Sparnotheriodontidae Anisolambdidae, , Soria, 2001

Püschel, Hans P., Shelley, Sarah L., Williamson, Thomas E., Perini, Fernando A., Wible, John R. & Brusatte, Stephen L., 2024, A new dentition-based phylogeny of Litopterna (Mammalia: Placentalia) and ‘ archaic’ South American ungulates, Zoological Journal of the Linnean Society 202 (1), pp. 1-50 : 13-15

Sparnotheriodontidae ( Figs 4L, 5L)

Sparnotheriodontidae currently has five accepted genera, all with a fossil record exclusively in the Palaeogene ( Fig. 2; Supporting information, Table S1 ). The family Sparnotheriodontidae was initially proposed by Soria (1980a) to include Sparnotheriodon epsilonoides Soria, 1980a from Cañadón Vaca Member, Sarmiento Formation, Argentina, which was classified as incertae sedis within the order Notoungulata . The family was later removed from Notoungulata and incorporated into the order Litopterna , being expanded by Soria (1980b) adding the genera Victorlemoinea and Phoradiadus Simpson et al. 1962 . Victorlemoinea was previously considered an early macraucheniid (Simpson 1945, 1948), and Phoradiadus was previously considered with doubts as a proterotheriid (Simpson et al. 1962). Soria’s (1980b) proposal was based on shared similarities in the upper and lower molars, such as the presence of a lophoid metaconule and a closed trigon basin in M2. Phoradiadus includes one species, Phoradiadus divortiensis , from Divisadero Largo Formation, Mendoza, Argentina (age uncertain; Cerdeño et al. 2008, López 2010, Woodburne et al. 2014a). An interesting feature of the teeth of the Sparnotheriodontidae is that they possess vertical Hunter–Schreger bands, a feature unknown in any other litoptern ( Bond et al. 2006).

Simpson (1948) tentatively accepted three Patagonian species of the genus Victorlemoinea : Victorlemoinea labyrinthica Ameghino, 1901 , Victorlemoinea emarginata Ameghino, 1901 , and Victorlemoinea longidens Ameghino, 1901 , as they were represented only by isolated teeth, the first two being represented by upper teeth, and the last by lower teeth. This means that V. longidens could correspond to teeth from the lower dentition of either V. labyrinthica or V. emarginata , in particular the mandibular fragment with m1–m2 of the holotype (MACN A-10670), as the allegedly associated premolars of this specimen seem to correspond to an isotemnid notoungulate ( Bond et al. 2006). In the genus Victorlemoinea, Bond et al. (2006) included only V. labyrinthica in their list of formally recognized sparnotheriodontid species, and more recently some authors have mentioned only V. labyrinthica for the Riochican and Vacan Patagonian faunas, omitting V. emarginata and V. longidens (e.g., Reguero et al. 2014, Gelfo 2016, Gelfo et al. 2019), which can be interpreted as implicitly synonymizing them with V. labyrinthica . In addition, there is one Victorlemoinea species from Itaboraí, Brazil, Victorlemoinea prototypica Paula Couto, 1952 , also known mostly from dental material.

Cifelli (1983b) indirectly assigned some tarsals (i.e., calcaneum and astragalus) to V. prototypica , based on their relative size and abundance. These tarsals were large and scarce, so considering linear regressions with the dentition (m2 area), he argued it could only correspond to either the sparnotheriodontid V. prototypica or the didolodontid Lamegoia conodonta . Considering the similarities of the tarsals assigned to V. prototypica and others indirectly assigned to didolodontids, such as presenting a medial malleolar facet of the astragalus extending onto the neck and the presence of a dorsal beak on the distal end of the calcaneum,Cifelli (1983a) grouped Sparnotheriodontidae and Didolodontidae under the superfamily Didolodontoidea in the order Condylarthra ( Table 1 ). This hypothesis found some phylogenetic support ( Cifelli 1993, Bergqvist 1996; Fig. 1C), and has been followed by some authors (e.g., Bergqvist 2008). However, other authors have questioned the indirect anatomical assignment of these tarsals to V. prototypica (e.g., Hoffstetter and Soria 1986, Soria 2001, Lorente 2015), and even suggested an affinity of these tarsals with Notoungulata (Soria 2001) or with Astrapotheria ( Lorente 2015) .

Among the different litoptern families, Soria (2001) considered that sparnotheriodontids were more closely related to the family Anisolambdidae due to overall dental similarities between both families, and also considering an isolated M1 (MNRJ 1479- V) from Itaboraí showing an intermediate anatomy between sparnotheriodontids and anisolambdids. Therefore, he created the suborder Eolitopterna to include both families ( Table 1 ). In addition, Soria (2001) tentatively added Heteroglyphis dewoletzky Roth, 1899 to Sparnotheriodontidae without any detailed anatomical justification; this species is known from a broken upper molar probably from Cerro del Humo, Argentina. This taxon was previously referred to Proterotheriidae by Simpson (1948).

More recently, one genus and two new species from the Eocene La Meseta and Submeseta formations, Antarctica, have been added to the family Sparnotheriodontidae , Notiolofos arquinotiensis Bond et al. 2006 and Notiolofos regueroi Gelfo, López & Santillana, 2017 . They are differentiated mainly by their size and represented by isolated teeth (the latter species only by a m3; Gelfo et al. 2019). Notiolofos arquinotiensis is particularly interesting as it is the most abundant land mammal in the Palaeogene of the La Meseta and Submeseta formations, presenting a continuous fossil record of at least 10 Mya [44–34 Mya ( Douglas et al. 2014, Amenábar et al. 2020), but see Gelfo (2016) for a different age interpretation], which has been interpreted as a case of morphological stasis (Gelfo 2016). The presence of the sparnotheriodontid Notiolofos in Antarctica raises important biogeographic questions, as no other litoptern lineage seems to have crossed to Antarctica, and among SANUs excluding Notiolofos , only the astrapothere Antarctodon sobrali ( Bond et al. 2011) has been found thus far on this continent. However, as the preservation of mammals is allochthonous in the marine sediments of La Meseta and Submeseta formations a sampling issue cannot be disregarded for explaining the absence of other SANUs ( Gelfo et al. 2019).