The first Peruvian record of Enchodus (Actinopterygii, Aulopiformes, Enchodontidae) in the Upper Cretaceous Vivian Formation

We describe isolated teleostean teeth found in no association with the jaw bone. The specimens have been recovered in Late Cretaceous marine deposits of the Vivian Formation in the Peruvian Sub-Andean Region. The deposition sequence from where the teeth come is interpreted as a shallowing-upward sequence of low salinity. The fish material is identified as Enchodus aff. E. gladiolus based on the presence of a small but well-developed post-apical barb, an anterior cutting edge, the crown is symmetrical in cross-section, have a sigmoidal profile, and bears strong ridges (=striations). The Peruvian material differs from the typical E. gladiolus teeth in having a faintly serrated anterior cutting edge which is absent in most specimens referred to E. gladiolus. We also highlight that taxonomic assignments made based on isolated teeth must be taken with care. Despite scarce, the material recovered denotes that the marine units of Peru can give valuable information about the Pacific fish fauna during the Late Cretaceous.


Introduction
The Mesozoic fish record is extremely scarce in Peru, especially compared with coeval fossil bearing localities in neighboring countries of western South America such as Bolivia and Chile (e.g., Kriwet and Klug, 2012;Suárez, 2003;Arratia, 2015).
One of the most striking features of Enchodus species is the presence of long palatine bones with a single, large tooth (= fang), and a corresponding dentary tooth near the mandibular symphysis. These large teeth and others have very good potential of preservation and are relatively diagnostic (Grandstaff and Parris, 1990). Moreover, isolated palatine, lower and upper jaw teeth and isolated fragmentary bones have been used to identify material at species level (e.g., Carbot-Chanona and Than-Marchese, 2013) and considered type material (e.g., Holloway et al., 2017). However, the definition of species based on isolated teeth and isolated skeletal elements has been questioned (Goody, 1976;Chalifa, 1996).
The aim of this study is to describe and discuss the taxonomic assignment of isolated teeth recovered in Upper Cretaceous levels of Vivian Formation in Perú, South America. Additionally, we comment about the paleogeographic distribution of the genus Enchodus.

Material
Material consist of eight teeth -none of them associated with bone-housed at the INGEMMET (Instituto Geológico, Minero y Metalúrgico), Lima, Perú.

Methods
Teeth were studied and photographed under a binocular microscope. Also, a Scanning Electron Microscope FEI ESEM Quanta 200 with electron source from a tungsten filament, with accelerating voltage of 200 V-30 kV of the SeMFi-LIMF (Servicio de Microscopía Electrónica de Barrido y Microanálisis del Laboratorio de Investigaciones de Metalurgia Física "Ing. Gregorio Cusminsky", Facultad de Ingeniería, Universidad Nacional de La Plata) was used. The Energy Dispersive X-ray spectroscopy (EDAX) analysis was performed using an EDAX SDD Apollo 40 in the SEM. This equipment allows the detection of light elements of boron, with a resolution <135 eV, with a qualitative, semiquantitative and with the possibility of quantitative analysis on chemical elements in a microanalysis using a 1 mm 3 sample.
In the jungle area of Ucayali, Junín, and Pasco departments, rocks of the Vivian Formation are outcropping in isolated sections of low thickness over the margins of large river courses, as occurs in the Río Apurucayali (Fig. 1B). In this area, the logged section of the Vivian Formation ( Fig. 1C) is approximately 100 meters thick. Its general strike direction is NNE-SSW dipping up to 10 degrees west. The lower part of the Vivian Formation is characterized by fine, very well-sorted, white, quartzose arenites named "White-sugar Vivian" that become red quartzites, with reddish pelitic levels, to their upper part named as "Red Vivian". In the lower part of the Vivian Formation in the Apurucayali River profile (Fig. 1C), eight sedimentary facies have been identified, which are represented by sandy, heterolithic, and shale facies (Fig. 1C). The main psammitic facies consist of whitish, pebbly-sandstones, massive (SGm) and tangential I N P R E S S crossbedding (SGta), and whitish sandstone, with trough crossbedding (St) tangential crossbedding (Sta), hummocky cross-stratification (Shcs) and massive sandstone (Sm). The heterolithic deposits are composed of normal heterolithic facies (Ht) and very sandrich heterolithic facies (Htvs). The fine, dark gray to black, shales, facies are laminated (Fl) or massive (Fm) (Poiré et al., 2016).

I N P R E S S
Paleontological content in the lower part of the Vivian Formation at Río Apurucayali section (Fig. 1C) comprises some remains of isolated invertebrates (mollusks), vertebrate remains (fishes, reptiles: cf. Plesiosauria, coprolites) and some plant debris and carbonaceous material (Poiré et al., 2016). Trace fossils are represented by the ichnogenera Palaeophycus, Skolithos, and Chondrites, which are distributed as ichnofacies of Cruziana in its lower part and Cruziana-Skolithos Impoverished ichnofacies in its middle and upper part. The former is poorly developed, has centimeter-size diameter traces and a low bioturbation index (BI = 1 to 2). On the other hand, these same ichnogenera in the Impoverished ichnofacies are well-developed but present smaller sizes (millimeters) and a high degree of bioturbation (BI = 4 to 6) (Poiré et al., 2016).
In general, the sedimentary rocks of the Vivian Formation have been considered as the product of a deposition in marine littoral (Jaillard et al., 1995), distal fluvial (Rodríguez et al., 2017), and fluvial-estuarine or fluvial-deltaic environments (Navarro, 2005). In this sense, although the presence of fossil traces of the Cruziana-Skolithos ichnofacies, very common in this unit, indicates a shallow marine environment, it is not present in the fossil level (lag), where the impoverished ichnofacies indicates deposition of sediments in brackish waters typical of an estuary (or a tide-dominated delta) (Poiré et al., 2016). It should be noted that the former is from normal salinity marine waters and pass to the same fossil but tiny traces, typical of an impoverished brackish water ichnofacies (Howard and Frey, 1975).

Description
The material consists of eight fang-like, small teeth. Few teeth are complete but some are incompletely preserved. Despite no teeth were connected to bone, CPI 7029a ( Fig. 2A) greatly agrees in morphology with palatine fang-like teeth. CPI 7029a is about 8 mm long and 2 mm antero-posteriorly wide (measured at the tooth base). The tooth is I N P R E S S distally compressed and slightly sigmoidal with an anterior cutting edge and a small posterior post-apical barb ( Fig. 2A-B). The acrodine cap of the apical barb seems to be delimited by a faint outline (Fig. 2B). Since the acrodine cap shows different gray tonalities, it could be possible that the enameloid was restricted to the tip of the cap (Fig. 2B). However, a detailed histological study is necessary to allow accurate identification of the tissues. The teeth shaft has several -at least twelve-very welldeveloped longitudinal and undivided ridges or folds (= striations; Fig. 2A). The anterior margin of the teeth bears a slightly wavy and well-developed keel ( Fig. 2A, D). In cross section, teeth are compressed near the tip and symmetrical and drop-like more basally (Fig. 2C). The anterior margin has a faint irregular serration that is only observed under SEM (Fig. 2 D-E). A natural fracture in one tooth (CPI 7029b) shows that the pulp cavity at this height is filled with osteodentine (sensu Ørvig, 1951;Fig. 2F). The complex structure is interpreted as dentine trabeculae and denteons formed around vascular canals. Radiating tubules arise from these canals (Fig. 2F). The enameloid is not clearly distinguished in the natural section but it should be located distally to the parallel series of denteons (Fig. 2F).

Taxonomic remarks
Enchodus is the genus of the family with more nominal species, longest stratigraphic range, and widest geographical distribution (see above). The teeth described herein resemble those of Enchodus gladiolus (Cope, 1872) in size and because they are symmetrical in cross section, have strong basoapical ridges or folds (= striations), a sigmoidal curvature of crown, an anterior cutting edge, and a small but well-developed post apical barb. Moreover, a phylogenetic analysis recently published (Díaz-Cruz et al., 2019, fig. 7), shows that the sigmoid dermopalatine tooth is a synapomorphy of a clade gathering the species: E. gladiolus, E. dirus, the "Enchodus" from Gavdos (Cavin et al., 2012), and E., tineidae (Holloway et al., 2017). Noteworthy, "among these species the dermopalatine tooth of E. gladiolus shows a peculiar condition, a terminal wing or barb in the posterior end of its tip" (Alvarado-Ortega et al., 2020, p. 12).
Teeth from Peru greatly resemble those identified as palatine teeth of Enchodus cf. Notwithstanding that Cope (1872 p. 353) mentioned that the barb was absent in the holotype (only specimen known to him and lost), Goody (1976 p.103), based on more complete materials, did a detailed review and assigned the species of Cope (1872) to the genus Enchodus mentioning the presence of a posterior apical barb. Since then the post-apical barb was described as commonly present in the species (e.g. Becker et al., 2010;Shimada et al., 2006 fig. 2Q-T), assigned an isolated palatine tooth from the Maastrichtian of the Ocozocoautla Formation (Mexico) to Enchodus gladiolus. Notably, the tooth illustrated has an even more heavily serrated anterior margin than the Peruvian teeth reported herein. We agree with previous authors (see above) in that taxonomic assignment of isolated teeth needs to be taken carefully, and more and betterpreserved material is always desirable to allow an accurate assignment. Considering that some specimens, such as those described here, show characters that are not typical of one species, the specimens must be tentatively referred to specific level. We decided to refer the Peruvian specimens as Enchodus aff. E. gladiolus denoting that probably belongs to a new species but the available material does not allow a better assignation.

Remarks on lifestyle
According to the morphology of the skull and due to the mouth gape, it has been postulated that Enchodus was a predator occupying a middle position in the food chain (Fielitz, 2004). Quite probably, Enchodus feeded on cephalopods and small fishes (Grandstaff and Parris, 1990) and was consumed by bigger teleosts (e.g., Cavin, 1999), sharks, and even plesiosaurs (Cicimurri and Everhart, 2001).
The genus Enchodus clusters medium-sized, predatory fishes. The largest forms might be epipelagic but most of Enchodus species seem to have been shallow-dwelling fishes inhabiting marine shelf or coastal waters (i.e, above 200 m). Today, shelf and coastal waters have the greatest marine fish diversity and a high productivity; moreover, they are biogeographically distinct from the adjacent high seas and deep benthic environments (e.g., Spalding et al., 2007).

4.4.
Preservational remarks The specimens here described were found in the shallowing-upward facies of psamites that pass to a mixed composition in the lower portion of the Vivian Formation. This shallowing-upward sequence combines the features of the low salinity proximal environment.
The facies seem to have allowed the mineralogical preservation of the crystalline material that composes the teeth. This is confirmed by the analysis of major chemical elements carried out at the tooth crown surface (Fig. 3A) and tooth pulp cavity (Fig. 3B).
The EDX-spectrum of the tooth crown surface results in an almost negligible presence of C indicating the absence or poor representation of CaCO3. We interpret that there was a little fossil-diagenetic alteration of the teeth with respect to the bearing sediments. Moreover, the EDX-spectrum of the tooth crown surface shows three main peaks -phosphorus, carbon, and oxygen-with minor peaks of calcium and chlorine. Those elements are the constituent of Calcite [CaCo3] and Hydroxyapatite [Ca10(PO4)6(OH)2] being the hydroxyapatite the major mineral component of teeth and bones (Carter, 1990). Likewise, the spectrum of the external surface part of the tooth would be reflecting a higher content of C from the carrier rock (black massive mudstone), see Figs. 3A and 3B.

I N P R E S S
We conclude that the crystallinity of the dental material without an obvious alteration and the presence of Carbon in the apical surface of the tooth would indicate a fossil-diagenetic process that left an imprint on the phosphate material. Without an obvious replacement, the presence of elements that make up the carbonate fossil would be indicative of a protective nature of the fossil-bearing sedimentary level.

4.5.
Comments on the fossil record of Enchodus gladiolus Enchodus gladiolus ranges from the Cenomanian to Maastrichtian. Most of the reports come from the Western Interior Seaway of North America. The lost type material (Cope, 1872) came from the Niobrara Formation (Coniacian-Santonian of Kansas; Goody, 1976). The oldest reports are Cenomanian (e.g. Cumbaa et al., 2010;Nagrodski et al., 2012;Gallardo et al., 2013;Jansen et al., 2012). Other records are Turonian (Nebraska; Ouroumova et al., 2016) and Maastrichtian (Dakota;Becker et al., 2010. In the southern part of North America, the species was reported from the Maastrichtian of Chiapas, Mexico (Carbot-Chanona and Than-Marchese 2013). In extracontinental areas, two fragmentary teeth assigned to Enchodus cf. gladiolus were found in Cenomanian beds of Nigeria (Vullo and Courville, 2014). However, the material is too fragmentary and should be compared with teeth of African species (see Arambourg, 1952). Teeth resembling E. gladiolus were also reported from the Campanian of Sweden (Bazzi et al., 2015). However, those teeth do not have the postapical barb and the ridges are extremely faint (see Bazzi et al., 2015 fig 5C). Bogan and Agnolin (2010 p. 185) identified one tooth from a marine Atlantic encroachment of Maastrichtian age in northern Patagonia as "Enchodus gladiolus type" sensu Goody, 1976 aff. E. gladiolus Cope, 1872". However, the tooth figured by Bogan and Agnolin (2010 fig. 6E-F) lacks typical characters such as well-developed ridges or folds (= striations), it is highly compressed, and it is not sigmoidal. Consequently, we consider that all the previous reports of Enchodus gladiolus outside North America are tenuous and must be reviewed when more and better-preserved material became available.

Paleobiogeography
Due to the N-S development of the epicontinental sea during the Late Cretaceous, the latitudinal distribution of the Enchodus species was very extended -reaching Antarctica-(e.g., Cione et al., 2018;Fig. 4). This might be related with the higher average sea temperature and lower latitudinal gradient of temperature during the Late Cretaceous (Cavin et al., 2012). Also, the widening of the Northern South Atlantic Ocean during the beginning of the Late Cretaceous improved the water circulation allowing the expansion of the enchodontids distribution (Silva and Gallo, 2016).
In South America, materials referred to the genus Enchodus have been reported in the Turonian and Maastrichtian of Brazil (Cope, 1886;Woodward, 1907;Oliveira and Silva Santos, 1950;Rebouças and Silva Santos, 1956;Silva, 2007;Silva Santos and Salgado, 1969;Gallo et al., 2006). The South American Maastrichtian record also comprises an isolated mention in Chile (Quiriquina Formation, Suárez, 2001Suárez, , 2003, a figuration of a tooth (El Molino Formation of Bolivia; Gayet et al., 1993), and a formal I N P R E S S description in Argentina (Jagüel Formation in Río Negro, Bogan and Agnolin, 2010). However, as was pointed out above for E. gladiolus, many of these records need to be reviewed carefully and are in need of more and better-preserved specimens. Isolated teeth of Enchodus were recorded in the Campanian-Maastricthian of Antarctica (Kriwet et al., 2006;Cione et al., 2018). During the Late Cretaceous the radiation of Enchodus occurred in the United States -comprising part of the Western Interior Seaway and even New Jersey-from where four species have been reported (i.e., E. gladiolus, E. shumardi Leidy, 1856, E. petrosus Cope 1874, and E. dirus Leidy, 1857, see Goody, 1976and Cavin 2008. However, in relation to the recent information (Alvarado-Ortega et al., 2009;Fielitz and González-Rodríguez, 2010;Díaz-Cruz et al., 2016, 2019, the radiation of Enchodus could have occurred outside the United States, including the present territory of Mexico. Although, as was mentioned by Díaz-Cruz et al. (2019), the biogeographical history of Enchodontidae is difficult to establish due to the short life span and wide inhabit region (the whole Tethyan Sea) of most of the genera.
The evidence highlights that -particularlythe Western portion of the Tethys Sea was crucial for the evolution of the Enchodontidae (see Díaz-Cruz et al., 2019). Most of the Enchodus species are restricted to epicontinental seas of temperate to tropical waters.

I N P R E S S
The report from Peru might correspond to the connections of shelf areas of the western part of South America with those of North America during the Late Cretaceous. Finally, the few Cenozoic reports of enchodontids need to be confirmed (e.g., Arambourg, 1952 mentioned but neither figured nor described rare and very small isolated teeth of E. elegans from Paleocene beds of Morocco; Rana et al., 2005 identified a tooth from the Eocene of India as Enchodus sp. However, it does not show clear diagnostic features). Thus, enchodontids seems to became extinct at the K/Pg extinction event and apparently, no other osteichthyan occupied immediately the ecospace in shelf areas. Chalifa, Y. 1989