Aristonectes

The first Aristonectes fossil was discovered long before the genus was named by Ángel Cabrera in 1941.[1] In 1848, the Franco-Chilean naturalist Claude Gay described the first known plesiosaur from South America, Plesiosaurus chilensis, on the basis of a single caudal vertebra discovered in Quiriquina Island, in Concepción Province, Chile.[2] In a book published in 1889, Richard Lydekker placed this species in the genus Cimoliosaurus.[3] In 1895, Wilhelm Deecke moved it to the pliosaurid genus Pliosaurus as Pliosaurus chilensis, and referred other fossils that have been discovered in the same locality to it. Among these fossils are some vertebrae, a femur and fragments of an ischium as well as ribs.[4] In 1918, the Russian paleontologist Pavel A. Pravoslavlev moved it into the related genus Elasmosaurus, as Elasmosaurus chilensis.[5] In 1941, Cabrera described a different specimen from Chubut Province, Argentina as the new genus and species Aristonectes parvidens. In the same publication, he also moved Elasmosaurus chilensis back into Plesiosaurus with doubts (therefore resurrecting its original name "Plesiosaurus" chilensis), and compared it with his new species A. parvidens.[6] In 1949, Edwin H. Colbert found that the holotype specimen of "P." chilensis (the caudal vertebra originally described by Gay) belongs to a pliosauroid, but seemed uncertain about its generic placement. Also in the article, he considers that the other fossil material previously attributed to "P." chilensis could come from elasmosaurids.[7][8] After a further analysis published in 2013 by José P. O'Gorman and colleagues, the holotype specimen of "P." chilensis was recognized as potentially belonging to A. parvidens. As this specimen merely consists of a single caudal vertebra, it was referred to as A. cf. parvidens to indicate uncertainty in its assignment.[1]

The first specimen formally identified as A. parvidens was collected by Cristian S. Petersen in collaboration with local resident Victor Saldivia, in the Cañadón de los Loros, near the locality of Paso del Sapo in Chubut Province, Argentina.[6] The specimen was discovered in the Lefipán Formation and was dated to the Maastrichtian.[9][10][11] In September 1940, Pablo Groeber sent this specimen to the Museo de La Plata, as a donation for the Dirección de Minas y Geología del Ministerio de Agricultura (Directorate of Mines and Geology of the Ministry of Agriculture). A vertebra and incomplete phalanges from the same region had previously been donated to the Museo de La Plata by Mario Reguiló, and were subsequently referred to the holotype, catalogued as MLP 40-XI-14-6, as they most likely came from the same specimen.[6] The holotype consists of a partial skull attached to the mandible, the atlas and axis, 21 cervical vertebrae (including 16 anterior cervicals which are articulated), 8 caudal vertebrae and an incomplete left forelimb. The specimen is interpreted as an adult, or even an elderly individual.[9][12] Cabrera described the animal as Aristonectes parvidens in his article published in 1941 by the museum that houses the fossil material.[6] The genus name Aristonectes comes from the Ancient Greek words ἄριστος (áristos, "best", "superior") and νηκτός (nêktós, "swimmer"), and may be translated as "best swimmer", so named to indicate the best specimen of plesiosaur found at that time in Argentina. The specific epithet parvidens literally means "small teeth", in reference to the rather small size of the tooth sockets of this taxon.[13]

Since the unknown parts of the original specimen were reconstructed from added casts of the fossils, the precise anatomy of Aristonectes remained uncertain throughout the 20th century.[14][15][16] In 2003, Spanish paleontologist Zulma Gasparini and colleagues re-prepared the holotype skull along with the atlas and axis, in an attempt to clarify its anatomy.[9] In a doctoral thesis published in 2013, O'Gorman informally assigns several caudal vertebrae to A. parvidens.[10] In 2016, the same author revised the holotype again and described new details not reported in previous studies concerning this species.[11]

The first discovery of the second known species, A. quiriquinensis, dates back to the late 1950s, when Argentinian paleontologist Rodolfo Casamiquela discovered a partial skeleton in Las Tablas Bay, located north of Quiriquina Island, Chile. The specimen, cataloged as SGO.PV.260, was found in Maastrichtian beds of the Quiriquina Formation, and first consisted of five vertebrae, a humerus fragment, and the distal end of a limb. The remarkable hardness of the rocks containing the fossils made subsequent preparation difficult. After preparation, the fossil contains all four limbs, a posterior portion of the neck, most of the trunk, and a complete tail. It is interpreted as a juvenile. This specimen was first analyzed in 2012 and identified as an indeterminate aristonectine,[17] before being assigned to the genus Aristonectes in 2013.[18] In 2014, the specimen was referred to as A. quiriquinensis.[12]

In 2001, Chilean paleontologist Mario E. Suárez collected a partial skull, mandibular fragments and 12 anterior cervical vertebrae from a beach near Cocholgüe, located north of Tomé. This material was described in the following year and attributed to Aristonectes, but identification as a separate species was then impossible due to lack of preparation.[19] In early 2009, a second excavation was carried out independently at the same site by a team from the University of Concepción of Chile and the University of Heidelberg of Germany. This excavation recovered 119 blocks of sandstone, most of them with bone material, but some were damaged due to the degradation of the bones by the periodic immersion of sea water which transformed the most delicate parts into brittle surfaces. Additionally, several anatomical contacts were lost as blocks of sandstone were cut from the beach using a rock saw at low tide. The precise location of the find and the taphonomic distribution of the fossils show that they belong to the same specimen that was discovered in 2001. The bones were sent to the Chilean National Museum of Natural History for the preparation and scientific analysis. The completely prepared fossil consists of the skull, the axis and the atlas, 12 anterior cervical vertebrae, 23 middle to posterior cervical vertebrae, most of the trunk, the two almost complete forelimbs and a significant part of the right hind limb. The specimen was likely a young adult. After comparison with A. parvidens, the specimen, which is cataloged as SGO.PV.957, became the holotype of the new species A. quiriquinensis. The specific epithet is a reference to the formation of Quiriquina, which is the type locality of this taxon.[12]

In 2015, two other specimens from Quiriquina Island were referred to this species. These are composed of a complete left femur and a proximal part of a humerus, cataloged respectively as SGO.PV.135 and SGO.PV.169, which were previously referred to the now dubious genus Mauisaurus.[20] In 2018, the holotype specimen was redescribed after further preparation. In the same publication, another specimen, SGO.PV.94, consisting of a series of anterior caudal vertebrae, was also referred to A. quiriquensis.[21]

In the two known species, the skull of Aristonectes is flattened and ogive-shaped.[9][11][12][21] The size of the skull varies very slightly between the two species, that of the holotype of A. quiriquinensis having a proposed size between 65–70 cm (26–28 in), while from the holotype of A. parvidens measures 73.5 cm (28.9 in).[12] In A. quiriquinensis the squamosal bones extend well posterior to the occipital condyle. The pterygoid bones are large and project posterior to the neurocranium. The posterior extension of the pterygoid bones surrounding the axis and the atlas, this reduces the space between the skull and the vertebrae, and would therefore limit lateral movement at their joints.[21] Around 13 teeth were present in the premaxilla and 50 teeth present in the maxilla.[12][11] Among the unique characters of the skull of A. quiriquinensis is the presence of a mental boss (a ridge that appears in the mandibular symphysis), a feature not observed in the holotype of A. parvidens.[12]

The mandibular symphysis of Aristonectes is short.[6][9][11][12][21] The dentary bones, although being partially preserved in both species, would have a total of 50 teeth in A. quiriquinensis and 63 to more in A. parvidens according to estimates.[9][12][24][11][lower-alpha 2] The mandibular symphysis of the dentary bones in A. quiriquinensis is comparatively thicker and lacks the deep groove seen in ventral view in A. parvidens. The most posterior teeth of A. parvidens were about 60 mm (2.4 in) rostrally from the coronoid process. In A. quiriquinensis, there are no tooth sockets present although 130 mm (5.1 in) of the mandible is preserved rostral to the coronoid process. Therefore, it is likely that A. quiriquinensis would have around fewer teeth than A. parvidens. In A. parvidens, the labial contact between the angular bone and the surangular, anterior to the glenoid fossa, bears a deep cavity which is not present in A. quiriquinensis.[12]

The teeth of A. quiriquinensis have an oval cross-section, being very similar to elasmosaurids of the Northern Hemisphere and Weddellian Province, as in Kaiwhekea. The length of the root of complete teeth is slightly longer than that of the crown. The teeth of A. quiriquinensis, and most likely A. parvidens,[11][lower-alpha 3] is homodont, wich means that the teeth are all of identical shape in the jaw. The largest known tooth is very thin and pointed, having a crown bearing ridges on the lingual side. The smaller tooth is a short-rooted replacement tooth, having a similar morphology to the larger one.[12] All teeth are inclined toward the front and would not fit together when the jaws are closed.[21]

Although some postcranial remains are known in both species, they are mainly documented in A. quiriquinensis, the holotype specimen of this latter being significantly more complete than that of A. parvidens.[12][21] In 2018, Otero and his colleagues estimate that A. quiriquinensis would have had a total of 109 vertebrae, among which would be 43 cervical vertebrae, 3 pectoral vertebrae, 24 dorsal vertebrae, 3 sacral vertebrae and 35 caudal vertebrae.[21] The anterior cervical ribs are recurved in A. parvidens and fused distally,[6] while those of A. quiriquinensis are short and lack contacts.[12] The cervical ribs and cervical neural spines of Aristonectes are curved towards the head, similar to other members of the subfamily Aristonectinae.[12][21] This morphology would have allowed for maintaining the trunk higher than the head.[21] The neural spines present on the middle of the cervical part look like blades and are enlarged distally.[12]

The dorsal vertebrae are larger and taller than the cervical vertebrae and are also taller than they are wide. Dorsal vertebrae have centra which exhibit constrictions and which have amphicelous articular surfaces.[12] Between the glenoid cavities, the pectoral girdle of A. quiriquinensis have a width of nearly one meter. The ribs appear thickened on their distal sides. The gastralia are pachyostotic and have a moderate length. The caudal vertebrae of A. quiriquinensis are wider than tall or long. The neural spines of the latter become progressively blunt and short, possessing a thick dorsal apex with a flat surface, this suggesting the probable existence of a strong ligament all along the dorsal part of the tail.[21]

The limbs of A. quiriquinensis are very long for an elasmosaurid, the best-preserved swim paddle being about 3 m (9.8 ft) long. Combined with the aforementioned width of the pectoral girdle, this indicates a total wingspan of 7 m (23 ft). In aquatic tetrapods, such proportions in the limbs are only observable in the rhomaleosaurid from the Early Jurassic Meyerasaurus and in the modern humpback whale.[21] A. parvidens would also have had similar fin size. In both species, the phalanges of their limbs have large articular facets, featuring elongated coil-shaped bones. In A. quiriquinensis the ulna is slightly shorter than the radius and the tibia is slightly wider than the fibula.[12]

As with most other plesiosaur fossils, gastroliths have been found in some specimens of Aristonectes. Among these are 5 elements interpreted as gastroliths who have been discovered during the second exhumation of the holotype of A. quiriquinensis,[12] and over 793 listed in the largest known specimen of the genus, MLP 89-III-3-1, discovered in Antarctica.[32][lower-alpha 4] The function of gastroliths in plesiosaurs is still controversial and has generated several hypotheses about their function.[34] The most frequently cited hypotheses are either their use in buoyancy control or in the digestive process, the latter being the most widely accepted by recent studies.[35][36][37] In the analysis carried out on the specimen MLP 89-III-3-1, the researchers conclude that Aristonectes would not ingest the gastroliths individually, but would pick them up randomly, an affirmation which is based according to their consistencies. However, as this specimen was disarticulated when it was discovered (i.e., the bones had moved out of their original anatomical position), it is likely that not all of its gastrolithes have been found. The gastroliths do not indicate whether they were used in this genus for digestion.[32]

A. quiriquinensis has a morphology of the swimming paddles and the tail close to various modern cetaceans, and postcranial characteristics which would limit the lateral movement of the head. In addition, the animal's skull has an enlarged mouth that would have allowed the engulfment of a large volume of water. This indicates that A. quiriquinensis fed in benthic zones, mixing prey and sediment at the same time. This type of feeding pattern is also documented in modern gray whales. The presence of decapod crustaceans and fishes within the Quiriquina Formation further reinforces this assertion.[21]

After the discovery of A. quiriquinensis in 2014, several specimens were subsequently referred to this latter. In a study published in 2015 by Otero and his colleagues, it was noted that during their growth, the humerus and femur go from a flat capitulum in juveniles to hemispherical shaped heads of the two bones at adulthood.[20]