Sauropod teeth are often described within a taxonomic framework, with few studies focusing on heterodonty or intraspecific morphological variation. The Lagerstätte of Angeac-Charente has yielded 167 heart-shaped isolated teeth from a single turiasaurian species. Observations of this material during its collection in the field, preparation, and recording in the collections suggest the presence of possible heterodonty in this species. In this study, we analyze this heterodonty using both comparative anatomy and 3D geometric morphometric analyses coupled with machine learning algorithms. Our principal objective is to determine whether one of the two approaches-quantitative or qualitative-is more effective than the other and whether they can be complementary studying heterodonty. The results confirm the presence of heterodonty, with five different morphotypes, which correspond to premaxillary, maxillary anterior, maxillary posterior, anterior dentary, and posterior dentary teeth. An anteroposteriorly decreasing size gradient in the jaw is observed. Qualitative comparative anatomy is the quickest method, with few conflicts in morphotyping, while quantitative approaches might resolve ambiguities through more precise and replicable analyses. Together, these methods provide complementary insights and allow retrospective positioning of teeth within the jaw. This study, therefore, allows for the reliable assignment of teeth to specific regions or positions in turiasaurian sauropods. Such assignment work is also necessary for many other taxa, and this method will, therefore, be useful for future research. Indeed, it is fully applicable to other sauropod taxa, theropods, or crocodilians.

Passive Acoustic Monitoring (PAM) is widely used for biodiversity assessment. Its application in African tropical forests is limited by scarce annotated data, reducing the performance of general-purpose ecoacoustic models on underrepresented taxa. In this study, we introduce DeepForestSound (DFS), a multi-species automatic detection model designed for PAM in African tropical forests. DFS relies on a semi-supervised pipeline combining clustering of unannotated recordings with manual validation, followed by supervised fine-tuning of an Audio Spectrogram Transformer (AST) using low-rank adaptation, which is compared to a frozen-backbone linear baseline (DFS-Linear). The framework supports the detection of multiple taxonomic groups, including birds, primates, and elephants, from long-term acoustic recordings. DFS was trained on acoustic data collected in the Sebitoli area, in Kibale National Park, Uganda, and evaluated on an independent dataset recorded two years later at different locations within the same forest. This evaluation therefore assesses generalization across time and recording sites within a single tropical forest ecosystem. Across 8 out of 12 taxons, DFS outperforms existing automatic detection tools, particularly for non-avian taxa, achieving average AP values of 0.964 for primates and 0.961 for elephants. Results further show that LoRA-based fine-tuning substantially outperforms linear probing across taxa. Overall, these results demonstrate that task-oriented, region-specific training substantially improves detection performance in acoustically complex tropical environments, and highlight the potential of DFS as a practical tool for biodiversity monitoring and conservation in African rainforests.

The rodent genus Praomys, endemic to the Afrotropical region, is characterized by significant diversity including cryptic species that complicate its taxonomic classification. Recent fieldwork conducted in Fazao-Malfakassa National Park in Togo, combined with molecular and morphometric analyses of newly collected individuals and museum specimens, has led to the discovery and formal description of a new species within the Praomys tullbergi complex endemic to Togo and Benin. Morphological analyses, based on 22 cranial and dental measurements from 263 adult specimens of the P. tullbergi complex, revealed clear differentiation of P. sp. nov. from other members of this complex. The new species can be distinguished by several external and cranio-mandibular characters. Molecular phylogenetic analyses and diagnostic nucleotide characters using mitochondrial cytochrome b sequences confirmed the genetic distinctiveness of P. sp. nov. within the P. tullbergi complex. In Togo and Benin, the new species co-occurs with another species of the P. tullbergi complex, P. misonnei, but the 2 species are separated by morphological, molecular, and ecological traits. The formal description of P. sp. nov. broadens and clarifies the taxonomy of the P. tullbergi complex.Nomenclatural statement: A Life Science Identifier (LSID) number was obtained for this publication: urn: lsid: zoobank.org: pub: E9170039-0D79-4125-9643-05501E1B4F4F

High species richness in tropical West Africa has been demonstrated for many species groups, among which shrews are no exception. Within the Crocidura poensis species complex, six species are currently described and recognized in West Africa, but a recent study suggested the existence of an additional cryptic species based on dorsal skull morphology. Here, an integrative approach combining the complete mitochondrial genome, eight nuclear markers, external morphology and geometric morphometrics methods on the skull and mandible of the C. poensis species complex distributed across West Africa is used to test the validity of the new candidate species. Species delimitation analyses performed separately on nuclear and mitochondrial DNA support the existence of seven species. Despite low genetic distance with its closest relative, the new species can be distinguished by several diagnostic nucleotide characters using cytochrome b sequences, by morphometric analyses on the skull and mandible as well as traditional external measurements. This allowed us to describe the new species as Crocidura pediculus Voet, Cornette & Nicolas sp. nov.

The forelimbs of mammals are involved in many crucial behaviours for an animal's ecology, including locomotion. It has been shown that forelimb morphology and locomotor mechanisms are greatly impacted by functional constraints induced by the properties of the media across or in which the animal moves. These functional constraints are thought to drive an important part of bone shape, as bone directly remodels in response to both muscle and external forces. Due to its anatomical particularities, the forelimb of fully fossorial moles is of particular interest to better understand fossorial adaptations and has already been studied extensively. Recently, some studies focusing on two European mole species, Talpa europaea and the recently described Talpa aquitania, highlighted inter and intraspecific variations in the inner ear and forelimb bones morphology, which could be linked to locomotor performances. To better understand the specificity of their fossorial adaptations, we focus, in the present study, on these two species' musculature. Performing anatomical dissections, and in accordance with the literature, we provided a redescription of 39 extrinsic and intrinsic muscles inserting on the forelimb bones. We also made a quantitative comparison of muscles features, both at the inter and intraspecific level. Especially, we focus on muscle Physiological Cross-Sectional Area (PCSA), a measure considered as a good estimator of the force-producing capacity of a muscle. Finally, to investigate relationships between bone shape and muscle force, we quantified covariations between shape data and muscle PCSA. Our results highlighted inter and intraspecific (in T. aquitania) variations in muscle force-producing capacity, which is consistent with the studies of the inner ear and forelimb bones. Focusing on the relationship between muscle PCSA and shape of ulna and humerus, we showed that shoulder extensors, carpal/digital extensors, and carpal/digital flexors' PCSA were highly integrated with humerus shape and that elbow extensors' PCSA were highly integrated with the ulna shape. These results are rather consistent with insertion sites of these muscles on bones.

Geographic patterns in head morphology can reveal important aspects of population structure and phenotypic differentiation in marine species. We investigated these patterns in the broad-nosed pipefish Syngnathus typhle Linnaeus, 1758, a coastal species widely distributed across distinct marine regions (Baltic, North, East Atlantic, and Mediterranean seas). Using landmark-based geometric morphometrics, we quantified head morphology variation in 241 live, wild-caught adults. Analyses revealed consistent geographic patterns in snout elongation, head depth, and eye position. Multivariate comparisons showed significant morphological separation among marine regions, with high classification success based solely on head morphology. These results indicate that S. typhle populations are spatially structured and not morphologically homogeneous across their range. Our findings demonstrate the utility of geometric morphometrics to detect subtle geographic variation in morphologically conservative taxa such as pipefish and provide a quantitative baseline for future research on phenotypic diversity and conservation management. Importantly, the study is based on non-invasive monitoring protocols using live animals.

Hypotheses linking climate change to hominin evolution are based on the observation that many events of speciation, extinction, and morphological and behavioural evolution coincided with periods of increased climatic variability and habitat instability. Fossil fauna recovered from palaeoanthropological sites can be used to reconstruct past climatic and environmental conditions associated with hominin assemblages and to explore correlative changes through time. However, temporal correlations between global climate change patterns and local environmental proxies from hominin-bearing deposits are not always verifiable. This study explores the use of machine learning algorithms for faunal-based palaeoenvironmental reconstructions in the Cradle of Humankind (CoH) over the past 3.5 million years. We develop a new method based on random forest models using the presence/absence data of rodents and bovids, two clades commonly employed in palaeoenvironmental reconstructions, to assess the evolution of six temperature and precipitation variables in the region. In contrast to continental-scale trends, we found no evidence for a gradual transition toward more arid conditions. Instead, all reconstructions indicate lower precipitation levels accompanied by more moderate interannual precipitation variability in the past. Although interpretations differ slightly when faunal indicators are analysed separately, the combined evidence from rodents and bovids shows that habitats, despite considerable climatic variability through time, consistently occupied the ecotone between grassland and savanna, closely resembling present-day conditions in the CoH. These findings challenge hypotheses that propose a progressive shift from closed, mesic environments to open, arid landscapes during the Pliocene and Pleistocene in the region.

Birds possess a unique balance organ, the lumbosacral organ (LSO), located in the lumbosacral region of the synsacrum. This organ surrounds the spinal cord and leaves distinct traces of its size and shape on the endocast of the vertebral canal. To date, many questions about the function of the LSO and its implications in bird biology remain. Here, we investigate whether the shape of the synsacral vertebral canal endocast, influenced by the LSO, is related to locomotor habits, pelvic morphology, and phylogeny. We used 2D and 3D geometric morphometrics to characterise the shape of the digital synsacral vertebral canal cast and to test whether its morphology is indicative of locomotor behaviour and pelvic morphology. We also quantified the phylogenetic signal to determine whether phylogeny has an impact on morphology. Our results suggest that the vertebral canal endocast is shaped by the LSO, particularly in predominantly perching birds, where it is proportionally larger than in other locomotor groups. We also show that the pelvic morphology covaries significantly with the vertebral canal morphology. A proportionally larger LSO corresponds to a shorter, wider pelvis, while a smaller LSO corresponds to a longer, more slender pelvis. Finally, in addition to a strong phylogenetic signal in vertebral canal morphology, we identify allometry, indicating that body size also influences LSO morphology.

Natural history collections are seen as treasure troves we need to both preserve and study. Campaigns of 2D and 3D digitization have emerged in numerous institutions as an opportunity to maximize specimen's diffusion while limiting the risk associated to their manipulation. 2D and especially 3D models can be used for various scientific purposes. Because of different obstacles (time, technical limitations, cost, etc.), the digitization of small and numerous objects, like insect specimens, remains to be improved. Among the existing options, photogrammetry is generally less expensive than µCT-scan, two of the main methods for digitizing objects, but it remains time-consuming for small objects because focus staking-which involves a multiplication of shots-is strongly recommended to increase the depth of field. Here, we present a fast and inexpensive photogrammetric pipeline that generates 3D models of cockroaches of sufficient quality for morphometric geometric analyses. By focusing on a region of interest in the specimens-identified according to the goal of the digitization-the depth of field is reduced by comparison with the one encompassing the whole specimen. Thus, we eliminated the need for focus stacking. We produced 3D models for 62 species and compared 13 of the photogrammetric 3D models qualitatively and quantitatively with those obtained from µCT-scans of the same 13 species. We conclude that the 3D models produced with our pipeline are of sufficient quality to perform geometric morphometric analyses, which will be published elsewhere in a companion paper. Despite a few limitations, we hope that our pipeline will generate opportunities for the study of small objects like insects, one of the most species-rich group on Earth and in natural history collections.

Investigating forms specialized for their ecology (i.e. ecomorphs) is a major theme in ecology and evolution. Among the remarkable ecomorphs, those associated with fossorial habits have received increasing attention in vertebrates, but less so in invertebrates. Here, we focus on cockroaches, insects whose burrowing ecomorphs have long been observed but never formally investigated at a large taxonomic scale. Our goal is to characterize different cockroach ecomorphs, with a particular focus on fossorial organisms in the broadest sense, that is, including cockroaches digging galleries and feeding in deadwood. The study of these wood-eating species is particularly important because of the role of this diet in the evolution of eusociality in Blattodea (i.e. cockroaches and termites). We used two sets of morphological data while taking into account the phylogenetic relationships of 62 species of cockroaches, sampled from nine families and covering three fossorial (hollow trunk, sand, dead wood) and two non-fossorial microhabitats (litter and under bark). The first set of data corresponded to linear measurements for 13 morphological traits, while the second set comprised 3D models of pronotums obtained by photogrammetry and microtomography and analysed by geometric morphometrics. To account for shared ancestry between species, phylogenetic comparative analyses were then carried out, including predictive analyses for six species whose microhabitat are unknown. We identified morphological differences associated with different burrowing lifestyles while highlighting the importance of considering shared ancestry. Among the most discriminating morphological traits are body length and anterior tibial spine length, which distinguish species burrowing in deadwood from those digging in sand. 3D analyses of the pronotum also revealed interesting differences related to the microhabitat. Despite the ability to characterize ecomorphs from both sets of data, predictive analyses, which would be particularly valuable for all the living and fossil species whose microhabitats is unknown-that is, the vast majority of cockroach species-need further refinement.Read the free for this article on the Journal blog.

Tropical forests are rich in biodiversity but face the rapid loss of their wildlife due to increasing anthropogenic pressure, underscoring the urgent need for effective monitoring. Remote-sensing tools such as camera traps offer faster, less invasive alternatives to human observations. These technologies can provide complementary insights into elusive species, especially in habitats that are difficult to access through direct observation. However, analysing the large volumes of data they produce is labour-intensive, often leaving datasets underutilised due to limited human resources. Deep learning algorithms can automate aspects of data analysis, but their value to research and conservation efforts depends on their ability to reliably identify target species and be easily deployed in field conditions. To improve wildlife monitoring in African forests using camera trap data, we develop DeepForestVision, the first deep learning algorithm tailored to these challenging habitats that can be used in the field to process both photographs and videos. DeepForestVision was trained on an unprecedented dataset of 2,775,671 photographs and 221,982 videos gathered from camera traps from more than 63 research sites across 11 African countries. It identifies 33 non-human vertebrate taxa, including 31 mammal taxa, from the most common to the most threatened ones observed on ground-level camera traps, as well as humans, vehicles and blank photographs or videos. Classification tests demonstrate that DeepForestVision achieves an accuracy of 87.7% on the video test set with 23 taxa. It outperforms the three existing species identification algorithms applicable to these environments: Zamba by 13.1%, Mbaza by 45.0% and SpeciesNet by 37.7%. We provide the model weights for researchers and developers and offer DeepForestVision through a free offline interface. The interface is designed to function in the field in a low-resource setting and requires no programming expertise. Solution: DeepForestVision is a reliable field tool for monitoring species observed on camera trap photos and videos in African tropical forests. Used by research, conservation, private or political actors, it can guide conservation strategies inside and outside of protected areas, and thus contribute to reducing the loss of biodiversity in African tropical forests. Les for & ecirc;ts tropicales africaines sont les habitats terrestres les plus riches en biodiversit & eacute;, mais elles subissent une perte rapide de leur faune en raison de la pression anthropique croissante. Afin de r & eacute;pondre & agrave; l'urgence d'un suivi efficace de cette biodiversit & eacute;, les outils de t & eacute;l & eacute;d & eacute;tection comme les cam & eacute;ras & agrave; d & eacute;tection automatique offrent des alternatives plus rapides et moins invasives que les observations humaines, en particulier pour des esp & egrave;ces cryptiques et dans des habitats difficiles d'acc & egrave;s. N & eacute;anmoins, l'analyse des grands volumes de donn & eacute;es issues de ces & eacute;quipements est chronophage, et les donn & eacute;es acquises sont fr & eacute;quemment sous-utilis & eacute;es en raison de ressources humaines limit & eacute;es. Des algorithmes d'intelligence artificielle permettent d'automatiser certains aspects de l'analyse, mais leur utilit & eacute; pour la recherche et la conservation d & eacute;pend de leur fiabilit & eacute; pour identifier les esp & egrave;ces cibles et de leur capacit & eacute; & agrave; & ecirc;tre facilement d & eacute;ploy & eacute;s sur le terrain. Dans le cadre de l'initiative One Forest Vision (https://www.oneforestvision.org/), nous avons d & eacute;velopp & eacute; DeepForestVision, le premier algorithme d'apprentissage profond adapt & eacute; & agrave; ces habitats exigeants, capable de traiter sur le terrain & agrave; la fois des photos et des vid & eacute;os. DeepForestVision a & eacute;t & eacute; entra & icirc;n & eacute; sur un jeu de donn & eacute;es sans pr & eacute;c & eacute;dent comprenant 2,775,671 photos et 221,982 vid & eacute;os issues de cam & eacute;ras automatiques collect & eacute;es sur 63 sites de recherche r & eacute;partis dans 11 pays africains. Il identifie 33 taxons de vert & eacute;br & eacute;s non humains, dont 31 taxons de mammif & egrave;res, allant des plus communs aux plus menac & eacute;s observ & eacute;s au niveau du sol. Test & eacute; sur un jeu de donn & eacute;es vid & eacute;os du Sebitoli Chimpanzee Project dans le parc national de Kibale en Ouganda, DeepForestVision r & eacute;alise 87.7% de pr & eacute;dictions correctes pour les 23 taxons pr & eacute;sents, surpassant les trois algorithmes existants applicables & agrave; ces environnements: Zamba (+13.1%), Mbaza (+45.0%) et SpeciesNet (+37.7%). DeepForestVision est accessible gratuitement dans AddaxAI, une interface hors ligne con & ccedil;ue pour fonctionner sur le terrain dans un contexte & agrave; faibles ressources et ne n & eacute;cessitant aucune comp & eacute;tence informatique. Nous mettons & eacute;galement & agrave; disposition les poids du mod & egrave;le pour la communaut & eacute; scientifique. Solution: DeepForestVision est un outil fiable pour le suivi des esp & egrave;ces observ & eacute;es sur les photos et vid & eacute;os issues de cam & eacute;ras automatiques dans les for & ecirc;ts tropicales africaines. Utilis & eacute; par des acteurs de la recherche, de la conservation, du secteur priv & eacute; ou des politiques publiques, il pourra orienter les strat & eacute;gies de conservation & agrave; l'int & eacute;rieur et & agrave; l'ext & eacute;rieur des aires prot & eacute;g & eacute;es, et ainsi contribuer & agrave; la pr & eacute;servation de la biodiversit & eacute; animale.

The jaw system in mammals is complex and different muscle morphotypes have been documented. Pigs are an interesting group of animals as they are omnivorous and have a bunodont crushing dentition. Moreover, they have interacted with humans for over 10,000 years and grow nearly two orders of magnitude in size. Despite being a model system for studies on cranial form and function, data on the growth of the jaw adductor muscles are scant. Moreover, whether captivity impacts the growth and architecture of the jaw adductors remains unknown. Based on dissection data of the jaw adductors of 45 animals ranging from less than 1 kg to almost 100 kg, we show that muscle masses, muscle fiber lengths, and cross-sectional areas scale as predicted for geometrically similar systems or with slight negative allometry. Only the fiber length of the lateral pterygoid muscle grew with slight positive allometry. Animals raised in captivity in stalls or in an enclosure were overall very similar to wild animals. However, some muscles were larger in captive animals. Interestingly, variation in bite force in captive animals was well predicted by the variation in the size of the superficial masseter muscle relative to the overall jaw adductor mass. We show that muscle masses, muscle fiber lengths, and cross-sectional areas in wild boar scale as predicted for geometrically similar systems or with slight negative allometry. Animals raised in captivity in stalls or in an enclosure were overall very similar to wild animals.image

The Cradle of Humankind (Gauteng, South Africa) provides an important fossil record of the evolutionary history of PlioPleistocene hominins. Cooper's Cave deposits have yielded a rich fossil faunal assemblage, as well as six remains attributed to Paranthropus robustus. This study provides the first taxonomic, taphonomic and palaeoecological description of the micromammal material from the 1.4 Ma assemblage of Cooper's D. The taphonomic signature of the assemblage indicates an accumulation by tytonid owls (probably Tyto alba) and advanced postdepositional disturbance probably related to trampling by the occupants of the cave, sorting of the bones along slope, and burying. The taxonomic analysis undertaken here at genus level describes at least 22 taxa of small mammals, including one extinct genus Proodontomys. This assemblage is dominated by Mystromys and Otomys, two rodent genera adapted to grassland habitats which are among the most common among pliopleistocene micromammal faunas from the region. The palaeoecological analysis suggests an open landscape with a predominance of grassland and savanna vegetation, and the proximity of rocky outcrops and a perennial river. These results support previous indications of a shift in the African climate and vegetation towards more open habitats during the Early Pleistocene.

African elephants have a wide range of abilities using their trunk. As a muscular hydrostat, and thanks to the two finger-like processes at its tip, this proboscis can both precisely grasp and exert considerable force by wrapping. Yet few studies have attempted to quantify its distal grasping force. Thus, using a device equipped with force sensors and an automatic reward system, the trunk tip pinch force has been quantified in five captive female African savanna elephants. Results showed that the maximum pinch force of the trunk was 86.4 N, which may suggest that this part of the trunk is mainly dedicated to precision grasping. We also highlighted for the first time a difference in force between the two fingers of the trunk, with the dorsal finger predominantly stronger than the ventral finger. Finally, we showed that the position of the trunk, particularly the torsion, influences its force and distribution between the two trunk fingers. All these results are discussed in the light of the trunk’s anatomy, and open up new avenues for evolutionary reflection and soft robot grippers.

Aquatic birds represent diverse ecologies and locomotion types. Some became flightless or lost the ability for effective terrestrial locomotion, yet, certain species excel in water, on land, and in air, despite differing physical characteristics associated with each medium. In this exploratory study, we intend to quantitatively analyze the morphological variety of multiple limb bones of aquatic birds using 3D geometric morphometrics. Morphological variation is mainly driven by phylogeny, which also affects size and locomotion. However, the shape of the ulna, including the proportion and orientation of the epiphyses is influenced by size and aquatic propulsive techniques even when phylogeny is taken into consideration. Certain trends, possibly linked to functions, can be observed too in other bones, notably in cases where phylogenetic and functional signals are probably mixed when some taxa only englobe species with similar functional requirements: penguins exhibit the most distinctive wing bone morphologies, highly adapted to wing-propulsion; advanced foot-propellers exhibit femur morphology that reduces proximal mobility but supports stability; knee structures, like cnemial crests of varied sizes and orientations, are crucial for muscle attachments and efficient movement in water and on land; taxa relying on their feet in water but retaining terrestrial abilities share features enabling swimming and walking postures. Size-linked changes distinguish the wing bones of non-wing-propelled taxa. For hindlimbs, larger size relates to robust bones probably linked to terrestrial abilities, but robustness in femora can be connected to foot-propulsion. These results help us better understand birds' skeletal adaptation and can be useful inferring extinct species' ecology.

Nautiluses are key to understanding the evolution of cephalopods, as they are their only outer-shelled modern representatives. Their chitinous beaks possess a calcified tip on the upper beak called a rhyncholite. These rhyncholites are found abundantly in the fossil record from Middle Triassic onwards and are described as being morphologically stable. Because of their feeding function, they likely carry information related to their diet and environment. Yet, the morphological variability of these fossils was poorly investigated. Here, we characterized the morphology of rhyncholites through time, together with morphologically similar fossils from the Jurassic and Cretaceous, the rhynchoteuthis. We digitized 254 specimens using X-ray microtomography (mu CT) and performed a shape analysis using 3D geometric morphometrics. The morphological continuum observed between rhyncholites and rhynchoteuthis leads us to consider them as close relatives of nautilids, if not part of this group. The highest rhyncholites morphological disparity is observed after their apparition in Triassic. Triassic rhyncholites are morphologically closer to modern ones, probably sharing their opportunistic carnivorous diet and able to feed on hard items. This ability might have been advantageous during the Marine Mesozoic Revolution. During this arms race period, competition might have favoured the emergence of new forms like rhynchoteuthis, and the exploration of new niches as seen in Cretaceous morphologies indicating a possible diet shift toward softer prey items. Altogether, rhyncholites and rhynchoteuthis testify of a nautilids evolutionary history richer than what can be deduced from the shells only and carry important information that could improve reconstructions of past food-webs. Nautilus, square rhyncholites, rhynchoteuthis, 3D morphometrics, Marine Mesozoic Revolution