State, Mexico (26.655° N, 100.044° W); Vallecillo Member of the Agua Nueva Formation, middle -upper Pseudaspidoceras flexuosum Zone, lower Turonian, Upper Cretaceous (Fig. S 1) (32). This stratigraphic level is ca. 93 million years in age. Diagnosis of family, genus and species Medium-sized neoselachian shark that differs from all other selachimorphs in having hypertrophied, scythe-shaped plesodic pectoral fins whose span exceeds the total length of the animal. High number (~70) of anteriorly directed pectoral radials. Head short and broad, with wide and near-terminal mouth. Caudal fin markedly heterocercal. Caudal fin skeleton showing a high hypochordal ray angle (i.e., ventrally directed hypochordal rays). Caudal tip slender with no (or strongly reduced?) terminal lobe. Squamation strongly reduced (or completely absent?). Remark on Aquilolamnidae The family Aquilolamnidae (type and only known genus: Aquilolamna) possibly includes the enigmatic, poorly known Late Cretaceous genera Cretomanta (24) and Platylithophycus (27). Systematic and paleoecological assumptions combined with stratigraphic evidence (see main text) suggest that the tooth-based genus Cretomanta is a suitable candidate for being the dental elements of the skeleton-based genus Aquilolamna (for which teeth are unknown in the sole specimen available to date) and thus may be a senior synonym of the latter; however, both taxa must be considered distinct pending the discovery of overlapping material. Description The holotype is a 166-cm-long (total length TL) holomorphic specimen with a precaudal length (PCL) of 122 cm and an estimated pectoral fin span (PFS) of about 190 cm (158 cm as preserved) (PFS about 1.1 times TL and 1.6 times PCL); the specimen is preserved flattened dorsoventrally, with the exception of the tail exposed in lateral view (Fig. 1). The vertebral column consists of about 225 asterospondylic vertebrae, with the caudal vertebral count (~120) higher than the precaudal vertebral count (~105). Anteriorly, there is no synarcual element, indicating that this specimen cannot be classified as a batoid (9). Tessellated prismatic calcified cartilage is well preserved in various parts of the specimen, with tesserae showing the sub-hexagonal close-packing arrangement typical of chondrichthyans (27). Soft-tissue imprints (skin, muscles) are preserved, providing information on the body outline (Figs. 1 and 2 and Figs. S3 to S6). The skin seems to be overall devoid of dermal denticles, as in many myliobatiform rays (54). The head is severely crushed dorsoventrally; it is short, with an indistinct snout and a wide mouth (Fig. 2A and Fig. S3). As indicated by the position of the anteriormost vertebra, the poorly preserved chondrocranium represents about 10% of the precaudal length. The overall outline of the chondrocranium is not recognizable, although some structures can be tentatively interpreted as rostral cartilages and nasal capsules. The jaws are long, extending posterior to the occiput. With the exception of a disarticulated and displaced Meckelʾs cartilage (259 mm in length), jaw elements cannot be identified with certainty, as those of hyoid and branchial arches. No teeth are preserved, possibly due to rapid post-mortem disarticulation and scattering affecting the dentition, like in fossil skeletons of cetorhinid sharks (55, 56). The hyper-elongate pectoral fins are rather broad at their bases, slightly forward-directed, and gently recurved at their ends (Fig. 2D and Fig. S4). They show an extremely increased number of radial cartilages. At least 70 radials are present, probably supported by several rodlike metapterygial elements. These radials are directed towards the anterior margin of the pectoral fin and some of them are bifurcated anteriorly (Fig. S 5), as in the megamouth shark (7). The trunk is moderately slender. No stomach/gut contents are preserved. There are no traces of dorsal, pelvic or anal fins; this reflects either original morphology (true phylogenetic absence) or taphonomic loss (post-mortem disarticulation and degradation). The caudal fin skeleton is characterized by ventrally directed hypochordal rays (hypochordal ray angle = 120°). The heterocercal caudal fin shows an elongate and slender dorsal lobe (dorsal margin length = 51 cm) at a moderate angle above the body axis (heterocercal angle = 27°), and a rather strong ventral lobe (preventral margin length = 21 cm) at a high angle below the body axis (hypocercal angle = 55°) ["lamniform caudal fin type 2" sensu (13)]. No terminal lobe can be discerned at the caudal tip (Fig. 2 C and Fig. S 6). Remark on the caudal fin shape Interestingly, the caudal fin shape of sharks is generally related to their primary habitat and lifestyle (1, 57). Among modern sharks, four main types of caudal fin shape can be identified (57). The caudal fin type observed in Aquilolamna milarcae is present in most active-swimming littoral and oceanic species (e.g., Megachasmidae, Odontaspididae, Carcharhinidae) (1, 13, 57). High-speed pelagic sharks with a tuna-like morphology (Lamnidae) have a crescentic, rather symmetrical caudal fin (i.e., high heterocercal and hypocercal angles) (1, 13, 57). Deepwater sharks such as Squaliformes generally have a caudal fin with a moderate to low heterocercal angle, a well-developed epichordal (= epicaudal) lobe, and a large terminal lobe (1, 57). Lastly, bottom-dwelling sharks usually have a caudal fin characterized by a very low heterocercal angle and a poorly defined (or no separate) ventral lobe (1, 57) (Fig. S 7). In conclusion, the Megachasma -like caudal fin shape of Aquilolamna milarcae, well distinct from that of benthic species, strongly indicates that this shark was an active-swimming pelagic form (although not capable of high swimming speeds) (Fig. S 8). Remark on the presence/absence of dorsal, pelvic and anal fins As mentioned above, the apparent absence of dorsal, pelvic and anal fins in INAH 2544 P.F.17 reflects either original morphology or taphonomic loss. In modern sharks, the first dorsal fin is more or less developed (1, 9). Among filter-feeding Lamniformes, the basking shark (Cetorhinus maximus; derived tachypelagic form) has a high, well-developed first dorsal fin, whereas the megamouth shark (Megachasma pelagios; macroceanic form) has a low, moderately large first dorsal fin (16). Interestingly, this is inversely proportional to the development of their pectoral fins. In Cetorhinus and Megachasma, the pectoral fin span represents 54% and 70% of the precaudal length, respectively (see Table S 1). Furthermore, the pectoral area is about equal to the first dorsal fin area in Cetorhinus, whereas the pectoral area is about triple the first dorsal fin area in Megachasma (16). Similarly, pelagic cownose, eagle and devil rays (Rhinopteridae, Myliobatidae and Mobulidae; aquilopelagic forms) have enlarged winglike pectoral fins and a small dorsal fin (despite the lack of caudal fin) (1, 9). In analogy with modern elasmobranchs, the hyper-elongate pectoral fins of Aquilolamna would suggest that the first dorsal fin, if present, might have been small. 8 Whereas the anal fin was secondarily lost in many living sharks and all living batoids, pelvic fins are present in all modern elasmobranchs (9, 10). Pelvic fins were not secondary lost because males have paired copulatory organs (claspers) whose cartilaginous elements correspond to a modified portion of the pelvic fin skeleton (9). Therefore, the apparent lack of pelvic fins in the holotype of Aquilolamna is most likely due to taphonomic loss. Only the discovery of new material may help determine whether Aquilolamna actually possessed dorsal, pelvic and anal fins. Remark on the body size The holotype and single known specimen of Aquilolamna milarcae is a medium-sized individual; however, it is not possible to determine 1) whether this specimen represents a juvenile, subadult or adult individual, and 2) the maximum size attainable by this species. Unfortunately, the origin and evolutionary history of Aquilolamnidae are undocumented, and we do not know if these putative lamniform sharks could reach gigantic sizes like in the four modern lineages of filter-feeding elasmobranchs (i.e., Rhincodontidae, Megachasmidae, Cetorhinidae, Mobulidae) (3). If the enigmatic genus Platylithophycus (known by a single specimen representing a large individual) (27) can be proven to be actually an aquilolamnid, this would suggest an evolutionary pathway toward gigantism in this lineage of Late Cretaceous sharks. Similarly, it is not possible, based on the current knowledge, to assess whether aquilolamnids arose from medium-sized benthic ancestors (as did rhincodontids and mobulids) (3), large-sized pelagic ancestors (as did megachasmatids and cetorhinids) (3), or medium-sized pelagic ancestors. LDA results The first LDA including only shark species (correct classification = 88.5%; Table S 5) shows that Aquilolamna milarcae clearly falls outside the five specialized ecomorphotypes considered here (Fig. S 10). This is also true when a 250 cm hypothetical maximum total length is used for A. milarcae. This strongly suggests that the body proportions of Aquilolamna are distinctive among selachimorphs. The second LDA including both shark and batoid species (correct classification = 73.6%; Table S 6) shows that A. milarcae falls in the overlapping area between the aquilopelagic and rajobenthic ecomorphotypes (Fig. 3 and Fig. S 11). However, the LDA classifier assigns A. milarcae to the aquilopelagic group based on the minimal Mahalanobis distanceto the group mean. This is confirmed when a 250 cm hypothetical maximum total length is used for A. milarcae, with a new position located within the aquilopelagic ecomorphotype but outside the rajobenthic ecomorphotype. The partial overlap observed between the aquilopelagic and rajobenthic ecomorphotypes is due to the aquilopelagic-like disc proportions (i.e., disc markedly wider than long) of the butterfly rays (Gymnura spp.) (21, 58). Gymnurids were originally categorized as bottom-dwelling, undulatory rajobenthic rays (1) (ecomorphotype assignment followed here; Table S 2), but subsequently studies demonstrated that this group actually use fin movements intermediate between undulatory (like in typical rajobenthic forms) and oscillatory (like in typical aquilopelagic forms) locomotion (21, 58). Gymnurids exhibit an undulatory-based swimming mode near the sea bottom, whereas their locomotion is more oscillatory in the water column (21). Therefore, gymnurids should not be considered as a rajobenthic group sensu stricto. Arajobenthic ecomorphotype can be ruled out for A. milarcae, as the latter was likely an active-swimming pelagic form (see above) characterized by pectoral fins with a narrowly angular shape and gently recurved distal tips (unlike gymnurids but like aquilopelagic myliobatiforms). With the exception of its typical shark caudal fin, A. milarcae shows aquilopelagic-like morphology and proportions, clearly distinct from those observed in selachimorphs. Interestingly, the living demersal species Rhina ancylostoma (shark ray), which is similar to A. milarcae in having a shark -ray mixed appearance, is a rhinobenthic batoid that falls between the macroceanic and squatinobenthic shark species. Whereas the body of A. milarcae is wider than long, that of R. ancylostoma is clearly longer than wide, with precaudal length about 0.6 and 1.5 times pectoral fin span, respectively.