Functional space analyses reveal the function and evolution of the most bizarre theropod manual unguals

Abstract

Maniraptoran dinosaurs include the ancestors of birds, and most used their hands for grasping and in flight, but early-branching maniraptorans had extraordinary claws of mysterious function. Alvarezsauroids had short, strong arms and hands with a stout, rock-pick-like, single functional finger. Therizinosaurians had elongate fingers with slender and sickle-like unguals, sometimes over one metre long. Here we develop a comprehensive methodological framework to investigate what the functions of these most bizarre bony claws are and how they formed. Our analysis includes finite element analysis and a newly established functional-space analysis and also involves shape and size effects in an assessment of function and evolution. We find a distinct functional divergence among manual unguals of early-branching maniraptorans, and we identify a complex relationship between their structural strength, morphological specialisations, and size changes. Our analysis reveals that efficient digging capabilities only emerged in late-branching alvarezsauroid forelimbs, rejecting the hypothesis of functional vestigial structures like T. rex. Our results also support the statement that most therizinosaurians were herbivores. However, the bizarre, huge Therizinosaurus had sickle-like unguals of such length that no mechanical function has been identified; we suggest they were decorative and lengthened by peramorphic growth linked to increased body size.

Fig. 1. Key taxa and work pipeline use in this paper.

Silhouettes show the large and elongated forelimb of the late-branching therizinosaurian Therizinosaurus (a) and the overall body shape and highlighted forelimb of the late-branching alvarezsauroid Mononykus (b), scaled against an adult human (height ~1.8 m). The work pipeline demonstrated by an ungual model from the Jurassic alvarezsauroid Haplocheirus, includes processes of 3D model reconstruction (c); model smoothing, measurement, and morphological analysis (d); finite-element analysis, ‘intervals’ method and functional-space analysis (e); and total evidence functional assessment (f).

Fig. 2. Phylogenetic comparison of von Mises stress plots of dinosaur manual unguals under three functional scenarios.

Unguals are scaled to the same surface area and illustrated in lateral view. This phylogeny is based on studies by Qin et al. and Zanno et al.^,.

Fig. 3. Functional space plots illustrating the divergent functional performances of manual unguals.

Unguals from reference mammals (a), non-maniraptoran theropods (b), alvarezsauroids (c) and therizinosaurians (d), tested in three major ungual functional scenarios, piercing (marked as circles, demonstrated by manus of Guanlong), hook-and-pull (marked as triangles, demonstrated by manus of Therizinosaurus) and scratch-digging (marked as squares, demonstrated by manus of Linhenykus). Silhouettes roughly exhibit their body forms and sizes.

Fig. 4. Quantified manual ungual functional divergence and overall functional performance estimation though geological time scales.

Functional divergences are exhibited between hook-and-pull and piercing (a), and scratch-digging and piercing (b) within alvarezsauroid and therizinosaurian claws.

Fig. 5. Linear regressions of morphologies, functional performances and ungual sizes vs. body size of therizinosaurians.

Histogram of functional performance showing divergent evolutionary paths of unguals from late gigantic members.

Fig. 6. Linear regressions of average functional performance, morphologies, and ungual size vs. body size of alvarezsauroids.

(a) a good regression relationship only to the ungual sizes (R² = 0.83, P value = 0.005 < 0.05). Linear regressions of functional divergences vs. morphological divergences (b), morphological divergences vs. log-transformed ungual volume (c), functional divergences vs log-transformed ungual volume (d) and proportion of length of unguals to the manus through geological time (e), showing close relationship between functional and morphological divergences (R² = 0.65, P value = 0.029 < 0.05), but failing to support a strong relationship between functional performance and morphological specialisations to ungual volumes (R² = 0.04, P value = 0.071 > 0.05; R² = 0.09, P value = 0.510 > 0.05), and higher proportion of manual unguals in late-branching alvarezsauroids (R² = 0.70, P value = 0.019 < 0.05).