Zoological Journal of the Linnean Society, 2026, 206(3), zlag021 https://doi.org/10.1093/zoolinnean/zlag021 The locomotor behaviour of subfossil Malagasy sloth-lemurs (Strepsirrhini: Indriidae) and koala-lemurs (Strepsirrhini: Megaladapidae): new insights from limb trabecular bone Fabio Alfieri1,2,3,4,*,  , Julia Arias-Martorell5,  , Carla Argilés-Esturgó5,6,  , Damiano Marchi7,8,  1Institute of Ecology and Evolution, Universität Bern, Bern, 3012, Switzerland 2Department of Earth Sciences, University of Cambridge, Cambridge, CB2 3EQ, UK 3Institut für Biologie, Humboldt-Universität zu Berlin, Berlin, 10099, Germany 4Museum für Naturkunde, Leibniz-Institut für Evolutions- und Biodiversitätsforschung, Berlin, 10115, Germany 5Institut Català de Paleontologia Miquel Crusafont (ICP-CERCA), Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Barcelona, 08193, Spain 6Departament de Geologia, Universitat Autònoma de Barcelona, Barcelona, 08193, Spain 7Department of Biology, Università di Pisa, Pisa, 56126, Italy 8Centre for the Exploration of the Deep Human Journey, University of the Witwatersrand, Johannesburg, 2050, South Africa *Corresponding author. Institute of Ecology and Evolution, Universität Bern, Baltzerstrasse 6, Bern 3012, Switzerland. E-mail: fabio.alfieri@unibe.ch or fabio_alfieri@yahoo.it ABSTRACT The locomotion of Malagasy Quaternary subfossil lemurs, including palaeopropithecines (‘sloth-lemurs’) and megaladapids (‘koala-lemurs’), has been investigated on abundant postcranial remains. Proposed strategies include some that lack living primate parallels, such as sloth-like suspensory arboreality in palaeopropithecines, although the degree of suspensory behaviour in palaeopropithecines or locomotor diversity in koala-lemurs is poorly understood. Unlike the external morphology, internal bone structure in these taxa is largely unexplored. We compared the humeral and femoral trabecular architecture of sloth- and koala-lemurs with several extant mammals, studying spherical/hemispherical trabecular samples extracted from high-resolution scans. After defining locomotor categories from quantitative data, we tested links between trabecular parameters and locomotor modes through exploratory and multivariate analyses, accounting for body size and phylogeny. In extant mammals, only femoral trabecular traits, particularly the degree of anisotropy and bone volume fraction, were significantly associated with locomotion, distinguishing suspensory and bridging arboreal taxa from others. Using this model, we inferred suspensory adaptations in palaeopropithecines, especially Palaeopropithecus, confirming earlier reconstructions, but also in Megaladapis edwardsi, a striking result that would place it alongside extant orang- utans as the largest mammals known to adopt such habits. This work highlights the potential of internal bone structure for reconstructing primate locomotor evolution. Keywords: arboreal suspensory locomotion; cancellous bone; femur; humerus; Madagascar; Megaladapis; palaeopropithecines I N T RO D U CT I O N Studying the fauna of Madagascar offers a unique opportunity to address central questions in biology, owing to the exceptional biodiversity of the island and its distinctive geological and evolu- tionary history (e.g. Muldoon et al. 2009, Crowley 2010, Muldoon 2010). Lemurs, among the most iconic Malagasy organisms, are taxonomically diverse (representing ∼13% of living primates; Rowe 1996, Martin 2000) and ecologically varied (Martin 2000), particularly in their locomotor/postural behaviours. Among others, these include climbing and leaping, quadrupedalism (arboreal and terrestrial), vertical clinging, and terrestrial hopping (Walker 1974, 1979, Gebo 1987, Demes et al. 1996, Jungers et al. 2005, Furnell et al. 2015). When humans first reached Madagascar (Muldoon 2010, see also Douglass et al. 2019; ∼2300 years ago), they encountered not only the modern primate community but also additional extinct lemur groups, i.e. palaeopropithecines (‘sloth-lemurs’), archaeole- murines (‘monkey-lemurs’), megaladapids (‘koala-lemurs’), the Original Article Received 30 September 2025; revised 3 January 2026; accepted 12 January 2026 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. © The Author(s) 2026. Published by Oxford University Press on behalf of The Linnean Society of London. https://doi.org/10.1093/zoolinnean/zlag021 https://orcid.org/0000-0001-8241-7309 https://orcid.org/0000-0001-8110-2946 https://orcid.org/0009-0002-8279-8839 https://orcid.org/0000-0002-6331-8783 mailto:fabio.alfieri@unibe.ch mailto:fabio_alfieri@yahoo.it https://creativecommons.org/licenses/by/4.0/ 2  •  Alfieri et al. daubentoniid Daubentonia robusta (Lamberton, 1934), and the lemurid Pachylemur Filhol, 1874. These species persisted for cen- turies alongside humans, with some surviving until ∼500 years ago (Crowley 2010). Their recent extinction led to the preserva- tion of extensive skeletal material, which has been pivotal in recon- structing their biology, including body mass (e.g. Jungers et al. 2002, 2008), phylogenetic affinity (e.g. Yoder et al. 1999, Karanth et al. 2005, Kistler et al. 2015, Herrera and Dávalos 2016), and locomotor/positional behaviours (e.g. Godfrey et al. 1995, Jungers et  al. 1997, Shapiro et  al. 2005, see also below). Locomotor behaviours in particular were remarkably diverse, exhibiting a range of adaptations comparable to all the other primate clades (Walker 1974, Tattersall 1982, Godfrey et al. 1997, Jungers et al. 2002, 2005, Granatosky 2022). Notably, within subfossil lemurs, sloth-lemurs and koala-lemurs share inferred ecological and physiological traits, i.e. slow arbore- ality, low metabolic rates, and dominance of rest in daily activity (Walker et al. 2008, Hogg et al. 2015, Godfrey et al. 2016, Alfieri et al. 2021, 2023). For instance, the radius of curvature of their semicircular canals (a proxy for locomotor agility, with the canals of cautious taxa shown to be smaller than those of agile taxa after accounting for body mass; Spoor et al. 2007) indicates that they were the least agile among subfossil lemurs (Walker et al. 2008). Such ecological traits, shared with tree-sloths and koalas, might be associated with convergent morphologies, such as low cortical bone compactness in limb diaphyses (Alfieri et al. 2021). How- ever, other anatomical features suggest biomechanical divergence between sloth-lemurs and koala-lemurs, despite their broad eco- logical similarities (Alfieri 2022, Alfieri et al. 2023). Sloth lemurs [Palaeopropithecus Grandidier, 1899, Mesopro- pithecus Standing, 1908, Babakotia Godfrey, Simons, Chatrath & Rakotosamimanana, 1990, and Archaeoindris Standing, 1908 (Simons et al. 1992, Jungers et al. 1997, Godfrey and Jungers 2003, Gommery et al. 2009)] ranged from ∼11 to ∼160 kg in body mass (Granatosky 2022), a range revised downwards to between ∼7 and ∼128 kg by more recent estimates (Thompson et al. 2025). Extant indriids are their closest living relatives (Tattersall 1973, Tattersall and Schwartz 1974, Karanth et al. 2005, Kistler et al. 2015, Herrera and Dávalos 2016). Although the locomotion of Archaeoindris remains debated owing to limited fossil evidence (e.g. Godfrey and Jungers 2002, Godfrey et al. 2016), abundant studies of postcranial axial and appendicular skeletal traits have suggested that Palaeopropithecus, Mesopropithecus, and Babakotia probably showed specialized suspensory adaptations based on below-branch quadrupedalism. This specialization was potentially more extreme in Palaeopropithecus, followed by Babakotia, which would have been intermediate, and finally by the least-specialized Mesopropithecus ( Jungers 1980, Jungers et al. 1991, Shapiro et al. 1994, 2005, Godfrey et al. 1995, Wunderlich et al. 1996, Hamrick et al. 2000, Godfrey and Jungers 2003, Patel et al. 2013, Marchi et al. 2016, Granatosky 2018, 2020, 2022). Although there seems to be a certain consensus on the more extreme adaptations of Palaeopropithecus, the relative degree of specialization in suspen- sory behaviours of the other two taxa appears less clear. For instance, it was also proposed that Babakotia and Mesopropithecus (the smaller palaeopropithecines) were more similar to the two-toed sloth Choloepus Illiger, 1811, contrasting with Palaeo- propithecus, which was instead inferred to resemble more the three-toed sloth Bradypus Linnaeus, 1758 (e.g. Marchi et al. 2016). Also, Babakotia and Mesopropithecus are not distinct for some potentially functional traits (i.e. within-limb proportions), hence they do not clearly reflect the expected different degree of suspen- sory adaptations mentioned above (Marchi et al. 2016). Accord- ingly, Babakotia and Mesopropithecus have been described overall as taxa with a more diverse locomotor repertoire, in comparison to Palaeopropithecus (Granatosky 2022). Koala-lemurs, i.e. Megaladapis Major, 1893 spp. (Walker 1967, Tattersall 1975, Jungers 1977, 1980, Szalay and Delson 1979, Gebo 1986), include three species, with body mass ranging from the ∼45 kg of Megaladapis madagascariensis Forsyth Major, 1894 to the ∼85 kg of Megaladapis edwardsi Grandidier, 1899 and with the medium-sized Megaladapis grandidieri Standing, 1903 (∼75 kg) in between (Walker 1967, Tattersall 1975, Jungers 1978, 1977, Jungers et  al. 2008). More recent estimates have revised these values downwards to ∼31.5, ∼65, and ∼56 kg, respectively (Thompson et al. 2025), thereby confirming the variation from the smallest Megaladapis madagascariensis to the largest Megalada- pis edwardsi, with Megaladapis grandidieri in between. The phylo- genetic position of Megaladapis spp. has been controversial, i.e. close to Lepilemur Geoffroy Saint-Hilaire, 1851 (Tattersall and Schwartz 1974, Godfrey et al. 2010), to Lemuridae (Karanth et al. 2005, Orlando et al. 2008, Kistler et al. 2015), or diverging early from all non-daubentoniid lemurs (Herrera and Dávalos 2016). Here, we will consider koala-lemurs as the sister taxon of Lemu- ridae, because this has been confirmed by more recent studies addressing nuclear DNA sequences (Upham et al. 2019, Marciniak et al. 2021). The locomotor behaviour of Megaladapis spp. has also been long debated (Godfrey and Jungers 2002). Although there is now wide consensus on their arboreality, with behaviours dom- inated by forelimb activity (but without being characterized by extreme suspensory adaptations, such as in palaeopropithecines, see above), which is supported by a high intermembral index, powerful hands and feet, and body proportions reminiscent of koalas ( Jungers 1978, Wunderlich et al. 1996, Godfrey and Jungers 2002, Godfrey et al. 2016), some uncertainties remain regarding their specific postural and locomotor behaviours. A first traditional view holds that vertical climbing and clinging were the primary postures for all the Megaladapis taxa ( Jungers 1977, 1978; agree- ing with Carleton 1936). In this framework, the three species are considered merely as size and geographical variants, i.e. the largest Megaladapis edwardsi and the smallest Megaladapis madagascar- iensis representing the south-southwestern varieties, with the medium-sized Megaladapis grandidieri living in central Madagascar (Walker 1967, Tattersall 1975, Jungers 1977, 1978, Jungers et al. 2008). However, new fossil evidence has led to the recognition of two potential behaviourally distinct sublineages of koala-lemurs: the small to medium-sized and widely distributed Megaladapis (Megaladapis madagascariensis + Megaladapis grandidieri), char- acterized by vertical climbing combined with quadrupedal hang- ing and pedal suspension (Wunderlich et al. 1994, 1996, Jungers et al. 2002, 2008), and the large-sized and southern Peloriadapis Grandidier, 1899 (subgenus corresponding to Megaladapis edwardsi; Vuillame-Randriamanantena et  al. 1992, Wunderlich et al. 1996), proposed as relatively more terrestrial compared with Megaladapis madagascariensis and Megaladapis grandidieri (Wun- derlich et al. 1994, 1996, Jungers et al. 2002, 2008). Most of the reconstructions of locomotion in subfossil lemurs have focused on bone external morphology (e.g. Godfrey et al. Trabecular bone of subfossil lemurs    •  3 1995, Wunderlich et al. 1996, Marchi et al. 2016; although Marchi et al. also analysed diaphyseal cross-sectional properties). How- ever, outer morphology, especially concerning joint anatomy, is potentially more genetically (thus, less functionally) driven than internal bone structure, i.e. the inner repartition of osseous tissue within skeletal elements (Rafferty and Ruff 1994, Lieberman et al. 2001). This is especially relevant for the long bones of the appen- dicular skeleton, which are well known for their ability to inform on locomotor behaviour (Dunn 2018). The external morphology of long bones mainly reflects the potential locomotor capabilities of an organism: what it could do theoretically, e.g. the maximum range of rotation allowed at a joint without causing disarticulation or injury. In contrast, the internal structure of long bones more closely reflects the realized niche, i.e. the behaviours the organism performs regularly. This applies to all levels of inner architecture, such as diaphyseal compact bone and epiphyseal trabecular struc- ture. Particularly the latter has been shown (both experimentally, e.g. Biewener et al. 1996, Pontzer et al. 2006; and computationally, e.g. Huiskes et al. 2000, Keaveny et al. 2001) to adapt in response to biomechanical loadings (summarized by Kivell 2016). This characteristic highlights the potential of the trabecular structure of limb long bones to reflect the loading conditions to which they are subjected, which can, in turn, be linked to specific locomotor behaviours. Some studies have not found trabecular structural differences reflecting different locomotor habits (e.g. Carlson et al. 2008, Ryan and Walker 2010), which was explained by the fact that trabecular bone is also affected by non-functional aspects [e.g. systemic (Tsegai et al. 2018a), genetic (Ryan et al. 2017), ontoge- netic (Macho et al. 2005), and hormonal (Khosla et al. 2006)]. However, many comparative analyses across several postcranial regions have shown that trabecular bone co-varies with locomotor behaviour (e.g. Fajardo and Müller 2001, Chang et al. 2008, Sapa- rin et al. 2011, Ryan and Shaw 2012, Tsegai et al. 2013, Mielke et al. 2018, Georgiou et  al. 2019, Arias-Martorell et  al. 2021, Alfieri 2022, Alfieri et al. 2022), making the study of this anatomical level a common approach for inferring past locomotor habits in extinct primates (e.g. Ryan and Ketcham 2002a, Barak et al. 2013b, Kivell et al. 2018, Ryan et al. 2018). Despite its potential, the internal architecture of the long bones of the sloth-lemurs and koala-lemurs has not yet been analysed. Only recently, the trabecular structure of long bone epiphyses in palaeopropithecines and Megaladapis has been quantified, but primarily to investigate evolutionary pat- terns in slow arboreal mammals (Alfieri et al. 2023), rather than to infer their locomotor behaviour directly. To fill this gap, in this study we examine humeral and femoral trabecular architecture in sloth-lemurs and koala-lemurs, compar- ing them with those of a sample of extant mammals, encompassing potential ecological analogues of subfossil lemurs, inferred sister taxa of sloth-lemurs and koala-lemurs, and other relevant groups (Fig. 1). Initially, we used quantitative behavioural data on extant taxa to classify them in locomotor categories; we then analysed their humeral and femoral trabecular structure with the aim of understanding which skeletal region and trabecular trait discrim- inate extant mammals according to their locomotion. Hence, we inferred the most likely locomotor behaviour for extinct lemurs, using both an exploratory approach (i.e. assessing morphospaces) and an analytical approach [i.e. running a phylogenetically informed discriminant function analysis (pDFA) on the most informative trabecular traits]. This protocol allowed us to under- stand whether trabecular bone supports previous hypotheses regarding: (i) the particularly derived suspensory adaptations of Palaeopropithecus; (ii) the relatively intermediate degree of sus- pensory specialization in Mesopropithecus and Babakotia; and (iii) potential locomotor differences across the small/medium- sized Megaladapis madagascariensis and Megaladapis grandidieri, and the large Megaladapis edwardsi (see above). M AT E R I A L S A N D M ET H O D S Sample Subfossil lemurs We used high-resolution microcomputed tomography (μCT) data derived from humeri and femora of Malagasy subfossil lemurs collected in the Muséum national d’Histoire Naturelle (MNHN; Paris, France) and the Division of Fossil Primates, Duke Lemur Figure 1. A, principal component analysis on quantitative locomotor data (top, scatterplot; bottom, variable loading plot). Identified locomotor categories are as follows: AG, arboreal generalist; L, leaping; QSC, quadrupedal/scrambling/clambering; and SB, suspensory/bridging. B, phylogenetic relationships between the studied species with locomotor categories mapped on the phylogeny. 4  •  Alfieri et al. Center (renamed as the Duke Lemur Center Museum of Natural History), Durham (NC, USA) (e.g. Fig. 2A, B). Details regarding acquisition can be found in the paper by Alfieri et al. (2023). The studied data are derived from specimens belonging to Babakotia sp. (one complete humerus and one distal humerus), Palaeopro- pithecus sp. (four complete humeri, one proximal humerus, one distal humerus, four complete femora, one proximal femur, and two distal femora), Mesopropithecus dolichobrachion (one complete humerus), and Megaladapis sp. (three complete humeri, one prox- imal humerus, one distal humerus, two complete femora, two proximal femora, and two distal femora). Babakotia sp. was assigned to Babakotia radofilai, the only species recognized within the genus, and Palaeoproprithecus sp. and Megaladapis sp. were assigned to known species within the two genera, following Alfieri et al. (2023) and as summarized in the Supporting Information (Appendix S1). Extant primates We built a comparative sample of extant mammal species (Table 1) by collecting μCT data from previous studies (Georgiou et  al. 2018, 2019, Kivell et al. 2018, Ryan et al. 2018, Tsegai et al. 2018a,b, Arias-Martorell et al. 2021, Alfieri et al. 2022, 2023) and/or down- loading data from MorphoSource (https://www.morphosource. org; Supporting Information, Tables S1 and S2). These data are derived from 55 humeri and 53 femora collected across several institutions (Supporting Information, Appendix S2). We studied adult specimens, as determined by full epiphyseal fusion, and selected epiphyses with preserved trabecular bone for the extant sample. For the subfossil lemur epiphyseal data, perfect preserva- tion of trabeculae was not always possible; therefore, we either discarded these specimens or adopted procedures to analyse fragmented individuals (see below, section “Extraction of trabec- ular parameters”). Most of the studied extant mammals have been proposed as extant ecological analogues of extinct lemurs and/or were used in previous morphological comparative works of the studied subfossil species (e.g. Godfrey 1988, Godfrey et al. 1995, 2016, Wunderlich et  al. 1996, Jungers et  al. 1997, 2002, 2008, Hamrick et al. 2000, Shapiro et al. 2005, Walker et al. 2008, Grana- tosky et al. 2014, Boyer et al. 2015, Marchi et al. 2016, Amson and Nyakatura 2018, Granatosky 2022, Alfieri et al. 2021, 2023), i.e. tree-sloths (both three-toed sloths, Bradypodidae, and two-toed sloths, Choloepodidae), the koala (Phascolarctidae), lorisids (Lorisidae), and non-human great apes (Hominidae). We also included representatives from the two taxa inferred as the extant closest relatives of palaeopropithecines and Megaladapis, i.e. the subfamily Indriinae and the family Lemuridae, respectively (Baab et al. 2014, Herrera and Dávalos 2016, Upham et al. 2019, Mar- ciniak et al. 2021, Fleagle et al. 2025). Also, we included the platyr- rhine Alouatta caraya (Humboldt, 1812), as an example of a non-strepsirrhine primate adapted to an energy-saving slow arbo- real lifestyle (Bicca‐Marques and Calegaro‐Marques 1998, Prates et al. 2018), an ecology that has been inferred for both palaeopro- pithecines and Megaladapis (Walker et  al. 2008, Godfrey et  al. 2016, Alfieri et  al. 2021). Additonally, we included galagids, a strepsirrhine family that is not closely related to subfossil lemurs and whose members are specialized in vertical clinging and leap- ing (Walker 1979, Gebo 2011, Fleagle 2013), the most common and possibly ancestral behaviour in the strepsirrhine radiation (Boyer et al. 2017). By doing so, we were able to include a leaping group that is not closely related to extinct lemurs, contrary to Ind- riinae, which show leaping habits (see below, Results) but are also the extant sister taxon of palaeopropithecines (Fig. 1). All taxa Figure 2. A, B, three-dimensional models of two subfossil lemur specimens studied in this work: A, humerus of Megaladapis edwardsi MNHN-MAD 7777; and B, femur of Palaeopropithecus ingens MNHM-MAD 8808. C–G, proximal (C) and distal (D) humeral epiphyses, in addition to proximal (E), distal lateral condyle (F), and distal medial condyle (G) femoral epiphyses, from which trabecular bone was studied. The extraction of spherical and hemispherical volumes of interest (highlighted in red) is shown. C–G refer to a humerus of Perodicticus potto (Müller, 1766) (NMW 32674) (C, D) and a femur of Propithecus verreauxi Grandidier, 1867 (AMNH 170474) (E–G). Images are not to scale. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://www.morphosource.org https://www.morphosource.org https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data Trabecular bone of subfossil lemurs    •  5 included in the sample were taxonomically categorized at the spe- cies level (Table 1; for more details, see Supporting Information, Tables S1 and S2). Extant primates: locomotor diversity For 62% of the extant species included here (Fig. 1), we were able to use data on the relative frequency (expressed in percentage) of locomotor behaviours collected by Granatosky (2018). In collect- ing these quantitative data, we largely followed the approach of Monclús-Gonzalo et  al. (2023, 2025), i.e. we pooled the behaviours quantified by Granatosky (2018) in six broader loco- motor categories: (i) leaping; (ii) quadrupedalism; (iii) climbing; (iv) suspension; (v) bridging; and (vi) scrambling/clambering. The description of the locomotor categories can be found in the paper by Monclús-Gonzalo et  al. (2023). We transformed the frequency data through the arcsine square root, a common procedure when proportional ecological data are studied (Sokal and Rohlf 1995, Monclús-Gonzalo et al. 2023; Supporting Infor- mation, Table S3). For some species not listed by Granatosky (2018), we used quantitative data from closely related taxa (Sup- porting Information, Appendix S3). Given that we aimed to obtain quantitative data representative of species, in cases of mul- tiple datasets for a single taxon reported by Granatosky (2018) we averaged the data for each behaviour, resulting in a single value per locomotor category per species. The behavioural dataset was used to run a principal component analysis (PCA; Supporting Information, Table S4), from which we visually identified, on the principal component (PC)1–PC2 biplot and variable loading plot, how taxa are divided into subgroups owing to their locomo- tor behaviours. These subgroups represented the discrete locomo- tor categories used to assign species for which quantitative locomotor data (sensu Granatosky 2018) were unavailable. Table 1. Taxa studied in this work, with the sample size for each analysed epiphyseal joint from which trabecular structure was quantified. Taxon Proximal humeri (N) Distal humeri (N) Proximal femora (N) Femoral lateral condyles (N) Femoral medial condyles (N) Extant taxa Alouatta caraya 1 NA 1 NA NA Arctocebus calabarensis 1 1 1 1 1 Avahi laniger 3 3 3 3 3 Bradypus tridactylus 1 1 1 1 1 Bradypus variegatus 3 3 3 3 3 Choloepus didactylus 4 4 4 4 4 Eulemur albifrons 2 2 2 2 2 Euoticus elegantulus 1 1 1 1 1 Galago matschiei 1 1 1 1 1 Galago senegalensis 1 1 1 1 1 Gorilla beringei NA 1 NA NA NA Gorilla gorilla 1 1 1 1 1 Hapalemur griseus 2 2 2 2 2 Indri indri 3 3 3 3 3 Lemur catta 4 (1) 4 3 3 3 Loris tardigradus 1 (0) 1 1 (0) 1 1 Nycticebus bengalensis 1 1 1 1 1 Nycticebus coucang 3 (1) 3 3 (2) 3 (2) 3 (2) Otolemur crassicaudatus 1 (0) 1 (0) 1 1 1 Pan troglodytes 1 1 2 1 1 Perodicticus potto 5 5 5 (4) 5 (4) 5 Phascolarctos cinereus 5 (2) 5 5 (2) 5 5 Pongo abelii 1 NA 1 NA NA Pongo pygmaeus 1 1 NA 1 1 Propithecus diadema 1 1 1 1 1 Propithecus verreauxi 2 (1) 2 2 2 (1) 2 (1) Varecia variegata 1 1 1 1 1 Subfossil lemurs Babakotia radofilai 1 2 NA NA NA Megaladapis edwardsi 2 2 (1) 2 2 (1) 2 (1) Megaladapis madagascariensis 2 2 2 2 (1) 2 (1) Mesopropithecus dolichobrachion 1 1 (0) NA NA NA Palaeopropithecus ingens 4 4 4 (3) 5 (4) 5 (2) Palaeopropithecus kelyus 1 1 NA NA NA Palaeopropithecus maximus NA NA 1 (0) 1 1 Abbreviations: N, number of specimens; NA, joints for which data were not available owing to fragmentary specimens. In cases where specimens were discarded because of poor trabecular bone preservation or an insufficient number of trabeculae (see the Materials and Methods; Supporting Information, Appendices S5–S7), the number of specimens ultimately used in the analyses is given in parentheses. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data 6  •  Alfieri et al. Extraction of trabecular parameters Most of the collected μCT data (i.e. all species except apes and Alouatta caraya) consisted of spherical volumes of interest (VOIs) of trabecular bone. These VOIs were extracted in previous studies (Alfieri et al. 2022, 2023) from proximal and distal humeral and femoral epiphyses after orienting the humeri and femora in stan- dard anatomical positions, i.e. x-axis mediolaterally oriented, y-axis anteroposteriorly oriented, and z-axis proximodistally ori- ented (Supporting Information, Appendix S4). Each spherical VOI represents the largest sphere centred within the humeral head and capitulum, and within the femoral head, lateral condyle, and medial condyle (Fig.  2), such that only trabecular bone was included and cortical bone was entirely excluded. The femoral head VOIs were subsequently halved by removing the medial hemisphere (Fig. 2E). This modification was necessary because some taxa, including some of those newly analysed here [e.g. Alouatta caraya and Gorilla gorilla (Savage, 1847)], exhibit a pro- nounced fovea capitis on the medial side of the femoral head. Using a fully spherical VOI in these cases would have included a large empty space corresponding to the fovea capitis, thereby bias- ing trabecular measurements (Alfieri et al. 2022, 2023). As for the rest of the sample, VOIs were extracted using the same protocols. Ultimately, all the studied VOIs represent homologous regions. The VOIs were imported into ImageJ2 (Rueden et al. 2017) and segmented, isolating the trabecular bone fraction. For extant spe- cies, it generally corresponds to binarizing the μCT data, because only two fractions should be distinguished, i.e. trabecular bone from intertrabecular voids, to be assigned to foreground (i.e. one) and background (i.e. zero), respectively. Hence, for VOIs of almost all the extant species we used the automatic binarizing tool of ImageJ2 (‘Threshold’), yielding a result that we considered satis- fying by assessing, through a visual comparison, the respective non-binarized VOIs. For the other VOIs (the few from extant species and those from subfossil lemurs), we followed alternative procedures, based on manual cleaning and/or tools for automated recognition of non-osseous fractions (detailed in Supporting Information, Appendix S5). From binarized VOIs, we extracted trabecular parameters using BoneJ2 (ImageJ2 plugin; Domander et al. 2021), including degree of anisotropy (DA, unitless; averaged across three runs), mean trabecular thickness (Tb.Th, in millime- tres), average branch length (Av.Br.Len, in millimetres), bone volume fraction (BV/TV, unitless), bone surface density (BS/TV, per millimetre), and connectivity density (Conn.D, per millimetre cubed). Because BoneJ2 assumes cubic VOIs, BV/TV, BS/TV, and Conn.D were corrected for spherical (or hemispherical, for the femoral head) VOIs using volume derived from VOI diameter. Connectivity was used as a proxy for the number of trabeculae to exclude VOIs with fewer than ∼50 trabeculae (Mielke et al. 2018; see further details in Supporting Information, Appendix S6). Given our research focus on subfossil lemurs, we aimed to maxi- mize their representation in our sample. Therefore, we included some of their VOIs even if they contained peripheral regions with broken or missing trabeculae. To do so, we gradually reduced the size of these VOIs until the biased regions were excluded. Notably, this approach requires applying the same scaling factor to all the other taxa to ensure consistency, i.e. studying anatomically homol- ogous regions. However, reducing VOI sizes also decreases the number of trabeculae included. Thus, as justified in the Supporting Information (Appendix S7), for each epiphysis this procedure was balanced against the need to avoid under-representing some extant taxa owing to the discarding deriving from too low a number of trabeculae. Ultimately, the radii of the smallest analysed spheres (with the femoral head VOIs subsequently halved, see above) were as follows: proximal humerus = 0.74 mm [Euoticus elegantulus (LeConte, 1857)] (AMNH 241147), distal humerus = 0.74 mm (Galago matschiei Lorenz, 1917) (FMNH 148985), proximal femur = 0.89 mm [Loris tardigradus (Linnaeus, 1758)] (AMNH 269), femoral lateral condyle = 0.74 mm (Loris tardigradus AMNH 269), and femoral medial condyle = 0.75 mm [Arctocebus calaba- rensis (Smith, 1860)] (AMNH 212576). Statistical analyses Relationship between trabecular structure and locomotion: inferential tests Owing to the high intercorrelation among trabecular parameters (e.g. Ryan and Shaw 2012), we analysed them collectively as mul- tivariate datasets (i.e. five anatomically separated datasets were defined: proximal humerus, distal humerus, proximal femur, lat- eral condyle, and medial condyle) and through PCAs. PCA is commonly applied to datasets comprising intercorrelated vari- ables (e.g. cranial measurements; Nishimura et al. 2019, Landi et al. 2021). We conducted a first set of PCAs on centred and scaled datasets including only trabecular parameters (‘PCATP’, hereafter). The resulting PCTP1s–PCTP4s (explaining the 93.4%– 96.4% of the variance across the five datasets; Supporting Infor- mation, Tables S5–S9) were used to test for a relationship between trabecular structure and ecological categories while accounting for phylogenetic affinities among species. As a phylogeny, we used the time-calibrated tree deriving from the mammalian tree of Upham et al. (2019) adapted to our sample (Supporting Informa- tion, Appendix S8). We hence ran multivariate phylogenetic gen- eralized least squares (mvPGLS) regressions (‘mvgls’ function, ‘mvMORPH’ package; Clavel et al. 2015) in R v.4.3.1 (R Core Team 2023) and phylogenetic MANCOVAs (‘manova.gls’ func- tion, ‘mvMORPH’, Pillai’s trace tests) (Table 2). Beyond locomo- tion and phylogeny, trabecular bone is also potentially affected by allometry. To account for this source of variation too, we included a body mass proxy (BMp) as covariate in mvPGLSs. As detailed in the Supporting Information (Appendix S9), as BMp we took the centroid size, deriving from the landmark coordinates from Alfieri et al. (2022, 2023), predicted from a linear regression of BMp against bone metric measurements for specimens lacking centroid size values. To establish whether, in mvPGLSs, we also needed to account for ecology–BMp interactions, we preliminarily tested for a relationship between BMp and locomotor categories (phylogenetic ANOVA, ‘phylANOVA’ function, ‘phytools’ package; Revell 2012). Relationship between trabecular structure and locomotion: qualitative assessment We conducted an additional set of PCAs on centred and scaled datasets including trabecular parameters and BMp (‘PCATP+BM’, hereafter). The BMp was included to visualize the effects of body size on trabecular structure and, in doing so, we could observe https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data Trabecular bone of subfossil lemurs    •  7 locomotor patterns on biplots built with PCTP+BMs that are not primarily driven by BMp. It could be detected by assessing how BMp contributes to PCTP+BMs in the respective variable loading plots (i.e. through the length and direction of BMp vectors). More- over, by this approach we avoided potential biases associated with univariate size correction (e.g. residuals), which may not fully remove size effects in multivariate analyses (e.g. Alfieri et  al. 2025a). In PCATP+BMs, we also included data of subfossil lemurs to discuss their position on scatterplots. We built biplots with the first three PCTP+BMs (Figs 3–7; Supporting Information, Tables S10–S14), and on them we observed the distribution of locomo- tion categories and which specific locomotor groups are discrim- inated by which specific trabecular properties. This included assessing whether subfossil lemurs occupy regions preferentially associated with certain extant taxa, thereby guiding our choice of relevant locomotor categories for further analysis (see Supporting Information, Appendix S10). When we identified a trend related to a single PCTP+BM (i.e. locomotor categories, or combinations of them, are mainly discriminated along one PCTP+BM), we tested for significance through univariate PGLS and ANOVA (‘gls’ function, ‘nlme’ package; Pinheiro et al. 2020). Additionally, we showed the univariate distribution of these single PCTP+BMs through box- plots (Figs 5E, 6E, 7E). The contribution of trabecular traits and BMp to PCTP+BMs was assessed through eigenvector loadings and correlation analyses (Pearson coefficients and coefficients of determination, R2). To ascertain that the distribution of taxa on the PCATP+BM scatterplots was not primarily driven by the inclu- sion of BMp itself, we additionally generated scatterplots using the first three PCTPs (i.e. from PCAs performed on datasets devoid of BMp; Supporting Information, Figs S1–S5). Phylogenetic discriminant function analysis Once a subset of trabecular traits from different anatomical levels was identified as the most effective at discriminating specific loco- motor categories (Table 3), we compiled a single heterogeneous dataset to run a phylogenetic discriminant function analysis (pDFA). The pDFA preliminarily required a multivariate PGLS (‘mvgls’ function, ‘mvMORPH’) and phylogenetic MANOVA (‘manova.gls’ function, ‘mvMORPH’) on extant taxa to ensure that the selected traits captured significant locomotor variation. The multivariate PGLS fit was then used to run the pDFA model (‘mvgls.dfa’ function, ‘mvMORPH’; discriminant function, DF1, Fig. 8; for summary statistics, see Supporting Information, Tables S15–S17). The pDFA model was then tested by predicting the known locomotor categories of extant taxa (‘predict’ R function). Owing to low sample size (all analyses being phylogenetically informed, the observations are represented not by single speci- mens but by taxa) that could overestimate the performance of the pDFA, we used a leave-one-out cross-validation, which involved the exclusion of one random observation from the training set, then using the pDFA model to predict the class of the excluded observation. This procedure was iterated until all the taxa were excluded once, but avoiding the repeated exclusion of the same taxon, and allowed us to compute the misclassification rate of the pDFA model. Finally, the pDFA model was applied to subfossil lemurs. To account for uncertainties in the classification of extinct taxa, we followed the approach of Monclús-Gonzalo et al. (2023) (see also Widrig et al. 2025), discouraging the use of posterior probabilities as a reliable proxy for classification uncertainties (e.g. owing to overfitting; Qiao et al. 2009), and instead using subsam- pling techniques. Hence, for the subfossil taxa, the locomotor assignment was iterated several times, each time subsampling the training set used to train the pDFA model, and counting the times the extinct species were assigned to each class. As a resampling technique, we used leave-one-out cross-validation. R E SU LTS The PC1–PC2 scatterplot and variable loading plot deriving from the PCA run on quantitative locomotor data revealed four loco- motor groups (Fig. 1A): leaping (L), quadrupedal/scrambling/ clambering (QSC), suspensory/bridging (SB) and arboreal gen- eralists (AG). Accordingly, we framed, in these locomotor cate- gories, the taxa for which quantitative data were not available, thus obtaining the final ecological categorization (Fig. 1B; see further details in Supporting Information, Appendix S11). The four cat- egories were not significantly related to BMp (humerus, P = .162; femur, P = .476), implying that, in mvPGLS, the BMp–locomotion interaction effects should not be tested and that we could detect distinct effects of body mass and locomotion on trabecular data (Table 2). Humeral trabecular structure The humeral trabecular structure of the extant species was not correlated with locomotion. Indeed, from multivariate PGLSs and phylogenetic MANCOVAs, the humeral datasets were signifi- cantly correlated with BMp but not with locomotion (Table 2). Also, proximal and distal humerus PCATP+BM scatterplots and load- ing vectors showed that PC1TP+BMs are primarily related to BMp and to a set of variables (i.e. Tb.Th, Conn.D., Av.Br.Len, and BS/ TV). The strong body size effect of proximal and distal humeral PC1TP+BMs was clear owing to the position of great apes and Table 2. Results from multivariate phylogenetic generalized least squares regressions and phylogenetic MANCOVAs testing the relationship between trabecular structure (quantified through PCTP1–PCTP4 at different anatomical regions) and locomotion, while accounting for body mass effects (for details, see the Materials and Methods). Anatomical region (articulation) Principal components Proportion of variance (%) Body mass P-value Locomotion P-value Proximal humerus (humeral head) PCTP1–PCTP4 93.4 <.001 .21 Distal humerus (humeral capitulum) PCTP1–PCTP4 95.6 <.001 .46 Proximal femur (femoral head) PCTP1–PCTP4 95.4 <.001 .094 Distal femur (lateral condyle) PCTP1–PCTP4 96.8 <.001 .024 Distal femur (medial condyle) PCTP1–PCTP4 96.4 <.001 .019 https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data 8  •  Alfieri et al. subfossil lemurs (i.e. the largest taxa) on one extreme of PC1TP+BM values with galagids and lorisids (i.e. the smallest taxa) instead lying on the other extreme (Figs 3A, 4A). In both the humeral epiphyses, correlation tests confirmed the strong BMp effects on PC1TP+BMs, showing that Tb.Th, Conn.D., Av.Br.Len, and BS/TV contribute significantly to PC1TP+BMs (Supporting Information, Table S18) and that these trabecular traits are related to BMp themselves (Supporting Information, Table S19). Both proximal and distal humeral DA and BV/TV, instead, arose as not signifi- cantly related to the respective PC1TP+BMs (with the only exception of distal humeral PC1TP+BM–BV/TV, but with R2 = .29; Supporting Information, Table S18) and to body mass (Supporting Informa- tion, Table S19; Figs 3C, 4C). Humeral proximal and distal PC2TP+BMs and PC3TP+BMs, instead, do not show any significant relationship with BMp (Figs 3B, 4B; Supporting Information, Table S18). Loading vectors suggest that the PC2TP+BMs of both the proximal and distal humerus include information mainly related to DA and BV/TV (that do not con- tribute to PC1, see above; Figs  3D, 4D). It is confirmed by DA–PC2TP+BMs and BV/TV–PC2TP+BMs correlation tests, com- bined with the absence of significant correlation tests between any other trabecular variable and PC2TP+BMs. Proximal and distal humeral PC3TP+BMs contain information related to Av.Br.Len (pre- cisely to its portion of variance not related to BMp, differently from the one driving PC1, see above), because this parameter yielded the longest vector oriented towards PC3TP+BMs (Figs 3D, 4D) and showed a significant correlation with PC3TP+BMs (Supporting Information, Table S18). Although a few other variables occasion- ally showed a significant correlation with PC3TP+BMs too (DA in the proximal humerus, and Conn.D in the distal humerus), their low R2 (.13 for DA and .20 for Conn.D; Supporting Information, Table S18) led us to consider their contribution to the respective PC3TP+BM as minor. The humeral PC2TP+BM–PC3TP+BM scatterplots yielded a strong overlap of locomotor categories (especially clear for distal humeral data, with almost all the taxa clustering in the same region of the morphospace, i.e. positive PC3TP+BM scores and PC2TP+BM scores ranging between zero and one; Figs 3B, 4B). This, in addition to the non-significant relationship between PC1–PC4 and locomotion from multivariate PGLSs and phylogenetic MANCOVAs (Table 2), led us not to use humeral trabecular prop- erties in the following steps of the analysis. In the humeral PC2TP+BM–PC3TP+BM morphospaces (in which allometric effects are minimized, see above), subfossil lemurs overall cluster in the same regions (i.e. positive PC2TP+BM–negative PC3TP+BM scores for the proximal humerus, and positive PC2TP+BM scores for the distal humerus). In the range of variation of sloth-lemurs, Palaeopropithecus spp. tend to occupy the most peripheral position (with the exception of Palaeopropithecus ingens in the distal humerus), Babakotia lies the most closely to the other extant taxa, and Mesopropithecus (data available only for the proximal humerus) lies in another peripheral region of the morphospace (close to Arctocebus calabarensis; Figs 3B, 4B). Figure 3. Proximal humerus results of principal component analyses (PCA) on centred and scaled datasets including trabecular parameters and body mass proxy (PCATP+BM). A, B, PC1–PC2 (A) and PC2–PC3 (B) scatterplots, with extant species grouped by locomotor category and subfossil lemur data shown with purple rhombi. C, D, PC1–PC2 (C) and PC2–PC3 (D) variable loading plots. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data Trabecular bone of subfossil lemurs    •  9 Megaladapis spp. tend to lie close to Babakotia. Interestingly, in the humeral PC2TP+BM–PC3TP+BM morphospaces, most subfossil lemurs tend to occupy distinct regions, i.e. they do not overlap with those of extant mammals. Indeed, in the proximal humerus, Palaeopropithecus spp., Mesopropithecus dolichobrachion Simons et al., 1995, and Megaladapis madagascariensis lie outside the con- vex hulls of extant mammal locomotor categories (Fig. 3B). In the distal humerus, a similar pattern is observed for Palaeopro- pithecus kelyus and Megaladapis edwardsi, whereas Babakotia and Megaladapis madagascariensis occupy peripheral positions within the SB area and trend towards Palaeopropithecus kelyus and Megal- adapis edwardsi (Fig. 4B). Both the proximal and distal humeral outlying patterns of subfossil lemurs are mainly driven by high BV/TV (Figs 3D, 4D). Femoral trabecular structure Femoral data yielded patterns related to locomotor categories. Indeed, trabecular data from the lateral and medial condyles proved to be significantly related to locomotion, as shown by mul- tivariate PGLSs and phylogenetic MANCOVAs (Table 2). These epiphyseal regions, such as the proximal femoral, are also signifi- cantly related to body mass (Table 2) but, given that locomotor and allometric effects are separated in PGLSs (see above), lateral and medial condyle size-corrected data are significantly related to locomotor categories. Although not reaching significance, proxi- mal femoral data yielded a particularly low P-value for a relation- ship with locomotion (.09; Table 2). The PC1TP+BMs across three femoral datasets are explained by body mass and collect the allometric effects, as shown by the posi- tion of great apes/subfossil lemurs (i.e. the largest taxa) and galagids/lorisids (i.e. the smallest taxa) at the two extremes of PC1TP+BM scores (Figs 5A, 6A, 7A). This is also suggested by the directions of BMp loading vectors (Figs 5C, 6C, 7C), additionally indicating that PC1TP+BMs are driven by a portion of variance of a subset of trabecular variables that are likely to include size effects themselves. Through correlation tests across the femoral datasets, we confirmed the significant relationship between PC1TP+BMs and BMp, we identified the trabecular variables driving PC1TP+BMs (i.e. Tb.Th., Conn.D., BS/TV, and Av.Br.Len; Supporting Information, Table S18), and we found that they drive PC1TP+BM through their relationships with BMp (Supporting Information, Table S19). In the proximal femur and in the medial condyle, a portion of BV/ TV variance also is correlated with PC1TP+BM but with low strength (R prox 2  = .24, R med con. 2  = .24; Supporting Information, Table S18). Likewise, BV/TV significantly but weakly relates to BMp in these two levels (R prox 2  = .21, R med con. 2  = .23; Supporting Information, Table S19). DA makes a minor contribution to femoral PC1TP+BMs, because it is not correlated or is weakly correlated with PC1TP+BMs (R lat con. 2  = .14; Supporting Information, Table S18) and is not cor- related with BMp (Supporting Information, Table S19). Femoral PC2TP+BMs and PC3TP+BMs do not include size effects, as suggested by BMp vectors in the loading plots (Figs 5D, 6D, 7D) and PC2TP+BMs and PC3TP+BMs not being correlated with BMp (Supporting Information, Table S18). Correlation tests Figure 4. Distal humerus results of principal component analyses (PCA) on centred and scaled datasets including trabecular parameters and body mass proxy (PCATP+BM). A, B, PC1–PC2 (A) and PC2–PC3 (B) scatterplots, with extant species grouped by locomotor category and subfossil lemur data shown with purple rhombi. C, D, PC1–PC2 (C) and PC2–PC3 (D) variable loading plots. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data 10  •  Alfieri et al. show that proximal femoral PCTP+BM2 is mainly driven by DA, especially (R2 = .69), and BV/TV (R2 = .47), and no other vari- ables (Supporting Information, Table S18). Both DA and BV/ TV contribute negatively to proximal femoral PC2TP+BM, i.e. higher PC2TP+BM scores are yielded for lower DA and BV/TV (Fig. 5D; Table 3). Correlation tests also tell us that lateral and medial condyle PC2TP+BMs are driven by BV/TV primarily, less strongly by Av.Br.Len, and not by other variables (no P < .05, with the only exception of DA in the medial condyle, P = .02, but with weak correlation, R2 = .19; Supporting Information, Table S18). In the proximal femur, PC3TP+BM captures informa- tion on Av.Br.Len and, albeit weakly, on Conn.D (R2 = .18), and no other variables (Fig.  5D; Supporting Information, Table S18). Correlation tests show that the lateral and medial condyle PC3TP+BMs are mainly driven by DA and no other variables (Sup- porting Information, Table S18; Figs 6D, 7D). DA contributes negatively to lateral and medial condyles PC3TP+BMs, i.e. lower DA is shown by taxa yielding higher PC3TP+BM scores (Figs 6D, 7D; Table 3). On the femoral PC2TP+BM–PC3TP+BM morphospaces, we visually identified a pattern of discrimination of ‘SB’ from all the other taxa (‘others’ hereafter) that possibly caused the significant relationship with locomotion arising from multivariate PGLSs and phyloge- netic MANCOVAs for two femoral epiphyseal regions (Table 2). We focused on the discrimination between ‘SB’ and ‘others’, because in all the femoral anatomical levels, subfossil lemurs tended to lie in the morphospace areas occupied by ‘SB’, making this trend of interest for the focus of this work (Figs 5B, 6B, 7B). In the proximal femur, ‘SB’ tend to yield high PC2TP+BM scores (although Perodicticus potto, Pongo abelii Lesson, 1827, and Arcto- cebus calabarensis lie within ‘others’; Fig. 5B). In the lateral and medial condyle, ‘SB’ yielded higher PC3TP+BM scores [with some exceptions: for the lateral condyle, the ‘SB’ Pongo pygmaeus (Lin- naeus, 1760) lies within ‘others’, and Pan troglodytes (Blumenbach, 1775), belonging to ‘others’, instead lies within ‘SB’; for the medial condyle, the position of a few taxa belonging to ‘others’, i.e. Gorilla gorilla, Avahi laniger (Gmelin, 1788), and Pan troglodytes, cause overlap with the ‘SB’ range; Figs 6B, 7B]. The PC2TP+BM pattern for the proximal femur and the PC3TP+BM patterns for lateral and medial condyles (Figs 5E, 6E, 7E) were all confirmed qualitatively through PGLSs and ANOVAs (Table 3). All the patterns identified in the humeral and femoral PCATP+BM scatterplots outlined above are not driven primarily by the inclusion of BMp in the PCAs. Indeed, the distribution of locomotor groups, the positions of subfossil lemurs, and the vari- able loading plots remain largely unaltered in the PCATP scatter- plots (cf. Figs 3–7 and Supporting Information, Figs S1–S5). Phylogenetic discriminant function analysis As detailed above, the discrimination between ‘SB’ and ‘others’ was the only pattern of interest yielded by trabecular structure. It was caused only by femoral traits and, specifically, by higher PC2TP+BM (primarily corresponding to lower BV/TV and DA) in the femoral head PCATP+BM (Fig. 5E) and by higher PC3TP+BM Figure 5. Proximal femur results of principal component analyses (PCA) on centred and scaled datasets including trabecular parameters and body mass proxy (PCATP+BM). A, B, PC1–PC2 (A) and PC2–PC3 (B) scatterplots, with extant species grouped by locomotor category and subfossil lemur data shown with purple rhombi. C, D, PC1–PC2 (C) and PC2–PC3 (D) variable loading plots. E, the univariate distribution of PC2TP+BM, along which we identified the separation of suspensory/bridging (SB) from ‘others’, is shown through a boxplot. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data Trabecular bone of subfossil lemurs    •  11 (mainly corresponding to lower DA) in both the lateral and the medial condyle PCATP+BMs (Figs  6E, 7E). Hence, we took the trabecular traits primarily contributing to these femoral PCTP+BMs (i.e. BV/TVprox, DAprox, DAlat.con, and DAmed.con, identified as detailed above) and we built a multivariate heterogeneous dataset including only extant taxa (i.e. the training set), on which we trained a pDFA model that is, accordingly, potentially able to classify taxa in ‘SB’ or ‘others’. The heterogeneous dataset con- firmed the potential to discriminate ‘SB’ from ‘others’ (multivar- iate PGLS and phylogenetic MANOVA, P = .002), and the pDFA model trained on the training set yielded a misclassification rate of .14. We considered this value as sufficiently low (i.e. the model correctly predicted locomotion for 86% of the observations excluded during leave-one-out cross-validation). The pDFA gen- erated a single discriminant function (i.e. DF1), because the model accounted for only two classes (i.e. ‘SB’ vs. ‘others’). DF1 was positively correlated with all the variables on which the model was run (Supporting Information, Table S16) and, as shown with a boxplot, it neatly separated ‘SB’ (lower DF1, i.e. thus lower femoral BV/TVprox, DAprox, DAlat.con, and DAmed.con) from ‘others’ (Fig. 8). The pDFA model, run on Megaladapis mad- agascariensis, Megaladapis edwardsi, and Palaeopropithecus ingens Grandidier, 1899 (the three subfossil lemur species for which we had data from the proximal femur, lateral condyle, and medial condyle), classified Megaladapis edwardsi and Palaeopropithecus ingens as ‘SB’ in 100% of the iterations, whereas Megaladapis mad- agascariensis was never classified as ‘SB’. D I S C U S S I O N Locomotor signal in extant mammal trabecular structure When assessing the locomotor categories that we established mapped onto the phylogeny, it becomes evident that, although some clades are characterized by displaying a single locomotor category (e.g. all galagids are classified into the leaping category), the overall distribution reveals multiple independent acquisitions of the same (broadly defined) locomotor behaviour across distant lineages [e.g. SB in Pongo Lacépède, 1799, Bradypodidae, Cho- loepodidae, and Lorisidae; leaping in Galagidae, Indriinae, and Hapalemur griseus (Link, 1795); Fig. 1B]. This distribution sup- ports the interpretation that any trabecular similarity among taxa sharing the same locomotor category in different regions of the phylogeny may be attributed to functional adaptations related to locomotor behaviours. Also, the relationship between trabecular structure and locomotion was tested through phylogenetically informed analyses (e.g. PGLS and pDFA) that account for shared phylogenetic history of the studied taxa. It ensures that the locomotion-related patterns that we found and discuss below are not attributable to phylogenetic affinity. We aimed not only to exclude phylogenetically driven patterns but potentially to observe them, in case they were dominant (i.e. by observing closely related taxa clustering in a region of the morphospace, such as subfossil lemurs more closely resembling other Malagasy strepsirrhines). This rationale was at the basis of our choice to use traditional PCAs instead of other techniques that erase phylogenetic trends (e.g. phylogenetic PCA; Polly et al. 2013), for qualitative assessments. Figure 6. Lateral femoral condyle results of principal component analyses (PCA) on centred and scaled datasets including trabecular parame- ters and body mass proxy (PCATP+BM). A, B, PC1–PC2 (A) and PC2–PC3 (B) scatterplots, with extant species grouped by locomotor category and subfossil lemur data shown with purple rhombi. C, D, PC1–PC2 (C) and PC2–PC3 (D) variable loading plots. E, the univariate distribu- tion of PC3TP+BM, along which we identified the separation of suspensory/bridging (SB) from ‘others’, is shown through a boxplot. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data 12  •  Alfieri et al. Regarding allometric effects, in both humeral and femoral PCAs they are mostly captured in PC1TP+BMs (Figs 3C, 4C, 5C, 6C, 7C). Given that PC1 normally accumulates most of the vari- ance in a PCA, a strong allometric effect on PC1TP+BMs suggests that body mass is a major driving factor of trabecular bone in the sample studied. This was also evident from PGLSs (Table 2), and it is consistent with previous findings in mammals (Doube et al. 2011, Amson et al. 2017), especially primates (Ryan and Shaw 2013, Alfieri et  al. 2025a), but also in other amniotes, both in broad amniote samples (e.g. Gônet et al. 2023) and in specific clades, such as birds (Doube et al. 2011, Alfieri et al. 2025b) and non-avian reptiles (Plasse et al. 2019). Investigation of detailed scaling patterns is outside the scope of this work, but it is interest- ing to note that, for instance, Conn.D and BS/TV positively co-vary in driving PC1TP+BMs across all the studied regions (see the directions of loading vectors in Figs  3C, 4C, 5C, 6C, 7C), which is consistent with previous works (e.g. Gônet et al. 2023) and expected structural patterns (i.e. the higher Conn.D., the higher the number of trabeculae, and the greater the bone surface). In all the PGLSs that we ran across the studied anatomical regions, effects of body mass were separated from locomotor ones. Hence, the effects of locomotion found in the PGLSs (Table 2) were already isolated from allometry. Also, all the locomotor pat- terns discussed below were derived from PC2TP+BMs or PC3TP+BMs and from two variables (BV/TV and DA) that largely did not include body mass effects. Indeed, no significant correlation with BMp was found for these parameters in the femoral datasets, with the only exceptions of BV/TV in the head and lateral condyle that were weakly correlated with BMp (Supporting Information, Table S19), and with BV/TV from lateral condyle that, importantly, was not used in the locomotor reconstructions. DA and/or BV/TV not being correlated to body mass aligns with many previous works (e.g. Cotter et al. 2009, Doube et al. 2011, Barak et al. 2013a, Fajardo et al. 2013, Rolvien et al. 2017, Kivell et al. 2018, Plasse Figure 7. Medial femoral condyle results of principal component analyses (PCA) on centred and scaled datasets including trabecular parame- ters and body mass proxy (PCATP+BM). A, B, PC1–PC2 (A) and PC2–PC3 (B) scatterplots, with extant species grouped by locomotor category and subfossil lemur data shown with purple rhombi. C, D, PC1–PC2 (C) and PC2–PC3 (D) variable loading plots. E, the univariate distribu- tion of PC3TP+BM, along which we identified the separation of suspensory/bridging (SB) from ‘others’, is shown through a boxplot. Figure 8. Distribution of discriminant function (DF1), deriving from a phylogenetic discriminant function analysis run on four predictor trabecular variables. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data Trabecular bone of subfossil lemurs    •  13 et  al. 2019, Gônet et  al. 2023) and indicates that the locomotion-related patterns we propose are independent of allo- metric effects. Moreover, the successful isolation of these two allometrically independent variables supports our choice to account for body mass effects by excluding PC1s (common prac- tice in shape analyses; e.g. Somers 1989, Klingenberg 2022) instead of using other size-correction techniques, such as those based on regression residuals (which previously proved to not to exclude allometric effects fully from trabecular variables; e.g. Alfi- eri et al. 2025a). Concerning allometric effects, it is also important to note that, because VOIs represent the largest spheres or hemi- spheres containing only trabecular bone in each studied epiphysis, differences in VOI size reflect differences in epiphyseal joint size. As detailed in the Supporting Information (Appendix S9), epiph- yseal metric measurements were used to predict the BMp for frag- mentary specimens, which could not yield centroid size estimates because they were not analysed using three-dimensional geomet- ric morphometrics by Alfieri et al. (2023). This prediction was possible owing to the strong correlation between values derived from epiphyseal metric measurements and BMp (Alfieri et  al. 2023). Consequently, because VOI size depends on epiphyseal joint size and joint size is indirectly related to BMp, VOI size can reasonably be assumed to reflect differences in body mass. Differ- ences in VOI size across the sample are therefore expected to have affected the results in a manner similar to body mass, with PC1T- P+BMs probably capturing most allometric effects, in addition to those associated with VOI size variation, and largely excluding these effects from PC2-3TP+BMs, on which our functional inference is based. The femoral trabecular structure yielded a relationship to loco- motor behaviour, i.e. significant correlation between trabecular parameters of the femoral condyles and locomotor categories (Table 2, with proximal femoral trabecular properties yielding a P-value approaching significance). The PC2TP+BM–PC3TP+BM mor- phospaces suggested that overall femoral trabecular structure is able to discriminate suspensory/bridging mammals (SB, i.e. lorisids, tree-sloths, and orangutans) from all the other extant mammals. Specifically, SB mammals are distinct for showing, on average, higher proximal femur PC2TP+BM scores (Fig. 5B), higher lateral condyle PC3TP+BM scores (Fig. 6B), and higher medial con- dyle PC3TP+BM scores (Fig. 7B). These patterns were confirmed quantitatively (see also Figs 5E, 6E, 7E) and, as indicated by load- ing vectors, are primarily driven by low BV/TVprox, DAprox, DAlat. con, and DAmed.con (Figs 5D, 6D, 7D; Table 3). Accordingly, when these four femoral trabecular parameters were analysed, they sig- nificantly discriminated SB from other taxa and, when used in a pDFA, yielded a DF that neatly discriminated SB from other mam- mals (Fig. 8) and that correctly classified them with a high success rate (i.e. 86%). Thus, multiple lines of evidence suggested that SB mammals differ from others by exhibiting lower bone volume (i.e. low BV/TV) in the proximal femur and a more isotropic trabec- ular structure (i.e. low DA) across all femoral levels. Proximal femur trabecular structure proved to be related to locomotor activ- ities in many previous studies mostly focusing on the mammal, especially primate, femoral head (e.g. Fajardo and Müller 2001, MacLatchy and Müller 2002, Ryan and Ketcham 2002b, 2005, Ryan and Krovitz 2006, Saparin et al. 2011, Ryan and Shaw 2012, Mielke et al. 2018, Georgiou et al. 2019, 2020, Alfieri et al. 2022), aligning with our results. Femoral low DA and BV/TV in SB mammals: functional interpretations A functional interpretation for lower femoral DA and BV/TV in SB mammals compared with others is consistent with the fact that these two parameters can generally provide sufficient information on the biomechanical regime and, accordingly, on locomotor hab- its, because cumulatively they explain ≤98% of bone stiffness (through their contribution to Young’s modulus; Stauber et al. 2006, Maquer et al. 2015). As shown experimentally, trabecular orientation tends to follow loading directions (Biewener et  al. 1996, Pontzer et  al. 2006), which can, in turn, be informative on locomotor habits. DA cap- tures the degree of preferential alignment of trabeculae, hence not the specific directions (e.g. in degrees of arc) of trabeculae but the extent to which they are uniformly oriented (as computed here, from DA = 0, i.e. highly multidirectional or fully isotropic struc- ture, to DA = 1, highly directional or fully anisotropic structure; Harrigan and Mann 1984). Thus, DA is positively related to direc- tional stereotypy of movements involved in locomotor activities, e.g. high DA is caused by fossoriality (Amson et al. 2017) and bipedalism (Ryan et al. 2018, Georgiou et al. 2019), whereas low DA derives from multidirectional locomotor repertoires. Arbore- ality represents a paradigmatic example of the latter condition, because moving in trees results in highly variable directions of biomechanical loadings (Carlson 2005, Demes et al. 2006). Our sample is represented by taxa that, to several extents, show arbo- real habits (Fig. 1), hence the lower DA in both the proximal and distal femur of SB mammals possibly reflects their outstandingly broader mobility in the hip and knee joints (Ishida et al. 1992, Runestad 1997, Thorpe and Crompton 2005, Zihlman et al. 2011, Nyakatura 2012, Marshall et al. 2021, Youlatos et al. 2025), even in comparison to other arboreal taxa, whose joints are also expected to be relatively mobile. An exception is represented by leaping taxa that, although being fully arboreal, are characterized by a distinctive behaviour related to powerful leaps and take-offs that cause high directional stereotypy and, as expected, high DA Table 3. Results from phylogenetic generalized least squares regressions and phylogenetic ANOVAs testing the discrimination of ‘SB’ from ‘others’, run with the PCTP+BMs (listed in the table, besides the explained portion of variance and by which traits they are mainly driven) that were previously identified as potentially yielding this locomotor information (for details, see the Materials and Methods). Anatomical region (articulation) Principal components Proportion of variance (%) Mainly related to ‘SB’–‘others’ P-value Proximal femur (femoral head) PCTP+BM2 19.18 BV/TV (inversely) .028 DA (inversely) Distal femur (lateral condyle) PCTP+BM3 14.34 DA (inversely) .001 Distal femur (medial condyle) PCTP+BM3 14.48 DA (inversely) <.001 https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data 14  •  Alfieri et al. (as found by Ryan and Ketcham 2002b, 2005). An extreme mobil- ity in hip and knee joints, related to SB locomotion, has been reported, or derived from joint/muscle configuration, for orang- utans (Thorpe and Crompton 2005, Zihlman et al. 2011), lorisids (Ishida et  al. 1992, Runestad 1997, Youlatos et  al. 2025), and tree-sloths (Nyakatura 2012, Marshall et al. 2021). Thus, SB taxa possibly occupy a positive extreme on the arboreal range of hind- limb joint mobility, and the consequent substantially multidirec- tional loadings acting on the hip and knee result in particularly isotropic trabeculae, i.e. low DA. BV/TV measures the proportion of inner epiphysis occupied by ossified fraction and informs on trabecular bone compactness. Increased loadings are expected to unbalance bone modelling (sensu Barak 2020) in favour of bone deposition at the expense of bone reabsorption and resulting in a greater bone fraction, whereas the opposite is expected under decreased loading regimes (Pearson and Lieberman 2004). Within an epiphysis, it is reflected by BV/ TV and, accordingly, a positive relationship between trabecular bone fraction and loadings has been found or hypothesized (e.g. Chang et al. 2008, Modlesky et al. 2008, Ryan and Shaw 2015). In this framework, the lower femoral BV/TV shown by SB taxa is possibly justified by particularly low hindlimb loadings related to arboreality based on SB. This condition has been consistently dis- cussed for tree-sloths, owing to their outstandingly cautious life- style (e.g. Alfieri et  al. 2021, 2022). Moreover, low hindlimb loadings have been proposed for lorisids, which in this regard are considered remarkable exceptions to primate hindlimb dominance (see above) (Ishida et al. 1990, Hanna et al. 2017, Yapuncich and Granatosky 2021). The same is likely to apply to orangutans, which, accordingly, show poorly robust hindlimb diaphyses (e.g. Ruff 1988). Based on these functional interpretations, we posit that low DA and BV/TV in the respective femoral epiphyses of subfossil lemurs are indicators of SB in these extinct taxa. The locomotor behaviour of palaeopropithecines We found strong confirmation for Palaeopropithecus as the most extremely suspensory sloth-lemur. Indeed, Palaeopropithecus ingens and/or Palaeopropithecus maximus always felt within the range that we interpreted as characterizing SB (i.e. higher femoral head PC2 scores and femoral condyles PC3 scores; Figs 5B, 6B, 7B). Furthermore, Palaeopropithecus ingens, the only sloth-lemur species for which we had data from all the studied femoral epiph- yseal regions, was classified as SB in 100% of the iterations by the pDFA model. Thus, we add femoral trabecular structure to the growing list of anatomical traits (e.g. Jungers 1980, Jungers et al. 1991, Shapiro et al. 1994, 2005, Godfrey et al. 1995, Wunderlich et al. 1996, Hamrick et al. 2000, Godfrey and Jungers 2003, Patel et al. 2013, Marchi et al. 2016, Granatosky 2018, 2020, 2022) indi- cating that Palaeopropithecus spp. were large lemurs exhibiting highly specialized suspensory arboreality. Previous works have proposed suspensory habits in Babakotia and Mesopropithecus too, but less extreme in comparison to Palae- opropithecus ( Jungers et al. 1991, 2002, Simons et al. 1992, God- frey et al. 1995, Hamrick et al. 2000, Godfrey and Jungers 2003, Shapiro et al. 2005, Marchi et al. 2016). Our results do not allow us to contribute to this view, owing to the availability/preservation of the specimens and the informative power that arose from the study of extant taxa, i.e. only femoral trabecular structure can be used for locomotor inference. Indeed, the fact that we had only humeral data for Babakotia and Mesopropithecus limits our infer- ential power on these smaller sloth-lemurs. Thus, we encourage further work to study femoral trabecular data on Babakotia and Mesopropithecus to test whether their DA and BV/TV are low enough to indicate suspensory/bridging habits for these two taxa too. The locomotor behaviour of Megaladapis spp. Despite early reconstructions not necessarily inferring an arboreal lifestyle for Megaladapis, with even an aquatic ecology having been suggested following the first fossil discoveries (summarized in Godfrey and Jungers 2002), in the last decades evidence from many postcranial skeletal traits indicates an arboreal lifestyle for koala-lemurs ( Jungers 1978, Wunderlich et al. 1996, Godfrey and Jungers 2002, Godfrey et al. 2016). However, doubts remain con- cerning the type of arboreal behaviour that koala-lemurs engaged in across species. According to an earlier view, all the species belonging to the genus Megaladapis engaged in vertical climbing and clinging, Hence, in this framework, Megaladapis spp. would be represented by size and geographical variants (Carleton 1936, Jungers 1977, 1978). However, more recent inferences hypothe- sized the presence of two sublineages: Megaladapis, including the medium-small sized species Megaladapis madagascariensis and Megaladapis grandidieri, widely distributed in Madagascar, exhib- iting more arboreal habits and mainly featuring vertical climbing (with additional potential behaviours, such as quadrupedal hang- ing and pedal suspension; Wunderlich et al. 1994, 1996, Jungers et al. 2002, 2008), and Peloriadapis, corresponding to Megaladapis edwardsi that, mainly owing to its large body size, has been pro- posed as relatively more terrestrial in comparison to other Megal- adapis species (Wunderlich et al. 1994, 1996, Jungers et al. 2002, 2008). Our results did not yield the informative power allowing us to identify vertical climbing from femoral trabecular bone. This locomotor habit can be ascribed to the arboreal generalist (AG) and/or quadrupedalism/scrambling/clambering (QSC) within our categories [especially the latter, remarkably representing the koala Phascolarctos cinereus (Goldfuss, 1817)], but the trabecular data that we analysed did not reveal differences related to these locomotor styles. Instead, our results provide support for the exis- tence of locomotor differences between small/medium and large koala-lemurs. Focusing on the femoral data, the smallest (Megal- adapis madagascariensis) and the largest (Megaladapis edwardsi) koala-lemurs are distant in the femoral morphospaces, with Megal- adapis edwardsi lying within the SB range of variation [i.e. higher PC2TP+BM scores in the head (Fig. 5B) and higher PC3TP+BM scores in the medial condyle, although close to the ‘non-SB’ range (Fig. 7B)] or even yielding the most extreme score in the direction interpreted as characterizing SB (i.e. extremely high PC3TP+BM score in the lateral condyle; Fig. 6B). In this regard, the results of pDFA for Megaladapis madagascariensis and Megaladapis edwardsi are striking. The smaller taxon Megaladapis madagascariensis was never classified as ‘SB’ across all the iterations, whereas the larger Megaladapis edwardsi was classified as SB in 100% of the iterations, mirroring Palaeopropithecus (see above). The remarkable aspect of this outcome mainly concerns the reconstruction of Trabecular bone of subfossil lemurs    •  15 Megaladapis edwardsi as an extremely SB Palaeopropithecus-like taxon. This represents a new view that, to the best of our knowl- edge, has never been proposed before. In the traditional locomotor reconstruction of koala-lemurs, quadrupedal hanging and pedal suspension have been proposed in addition to the dominant ver- tical climbing (Wunderlich et al. 1994, 1996, Jungers et al. 2002, 2008), but this description did not refer to interspecific differences within Megaladapis, i.e. the wide arboreal repertoire was proposed for the entire genus. Here, instead, we highlight the intrageneric locomotor differences in Megaladapis but not aligning to the pre- viously proposed relatively higher terrestriality of Megaladapis edwardsi (Wunderlich et al. 1994, 1996, Jungers et al. 2002, 2008). Here, the latter instead appears to be adapted substantially to SB arboreal habits; an adaptation that involves greater exploitation of trees, rather than reduced use. We are aware that this is a pre- liminary result, which should be confirmed by future work repre- senting more Megaladapis edwardsi data and more directly testing this hypothesis (e.g. studying femoral diaphyseal cross-sections of koala-lemurs) and owing to the contradictory humeral results. However, Megaladapis edwardsi being a fully suspensory/bridging arboreal species is remarkable considering the reconstructed body mass for this taxon, ∼85 kg ( Jungers et al. 2008), with more recent estimates suggesting an average of ∼65 kg but estimates for single specimens reaching up to ∼76 kg (Thompson et al. 2025). As of today, orangutans are the largest known extinct or extant mam- mals showing SB arboreal habits, with adult males of the largest orangutan species (i.e. Bornean orangutan Pongo pygmaeus) reach- ing 74 ± 9.78 kg (Rayadin and Spehar 2015); estimates that are consistent with older body mass estimates (e.g. 78 kg; Smith and Jungers 1997). Other paradigmatic arboreal SB mammals are char- acterized by smaller body mass (e.g. tree-sloths, 4–6 kg; Nowak 1999), even when other possibly extremely suspensory/bridging large-sized extinct lemurs are considered [Palaeopropithecus spp., ∼41–46 kg ( Jungers et al. 2008); ∼30 kg (Thompson et al. 2025)]. Thus, the extreme SB adaptations of Megaladapis edwardsi would place it within the body size range of Bornean orangutans, making it one of the largest mammals capable of this behaviour, along with all the distinctive morphophysiological adaptations required to sustain such a specialized lifestyle (Granatosky 2022, Nyakatura et al. in press). Archaeoindris is another sloth-lemur (not studied in this work), and it is the largest reconstructed subfossil lemur [i.e. ∼161 kg ( Jungers et  al. 2008); ∼128 kg (Thompson et  al. 2025)]. However, differently from other sloth-lemurs, such as Palaeopropithecus, its locomotion is still poorly understood owing to the few fragmented preserved postcranial elements for which locomotor inferences have been attempted: e.g. slow-moving arboreality, arboreal climbing and clinging, or ground sloth-like scansoriality (Godfrey and Jungers 2002, Godfrey et al. 2016). Retrieving and studying additional postcranial material of Archae- oindris might help to identify new instances of exceptionally large mammals with varying degrees of arboreal adaptation. Final remarks The absent locomotor signal in humeral trabecular bone, as revealed by multivariate inferential analyses (Table 2) and the fact that no visual distinctions between locomotor categories could be found in PC2TP+BM–PC3TP+BM morphospaces (Figs  3B, 4B), contrasts with the femoral trabecular pattern and, in this way, aligns to previous evidence. A discrepant pattern between humeral and femoral anatomy, with the latter being more informative on locomotor habits, arose in the study by Ryan and Shaw (2012), agreeing with outcomes from primate diaphyseal principal moments of area (Carlson 2005). Also, Ryan and Walker (2010) highlighted different patterns between primate humeral and fem- oral trabecular bone. These interlimb trends were previously jus- tified by the fact that in primates the vertical peak ground reaction forces to which hindlimbs are subjected are higher than those experienced by forelimbs (Scherf et al. 2013). This evidence is consistent with indications of higher hindlimb loadings provided by Ryan and Walker (2010) and has been built by studying differ- ent quadrupedal species from almost all primate subgroups (e.g. Kimura et al. 1979, Kimura 1985, 1992, Demes et al. 1994, Schmitt and Hanna 2004, Young 2012). Hence it can be considered a gen- eral primate condition (‘hindlimb dominance’) that is opposite to the one shown by other mammals (Larson 1998, Young 2012). If we consider that our extant mammal sample is largely composed of primates, hindlimb dominance might potentially explain the discrepant results between femoral and humeral trabecular struc- ture. The composition of our sample might be related to another explanation for the absence of locomotor signal in humeral tra- becular bone. Indeed, many studies that found an absent/weak locomotor signal in humeral trabecular structure (see above) focused on distantly related species, representing different phylo- genetic subgroups (primates in the work by Ryan and Walker 2010, Ryan and Shaw 2012), in this regard mirroring our sample (Fig. 1). Hence, humeral trabecular structure might not co-vary strongly with locomotor behaviour when it is analysed across phy- logenetically broad samples. This condition might not allow us to distinguish even highly divergent locomotor styles (e.g. see the strikingly similar humeral trabecular bone in bipeds and quadru- manous climbers in the study by Ryan and Shaw 2012). On the contrary, when restricted taxonomic contexts are analysed, broad-scale effects (e.g. phylogenetic, allometric) might be mini- mized, and functional aspects might be highlighted in forelimb trabecular structure. It would be consistent with the apparent locomotor signal found in humeral trabecular bone when restricted groups, e.g. hominids, were analysed (Scherf et al. 2013, Kivell et al. 2018), contrasting results from works focusing on dis- tantly related taxa (Ryan and Walker 2010, Ryan and Shaw 2012). A similar explanation might justify the absence or limited loco- motor signal in the trabecular structure of other skeletal regions, again in primates (e.g. distal fibula; Alfieri et al. 2025a) and even outside mammals [e.g. squamates, turtles, and crocodiles (Plasse et al. 2019); extant birds (Alfieri et al. 2025b); broad amniote sam- ple (Reinecke and Angielczyk 2025)]. Future work on humeral vs. femoral trabecular structure of more extensive and diverse samples might help to elucidate the locomotor effects in taxonom- ically and phylogenetically broad contexts and how humeral/ femoral patterns differ between primates and non-primates. Apart from these considerations, it should be noted that in the proximal humeral PC2TP+BP–PC3TP+BP biplot, Pongo pygmaeus occupies an outlying position in comparison to the other SB taxa, which tend to show higher PC2TP+BP scores (Fig. 3B). This suggests that the outlying proximal humeral results for this SB species alone 16  •  Alfieri et al. might have prevented detection of significant locomotor effects from PGLSs (P = .21; Table 2). Remarkably, Pongo pygmaeus prox- imal humerus data come from a single specimen (i.e. Pongo pyg- maeus ZSM 1907-609; Supporting Information, Table S1), whose patterns might not be representative of the species (a point that future studies might clarify). Thus, proximal humeral trabecular bone might have informed us about locomotion in a way similar to femoral trabecular traits if another individual or more individ- uals were chosen. Namely, SB would have been distinct compared with all the other taxa, through one single PCTP+BP that is driven by low DA (similarly to femoral data) and by high BV/TV. The latter would represent the opposite pattern compared with the femoral one (i.e. low BV/TV in SB; see above). However, consid- ering the strong loads experienced by the forelimbs in suspensory taxa, even in primates, where this may reverse the typical ‘hindlimb dominance’ characteristic of the group (Granatosky 2016), the humeral pattern could also be interpreted in functional terms (also considering that proximal humeral BV/TV has no relationship with BMp, see Results). In this framework, all subfossil lemurs would occupy the extreme peripheral range of the SB morpho- space region. This would occur especially for Palaepropithecus (both Palaeopropithecus ingens and Palaeopropithecus kelyus), whereas Mesopropithecus and Babakotia would occupy less distinc- tive regions. If more extreme positions in the SB region correspond to more extreme suspensory/bridging adaptations, then the humeral data results for subfossil lemurs would be consistent with previous locomotor reconstruction ( Jungers et  al. 1991, 2002, Simons et al. 1992, Godfrey et al. 1995, Hamrick et al. 2000, God- frey and Jungers 2003, Shapiro et al. 2005, Marchi et al. 2016). Likewise, the more extreme position of Megaladapis madagascar- iensis compared with Megaladapis edwardsi would suggest more extreme suspensory/bridging arboreal habits in the former, small koala-lemur compared with the latter, large koala-lemur. However, this would partly contradict what we proposed based on femoral trabecular structure (see above). Future work investigating already available or newly unearthed postcranial material of subfossil lemurs is needed to clarify these puzzling patterns. It might be time to consider that the exceptional features of subfossil lemurs, such as their unusually large body mass relative to strepsirrhine stan- dards, might make locomotor inference based solely on extant analogues limiting, as has been suggested for other peculiar mam- malian groups (e.g. xenarthrans; Vizcaíno et al. 2018). Finally, the single observations representing exceptions to the main patterns for femoral trabecular structure that we found (Figs  5B, 6B, 7B and above) should briefly be discussed. For instance, the chimpanzee Pan troglodytes, in this work categorized as QSC, often lies within the range interpreted as characterizing SB for femoral head (i.e. higher PC2TP+BP scores), lateral condyle (i.e. higher PC3TP+BP scores) and medial condyle (i.e. higher PC3T- P+BP scores). The chimpanzee shows some SB in its locomotor repertoire (Supporting Information, Table S3; Granatosky 2018, Hunt 2022), which explains its position in the locomotor behaviour PCA, i.e. tending more, compared with Gorilla gorilla, to high PC1 scores, for which higher portions are occupied by SB (Fig. 1A). Also, the black howler Alouatta caraya (AG category, in this work) and the eastern woolly lemur Avahi laniger (L category, in this work) lie within the SB range of variation for proximal femoral and medial condyle data, respectively (Figs 5B, 6B, 7B), the latter also mirrored by Gorilla gorilla (QSC category, in this work). All these taxa, although assigned to other locomotor cate- gories, rarely exhibit either bridging behaviour (Alouatta caraya and Avahi laniger) or suspensory behaviour (e.g. Gorilla gorilla), but not both, as is the case for Pan troglodytes (Supporting Infor- mation, Table S3), which, accordingly, lies within the SB range more often (see above). On the one hand, these patterns support the relationship between femoral trabecular properties and SB locomotion; on the other hand, they highlight again the common issue of locomotor discrete categories. In this work, we designed categories based on distribution on a PCA biplot, in turn deriving from locomotor quantitative data (Fig. 1A), because ultimately we needed discrete categories for the classification technique that we used, i.e. pDFA. However, assigning taxa to a single group inev- itably obscures aspects of their locomotor diversity, such as the suspensory and/or bridging behaviours of African great apes. The latter are particularly difficult to categorize in broad mammalian contexts, because they display distinctive and nearly unique loco- motor adaptations, such as knuckle-walking (Tarrega-Saunders et al. 2021), that resist classification into broad discrete categories. The complexity related to African great ape locomotor categori- zation is also reflected by the fact that their category in this work, i.e. QSC, often occupies wide morphospace regions, diagnostic of broad intra-category variability. Future works should aim to include quantitative behavioural data and analyse them as such, directly analysing the covariation between morphological and locomotor matrices. CO N CLU S I O N In this work, we analysed the humeral and femoral trabecular structure of sloth- and koala-lemurs, addressing this aspect of their anatomy for the first time in a morphofunctional and comparative context, using a broad sample of extant mammals. We found that femoral trabecular structure distinguishes arboreal mammals that predominantly display suspensory and bridging behaviours, such as tree-sloths, lorisids, and orangutans. These taxa exhibit substan- tially more isotropic trabeculae and lower bone volume fraction; traits that can be interpreted functionally as adaptations to their specialized arboreal habits. Accordingly, we found that the largest palaeopropithecine, Palaeopropithecus, was strongly adapted to a suspensory/bridging arboreal repertoire, aligning with previous reconstructions. How- ever, incomplete or missing data prevented confirmation of pre- viously reconstructed behaviours for other palaeopropithecines, namely Babakotia and Mesopropithecus. Regarding koala-lemurs, we could not provide a detailed reconstruction for the small Megaladapis madagascariensis, here only inferred as non- suspensory/bridging. Strikingly, however, we found that the larg- est species, Megaladapis edwardsi, was a specialized suspensory and bridging arboreal mammal. More studies are needed; never- theless, our results provide evidence for one of the largest mam- mals ever known to show adaptations to an arboreal lifestyle characterized by the use of suspensory and bridging behaviour, alongside extant orangutans. These new insights substantiate the use of trabecular bone structure to clarify past locomotor habits https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data Trabecular bone of subfossil lemurs    •  17 and highlight the importance of studying or recovering additional postcranial elements from extinct primates. A CKO W L E D G E M E N TS We thank two anonymous reviewers for their insightful sugges- tions and comments. For allowing visits to museum collections, CT scanning, and providing and/or allowing the download of μCT data we thank: Matthew Skinner, Timothy Ryan (at the time of data acquisition funded by the NSF grant number BCS-0617097), Heiko Temming, the AMNH (New York) Mam- malogy Department, and Joshua Wisor, Frieder Mayer, Christiane Funk, Anna Rosemann, Kristin Mahlow, Martin Kirchner (Mf N, Berlin, Germany), Eva Bärmann and Jan Decher (ZFMK, Bonn, Germany), Frank Zachos and Alexander Bibl (NMW, Vienna, Austria), Guillaume Billet (MNHN, Paris, France), Renaud Leb- run (MRI-ISEM, Montpellier, France), Neil Duncan and Sara Ketelsen (AMNH, New York, USA), Adam Ferguson, William Simpson (FMNH), April Isch Neander and Zhe-Xi Luo (Univer- sity of Chicago, Chicago, IL, USA), Matt Borths, Catherine Riddle (Duke Lemur Center Museum of Natural History) and Justin Gladman (SMIF) (Durham, NC, USA), Rachel Jennings (PC-MER, Birchington-on-Sea, UK), and Linda Gordon (Smith- sonian NMNH). We additionally thank the Bavarian State Col- lection of Zoology (ZSM, Munich, Germany), in particular Anneke H. van Heteren and Gerhard Haszprunar, for access to the objects and the Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany ( Jean-Jacques Hublin), and the Tai Chimpanzee Project (Chris- tophe Boesch). Moreover, we thank the MRI platform member of the national infrastructure France—BioImaging supported by the French National Research Agency (ANR-10-INBS-04, ‘Invest- ment for the future’), the labex CEMEB (ANR-10-LABX-0004), and NUMEV (ANR-10-LABX-0020). We acknowledge John A. Nyakatura (Humboldt-Universität zu Berlin, Germany) and Eli Amson (Staatliches Museum für Naturkunde Stuttgart) for the supervision and support of the PhD project of Fabio Alfieri, which provides key concepts and most of the data for this work. This work was performed in part at the Duke University Shared Mate- rials Instrumentation Facility (SMIF), a member of the North Carolina Research Triangle Nanotechnology Network (RTNN), which is supported by the National Science Foundation (award number ECCS-2025064) as part of the National Nanotechnology Coordinated Infrastructure (NNCI). During part of this work, Fabio Alfieri was part of the Department of Earth Sciences (Uni- versity of Cambridge, Cambridge, UK). AU T H O R CO N T R I B U T I O N S Fabio Alfieri (collection, processing, and analysis of μCT data; per- forming statistical analyses; drafting, writing, and editing of the man- uscript; conceiving the study; designing the sample; planning statistical analysis steps; and interpreting results), Julia Arias-Martorell, Carla Argilés-Esturgó, and Damiano Marchi (conceiving the study; designing the sample; planning statistical analysis steps; interpreting results; and contributing to writing and editing of the manuscript). SU P P L E M E N TA RY DATA Supplementary data is available at Zoological Journal of the Linnean Society online. CO N F L I CT O F I N T E R E ST None declared. F U N D I N G Data acquisition for the PhD project of Fabio Alfieri (which pro- vides most of the data for this work) was financed by Elsa-Neumann-Stipendium des Landes Berlin (Germany), Ger- man Research Council (Deutsche Forschungsgemeinschaft; grant number AM 517/1-1) and Kickstarter Program from RTNN (NC, USA). During this work, Fabio Alfieri was financially sup- ported by the project grant TMPFP3_217022 from the Swiss National Science Foundation (https://www.snf.ch; SNSF Swiss Postdoctoral Fellowships, SPF). Research for this paper was also partly funded by the projects PID2020-116908GB-100/ AEI/10.13039/501100011033/and PID2024-159434NB-I00 granted by the Agencia Estatal de Investigación of the Ministerio de Ciencia e Innovación from Spain. DATA AVA I L A B I L I T Y Raw results and additional information on the studied specimens (Supporting Information, Tables S1 and S2, including Morpho- Source ARK identified to download CT scan data), quantitative locomotor data on the studied species (Supporting Information, Table S3), summary statistics for all the PCAs (Supporting Infor- mation, Tables S4–S14), and additional methods and results (Supporting Information) can be downloaded from https://doi. org/10.6084/m9.figshare.30187675.v1. R E F E R E N CE S Alfieri F. Convergent evolution of humeral and femoral functional mor- phology in slow arboreal mammals. PhD dissertation, Humboldt- Universität zu Berlin, HU, Germany, 2022. Alfieri F, Botton-Divet L, Nyakatura JA et al. Integrative approach uncovers new patterns of ecomorphological convergence in slow arboreal xenar- thrans. Journal of Mammalian Evolution 2022;29:283–312. Alfieri F, Botton-Divet L, Wölfer J et  al. A macroevolutionary common-garden experiment reveals differentially evolvable bone orga- nization levels in slow arboreal mammals. Communications Biology 2023;6:995. Alfieri F, Demuth OE, Steell EM et al. Constrained diversification of avian wing bone internal architecture. Functional Ecology 2025b;39:3611–27. https://doi.org/10.1111/1365-2435.70178 Alfieri F, Nyakatura JA, Amson E. Evolution of bone cortical compactness in slow arboreal mammals. Evolution; International Journal of Organic Evolution 2021;75:542–54. Alfieri F, Veneziano A, Panetta D et al. The relationship between primate distal fibula trabecular architecture and arboreality, phylogeny and size. Journal of Anatomy 2025a;246:907–35. Amson E, Arnold P, van Heteren AH et al. Trabecular architecture in the forelimb epiphyses of extant xenarthrans (Mammalia). Frontiers in Zoology 2017;14:52. https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://www.snf.ch https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://academic.oup.com/zoolinnean/article-lookup/10.1093/zoolinnean/zlag021/#supplementary-data https://doi.org/10.6084/m9.figshare.30187675.v1 https://doi.org/10.6084/m9.figshare.30187675.v1 https://doi.org/10.1111/1365-2435.70178 18  •  Alfieri et al. Amson E, Nyakatura JA. The postcranial musculoskeletal system of xenar- thrans: insights from over two centuries of research and future direc- tions. Journal of Mammalian Evolution 2018;25:459–84. Arias-Martorell J, Zeininger A, Kivell TL. Trabecular structure of the elbow reveals divergence in knuckle-walking biomechanical strategies of Afri- can apes. Evolution; International Journal of Organic Evolution 2021;75:2959–71. Baab KL, Perry JMG, Rohlf FJ et al. Phylogenetic, ecological, and allometric correlates of cranial shape in Malagasy lemuriforms. Evolution; Interna- tional Journal of Organic Evolution 2014;68:1450–68. Barak MM. Bone modeling or bone remodeling: that is the question. Amer- ican Journal of Physical Anthropology 2020;172:153–5. Barak MM, Lieberman DE, Hublin JJ. Of mice, rats and men: trabecular bone architecture in mammals scales to body mass with negative allom- etry. Journal of Structural Biology 2013a;183:123–31. Barak MM, Lieberman DE, Raichlen D et al. Trabecular evidence for a human-like gait in Australopithecus africanus. PLoS One 2013b;8:e77687. Bicca‐Marques JC, Calegaro‐Marques C. Behavioral thermoregulation in a sexually and developmentally dichromatic neotropical primate, the black‐and‐gold howling monkey (Alouatta caraya). American Journal of Physical Anthropology 1998;106:533–46. Biewener AA, Fazzalari NL, Konieczynski DD et al. Adaptive changes in trabecular architecture in relation to functional strain patterns and dis- use. Bone 1996;19:1–8. Boyer DM, Toussaint S, Godinot M. Postcrania of the most primitive eupri- mate and implications for primate origins. Journal of Human Evolution 2017;111:202–15. Boyer DM, Yapuncich GS, Butler JE et al. Evolution of postural diversity in primates as reflected by the size and shape of the medial tibial facet of the talus: evolution of postural diversity in primates. American Journal of Physical Anthropology 2015;157:134–77. Carleton A. The limbbones and vertebrae of the extinct lemurs of Madagascar. Proceedings of the Zoological Society of London 1936;106:281–307. Carlson KJ. Investigating the form-function interface in African apes: Rela- tionships between principal moments of area and positional behaviors in femoral and humeral diaphyses. American Journal of Physical Anthro- pology 2005;127:312–34. Carlson KJ, Lublinsky S, Judex S. Do different locomotor modes during growth modulate trabecular architecture in the murine hind limb? Inte- grative and Comparative Biology 2008;48:385–93. Chang G, Pakin SK, Schweitzer ME et al. Adaptations in trabecular bone microarchitecture in Olympic athletes determined by 7T MRI. Journal of Magnetic Resonance Imaging: JMRI 2008;27:1089–95. Clavel J, Escarguel G, Merceron G. mvmorph: an R package for fitting mul- tivariate evolutionary models to morphometric data. Methods in Ecology and Evolution 2015;6:1311–9. Cotter MM, Simpson SW, Latimer BM et al. Trabecular microarchitecture of hominoid thoracic vertebrae. Anatomical Record (Hoboken, N.J.: 2007) 2009;292:1098–106. Crowley BE. A refined chronology of prehistoric Madagascar and the demise of the megafauna. Quaternary Science Reviews 2010;29:2591–603. Demes B, Carlson KJ, Franz TM. Cutting corners: the dynamics of turning behaviors in two primate species. The Journal of Experimental Biology 2006;209:927–37. Demes B, Jungers WL, Wunderlich RE et al. Body size and leaping kine- matics in Malagasy vertical clingers and leapers. Journal of Human Evo- lution 1996;31:367–88. Demes B, Larson SG, Stern JTJr, et al. The kinetics of primate quadrupe- dalism: “hindlimb drive” reconsidered. Journal of Human Evolution 1994;26:353–74. Domander R, Felder AA, Doube M. BoneJ2 - refactoring established research software. Wellcome Open Research 2021;6:37. Doube M, Klosowski MM, Wiktorowicz-Conroy AM et al. Trabecular bone scales allometrically in mammals and birds. Proceedings. Biological Sci- ences 2011;278:3067–73. Douglass K, Hixon S, Wright HT et al. A critical review of radiocarbon dates clarifies the human settlement of Madagascar. Quaternary Science Reviews 2019;221:105878. Dunn RH. Functional morphology of the postcranial skeleton. In: Croft DA, Su DF, Simpson SW (eds.), Methods in Paleoecology. Reconstructing Cenozoic Terrestrial Environments and Ecological Communities. Cham: Springer, 2018, 23–36. Fajardo RJ, DeSilva JM, Manoharan RK et al. Lumbar vertebral body bone microstructural scaling in small to medium‐sized strepsirhines. Ana- tomical Record (Hoboken, N.J.: 2007) 2013;296:210–26. Fajardo RJ, Müller R. Three-dimensional analysis of nonhuman primate trabecular architecture using micro-computed tomography: primate trabecular bone and the 3-D μCT method. American Journal of Physical Anthropology 2001;115:327–36. Fleagle JG. Primate Adaptation and Evolution, 3rd edn. Academic Press, 2013. Fleagle JG, Baden AL, Gilbert CC. Primate Adaptation and Evolution, 4th edn. Academic Press, 2025. Furnell S, Blanchard ML, Crompton RH et al. Locomotor ecology of Pro- pithecus verreauxi in Kirindy Mitea National Park. Folia Primatologica 2015;86:223–30. Gebo DL. The anatomy of the prosimian foot and its application to the primate fossil record. Ph.D. dissertation, Duke University, 1986. Gebo DL. Locomotor diversity in prosimian primates. American Journal of Primatology 1987;13:271–81. Gebo DL. Vertical clinging and leaping revisited: vertical support use as the ancestral condition of strepsirrhine primates. American Journal of Physical Anthropology 2011;146:323–35. Georgiou L, Dunmore CJ, Bardo A et al. Evidence for habitual climbing in a Pleistocene hominin in South Africa. Proceedings of the National Acad- emy of Sciences of the United States of America 2020;117:8416–23. Georgiou L, Kivell TL, Pahr DH et al. Trabecular bone patterning in the hominoid distal femur. PeerJ 2018;6:e5156. Georgiou L, Kivell TL, Pahr DH et al. Trabecular architecture of the great ape and human femoral head. Journal of Anatomy 2019;234:679–93. Godfrey LR. Adaptive diversification of Malagasy strepsirrhines. Journal of Human Evolution 1988;17:93–134. Godfrey LR, Granatosky MC, Jungers WL. The hands of subfossil lemurs. In: Kivell TL, Lemelin P, Richmond BG et al. (eds.), The Evolution of the Primate Hand. Anatomical, Developmental, Functional, and Paleonto- logical Evidence. New York, NY: Springer, 2016, 421–53. Godfrey LR, Jungers WL. The extinct sloth lemurs of Madagascar. Evolu- tionary Anthropology: Issues, News, and Reviews 2003;12:252–63. Godfrey LR, Jungers WL. Quaternary fossil lemurs. In: Hartwig WC (ed.), The Primate Fossil Record. Cambridge: Cambridge University Press, 2002, 97–121. Godfrey LR, Jungers WL, Burney DA. Subfossil lemurs of Madagascar. In: Werdelin L, Sanders WJ (eds.), Cenozoic Mammals of Africa. University of California Press, 2010, 351–67. Godfrey LR, Jungers WL, Reed KE et al. Subfossil lemurs: inferences about past and present primate communities. In: Goodman SM, Patterson BD (eds.), Natural Change and Human Impact in Madagascar. Wash- ington, DC: Smithsonian Institution Press, 1997, 218–56. Godfrey LR, Sutherland MR, Paine RR et al. Limb joint surface areas and their ratios in Malagasy lemurs and other mammals. American Journal of Physical Anthropology 1995;97:11–36. Gommery D, Ramanivosoa B, Tombomiadana-Raveloson S et al. Une nou- velle espèce de lémurien géant subfossile du Nord-Ouest de Madagas- car (Palaeopropithecus kelyus, Primates). Comptes Rendus Palevol 2009;8:471–80. Gônet J, Laurin M, Hutchinson JR. Evolution of posture in amniotes–Diving into the trabecular architecture of the femoral head. Journal of Evolu- tionary Biology 2023;36:1150–65. Granatosky MC. A mechanical analysis of suspensory locomotion in pri- mates and other mammals. Doctoral dissertation, Duke University, NC, USA, 2016. Trabecular bone of subfossil lemurs    •  19 Granatosky MC. A review of locomotor diversity in mammals with analyses exploring the influence of substrate use, body mass and intermembral index in primates. Journal of Zoology 2018;306:207–16. Granatosky MC. Primate locomotion. In: Vonk J, Shackelford T (eds.), Encyclopedia of Animal Cognition and Behavior. Cham: Springer Inter- national, 2020, 1–7. Granatosky MC. Pedal morphology and locomotor behavior of the sub- fossil lemurs of Madagascar. In: Zeininger A, Hatala KG, Wunderlich RE et al. (eds.), The Evolution of the Primate Foot: Anatomy, Function, and Palaeontological Evidence. Cham: Springer International, 2022, 415–40. Granatosky MC, Miller CE, Boyer DM et al. Lumbar vertebral morphology of flying, gliding, and suspensory mammals: implications for the loco- motor behavior of the subfossil lemurs Palaeopropithecus and Babakotia. Journal of Human Evolution 2014;75:40–52. Hamrick MW, Simons EL, Jungers WL. New wrist bones of Malagasy giant subfossil lemurs. Journal of Human Evolution 2000;38:635–50. Hanna JB, Granatosky MC, Rana P et al. The evolution of vertical climbing in primates: evidence from reaction forces. The Journal of Experimental Biology 2017;220:3039–52. Harrigan TP, Mann RW. Characterization of microstructural anisotropy in orthotropic materials using a second rank tensor. Journal of Materials Science 1984;19:761–7. Herrera JP, Dávalos LM. Phylogeny and divergence times of lemurs inferred with recent and ancient fossils in the tree. Systematic Biology 2016;65:772–91. Hogg RT, Godfrey LR, Schwartz GT et al. Lemur biorhythms and life his- tory evolution. PLoS One 2015;10:e0134210. Huiskes R, Ruimerman R, van Lenthe GH et al. Effects of mechanical forces on maintenance and adaptation of form in trabecular bone. Nature 2000;405:704–6. Hunt KD. Primate suspensory behavior. In: Vonk J, Shackelford TK (eds), Encyclopedia of Animal Cognition and Behavior. Cham: Springer Inter- national, 2022, 5598–604. Ishida H, Hirasaki E, Matano S. Locomotion of the slow loris between discontinuous substrates. In: Matano S, Tuttle RH, Ishida H et al. (eds.), Topics in Primatology. Tokyo: University of Tokyo Press, 1992, 139–52. Ishida H, Jouffroy F, Nakano Y. Comparative dynamics of pronograde and upside down horizontal quadrupedalism in the slow loris (Nycticebus coucang). In: Jouffroy F, Stack M, Niemitz C (eds.), Gravity, Posture and Locomotion in Primates. Firenze: Il Sedicesimo, 1990, 209–20. Jungers WL. Hindlimb and pelvic adaptations to vertical climbing and clinging in Megaladapis, a giant subfossil prosimian from Madagascar. Yearbook of Physical Anthropology 1977;20:508–24. Jungers WL. The functional significance of skeletal allometry in Megalada- pis in comparison to living prosimians. American Journal of Physical Anthropology 1978;49:303–14. Jungers WL. Adaptive diversity in subfossil Malagasy prosimians. Zeitschrift fur Morphologie und Anthropologie 1980;71:177–86. Jungers WL, Demes B, Godfrey LR, How big were the “giant” extinct lemurs of Madagascar? In: Fleagle JG, Gilbert CC (eds.), Elwyn Simons: A Search for Origins. New York: Springer, 2008, 343–60. Jungers WL, Godfrey LF, Simons EL et al. Phylogenetic and functional affinities of Babakotia (Primates), a fossil lemur from northern Mada- gascar. Proceedings of the National Academy of Sciences of the United States of America 1991;88:9082–6. Jungers WL, Godfrey LR, Simons EL et al. Phalangeal curvature and posi- tional behavior in extinct sloth lemurs (Primates, Palaeopropithecidae). Proceedings of the National Academy of Sciences of the United States of America 1997;94:11998–2001. Jungers WL, Godfrey LR, Simons EL et al. Ecomorphology and behavior of giant extinct lemurs from Madagascar. In: Plavcan JM, Kay RF, Jungers WL et al. (eds.), Reconstructing Behavior in the Primate Fossil Record. New York, NY: Kluwer Academic/Plenum, 2002, 371–411. Jungers WL, Lemelin P, Godfrey LR et al. The hands and feet of Archaeole- mur: metrical affinities and their functional significance. Journal of Human Evolution 2005;49:36–55. Karanth KP, Delefosse T, Rakotosamimanana B et al. Ancient DNA from giant extinct lemurs confirms single origin of Malagasy primates. Pro- ceedings of the National Academy of Sciences of the United States of America 2005;102:5090–5. Keaveny TM, Morgan EF, Niebur GL et al. Biomechanics of trabecular bone. Annual Review of Biomedical Engineering 2001;3:307–33. Khosla S, Melton JL, Achenbach SJ et al. Hormonal and biochemical deter- minants of trabecular microstructure at the ultradistal radius in women and men. The Journal of Clinical Endocrinology and Metabolism 2006;91:885–91. Kimura T. Bipedal and quadrupedal walking of primates: comparative dynamics. In: Kondo S (ed.), Primate Morphophysiology, Locomotor Analyses and Human Bipedalism. Tokyo: University of Tokyo Press, 1985, 81–104. Kimura T. Hindlimb dominance during primate high-speed locomotion. Primates 1992;33:465–76. Kimura T, Okada M, Ishida H. Kinesiological characteristics of primate walking: its significance in human walking. In: Morbeck ME, Preus- choft H, Gomberg N (eds.), Environment, Behavior and Morphology: Dynamic Interactions in Primates. New York, NY: G. Fischer, 1979, 297–311. Kistler L, Ratan A, Godfrey LR et al. Comparative and population mitog- enomic analyses of Madagascar’s extinct, giant ‘subfossil’ lemurs. Journal of Human Evolution 2015;79:45–54. Kivell TL. A review of trabecular bone functional adaptation: what have we learned from trabecular analyses in extant hominoids and what can we apply to fossils? Journal of Anatomy 2016;228:569–94. Kivell TL, Davenport R, Hublin J-J et al. Trabecular architecture and joint loading of the proximal humerus in extant hominoids, Ateles, and Aus- tralopithecus africanus. American Journal of Physical Anthropology 2018;167:348–65. Klingenberg CP. Methods for studying allometry in geometric morpho- metrics: a comparison of performance. Evolutionary Ecology 2022;36:439–70. Landi F, Alfieri F, Towle I et al. Fluctuating asymmetry and stress in Macaca fuscata: does captivity affect morphology? Applied Sciences 2021;11:7879. Larson SG. Unique aspects of quadrupedal locomotion in nonhuman pri- mates. In: Strasser E, Fleagle J, Rosenberger A et al. (eds.), Primate Locomotion. New York, NY: Plenum Press, 1998, 157–73. Lieberman DE, Devlin MJ, Pearson OM. Articular area responses to mechanical loading: effects of exercise, age, and skeletal location. Amer- ican Journal of Physical Anthropology 2001;116:266–77. Macho GA, Abel RL, Schutkowski H. Age changes in bone microstructure: do they occur uniformly? International Journal of Osteoarchaeology 2005;15:421–30. MacLatchy L, Müller R. A comparison of the femoral head and neck tra- becular architecture of Galago and Perodicticus using micro-computed tomography (μCT). Journal of Human Evolution 2002;43:89–105. Maquer G, Musy SN, Wandel J et al. Bone volume fraction and fabric anisot- ropy are better determinants of trabecular bone stiffness than other morphological variables. Journal of Bone and Mineral Research 2015;30:1000–8. Marchi D, Ruff CB, Capobianco A et al. The locomotion of Babakotia rad- ofilai inferred from epiphyseal and diaphyseal morphology of the humerus and femur. Journal of Morphology 2016;277:1199–218. Marciniak S, Mughal MR, Godfrey LR et al. Evolutionary and phylogenetic insights from a nuclear genome sequence of the extinct, giant, “subfos- sil” koala lemur Megaladapis edwardsi. Proceedings of the National Acad- emy of Sciences of the United States of America 2021;118:e2022117118. Marshall SK, Spainhower KB, Sinn BT et al. Hind limb bone proportions reveal unexpected morphofunctional diversification in xenarthrans. Journal of Mammalian Evolution 2021;28:599–619. Martin RD. Origins, diversity and relationships of lemurs. International Journal of Primatology 2000;21:1021–49. Mielke M, Wölfer J, Arnold P et al. Trabecular architecture in the sciuro- morph femoral head: allometry and functional adaptation. Zoological Letters 2018;4:10. 20  •  Alfieri et al. Modlesky CM, Subramanian P, Miller F. Underdeveloped trabecular bone microarchitecture is detected in children with cerebral palsy using high-resolution magnetic resonance imaging. Osteoporosis International 2008;19:169–76. Monclús-Gonzalo O, Alba DM, Duhamel A et al. Early euprimates already had a diverse locomotor repertoire: evidence from ankle bone mor- phology. Journal of Human Evolution 2023;181:103395. Monclús-Gonzalo O, Alba DM, Fabre AC et al. Reconstruction of the loco- motor repertoire of early primates in the light of astragalar and calcaneal shape. Journal of Human Evolution 2025;206:103730. Muldoon KM. Paleoenvironment of Ankilitelo Cave (late Holocene, south- western Madagascar): implications for the extinction of giant lemurs. Journal of Human Evolution 2010;58:338–52. Muldoon KM, Godfrey LR, Jungers WL et al. Geographic patterning in subfossil primate community dynamics in Madagascar. American Jour- nal of Physical Anthropology 2009;S48:195. Nishimura T, Morimoto N, Ito T. Shape variation in the facial part of the cranium in macaques and African papionins using geometric morpho- metrics. Primates; Journal of Primatology 2019;60:401–19. Nowak RM. Walker’s Mammals of the World (Vol.1). Baltimore: Johns Hop- kins University Press, 1999. Nyakatura JA. The convergent evolution of suspensory posture and locomotion in tree sloths. Journal of Mammalian Evolution 2012;19:225–34. Nyakatura JA, Alfieri F, Granatosky MC et al. Upside-down and in slow motion: the functional morphology and mechanics of sloth suspensory locomotion. In: Granatosky MC, , Toussaint S, Youlatos D (eds.), Func- tional Morphology and Biomechanics of Arboreal Locomotion in Tetrapods. Springer Nature Editions, in press. https://doi.org/10.13140/ RG.2.2.35257.89449 Orlando L, Calvignac S, Schnebelen C et al. DNA from extinct giant lemurs links archaeolemurids to extant indriids. BMC Evolutionary Biology 2008;8:121–9. Patel BA, Goodenberger KE, Boyer DM et al. Hallucal reduction in sloth lemurs and morphological convergence on orangutans by Palaopro- pithecus. American Journal of Physical Anthropology 2013;56(Suppl.):217. Pearson OM, Lieberman DE. The aging of Wolff ’s “law”: ontogeny and responses to mechanical loading in cortical bone. American Journal of Physical Anthropology 2004;(Suppl. 39):63–99. Pinheiro J, Bates D, DebRoy S, Sarkar D. nlme: linear and nonlinear mixed effects models. R package version 3.1-147, 2020. https://rdrr.io/ cran/nlme/. Plasse M, Amson E, Bardin J et al. Trabecular architecture in the humeral metaphyses of non‐avian reptiles (Crocodylia, Squamata and Testu- dines): lifestyle, allometry and phylogeny. Journal of Morphology 2019;280:982–98. Polly PD, Lawing AM, Fabre AC et al. Phylogenetic principal components analysis and geometric morphometrics. Hystrix, the Italian Journal of Mammalogy 2013;24:33–41. Pontzer H, Lieberman DE, Momin E et al. Trabecular bone in the bird knee responds with high sensitivity to changes in load orientation. The Jour- nal of Experimental Biology 2006;209:57–65. Prates HM, Hass GP, Bicca-Marques JC. Ranging behavior of black-and- gold howler monkeys (Alouatta caraya) in anthropogenic habitat patch in Southern Brazil. In: Urbani B, Kowalewski M, Cunha RGT et al. (eds.), La Primatología En Latinoamérica 2 - A Primatologia Na America Latina 2. Tomo I Argentina-Colombia. Ediciones IVIC. Caracas, Venezu- ela: Instituto Venezolano de Investigaciones Científicas (IVIC), 2018, 259–66. Qiao Z, Zhou L, Huang JZ. Sparse linear discriminant analysis with appli- cations to high dimensional low sample size data. International Journal of Applied Mathematics 2009;39:48–60. R Core Team. R: a Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. 2023. https:// www.R-project.org/. Rafferty KL, Ruff CB. Articular structure and function in Hylobates, Colo- bus, and Papio. American Journal of Physical Anthropology 1994;94:395–408. Rayadin Y, Spehar SN. Body mass of wild Bornean orangutans living in human‐dominated landscapes: implications for understanding their ecology and conservation. American Journal of Physical Anthropology 2015;157:339–46. Reinecke TE, Angielczyk KD. Exploring functional and phylogenetic vari- ation in limb trabecular structure across amniote taxa. Biological Journal of the Linnean Society 2025;146:blaf073. Revell LJ. phytools: an R package for phylogenetic comparative biology (and other things). Methods in Ecology and Evolution 2012;3:217–23. Rolvien T, Hahn M, Siebert U et al. Vertebral bone microarchitecture and osteocyte characteristics of three toothed whale species with varying diving behaviour. Scientific Reports 2017;7:1604. Rook L, Bondioli L, Köhler M et al. Oreopithecus was a bipedal ape after all: evidence from the iliac cancellous architecture. Proceedings of the National Academy of Sciences of the United States of America 1999;96:8795–9. Rowe N. The Pictorial Guide to the Living Primates. East Hampton, NY: Pogonias Press, 1996. Rueden CT, Schindelin J, Hiner MC et al. ImageJ2: ImageJ for the next generation of scientific image data. BMC Bioinformatics 2017;18:529. Ruff C. Hindlimb articular surface allometry in hominoidea and Macaca, with comparisons to diaphyseal scaling. Journal of Human Evolution 1988;17:687–714. Runestad JA. Postcranial adaptations for climbing in Loridae (Primates). Journal of Zoology 1997;242:261–90. Ryan TM, Carlson KJ, Gordon AD et al. Human-like hip joint loading in Australopithecus africanus and Paranthropus robustus. Journal of Human Evolution 2018;121:12–24. Ryan TM, Ketcham RA. Femoral head trabecular bone structure in two omomyid primates. Journal of Human Evolution 2002a;43:241–63. Ryan TM, Ketcham RA. The three-dimensional structure of trabecular bone in the femoral head of strepsirrhine primates. Journal of Human Evolution 2002b;43:1–26. Ryan TM, Ketcham RA. Angular orientation of trabecular bone in the fem- oral head and its relationship to hip joint loads in leaping primates. Journal of Morphology 2005;265:249–63. Ryan TM, Krovitz GE. Trabecular bone ontogeny in the human proximal femur. Journal of Human Evolution 2006;51:591–602. Ryan TM, Raichlen DA, Gosman JH, Structural and mechanical changes in trabecular bone during early development in the human femur and humerus. In: Percival CJ, Richtsmeier JT (eds.), Building Bones: Bone Formation and Development in Anthropology. Cambridge: Cambridge University Press, 2017, 281–302. Ryan TM, Shaw CN. Unique suites of trabecular bone features characterize locomotor behavior in human and non-human anthropoid primates. PLoS One 2012;7:e41037. Ryan TM, Shaw CN. Trabecular bone microstructure scales allometrically in the primate humerus and femur. Proceedings. Biological Sciences 2013;280:20130172. Ryan TM, Shaw CN. Gracility of the modern Homo sapiens skeleton is the result of decreased biomechanical loading. Proceedings of the National Academy of Sciences of the United States of America 2015;112:372–7. Ryan TM, Walker A. Trabecular bone structure in the humeral and femoral heads of anthropoid primates. Anatomical Record (Hoboken, N.J.: 2007) 2010;293:719–29. Saparin P, Scherf H, Hublin JJ et al. Structural adaptation of trabecular bone revealed by position resolved analysis of proximal femora of different primates. Anatomical Record (Hoboken, N.J.: 2007) 2011;294:55–67. Scherf H, Harvati K, Hublin J-J. A comparison of proximal humeral can- cellous bone of great apes and humans. Journal of Human Evolution 2013;65:29–38. Schmitt D, Hanna JB. Substrate alters forelimb to hindlimb peak force ratios in primates. Journal of Human Evolution 2004;46:239–54. Shapiro LJ, Jungers WL, Godfrey LR et al. Vertebral morphology of extinct lemurs. American Journal of Physical Anthropology 1994;37(Suppl):179–80. Shapiro LJ, Seiffert CVM, Godfrey LR et al. Morphometric analysis of lum- bar vertebrae in extinct Malagasy strepsirrhines. American Journal of Physical Anthropology 2005;128:823–39. https://doi.org/10.13140/RG.2.2.35257.89449 https://doi.org/10.13140/RG.2.2.35257.89449 https://rdrr.io/cran/nlme/ https://rdrr.io/cran/nlme/ https://www.R-project.org/ https://www.R-project.org/ Trabecular bone of subfossil lemurs    •  21 Simons EL, Godfrey LR, Jungers WL et al. A new giant subfossil lemur, Babakotia, and the evolution of the sloth lemurs. Folia Primatologica 1992;58:197–203. Smith R, Jungers W. Body mass in comparative primatology. Journal of Human Evolution 1997;32:523–59. Sokal R, Rohlf F. Biometry: The Principles and Practice of Statistics in Biolog- ical Research, 3rd edn. New York, NY: Freeman, 1995. Somers KM. Allometry, isometry and shape in principal components anal- ysis. Systematic Zoology 1989;38:169–73. Spoor F, Garland T, Krovitz G et al. The primate semicircular canal system and locomotion. Proceedings of the National Academy of Sciences of the United States of America 2007;104:10808–12. Stauber M, Rapillard L, van Lenthe GH et al. Importance of individual rods and plates in the assessment of bone quality and their contribu- tion to bone stiffness. Journal of Bone and Mineral Research 2006;21: 586–95. Szalay FS, Delson E. Evolutionary History of the Primates. New York, NY: Academic Press, 1979. Tarrega-Saunders ELR, King C, Roberts AM et al. Knuckle-walking and behavioural flexibility in great apes. Revue de Primatologie 2021;12. Tattersall I. Subfossil lemuroids and the “adaptive radiation” of the Malagasy lemurs. Transactions of the New York Academy of Sciences 1973;35:314–24. Tattersall I. Notes on the cranial anatomy of the subfossil malagasy lemurs. In: Tattersall I, Sussman RW (eds.), Lemur Biology. New York, NY: Plenum Press, 1975, 111–24. Tattersall I. The Primates of Madagascar. New York, NY: Columbia Univer- sity Press, 1982. Tattersall I, Schwartz JH. Craniodental morphology and systematics of the Malagasy lemurs (Primates, Prosimii). Anthropological Papers of the American Museum of Natural History 1974;52:137–92. Thompson KE, Rothman RS, Polk JD et al. Revised body mass estimates for extinct lemurs. American Journal of Biological Anthropology 2025;188:e70158. Thorpe SK, Crompton RH. Locomotor ecology of wild orangutans (Pongo pygmaeus abelii) in the Gunung Leuser Ecosystem, Sumatra, Indonesia: a multivariate analysis using log‐linear modelling. American Journal of Physical Anthropology 2005;127:58–78. Tsegai ZJ, Kivell TL, Gross T et al. Trabecular bone structure correlates with hand posture and use in hominoids. PLoS One 2013;8:e78781. Tsegai ZJ, Skinner MM, Pahr DH et al. Systemic patterns of trabecular bone across the human and chimpanzee skeleton. Journal of Anatomy 2018a;232:641–56. Tsegai ZJ, Skinner MM, Pahr DH et al. Ontogeny and variability of trabec- ular bone in the chimpanzee humerus, femur and tibia. American Journal of Physical Anthropology 2018b;167:713–36. Upham NS, Esselstyn JA, Jetz W. Inferring the mammal tree: species-level sets of phylogenies for questions in ecology, evolution, and conserva- tion. PLoS Biology 2019;17:e3000494. Vizcaíno SF, Toledo N, Bargo MS. Advantages and limitations in the use of extant xenarthrans (Mammalia) as morphological models for paleobio- logical reconstruction. Journal of Mammalian Evolution 2018;25:495–505. Vuillame-Randriamanantena M, Godfrey LR, Jungers WL et al. Morphol- ogy, taxonomy and distribution of Megaladapis—giant subfossil lemur from Madagascar. Comptes Rendus de l’Académie des Sciences Série III 1992;315:1835–42. Walker A. Locomotor adaptations in recent and fossil Madagascan lemurs. PhD Thesis, University of London, 1967. Walker A. Locomotor adaptations in past and present prosimian primates. In: Jenkins FA (ed.), Primate Locomotion. New York, NY: Academic Press, 1974, 349–81. Walker A. Prosimian locomotor behavior. In: Doyle GA, Martin RD (eds.), The Study of Prosimian Behavior. New York: Academic Press, 1979, 543–65. Walker A, Ryan TM, Silcox MT et al. The semicircular canal system and locomotion: the case of extinct lemuroids and lorisoids. Evolutionary Anthropology: Issues, News, and Reviews 2008;17:135–45. Widrig K, Alfieri F, Kuo PC et al. Quantitative analysis of stem-palaeognath flight capabilities sheds light on ratite dispersal and flight loss. Biology Letters 2025;21:20250320. Wunderlich RE, Jungers WL, Godfrey LR et al. Pedal form and function in subfossil indroids. American Journal of Physical Anthropology 1994;Suppl 18:211. Wunderlich RE, Simons EL, Jungers WL. New pedal remains of Megalada- pis and their functional significance. American Journal of Physical Anthro- pology 1996;100:115–39. Yapuncich GS, Granatosky MC. Footloose: articular surface morphology and joint movement potential in the ankles of lorisids and cheirogaleids. American Journal of Physical Anthropology 2021;175:876–94. Yoder AD, Rakotosamimanana B, Parsons TJ. Ancient DNA in subfossil lemurs. In: B Rakotosamimanana, Rasamimanana H, Ganzhorn JU et al. (eds.), New Directions in Lemur Studies. New York: Springer, 1999, 1–17. Youlatos D, Pylarinos D, Karantanis NE et al. Locomotion, postures, and substrate use in captive southern pygmy slow lorises (Strepsirrhini, Primates): implications for conservation. Animals 2025;15:1576. Young JW. Ontogeny of limb force distribution in squirrel monkeys (Saimiri boliviensis): insights into the mechanical bases of primate hind limb dominance. Journal of Human Evolution 2012;62:473–85. Zihlman AL, Mcfarland RK, Underwood CE. Functional anatomy and adaptation of male gorillas (Gorilla gorilla gorilla) with comparison to male orangutans (Pongo pygmaeus). Anatomical Record (Hoboken, N.J.: 2007) 2011;294:1842–55. Active Content List INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION AUTHOR CONTRIBUTIONS SUPPLEMENTARY DATA CONFLICT OF INTEREST FUNDING DATA AVAILABILITY