Exceptional preservation and foot structure reveal ecological transitions and lifestyles of early theropod flyers

The study addresses how foot morphology in early theropod flyers relates to ecological roles during the evolution of flight. In modern birds, features such as keratinised toe pads, scale types, toe pad alignment, claw shape/size, and the geometry of foot joints correlate with grasping ability, cursoriality, perching, wading, swimming, and prey capture. Two main toe pad alignments occur in living birds: arthral (pad over joint), common in raptorial taxa, and mesarthral (pad over phalanx), more typical of non-raptorial birds. Although arthral pads may activate the tendon-locking mechanism (TLM) more efficiently, the presence and relevance of the TLM in early flyers is uncertain and thus not used as a specialised grasping indicator here. Claw morphometrics have been used for decades to infer theropod ecology, with traditional morphometrics particularly effective at separating different raptorial strategies and non-raptorial modes. Joint ginglymoidy (hinge-like articulations) also reflects functional demands: strongly ginglymoid joints resist torsion and enhance grip (typical of raptors), whereas weakly ginglymoid joints allow greater mediolateral motion, consistent with cursoriality. The purpose of this study is to apply this modern ecomorphological framework to exceptionally preserved fossil feet from early theropod flyers, integrating soft tissue and joint evidence (via laser-stimulated fluorescence) with quantitative claw morphometrics to refine ecological profiles across key Jurassic and Early Cretaceous taxa.

Prior work has partially revealed early theropod flyer ecology through anatomy, diet, locomotion, and habitat analyses. In extant birds, toe pad morphology (flat, well-developed, protrusive) and plantar scale types (including reticulate spicules) correlate with perching, ground-dwelling, manipulation, and raptorial hunting. Toe pad alignment (arthral vs. mesarthral) relates to feeding behaviour rather than ancestry, though variation exists. The TLM is widespread among modern birds (excluding palaeognaths), and arthral pads may activate it more efficiently; however, evidence in early flyers is lacking. Traditional morphometrics of claw curvature and relative digit sizes effectively differentiate raptorial behaviours (restraint, strike, constrict, pierce) and non-raptorial modes (perching vs. ground), with increased curvature and enlarged opposing digits characterising raptorial strategies. Joint ginglymoidy correlates with increased grip strength (raptors) or cursorial function (weakly ginglymoid joints). Earlier studies examined claw curvature alone and various locomotor indices, but comprehensive datasets applying bone-based landmarks suitable for fossils and broader ecological sampling have shown greater discriminative power for ecological inference.

Study design combined qualitative assessment of preserved pedal soft tissues and joints using laser-stimulated fluorescence (LSF) with quantitative traditional morphometrics (TM) of claws, interpreted via principal component analysis (PCA) and linear discriminant analysis (LDA). Fossil sampling: Over 1000 early paravian fossils at the Shandong Tianyu Museum of Nature were screened under LSF, yielding 12 specimens with preserved toe pads, scales, and claws from Anchiornis, Confuciusornis, Sapeornis, Yanornis, and Microraptor. To broaden phylogenetic coverage, the Berlin and Thermopolis Archaeopteryx specimens were examined first-hand, and claw data for Ambopteryx (non-paravian glider) and Fortunguavis (enantiornithine) were incorporated from literature and photographs despite lacking preserved podotheca. Modern comparative sampling: Podothecae from 15 raptorial and 21 non-predatory bird species were examined at UNE and the Australian Museum; claws from 61 crown-bird taxa (especially raptors) were measured at CMNH and FMNH. LSF protocol: A 405 nm laser diode was used to induce fluorescence following safety protocols; long-exposure imaging in a dark room employed a Nikon D810 with a 425 nm blocking filter. Image post-processing (equalisation, saturation, colour balance) was applied uniformly. Integumentary terminology followed Lucas and Stettenheim (scutate, scutellate, reticulate scales) and published definitions for toe pad types (flat, well-developed, protrusional), furrows, and folds. Traditional morphometrics: Seven parameters measured—outer arc curvature (degrees) for digits I–IV and outer arc length ratios (I/III, II/III, IV/III). Because many fossils lack digit I, analyses were run with datasets including and excluding digit I. Linear measures used tape or callipers; angular measures were taken from lateral-view photographs tested for parallax. Tree and statistics: A composite phylogeny (modern backbone from birdtree.org; non-avian paravian branches grafted following published topologies) was used for phylogenetic mapping and honest significant differences (HSD) via RRPP. PCA used correlation matrices; LDA used discriminant analysis of principal components (DAPC) to address variable correlations; Pandion was excluded from LDA/HSD due to single-member group issues. Pairwise phylogenetic HSD tested differences among modern ecological groups; Kmult quantified phylogenetic signal in multivariate data. Fossil interpretation: All fossils were preserved in bedding plane orientation with limited crushing; ambiguous or overlapping soft-tissue outlines (or suboptimal orientations) were excluded from pad analysis. Joint ginglymoidy was assessed where articulation facets were observable. Ethics/permits: Collections were accessed under institutional guidelines; no specific permits required.

• Exceptional pedal soft-tissue preservation and joint morphology, combined with claw TM and multivariate analyses, reveal diverse ecological profiles among early theropod flyers.
• Ambopteryx: Claw TM LDA shows highest similarity to specialised constricting raptors (e.g., 0.629 posterior probability for constrict with digit I included), but given lack of opposed digits and weakly recurved claws, authors interpret a primarily ground-dwelling, scansorial forelimb/gliding ecology.
• Anchiornis: Toe pads arthral, ‘well-developed’; interphalangeal joints mostly weakly ginglymoid (cursorial signal). LDA indicates strong ground-bird affinities (e.g., up to 0.958 for ground when digit I excluded). Overall interpretation: primarily terrestrial with some raptorial capability for non-volant prey.
• Archaeopteryx: Claw curvature generally low; joints mostly weakly ginglymoid. LDA mixed signals—Berlin specimen shows some scavenging signal (e.g., 0.653 with digit I included), Thermopolis specimen shows constrict/restraint signals (e.g., 0.832 constrict with digit I included; 0.647 restraint without digit I). Overall: terrestrial-leaning and differing ecologies between specimens; potential no close modern analogue for Berlin specimen.
• Confuciusornis: Few low toe pads preserved; interphalangeal joints ginglymoid (grasp-related). LDA often aligns with specialised constrictors (e.g., 0.986 and 0.934 constrict with digit I included; 0.773 constrict without digit I), but other evidence (jaw mechanics resembling herbivores; limb metrics; habitat prevalence) supports a generalist ecology.
• Fortunguavis: Mixed ginglymoidy across digits; LDA supports perching affinity (e.g., 0.541 perch with digit I included; 0.437 perch without digit I). PCA positions it near macaw parrots, consistent with arboreal/perching and potential scansorial behaviours.
• Microraptor: Arthral pads with protrusive tarsal and proximal digit IV pads; strong interphalangeal ginglymoidy; reticulate scales including spicules; PCA/LDA show strong raptorial affinity—especially restraint/strike (e.g., 0.804 restraint for STM 5-75; 0.423 strike and 0.370 restraint for STM 5-109; 0.337 restraint for STM 5-172). Integrated evidence indicates a specialised aerial hunter using restraint/pinning strategies.
• Sapeornis: Well-developed arthral pads; strongly ginglymoid interphalangeal joints; reticulate scales; LDA indicates perching non-raptorial lifestyle (e.g., 0.670 perch with digit I included; 0.591 without). Interpretation: ecologically complex herbivorous thermal soarer, potentially supplementing diet with meat.
• Yanornis: Digit III pads mesarthral (unique among sampled taxa here), weakly ginglymoid joints; claws relatively straight. LDA suggests similarity to specialised constrictors and ground birds (e.g., 0.994 constrict with digit I included; 0.883 constrict without), but anatomical/functional evidence supports primarily terrestrial ecology with piscivory documented by gut contents.
• Statistical support: Phylogenetic HSD indicates significant claw-shape differences among certain ecological groups (e.g., ground vs. perch p=0.004; ground vs. restraint p=0.022 when digit I excluded), consistent with discriminability of ecological modes via TM.
• Temporal/ecological transition: Jurassic flyers (Ambopteryx, Anchiornis, Archaeopteryx) show more ground-dwelling profiles coincident with limited flight performance; Early Cretaceous flyers diversify into aerial niches, including specialised raptorial (Microraptor), perching (Fortunguavis), and generalist (Confuciusornis) roles.

Applying modern avian foot ecomorphology to exceptionally preserved fossils refines ecological interpretations of early theropod flyers. Arthral pad arrangements, presence of protrusive pads, and strongly ginglymoid joints correlate with enhanced grasping and raptorial capabilities, whereas weakly ginglymoid joints and mesarthral pads indicate cursorial or non-raptorial behaviours. Integrating LSF-revealed soft tissues and joint morphology with claw TM (PCA/LDA) and independent ecological evidence (gut contents, limb proportions, flight capability) resolves distinct ecological profiles by taxon. Jurassic forms (Ambopteryx, Anchiornis, Archaeopteryx) were more terrestrial, likely reflecting limited flight performance. Early Cretaceous taxa diversified: Microraptor exhibits a specialised aerial predatory niche (restraint/strike), Confuciusornis aligns with generalist strategies despite claw similarities to constrictors, Fortunguavis with arboreal perching, Sapeornis with complex herbivory-perching-soaring ecology, and Yanornis with primarily terrestrial habits and piscivory. These findings suggest that some specialised predatory roles in early ecosystems were occupied by non-avialan flyers (e.g., Microraptor) rather than early birds, and that certain early avialans had ecologies without close modern analogues.

This study demonstrates that early theropod flyers encompassed a broad spectrum of ecological strategies discernible from pedal soft tissues, joint morphology, and claw morphometrics. Jurassic flyers were predominantly ground-oriented, while Early Cretaceous taxa diversified into aerial specialists and generalists. Notably, Microraptor occupied a specialised aerial predatory niche, Confuciusornis functioned as an ecological generalist, Fortunguavis aligned with perching/arboreal habits, Yanornis was primarily terrestrial with piscivory, and both Berlin Archaeopteryx and Sapeornis exhibit ecologies lacking close modern analogues. These insights highlight ecological complexity during the evolution of flight and indicate that early specialists may have been more vulnerable during ecosystem crises. Future work should expand fossil sampling of pedal soft tissues, refine quantitative datasets (including digit I and non-ungual phalanges where possible), and further investigate taxa with ambiguous or unique ecological signals to better understand ecological turnover and the rise of modern birds.

• Preservation constraints: Many fossils lack complete podothecae or optimal orientations; some toe pads, scales, and joint facets are obscured or ambiguous, limiting pad-type recognition (e.g., Yanornis plantar pads not visible; Ambopteryx joints poorly preserved).
• Dataset constraints: Numerous fossil specimens lack digit I, necessitating separate analyses with/without digit I; non-ungual phalange measurements were excluded in TM due to disarticulation issues in modern skeletal specimens.
• Methodological limits: LDA/DAPC precision is affected by group representation (e.g., ‘pierce’ ecology represented by a single taxon, Pandion, excluded from LDA/HSD). Claw-based inferences may not fully capture complex behaviours (e.g., distinguishing climbers from perchers).
• Biological uncertainties: The tendon-locking mechanism’s distribution and function in early flyers are uncertain; therefore, arthral pad alignment is not taken as definitive evidence of specialised grasping.
• Taphonomic/interpretative issues: Two-dimensional preservation, crushing, and overlap can confound identification of soft-tissue boundaries; when ambiguous, features were excluded, potentially reducing sample breadth.