FMNH-P or FMNH-PM: Paleontological Collection, Field Museum of Natural History, Chicago, USA; MACN-Pv: “Museo Argentino de Ciencias Naturales Bernardino Rivadavia”, “Colección Paleontología de Vertebrados, Ciudad Autónoma de Buenos Aires”, Argentina; MCH-P: “Museo Arqueológico Provincial Condor Huasi Sección Paleontología”; Catamarca, Argentina; PVL: “Colección de Paleontología de Vertebrados Lillo”, Tucumán, Argentina.

The cranial nomenclature follows Cherem and Ferigolo (2012) and the dental nomenclature follows Pérez, 2010a and Pérez, 2010b. The endocast nomenclature follows Dozo, 1997a and Dozo, 1997b and Campos and Welker (1976).

Skull measurements. aP4: alveolar protuberance of P4; app: paraoccipital apophysis; arf: anterior retroarticular foramen; av: auditory-visual sulcus; bc: crest of basioccipital; Be: basisphenoidal; Bo: basioccipital; cnd: nuchal dorsal crest; cnv: nuchal ventral crest; co: occipital condyle; cs: sagittal crest; ec: ectosylvian convolution; Ect: ectotympanic; fem: stylomastoid foramen; fit: infratympanic fenestra; fmg: magnum foramen; Fr: frontal; gc: glenoidea cavity; if: incisor foramen; ihs: interhemispheric sulcus; iv: intervisual sulcus; La: lacrimal; ls: lateral sulcus; mae: external auditory meatus; mc: marginal convolution; Mx: maxilla; Na: nasal; Pa: parietal; pce: caudal process of squamosal; pet: epitympanic sinus; Pl: palatine; pmp: mastoid of petrous; Pre: presphenoid; prf: posterior retroarticular foramen; Prmx: premaxilla; Pt: pterygoid; ptf: pterygoid fossa; pts: temporal of supraoccipital process; rdm: dorsal root of the zygomatic process; rvm: ventral root of the zygomatic process; Sq: squamosal; ss: sylvian or suprasylvian sulcus; ssc: suprasylvian convolution; to: temporal orifice of temporal meatus; vep?: sphenopalatine vacuities.

Mandible measurements. df: dorsal fossa; ias: incisive alveolar sheath; nMpi: masseter medialis pars infraorbitalis muscle.

Dental abbreviations. al: anterior lobe; ap4: anterior projection of anterior lobe of p4; c: cement; dc: dentine crest; e: enamel; FHy: longitudinal furrow opposite to hypoflexus; Hy: hypoflexus; M1/m1: upper/lower first molar; M2/m2: upper/lower second molar; M3/m3: upper/lower third molar; P4/p4: upper/lower fourth premolar; pl: posterior lobe; pp: posterior projection of M3.

To test the affinities of P. prisca within Caviidae, a cladistic analysis using a modified version of the combined dataset of Madozzo-Jaén et al. (2018) was performed. P. prisca and three characters were added to the dataset. The combined matrix (Supplementary Online Material 1) for this analysis resulted in 65 taxa, 143 morphological characters, and 4014 characters from DNA sequences (Supplementary Online Material 2). The equally weighted parsimony analysis was conducted using TNT 1.1 (Goloboff and Catalano, 2016), performing a heuristic search of 1000 Wagner tree replicates, followed by TBR branch swapping, collapsing zero-length branches under the strictest criterion. Support values were calculated using bootstrap and jackknife.

P. prisca was described based on remains found in Andalhuala locality, Santa María Valley, Catamarca Province (Fig. 1) (see MACN sheet) with no accurate stratigraphic position (Araucanense, Rovereto, 1914). One dating (⁴⁰Ar–³⁹Ar) obtained in Santa María Valley, at the boundary between Las Arcas-Chiquimil formations (under the Andalhuala Formation) yielded an age of 6.88 Ma ± 0.06 (Georgieff et al., 2014 and Spagnuolo et al., 2015). The lower third of the Andalhuala Formation yielded an age of 6.02 ± 0.04 Ma (Marshall et al., 1979), and the upper limit was dated between 4.85 ± 0.70 Ma and 3.4 ± 0.6 Ma (Strecker et al., 1989; see Georgieff et al., 2017).

Certain remains tentatively assigned to P. prisca in this study present more precise stratigraphic data (Fig. 1 and Fig. 2): •

Hualfin Valley (Villavil, Catamarca): PVL 4930, PVL 4931, and MCH-P 96 were found in El Jarillal, Member of the Chiquimil Formation. These sediments were dated with ⁴⁰Ar–³⁹Ar, and yielded an age of 9.14 ± 0.09 Ma (Sasso, 1997) for El Áspero Member (under El Jarillal Member) in Villavil locality. In the same valley, at Puerta de Corral Quemado locality, dating was performed at the limit between Andalhuala and Chiquimil formations in 7.14 ± 0.02 Ma (Esteban et al., 2014, Latorre et al., 1997 and Sasso, 1997);

P. prisca is characterized by a unique combination of characters (autapomorphies marked with an asterisk): molariforms with longitudinal furrow opposite to the hypoflexus; HSE and HPE absent, as in other Dolichotinae. *P4 anteroposteriorly shorter than M1, M1 shorter than M2 and M2 shorter than M3; posterior projection of M3 posteriorly elongated, unlike Dolichotis and “Orthomyctera” chapalmalense. *Upper masseteric fossa deeper than Dolichotis and “Orthomyctera” chapalmalense; nasolacrimal foramina closed in lateral view of the rostrum as in other Dolichotinae and unlike Caviinae. Dorsal projection of the premaxilla extending anteriorly to the level of P4; apex of mesopterygoid fossa extends up to M2 or slightly anteriorly; temporal fossa deeper than Dolichotis and “Orthomyctera” chapalmalense.

Most of the anatomical knowledge about P. prisca is based on cranial remains (Fig. 3 and Fig. 4). The skull size is similar to Dolichotis salinicola, larger than cavies and smaller than D. patagonum and “Orthomyctera” chapalmalense.

Nasal. Only the posterior most part of the nasal is preserved. This bone is relatively wider than in Cavia aperea, Microcavia australis, Galea musteloides, and D. salinicola, and laterally contacts the premaxilla. Posteriorly, the nasal-frontal suture is anteriorly concave (U-shaped) and extends anteriorly to P4 at the same level of the root of the zygomatic arch (Fig. 3A). The suture differs from “Orthomyctera” chapalmalense and Dolichotis in which it is straight, hexagonal or curved; and from cavies in which it is quite variable (V-shaped, straight, hexagonal or curved).

Premaxilla. The dorsal process of the premaxilla is preserved, lacking most of the anterior part of the rostrum (Fig. 3 and Fig. 4). In lateral view, the dorsal process does not exceed the level of P4, in contrast to M. australis, Dolichotis and “Orthomyctera” chapalmalense, in which the projection extends posteriorly to P4. The dorsal process of P. prisca is posteriorly wider than in D. salinicola, and differs from D. patagonum in which it is not posteriorly expanded. The premaxilla-maxilla suture is V-shaped in the middle of the rostrum. The masseteric fossa for the insertion of the pars profunda of the masseter muscle (Fig. 3C) is deeper than in Dolichotis, “Orthomyctera” chapalmalense, and cavies are shallower than in H. hydrochaeris. In lateral view, the dorsal contour of the masseteric fossa is formed mostly by the premaxilla, unlike Dolichotis, in which this margin coincides with the premaxillary–maxillary suture.

In ventral view, the incisive foramen is sharp as in C. aperea, G. musteloides, Dolichotis and “Orthomyctera” chapalmalense. Sixty-four percent of this foramen extends inside the premaxilla in P. prisca (Fig. 3B), differing from “O.” chapalmalense and Dolichotis, in which this proportion is 74% and 83%, respectively.

Maxilla. The posterior portion of the rostrum, the anterior part of the zygomatic arch and a portion of palate are preserved. In lateral view, the masseteric fossa is widely extended in the rostrum and it is deeper than in Dolichotis, “Orthomyctera” chapalmalense, and cavies. The nasolacrimal foramen is not laterally open in the rostrum (Fig. 3C, D) as in Dolichotis, “O.” chapalmalense, and unlike Caviinae and Hydrochoerinae, in which it is open. The ventral margin of the posterior portion of the diastema is obliquely oriented, resembling that of C. aperea, Dolichotis, and “O.” chapalmalense.

In lateral view, the dorsal root of the zygomatic arch is obliquely oriented as in C. aperea and G. musteloides, but differs from “Orthomyctera” chapalmalense and Dolichotis in which it is dorsoventrally oriented. The dorsal root of the zygomatic arch is complete (not interrupted by the lacrimal, Fig. 3C) similarly to “O.” chapalmalense, D. salinicola, and C. aperea, but differing from G. musteloides, in which the lacrimal extends up to the anterior margin of the dorsal root of the zygomatic process; in D. patagonum this character is variable, the lacrimal is exposed anteriorly to the dorsal root of the zygomatic arch, or it is not (Kraglievich, 1930). Anteriorly to the ventral root of the zygomatic arch, the alveolar protuberance of P4 is smaller than that of D. patagonum. The groove for the passage of the infraorbital nerve in the maxilla is absent, as in Dolichotis and “O.” chapalmalense, and in contrast to Caviinae, in which it is present.

The orbital portion is poorly preserved, the holotype presents only a fragment of the side wall corresponding to the alveolar protuberance of M3 (Fig. 3C). This protuberance posteriorly extends at least up to the interorbital constriction, differing from Dolichotis and “Orthomyctera” chapalmalense, in which it extends anteriorly to the interorbital constriction.

In ventral view (Fig. 3 and Fig. 4) the posterior margin of the incisive foramen is sharpened, and in the holotype a fragment of vomer can be seen in such margin of the incisive foramen. The intermaxillary suture forms a small flange as in “Orthomyctera” chapalmalense and Dolichotis and the ventral root of the zygomatic arch is at the level of P4. The maxillary portion of the palate (Fig. 3 and Fig. 4) is small as in the other Dolichotinae, and differs from cavies, in which it is more posteriorly extended.

Frontal. In dorsal view, a fragment of frontal is preserved (Fig. 3 and Fig. 4). The anterior margin of the nasal process is triangular in shape and slightly smaller than that of D. patagonum. The frontal is plane as in “Orthomyctera” chapalmalense and Dolichotis, but differs from C. aperea and G. musteloides in which the posterior portion of the frontal is convex. The interorbital constriction seems to be wider than in cavies, but the lateral margin of the frontal is broken. In lateral view, the orbital portion is concave, but it is poorly preserved in all specimens (Fig. 3 and Fig. 4).

Lacrimal. This bone is not exposed anteriorly in the dorsal root of the zygomatic arch (Fig. 3C). Ventrally, it contacts the maxilla as in D. salinicola, “Orthomyctera” chapalmalense, and cavies; in some specimens of D. patagonum, the lacrimal contacts ventrally the jugal. The orbital portion is anteriorly concave, and the lacrimal-frontal suture and lacrimal foramen are not clearly preserved (Fig. 3C).

Parietal. The most anterior and posterior portions of this bone are preserved (Fig. 3 and Fig. 4). The frontal-parietal suture is straight and in the posterior portion, the sagittal crest is well-developed and extends up to the supraoccipital. The temporal fossa is deeper than in cavies, Dolichotis, and “Orthomyctera” chapalmalense. At the level of the external auditory meatus, a temporal foramen (temporal orifice of the temporal meatus sensu Popesko et al., 1992) is immediately above the parietal-squamosal suture (Fig. 3C, D).

Palatine. A fragment of palatine is preserved in the palate. It is V-shaped and the maxilla-palatine suture is not clear, but two foramina are present in the anterior apex; the apex of the mesopterygoid fossa extends up to M2 or slightly anteriorly (Fig. 3B) as in Dolichotis and “Orthomyctera” chapalmalense, but in most Caviinae it extends posteriorly up to M3. The outline of the sphenopalatine vacuity is not clear, but it is likely to be at the level of M3, although this cannot be confirmed in the type specimen of P. prisca (Fig. 3B). In that case, it is more anterior than in C. aperea, Dolichotis and “O.” chapalmalense, in which it is at the level of the pterygoid fossa (or alar canal sensu Popesko et al., 1992). The pterygoid fossa is triangular in outline, with the anterior and lateral portion formed by the palatine, and the medial part formed by the pterygoid (Fig. 3B).

Squamosal. This bone preserves the zygomatic and auditive portions (Fig. 3C, D). The caudal process is dorsal to the tympanic bullae and it is anteroposteriorly straight, with the posterior portion slightly curved ventrally. The squamosal-supraoccipital suture extends posteriorly to the auditory meatus (Fig. 3C, D). The caudal process of the squamosal (pce) is large, similarly to D. patagonum and “Orthomyctera” chapalmalense, but differs from D. salinicola in that it is anteroposteriorly shorter; the posterior portion is more curved and expanded than in P. prisca. The pce of cavies is more curved and dorsoventrally shorter than in P. prisca. Two retroarticular foramina between the pce and the ectotympanic are present in P. prisca. The anterior foramen is a fissure in the anterior portion of the tympanic bulla, and the posterior foramen is larger and oval-shaped as in Dolichotis (Fig. 3C, D).

Ectotympanic. The tympanic bullae are ovoid (Fig. 3B–E); in ventral view, the anteromedial-posterolateral axis of the tympanic bulla is approximately 33% of the length between the premaxillary-maxillary suture and the anterior border of the foramen magnum (Fig. 3B). This proportion is similar to D. salinicola and larger than D. patagonum and “Orthomyctera” chapalmalense. The Caviinae present the biggest tympanic bullae within Caviidae, except for C. aperea, in which such proportion is similar to that of D. patagonum (see Madozzo-Jaén et al., 2018).

In lateral view (Fig. 3C, D), the epitympanic portion of the petrous or epitympanic sinus (pet) is bullet-shaped as in Dolichotis and “Orthomyctera” chapalmalense, but differs from D. patagonum in that the pet is more gracile and less convex. The external auditory meatus is at the level of the dental series as in Dolichotis and “O.” chapalmalense, but differs from M. australis in which it is below the dental series. The circular infratympanic fenestra is located anteroventrally to the mae, and it is smaller; both foramina appear to be completely separated as in M. australis, but differ from Dolichotis and Cavia, in which both foramina can be connected. The stylomastoid foramen is at the same level as the mae, posteriorly to them, and smaller (Fig. 3C, D). The mastoid portion of the petrous is exposed laterally, dorsoventrally oriented; and it is plane in all extension as in Dolichotis and “O.” chapalmalense, differing from cavines in that its ventral region is convex and exposed laterally and posteriorly.

Presphenoid and basisphenoid. These bones are completely fused; both are long and narrow, wider at their posterior end, and the basisphenoid contacts posteriorly the basioccipital (Fig. 3B). The basisphenoid is dorsoventrally flattened and has two small posterior ridges. The presphenoid and basisphenoid form the middle and posterior part of the roof of the mesopterigoid fossa.

Occipital complex. The supraoccipital has a temporal process exposed dorsally, which contacts anteriorly the parietal, and lateroventrally with the caudal process of the squamosal as in Dolichotis (Fig. 3A,C). In posterior view, the dorsal nuchal and external occipital crests are small, but the ventral nuchal crest is developed with a triangle shape (Fig. 3E). The magnum foramen is straight in dorsal margin as some specimens of D. salinicola. There is a preserved portion of the paraoccipital process that is prolonged up to the level of the inferior border of the occipital condyle (Fig. 3D–E). Ventrally, the basioccipital is posteroventrally inclined and rectangular in section with a low ridge in the back half, and it is fused with the occipital condyles.

Endocast. MACN-PV 8348 preserves the neocortex (Fig. 5A). The endocast is giroencephalic with neocortical fissuration with the primary and secondary furrows. The telencephalic hemisphere is divided into three convolutions (marginal, suprasylvian, and ectosylvian) by the interhemispheric, lateral and sylvian or suprasylvian sulci (Dozo, 1997a) as in other caviids. The endocast of P. prisca is similar to that of Dolichotis but with less developed convolutions than that of Hydrochoerus (Fig. 5A–C, Campos and Welker, 1976 and Dozo, 1997b). In P. prisca, it is rhomboid as in other Cavioidea; the ectosylvian convolution is similar to that of Dolichotis, less developed than in Neocavia and Dolicavia (Fig. 5E–F, Dozo, 1997a). The lateral sulcus extends anteriorly up to the telencephalic margin, as in Dolichotis and Hydrochoerus, but unlike Cavia, Dolicavia, and Dasyprocta, in which it does not reach the anterior margin (Fig. 5B–G). Further, it differs from Dolicavia and Hydrochoerus in that the lateral sulcus is not caudally bifurcated. The suprasylvian sulcus (somatic-auditory sulcus sensu Campos and Welker, 1976) is anterolaterally-posteromedially oriented, and located between the somatic sensorial and auditory region, as in other caviids, differing from Cuniculus and Dasyprocta in which the suprasylvian sulcus is not present (Fig. 5G and H).

P. prisca has an intervisual sulcus between the lateral sulcus and the lateral margin of the endocast (Fig. 5A); this sulcus is present also in Dolichotis, Hydrochoerus, and Cuniculus but is absent in Neocavia, Dolicavia, and Dasyprocta (Fig. 5B–C, E–H). The telencephalon of Hydrochoerus and Dolichotis is more complex than that of other Caviidae (Fig. 5B and C). The auditory-visual sulcus is located between the auditory and visual regions and the intervisual sulci (iv2–iv3 sensu Campos and Welker, 1976: 253) in the visual region, more medially than the auditory-visual sulcus. In Cavia, the auditory-visual sulcus is present (Campos and Welker, 1976); nevertheless, certain authors do not differentiate it from the ectosylvian sulcus (Hatakeyama et al., 2017).

Upper molariforms. The dental series are anteriorly convergent (Fig. 6A–C). The upper molariforms are euhypsodont, with a constriction in the apex of each lobe and with a transverse dentine crest in the middle of the occlusal surface of each lobe, as in other Caviinae and Dolichotinae. The hypoflexus is funnel-shaped with cement; the enamel is continuous around the entire crown, but it is interrupted on the labial side, except in the furrow opposite to the hypoflexus, as in Dolichotis and “Orthomyctera” chapalmalense (Fig. 6A–C). The upper molariforms have a longitudinal furrow opposed to the hypoflexus, and the other flexi (HPE and HSE) are absent, as in Dolichotis and “O.” chapalmalense, differing from Caviinae in which the HSE (Cavia, Microcavia, Palaeocavia, Neocavia, Dolicavia) and HSE and HPE (Galea) are present.

The molariforms are double heart-shaped, but are less similar between each other than in Dolichotis and “Orthomyctera” chapalmalense. The anteroposterior length of P4 is larger than M1, M1 is lower than M2 (Table 1), unlike that of C. aperea, Dolichotis, and “O.” chapalmalense, in which M1 and M2 are approximately equal. The M3 is the largest tooth; and it presents an elongated posterior projection (Fig. 6A, B) similar to that of Dolicavia and Neocavia. It differs from M. australis, Dolichotis and “O.” chapalmalense in which the ppM3 is rounded as an incipient lobe, and from C. aperea in which the ppM3 is very short. The anteroposterior length of the anterior lobe of P4 is bigger than that of the posterior lobe, and the anteroposterior lengths of the posterior lobes of M1 and M2 are bigger than those of the anterior lobe (Table 1). It differs from Dolichotis and “O.” chapalmalense in which the lobes of M1–M2 are equal in size.

Mandible. So far, only cranial material has been described for P. prisca (see below, Remarks section). However, certain mandibles have been previously assigned to this species (FMNH-P 14335 and FMNH-PM 1095, Marshall and Patterson, 1981), or to Palaeocavia sp. (FMNH-P 14336, Marshall and Patterson, 1981), or to cf. Dolichotis (MCH-P 96, Esteban et al., 2014) and to Caviidae indet. (PVL 3302, included in the PVL collection donated by Peirano in 1939; PVL 4930, PVL, 4931, Powell et al., 1998) as well as to other specimens found in the revised collections that are restudied here, and that can be tentatively assigned to P. prisca due to the morphology and size of their molariforms (Table 1, (Fig. 7 and Fig. 8). It is similar in size to that of D. salinicola and more robust than that of cavies. In lateral view, the dorsal margin of the diastema (Fig. 8A) is similar to that of M. australis and more oblique than that of Dolichotis. The mental foramen is located anteriorly to p4 at half the dorsoventral height of the lateral surface of the dentary, as in other caviids. The notch for the insertion of the tendon of the masseter medialis pars infraorbitalis muscle is located between m1–m2, as in M. australis, Dolichotis and “Orthomyctera” chapalmalense, but it differs from C. aperea and G. musteloides in which the nMpi is between P4–m1. The horizontal crest (only preserved in PVL 3302 and PVL 4931; Fig. 8B–C) is well-developed, forming a laterally projected shelf that connects to the nMpi (Fig. 9B, C). The dorsal fossa of the horizontal crest is deeper than the nMpi (Fig. 8C and 9B).

In lingual view (Fig. 8D–F, K–N), the incisor extends backwards to the level of the anterior lobe of m2, as in C. aperea, Dolichotis, and “Orthomyctera” chapalmalense. The pterygoid fossa is well-developed, and the mylohyoid crest is not present (Fig. 8E–F, L) as in Dolichotis and “O.” chapalmalense. Alveolar protuberances are present on the ventral margin (Fig. 8G).

Lower molariforms. The lobes are more meso-lingually elongated (Table 1, Fig. 9) than in Dolichotis, “Orthomyctera” chapalmalense and other species of Orthomyctera, in which the lobes are more rounded, and similar in shape. In addition, m1 is anteroposteriorly shorter than m2, differing from Dolichotis and “O.” chapalmalense, in which the anteroposterior length of m1 and m2 is similar. The anterior lobe of p4 of FMNH-P 14336, MCH-P 96 and PVL 4930 (Fig. 9E–G) presents a rounded and elongated anterior projection (app4), similarly to D. salinicola. The PVL 3180 and PVL 3302 (Fig. 9) present a well-developed app4 separated from the anterior lobe by a well-marked flexid, and differs from D. patagonum, which has an incipient lobe-shaped projection in the app4, and from C. aperea, which has no app4. The m3 is the largest molar, as in Dolichotis, “O.” chapalmalense and other species of Orthomyctera.

Remarks . P. prisca was described by Rovereto (1914) on the basis of a skull (MACN 8348) from the Araucanense of Santa María Valley. Kraglievich (1932) referred a mandibular fragment (MACN 8519) from Huayquerías de San Carlos, south of Mendoza Province to P. prisca, but this material is currently missing and such collection number corresponds to a left mandible with p2–p4, and a right mandible with p4–m3 of the Notoungulate Pachyrukhos.

In this revision, I do not agree with certain assignments of materials to P. prisca (e.g., Marshall and Patterson, 1981 and Ubilla and Rinderknecht, 2003), detailed below.

FMNH-P 15288 is a mandible fragment with m1–m3 from Puerta de Corral Quemado, Catamarca without precise stratigraphic horizon (Marshall and Patterson, 1981). The anteroposterior length of m1 is similar to m2, and m3 is longer than m1 and m2, the lobes of molars are more rounded than those of Dolichotis, “Orthomyctera” chapalmalense and other species of Orthomyctera. It differs from P. prisca, in which the molar lobes are more elongated meso-lingually and the m1 is anteroposteriorly shorter than m2. For these reasons, I suggest that FMNH-P 15288 does not correspond to P. prisca.

MACN-Pv 12335 is a skull fragment with complete molariform series, coming from sediments of Río V, San Luis Province (Ubilla and Rinderknecht, 2003). This specimen shows anatomical differences with P. prisca. In MACN-Pv 12335, the nasolacrimal foramen is open in the lateral portion of the rostrum, unlike P. prisca, which has no foramen. The dorsal projection of the premaxilla extends behind the level of P4, unlike P. prisca in which this extension is anterior to the P4; the dorsal root of the zygomatic process of the maxilla is interrupted by the lacrimal, whereas in P. prisca, the lacrimal is allocated posteriorly to the maxilla. The palate is extended up to the anterior lobe of M3, and the apex of the mesopterygoid fossa is blunt and different from P. prisca in which the mesopterygoid fossa is V-shaped and the apex is at the level of M2. The dental series is shorter than in P. prisca. Based on these morphological characters and based on the diagnostic characters of P. prisca (see above) MACN-Pv 12335 does not belong to P. prisca.

On the basis of morphological and phylogenetic studies, the validity of P. prisca, and its basal position in relation to Dolichotis are corroborated (node B, Fig. 10). P. prisca is undoubtedly a Dolichotinae because it shares two synapomorphies with this clade. The absence of the nasolacrimal foramen opening laterally in the rostrum differs from Caviinae and Hydrochoerinae, in which the nasolacrimal foramen opens laterally in the rostrum. Another character shared with Dolichotinae is the presence of a longitudinal furrow opposite to the hypoflexus, a plesiomorphic character present in certain “eocardiids” (e.g., Eocardia robertoi, Schistomys erro).

The basal position of P. prisca is justified, since it lacks certain characters present in “Orthomyctera” chapalmalense and Dolichotis. The dorsal process of the premaxilla is extended posteriorly to the level of P4 in “O.” chapalmalense and Dolichotis; this character is shared with some Caviinae (e.g., M. australis, Dolicavia) and some Hydrochoerinae (e.g., Caviodon andalhualensis, Hydrochoerus hydrochaeris). As discussed in previous papers (Quintana, 1998 and Ubilla and Rinderknecht, 2003), the presence of an incipient lobe-shaped posterior projection in M3 could be a character independently acquired in the other lineages (M. australis), and the condition of P. prisca (posteriorly extended ppM3) would be a plesiomorphic state shared with the eocardids (e.g., Eocardia robusta, Guiomys unica).

Since the late Neogene, the Dolichotinae were already diversified in South America, with records in the North of Brazil, Southeast of Uruguay, and in the North and center of Argentina (Kerber et al., 2017, Madozzo-Jaén et al., 2013, Pascual, 1966, Pérez et al., 2017, Powell et al., 1998, Rodríguez-Brizuela and Tauber, 2006, Rovereto, 1914, Ubilla and Rinderknecht, 2003 and Vucetich and Pérez, 2011). Their record occurs up to the present in Argentina (Ameghino, 1889, Kerber et al., 2011, Lema et al., 2009, Ubilla et al., 2009 and Verzi and Quintana, 2005). Although the fossil diversity of Dolichotinae is greater than the present one, the actual number of species could be overestimated. These nominal species have been described on the basis of fragmentary material, and until now, none of them had been restudied, since only a few of their original descriptions were included in the phylogenetic analysis (Madozzo-Jaén, 2017, Pérez, 2010a and Pérez and Pol, 2012). For this reason, the diversity of Dolichotinae in the fossil record could be less than currently known, but certainly greater than the current diversity. The knowledge of the fossil Dolichotinae and the alpha taxonomy is crucial for the understanding of the evolutionary history and significance of the acquisition of morphological characters of this group.

As mentioned above, molecular studies estimated the divergence time of the split of Hydrochoerinae and Dolichotinae between 15.5–11.8 Ma (Opazo, 2005 and Upham and Patterson, 2015). Phylogenetic analyses based on molecular and anatomical characters of Cavioidea estimated a ghost lineage in the early evolution of Dolichotinae of approximately 9 Ma, from the middle Miocene (11.8–13.5 Ma) to the middle Pliocene (3.7 Ma; Madozzo-Jaén and Pérez, 2017, Pérez and Pol, 2012 and Vucetich and Pérez, 2011). However, due to the presence of remains (currently under study) of Dolichotinae in the middle Miocene of Colombia (Fields, 1957 and Walton, 1997), an earlier origin cannot be discarded. The revision of the alpha taxonomy of P. prisca and its incorporation in a phylogenetic analysis help completing a portion of the lineage of Dolichotinae, and sheds light on the evolution of living maras (Fig. 10B, Madozzo-Jaén and Pérez, 2017 and Pérez and Pol, 2012). The presence of P. prisca in the late Miocene-early Pliocene is based on records of the holotype in Santa María Valley (6–3.5 Ma) and referred cranial materials, and not from the ages assigned to the carrier level of the mandibular remains found in nearby intermontane valleys (5 to 10 Ma), because their assignment to P. prisca is still preliminary, and the diversity of maras of those times is currently under revision.

In rodents, certain studies of the endocast were performed on extant species (e.g., Campos and Welker, 1976 and Pilleri et al., 1984 and references therein), but up to date, only a few studies on fossil forms of caviomorphs have been accomplished (Dozo, 1997a, Dozo, 1997b and Dozo et al., 2004), and the existent studies focused mainly on basal forms of the Order Rodentia (Bertrand and Silcox, 2016, Bertrand et al., 2016 and Bertrand et al., 2018).

The patterns of the neuroanatomical sulcus in rodents are broad; from neocortex smooth with no sulcus or convolutions (lissencephalic) to fold into numerous convolutions (giroencephalic). Based on quantitative data, Pilleri et al. (1984) proposed a correlation between the development of the neocortical sulcus and body size. From descriptions and comparisons among the giroencephalic endocast of certain fossil and living members of Cavioidea, it can be observed that the Caviidae present well-defined body size patterns, which correlate with the complexity of the neocortical sulcus patterns. However, when such patterns are analyzed at the level of Cavioidea, exceptions can be found (Table 2).

Body and brain size of Dolichotinae are intermediate among Caviidae, as is their complexity in the neocortical sulcus patterns (Table 2). I estimate that P. prisca is smaller than D. salinicola, which is smaller than D. patagonum. In P. prisca, the pattern of development of the neocortical sulcus is less complex than that of Dolichotis; in turn, the pattern of the latter is less developed than that of Hydrochoerus, which presents the most developed neocortical sulcus and the largest body within Caviidae (Table 2).

Among Caviidae, cavies have the smallest body size and simplest neocortical sulcus pattern. Cavia and Dolicavia have a similar body size (Dozo, 1997a and Dozo, 1997b) but the endocast of Cavia exhibits a greater number of neocortical sulcus than Dolicavia, although it is less defined than the latter (Table 2). For this reason, certain authors have classified the brain of Cavia as lissencephalic (Musa et al., 2016 and Silva et al., 2016). Quantitative studies suggest that current mammals whose brain mass is less than five grams are lissencephalic (Macrini et al., 2007); however, Cavia has a brain mass of 4.5 grams and a girencephalic telencephalon (Table 2; Bertrand and Silcox, 2016).

When analyzing the variability within Cavioidea, medium body size has a complex relation with the fissurization pattern of the telencephalon (Table 2). I estimated that the body size of Dasyprocta is slightly larger than P. prisca, and Cuniculus and Dolichotis are similar to each other and larger than Dasyprocta and P. prisca. However, the development of neocortical sulcus in Dasyprocta and Cuniculus is much smaller than P. prisca and Dolichotis. Descriptive data of the telencephalon of Caviidae are here reported. However, future studies should be carried out in order to explore the potentially large amount of unexplored phylogenetic data (Macrini et al., 2007).