Glossotherium robustum ( Owen, 1842 )

Glossotherium robustum ( Owen, 1842)

Referred material: Cranium, dentary, 16 vertebrae (cervical, thoracic, and caudal), 25 postcranial elements (including ribs and phalanges), and four plates with osteoderms: PIMUZ A/ V 484 ( Figs. 3 View Fig , 10 View Fig , 11 View Fig ); cranium: PIMUZ A/ V 485; right upper fourth molariform: PIMUZ A/ V 486; right upper molariform: PIMUZ A/ V 487; lower third molariform: PIMUZ A/ V 488; left third lower molariform: PIMUZ A/ V 497; right first lower molariform: PIMUZ A/ V 498; right third lower molariform: PIMUZ A/ V 499; left fragmented tibia: PIMUZ A/ V 501; eight bones of the foot: PIMUZ A/ V 502 ( Fig. 7 View Fig ); incomplete right mandible: PIMUZ A/ V 4143; fragmented cranium in about 50 pieces: PIMUZ A/ V 4144.

Comment: Te Mylodontinae have been the subject of nomenclatural confusion for a long time, particularly with respect to Glossotherium and Mylodon (see De Iuliis et al., 2017; McAfee, 2009). Many recent studies have attempted to improve the diagnoses of multiple mylodont species (e.g., Boscaini et al., 2020b; Brambilla & Ibarra, 2018a; McAfee, 2009, 2016; Pitana et al., 2013). Initial diagnoses to distinguish mylodont species were based primarily on cranial characters including teeth (Brambilla & Ibarra, 2018a; McAfee, 2009). For Glossotherium , at least four species are recognized in South America, but only Gl. robustum is known from the Pleistocene of the Pampean formation associated with the presence of Glossotherium sp. ( Pitana et al., 2013). In the Roth collection at PIMUZ, four specimens have cranial remains. PIMUZ A/V 484 and PIMUZ A/V 485 are by far the most complete specimens. Te former shows a dental formula of 5/5 allowing to distinguish it from this specimen from M. darwinii ( McAfee, 2009) . In cross section, the upper caniniform is triangular, the Mf1 and Mf2 have a similar mesiodistal length, and the Mf4 is bilobate with a posterior lobe mediolaterally narrower than the anterior lobe. Unfortunately, the upper dentition of PIMUZ A/V 485 is not preserved. However, both specimens possess diagnostic cranial characters allowing to reject an assignment to M. darwinii and support an attribution to Gl. robustum ( McAfee, 2009) , including a short palatal length posterior to Mf4, a dome-shaped cranium with its maximum high at the level of the postorbital processes, narrow snout in front of the anterior margin of the orbit, and width of the nasal cavity greater than its height. PIMUZ A/V 484 and PIMUZ A/V 485 differ from Archaeomylodon sampedrinensis Brambilla & Ibarra, 2018 a, in having a less subcircular occiput and the absence of a prominent diastema between the Cf1 and Mf1 (Brambilla & Ibarra, 2018a). Specifically, the snouts of these specimens show the absence of a nasal elevation and well-marked fossa for the muscle buccinator, both features supporting their assignment to Gl. robustum rather than to A. sampedrinensis (Brambilla & Ibarra, 2018a). I therefore propose to reassign these two specimens to Gl. robustum . PIMUZ A/V 4144 is far too fragmentary to ensure a clear identification. Te little identifiable material is not different from the two most complete specimens and suggests a reassignment to Gl. robustum rather than M. darwinii . Te mandible of A. sampedrinensis is still unknown, and that of M. darwinii has the particularity of having a predental spout longer than the length of the tooth row. PIMUZ A/V 4143 shows a relative length of the tooth row longer than the predental spout, as in the mandible of PIMUZ A/V 484, which leads me to favor a reassignment to Gl. robustum . On the basis of the skull of PIMUZ A/V 484, I have evaluated the identification of the specimens represented only by isolated teeth. PIMUZ A/V 486 has a general shape closer to Mf1 than to Mf4, despite its initial determination (see Additional file 1: Table S1), but two sulci are visible on either side of the most distal part of the tooth, suggesting that the tooth is an Mf4. PIMUZ A/V 486 was probably a young specimen, implying that tooth lobation occurred progressively during tooth eruption and the early stages of tooth wear. Tis is a hypothesis to be explored that I suggest for both sloths and glyptodonts, two mammalian groups with hypso/hypselodonty and lobation (Bargo et al., 2006b; Vizcaíno et al., 2011). Because of its relatively young ontogenetic stage, a clear determination for PIMUZ A/V 486 is difficult to establish. However, the specimen perfectly matches the morphology of the Mf4 of a young Gl. robustum ( Pitana et al., 2013) . Terefore, I propose to assign PIMUZ A/V 486 to Gl. robustum while noting that this determination remains fragile. In contrast, PIMUZ A/V 487 exhibits an Mf4 bilobate with a posterior lobe narrower than the anterior lobe, suggesting an assignment to a small Gl. robustum . For the lower teeth, PIMUZ A/V 488 was identified as an mf3 but it does not show bilobation. It is therefore more likely that this tooth corresponds to an mf2. Te relatively subtriangular shape of the tooth suggests an attribution to Gl. robustum , as for PIMUZ A/V 497 and PIMUZ A/V 499, but these reassignments need to be considered with caution. PIMUZ A/V 498 corresponds to a well-developed caniniform compatible with Gl. robustum (see Bargo & Vizcaíno, 2008). Regarding the postcranial material, the recent work of McAfee (2016) allowed further characterizations of M. darwinii and Gl. robustum , but no postcranial material was described for A. sampedrinensis . Unfortunately, the description of the foot and pelvic girdle is not included in the diagnoses of M. darwinii and Gl. robustum (Cartelle et al., 2019; McAfee, 2009, 2016). PIMUZ A/V 502 corresponds to eight bones of the foot including the calcaneum, cuboid, metatarsus IV, and metatarsus V. Te portion of the foot preserved for this specimen matches with the full foot drawing proposed by Schulthess (1920) for Mylodon robustus Owen, 1842 , a synonym of Gl. robustum (see McAfee, 2009), and the shape of the calcaneum is relatively close to that of other Glossotherium species (e.g., Cartelle et al., 2019). Finally, PIMUZ A/V 501 shows a distal articular portion closer to Gl. robustum (see Cartelle et al., 2019) than to M. darwinii , as the medial portion of the astragalar articulation (= odontoid process facet in Boscaini et al., 2021) is relatively deeper, but mainly because this portion is well separated from the lateral portion for the articulation with the distal fibula (= distal fibular facet in Boscaini et al., 2021) according to the diagnosis of M. darwinii provided by McAfee (2016) (see also Roth, 1899). In the Roth collection at PIMUZ, the skeleton of Gl. robustum is relatively well represented. A particular feature of the collection is the presence of four association of osteoderms belonging to PIMUZ A/V 484, providing an opportunity to explore osteoderm development in xenarthrans (e.g., McDonald, 2018; Toledo et al., 2021).