The locomotor behavior of subfossil Malagasy sloth-lemurs (Strepsirrhini: Palaeopropithecidae) and koala-lemurs (Strepsirrhini: Megaladapidae): new insights from limb trabecular bone
Fabio AlfieriiD, Julia Arias‐MartorelliD, Carla Argilés-EsturgóiD, Damiano MarchiiD
Abstract
The locomotion of Malagasy Quaternary subfossil lemurs--including palaeopropithecids ( sloth-lemurs) and megaladapids ( koala-lemurs)--has been investigated on abundant postcranial remains. Proposed strategies include some without living primate parallels, including sloth-like suspensory arboreality in palaeopropithecids, although the degree of suspensory behaviour in palaeopropithecids, or locomotor diversity in koala-lemurs are poorly understood. Differently from external morphology, internal bone structure in these taxa is largely unexplored. We compared the humeral and femoral trabecular architecture of sloth- and koala-lemurs to several extant mammals, studying spherical 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 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 palaeopropithecids, especially Palaeopropithecus--confirming earlier reconstructions--but also in Megaladapis edwarsi, a striking result that would make M. edwarsi the largest mammal ever known to adopt such habits, approached only by extant orangutans. This work highlights the potential of internal bone structure for reconstructing primate locomotor evolution.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 3
- FWCI
- 11.70
- Field percentile
- top 2%
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (22 Oct 2025) · CC BY 4.0
Low femoral trabecular anisotropy and bone volume fraction mark suspensory/bridging mammals, and a phylogenetic discriminant model classifies extinct Palaeopropithecus ingens and Megaladapis edwarsi as suspensory.
Locomotor reconstructions of extinct Malagasy sloth-lemurs and koala-lemurs have rested mostly on external bone morphology, leaving the degree of palaeopropithecid suspension and locomotor differences among Megaladapis species unresolved. This work brings limb trabecular architecture, argued to reflect realized loading rather than potential joint range, into a phylogeny- and size-controlled comparison with extant mammals to test those hypotheses.
Key results
ShownIn extant mammals, only femoral trabecular data carried a locomotor signal. Humeral trabecular structure tracked body size but not locomotion once phylogeny and size were controlled.
mvPGLS and phylogenetic MANCOVA on PCTP1-PCTP4 (93.4%-96.4% of variance): humeral data related to the body mass proxy but not locomotion; lateral and medial condyle data significantly related to locomotion; proximal femur p=0.09. Locomotor categories were unrelated to the body mass proxy (p=0.162 humerus, p=0.476 femur). · Relationship between trabecular structure and…; Results; Humeral trabecular structure; Femoral trabecular structure
ShownSuspensory/bridging taxa (lorisids, tree sloths, orangutans) differ from other extant mammals in having lower proximal femoral bone volume fraction (BV/TV) and more isotropic trabeculae (lower degree of anisotropy, DA) at the femoral head and both condyles.
Proximal femoral PC2 driven by DA (R^2=0.69) and BV/TV (R^2=0.47); condylar PC3s driven mainly by DA. Univariate PGLS and phylogenetic ANOVA confirmed SB vs. others; the combined four-trait dataset separated SB from others (mvPGLS and phylogenetic MANOVA p=0.002). · Femoral trabecular structure; Phylogenetic discriminant function analysis; Locomotor signal in extant mammal trabecular…
ShownA phylogenetic discriminant function analysis on head BV/TV and DA plus lateral and medial condyle DA sorted extant taxa into suspensory/bridging versus other with good cross-validated accuracy.
Leave-one-out cross-validation misclassification rate 0.14 (86% of held-out taxa correctly predicted); one discriminant function (DF1), with lower values for SB taxa, positively correlated with all four predictors. · Phylogenetic Discriminant Function Analysis; Phylogenetic discriminant function analysis; Locomotor signal in extant mammal trabecular…
ShownFemoral trabecular data put Palaeopropithecus in the suspensory/bridging class, matching earlier reconstructions from external anatomy. The smaller sloth-lemurs could not be assessed.
P. ingens and/or P. maximus fell in the SB range in all femoral morphospaces. P. ingens, the only sloth-lemur with head and both condyle data, was classified SB in 100% of iterations. Babakotia and Mesopropithecus had only humeral data. · Phylogenetic discriminant function analysis; The locomotor behavior of palaeopropithecids
SuggestedWithin koala-lemurs, the large Megaladapis edwarsi was classified as suspensory/bridging and the small M. madagascariensis was not. This contradicts the earlier proposal that M. edwarsi was more terrestrial and would make it a suspensory mammal at or above orangutan size.
pDFA: M. edwarsi classified SB in 100% of iterations, M. madagascariensis never. M. edwarsi is reconstructed at ca. 85 kg vs. 74 ± 9.78 kg for adult male Bornean orangutans, 4-6 kg for tree sloths and ca. 41-46 kg for Palaeopropithecus. The authors call this result preliminary. · Phylogenetic discriminant function analysis; The locomotor behavior of Megaladapis spp
Methods, data and reproducibility
Spherical trabecular volumes of interest from μCT scans of humeral and femoral epiphyses were segmented in ImageJ2 and quantified with BoneJ2, reduced by PCA, and tested against PCA-derived locomotor categories using phylogenetic GLS and MANCOVA with a body mass covariate. Selected femoral traits fed a phylogenetic discriminant function analysis, which was then applied to subfossil lemurs.
- Subfossil Palaeopropithecus humerus and femur μCT scans (MNHN Paris; Duke Lemur Center Division of Fossil Primates), assigned to known species · Palaeopropithecus spp. (extinct sloth-lemur) · n = four complete humeri, four complete femora, plus partial bones
- Subfossil Babakotia humerus μCT scans (assigned to Babakotia radofilai); no femora available · Babakotia radofilai (extinct sloth-lemur) · n = one complete and one distal humerus
- Subfossil Mesopropithecus dolichobrachion humerus μCT scan; no femora available · Mesopropithecus dolichobrachion (extinct sloth-lemur) · n = one complete humerus
- Subfossil Megaladapis humerus and femur μCT scans, assigned to species including M. madagascariensis and M. edwarsi · Megaladapis spp. (extinct koala-lemur) · n = three complete humeri, two complete femora, plus partial bones
- Extant mammal humerus and femur μCT data (tree sloths, koala, lorisids, galagids, indriids, lemurids, great apes, Alouatta caraya) from earlier studies and MorphoSource · multiple extant mammals (mainly primates, plus sloths and koala) · n = 55 humeri and 53 femora · MorphoSource ARK identifiers listed in Tables S1-S2
- Quantitative locomotor behavior frequencies from Granatosky (2018), pooled into broader categories and arcsine square-root transformed · multiple extant mammal species · n = 62% of the extant species studied
- Time-calibrated mammalian phylogeny adapted to the study sample (Appendix S8) · multiple species
- Methods & tools
- micro-computed tomography (μCT), ImageJ2 (thresholding and segmentation), BoneJ2, R 4.3.1, mvMORPH (mvgls, manova.gls, mvgls.dfa), phytools (phylANOVA), nlme (gls), principal component analysis, phylogenetic discriminant function analysis, leave-one-out cross-validation, MorphoSource
- Reagents & environment
- R 4.3.1 with mvMORPH, phytools and nlme; ImageJ2 and BoneJ2 for segmentation and trabecular metrics (no package versions reported).
- Reproducibility gaps
- No analysis code repository is cited. Segmentation of subfossil volumes relied partly on manual cleaning, and the sizes of reduced volumes of interest are documented only in supplementary appendices (S5-S7).
Limitations (6)
- Babakotia and Mesopropithecus were represented only by humeri, and humeri showed no locomotor signal in extant mammals. Their degree of suspensory specialization therefore could not be evaluated. · stated by the authors
- The suspensory inference for M. edwarsi is called preliminary. It needs more M. edwarsi material and direct tests such as femoral diaphyseal cross-sections, and humeral data contradict it by placing M. madagascariensis in a more extreme position. · stated by the authors
- Proximal humeral data for Pongo pygmaeus came from a single specimen. Its outlying position may have hidden a humeral locomotor effect (PGLS p-value 0.21). · stated by the authors
- Discrete locomotor categories hide behavioral diversity: chimpanzees (QSC) often fall in the SB range. The authors also caution that subfossil lemurs' unusual body sizes may limit inference based only on extant analogues. · stated by the authors
- Fossil classification rests on very few femora: only three subfossil species had all femoral regions. Uncertainty was estimated by subsampling the extant training set, not by specimen-level variation within the extinct taxa. · noted in this summary
- Subfossil humeral outliers were driven by high BV/TV, and subfossil volumes needed separate segmentation procedures. Possible diagenetic or infilling effects on BV/TV and DA were not tested directly. · noted in this summary
Who should read:Read this if you infer locomotion in extinct primates or other mammals from trabecular bone, or work on the paleobiology of palaeopropithecid and Megaladapis lemurs.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Locomotor Trends Among Early Mammals Illuminated by Predictors of Arboreality from the Mammalian Appendicular Skeleton
Jonathan A. Nations, Lucas N. Weaver, David M. Grossnickle·bioRxiv·Jul 9, 2025
- Evolution of bone cortical compactness in slow arboreal mammals
F. Alfieri, J. Nyakatura, E. Amson·bioRxiv·Sep 2, 2020
- Palaeobiological inferences based on long bone epiphyseal and diaphyseal structure - the forelimb of xenarthrans (Mammalia)
Eli Amson, John A Nyakatura·bioRxiv·Sep 21, 2018
- Size and locomotor ecology have differing effects on the external and internal morphologies of squirrel (Rodentia: Sciuridae) limb bones
J. Rickman, A. E. Burtner, … C. J. Law·bioRxiv·Feb 8, 2023
- Morphological characteristics of the forelimb flexor digitorum muscles and their distinctive innervation pattern in koalas (Phascolarctos cinereus): implications for function and morphogenesis
Daiki Shigemitsu, Saori Anetai, … Sayaka Tojima·bioRxiv·Sep 12, 2026
- Kinematics and morphological correlates of descent strategies in arboreal mammals suggest early upright postures in euprimates
Séverine Toussaint, Dionisios Youlatos, John A. Nyakatura·bioRxiv·Jul 23, 2025