A new dentition-based phylogeny of Litopterna (Mammalia: Placentalia) and ‘ archaic’ South American ungulates Author Püschel, Hans P. Millenium Nucleus Early Evolutionary Transitions of Mammals (EVOTEM), Red Palentológica U-Chile, Santiago, 7800003, Chile & School of GeoSciences, University of Edinburgh, Grant Institute, James Hutton Road, Edinburgh, EH 9 3 FE, United Kingdom hpuschelr@gmail.com Author Shelley, Sarah L. School of GeoSciences, University of Edinburgh, Grant Institute, James Hutton Road, Edinburgh, EH 9 3 FE, United Kingdom Author Williamson, Thomas E. New Mexico Museum of Natural History and Science, Albuquerque, NM 87104, United States Author Perini, Fernando A. Instituto de Ciência Biológicas, UFMG, BH 31270 - 901, Brazil Author Wible, John R. Section of Mammals, Carnegie Museum of Natural History, 5800 Baum Boulevard, Pittsburgh, PA 15206, United States Author Brusatte, Stephen L. School of GeoSciences, University of Edinburgh, Grant Institute, James Hutton Road, Edinburgh, EH 9 3 FE, United Kingdom text Zoological Journal of the Linnean Society 2024 2024-09-07 202 1 1 50 http://dx.doi.org/10.1093/zoolinnean/zlae095 journal article 10.1093/zoolinnean/zlae095 0024-4082 14341218 Protolipternidae ( Figs 8H–J , 9F–H ) Protolipternidae currently includes three genera [or two if we excluded Asmithwoodwardia as Gelfo (2006) suggested] with an exclusively Palaeogene distribution ( Fig. 2 ; Supporting information, Table S1 ). This family was proposed by Cifelli (1983a) to classify specimens from Itaboraí (Itaboraian SALMA) with bunodont didolodontid-like dentition and litoptern-like tarsals. The tarsals show specializations that among SANUs are only present in litopterns, which include a spool-shaped astragalar body and a posterior astragalocalcaneal facet that allows rapid flexion of the crurotarsal joint (Cifelli 1983a). The taxonomic changes generated from this familial proposal included the consideration of Victorlemoinea prototypica [a former litoptern (Paula Couto 1952)] as a ‘condylarth’ closely related to didolodontids, and Miguelsoria parayirunhor [considered a didolodontid Ernestokokenia parayirunhor by Paula Couto (1952)] with Asmithwoodwardia scotti [considered a didolodontid by Paula Couto (1978)] as litopterns. Cifelli (1983a) grouped the last two, plus a new species Protolipterna ellipsodontoides Cifelli 1983a , within this family. However, the tarsals referred to Protolipterna ellipsodontoides and Miguelsoria parayirunhor were not found in association with their alleged dentition, being assigned indirectly by their relative size, abundance, and expected morphology (Cifelli 1983b). Some authors have accepted this interpretation (e.g., Cifelli 1985 , 1993 , Muizon and Cifelli 2000, Bergqvist 2008 ), although others have questioned it ( Hoffstetter and Soria 1986 , Soria 2001, Gelfo 2010 , Gelfo and Sigé 2011 , Gelfo and Lorente 2012 , Lorente 2015 ). For example, Soria (2001) suggested that the astragali and calcanea attributed by Cifelli (1983b) to didolodontids were most likely notoungulate elements. Indeed, a study that indirectly assigned isolated petrosals to the protolipternid Miguelsoria from Itaboraí failed to recover a monophyletic Litopterna when younger and better known litopterns (e.g., Macrauchenia and Diadiaphorus ) were included ( Billet et al. 2015 ). Considering that recent association models do not discard the association of litoptern-like tarsals to protolipternids ( Lorente 2015 ) made by Cifelli (1983b), some authors have started to consider the family Protolipternidae as a junior synonym of Didolodontidae ( Gelfo et al. 2016 , 2020 a, Croft and López 2020 ). In addition, based on anatomical observations and a phylogenetic analysis, Gelfo (2006) considered that Asmithwoodwardia scotti should be generically renamed as is it not recovered as a sister taxon of the Patagonian Asmithwoodwardia subtrigona Ameghino, 1901 . Apartfromearlyphylogeneticstudiesthatfoundprotolipternids closely related to other litopterns based on Cifelli’s (1983b) tarsal associations ( Cifelli 1993 , Bergqvist 1996 ; Fig. 1C ), other phylogenetic studies including protolipternids and didolodontids have usually failed to find them as distinctly separate groups [e.g., Muizon and Cifelli (2000); see also Didolodontidae section]. Other phylogenetic studies with a wide taxon sampling have found Protolipternidae : in a basal position in Laurasiatheria as a sister group of Didolodontidae (O’Leary et al . 2013) ; closely related to other more derived litopterns than to other SANUs (Muizon et al . 2015); among didolodontids at the stem of litopterns but not forming a monophyletic group ( Kramarz et al . 2021 ); and forming a monophyletic group as part of a polytomy that includes didolodontids, sparnotheriodontids, indaleciids, and North American phenacodontids (Zimicz et al . 2022). However, so far there has not been any phylogenetic analysis that includes protolipternids alongside representatives of all litoptern families, didolodontids, and the different SANU orders ( Table 2 ). If we follow Cifelli (1983a) and consider Protolipternidae as a valid family, its earliest members are Protolipterna ellipsodontoides , Miguelsoriaparayirunhor ,and Asmithwoodwardia scotti from Itaboraí, Brazil , and the youngest member would be Asmithwoodwardia subtrigona with its last records in Cañadón Vaca, Argentina ( Gelfo 2006 ), which gives the family a temporal interval of 56–42 Mya ( Fig. 2B ; Supporting information, Table S1 ).