Turtle body size evolution is determined by lineage-specific specializations rather than global trends

The complete mitogenome of Amazonian Brachyplatystoma filamentosum and the evolutionary history of body size in the order Siluriformes

The order Siluriformes (catfish) is one of the largest groups of fish. Diversity in the body size among its species, which range from a few centimeters to 4 meters, makes Siluriformes an interesting group to investigate the body size evolution. Here, we present the complete mitogenome of Brachyplatystoma filamentosum (Piraíba), the largest Amazonian catfish, to explore the evolutionary history of Siluriformes and their body size dynamics. The Piraíba's mtDNA is 16,566 bp long, with a GC content of 42.21% and a D-loop of 911 bp. Phylogenetic analysis was conducted using protein-coding sequences, tRNAs, and rRNAs from mtDNA of Piraíba and 137 other Siluriformes species. Time-calibrated maximum likelihood trees estimated the origin of the order Siluriformes to be ~118.4 Ma, with the Loricarioidei suborder diversifying first, followed by Diplomystoidei and Siluroidei. The Siluroidei suborder experienced rapid expansion around 94.1 Ma. Evolutionary dynamics revealed 16 positive and 11 negative directional body size changes in Siluriformes, with no global trend toward larger or smaller sizes, and with Piraíba showing a significant size increase (5.65 times over 40.8 Ma). We discuss how biological, ecological and environmental factors could have shaped the evolution of body size in this group.

Cladistic estimates of evolutionary rates focused on palaeontological datasets using TNT

We describe a protocol for estimating evolutionary rates from phylogenetic trees based on parsimony character optimization. The rate estimation is conducted through a TNT script and the results are analysed in a script for the software environment R. The TNT script allows analysing multiple optimal topologies, considering optimization ambiguity, and alternative time-calibrations or pre-calibrated trees. The R script summarizes estimated rates on a consensus tree and plots the variation of evolutionary rates through time, jointly with the phylogenetic diversity and a new metric (clade completeness index) that measures the distribution of missing data along the tree. We present results for simulated and empirical analyses, and evaluate the impact of missing data and alternative calibration methods in rate estimates. We found that while missing data can lower the nominal values of evolutionary rates, the overall pattern of rate variation through time remained robust. Empirical cases highlight different scenarios, such as datasets in which peaks of evolutionary rates can be coupled or decoupled from diversification dynamics (phylogenetic diversity) and cases in which missing data may influence the variation of estimated evolutionary rates. We conclude with recommendations for using this protocol and interpreting the results of parsimony-based rate estimates.

Shell Constraints on Evolutionary Body Size-Limb Size Allometry Can Explain Morphological Conservatism in the Turtle Body Plan

Turtles are a small clade of vertebrates despite having existed since the Late Triassic. Turtles have a conservative body plan relative to other amniotes, characterized by the presence of a shell and quadrupedality. This morphology is even retained in strong ecological specialists, such as sea turtles, which are secondarily adapted to marine locomotion by strong allometric scaling in their hands. It is possible that the body plan of turtles is strongly influenced by the presence of the shell, acting as a constraint to achieving greater diversity of body forms. Here, we explore the evolutionary allometric relationships of fore- and hindlimb stylopodia (i.e., humerus and femur) with one another as well as their relationship with shell size (carapace length) to assess evidence of constraint. All turtles, including Triassic shelled stem turtles, have near-isometric relationships that do not vary strongly between clades, and evolve at slow evolutionary rates. This indeed indicates that body proportions of turtles are constrained to a narrow range of possibilities. Minor allometric deviations are seen in highly aquatic sea turtles and softshell turtles, which modified their shells by bone losses. Our allometric regressions allow accurate body size estimations for fossils. Several independent sea turtle lineages converged on maximum sizes of 2.2 m of shell length, which may be a biological maximum for the group.

A systematic comparative description of extant turtle humeri, with comments on humerus disparity and evolution based on fossil comparisons

The humerus is central for locomotion in turtles as quadrupedal animals. Osteological variation across testudine clades remains poorly documented. Here, we systematically describe the humerus anatomy for all major extant turtle clades based on 38 species representing the phylogenetic and ecological diversity of crown turtles. Three Late Triassic species of shelled stem turtles (Testudindata) are included to establish the plesiomorphic humerus morphology. Our work is based on 3D models, establishing a publicly available digital database. Previously defined terms for anatomical sides of the humerus (e.g., dorsal, ventral) are often not aligned with the respective body sides in turtles and other quadrupedal animals with sprawling gait. We propose alternative anatomical directional terms to simplify communication: radial and ulnar (the sides articulating with the radius/ulna), capitular (the side bearing the humeral head), and intertubercular (opposite to capitular surface). Turtle humeri show low morphological variation with exceptions concentrated in locomotory specialists. We propose 15 discrete characters to summarize osteological variation for future phylogenetic studies. Disparity analyses comparing non-shelled and shelled turtles indicate that the presence of the shell constrains humerus variation. Flippered aquatic turtles are released from this constraint and significantly increase overall disparity. Ontogenetic changes of turtle humeri are related to increased ossification and pronunciation of the proximal processes, the distal articulation areas, and the closure of the ectepicondylar groove to a foramen. Some turtle species retain juvenile features into adulthood and provide evidence for paedomorphic evolution. We review major changes of turtle humerus morphology throughout the evolution of its stem group.

Shell biomechanics suggests an aquatic palaeoecology at the dawn of turtle evolution

The turtle shell is a remarkable structure that has intrigued not only evolutionary biologists but also engineering and material scientists because of its multi-scale complexity and various functions. Although protection is its most apparent role, the carapace and plastron are also related to many physiological functions and their shape influences hydrodynamics and self-righting ability. As such, analysing the functional morphology of the shell could help understanding the ecology of Triassic stem-turtles, which will contribute to the century-long debate on the evolutionary origins of turtles. Here, we used 3D imaging techniques to digitize the shells of two of the earliest stem-turtle taxa, Proganochelys and Proterochersis, and submitted their models to biomechanical and shape analyses. We analysed the strength performance under five predation scenarios and tested the function of two morphological traits found in stem-turtles, the epiplastral processes and an attached pelvic girdle. The latter, also present in the crown-lineage of side-necked turtles, has been suggested to increase load-bearing capacity of the shell or to improve swimming in pleurodires. Our results do not confirm the shell-strengthening hypothesis and, together with the results of our shape analyses, suggest that at least one of the first stem-turtles (Proterochersis) was an aquatic animal.

New insights into the early morphological evolution of sea turtles by re-investigation of Nichollsemys baieri, a three-dimensionally preserved fossil stem chelonioid from the Campanian of Alberta, Canada

The early evolution of Pan-Chelonioidea (sea turtles) is poorly understood. This is in part due to the rarity of undeformed skulls of definitive early stem chelonioids. In this work, we redescribe the holotype of Nichollsemys baieri using µCT scans and segmentations of the skull. This fossil is the best 3D preserved skull of any Campanian sea turtle, and includes partial "soft tissue" preservation. Nichollsemys is morphologically similar but clearly distinct from Toxochelys spp., and both show a mosaic of plesiomorphic and derived chelonioid features. The internal cranial anatomy documents the presence of derived characters in Nichollsemys baieri that are absent in Toxochelys spp., such as the loss of the epipterygoids and the rod-like shape of the rostrum basisphenoidale. Among the numerous plesiomorphic characters is the presence of a splenial bone, which was unnoticed before. An updated phylogenetic analysis retrieves Nichollsemys baieri as a non-protostegid early stem chelonioid in a slightly more crownward position than Toxochelys latiremis. Our phylogeny includes macrobaenids and protostegids as pan-chelonioids, and we find unorthodox results for dermochelyids. Thus, although Nichollsemys baieri provides important new insights into the early morphological evolution of sea turtles, much work remains to be done. As a completely 3D preserved specimen, we included Nichollsemys baieri into a recent landmark-based skull shape dataset of turtles. Morphospace analysis reveals an intermediate position between cryptodires and crown chelonioids. Based on these data, we also predict that Nichollsemys baieri was still capable of neck retraction, constraining the loss of this trait to more crownward pan-chelonioids.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13358-024-00323-8.

The latest freshwater giants: a new Peltocephalus (Pleurodira: Podocnemididae) turtle from the Late Pleistocene of the Brazilian Amazon

Overkill of large mammals is recognized as a key driver of Pleistocene megafaunal extinctions in the Americas and Australia. While this phenomenon primarily affected mega-mammals, its impact on large Quaternary reptiles has been debated. Freshwater turtles, due to the scarcity of giant forms in the Quaternary record, have been largely neglected in such discussions. Here we present a new giant podocnemidid turtle, Peltocephalus maturin sp. nov., from the Late Pleistocene Rio Madeira Formation in the Brazilian Amazon, that challenges this assumption. Morphological and phylogenetic analyses of the holotype, a massive partial lower jaw, reveal close affinities to extant Amazonian species and suggest an omnivorous diet. Body size regressions indicate Pe. maturin possibly reached about 180 cm in carapace length and is among the largest freshwater turtles ever found. This finding presents the latest known occurrence of giant freshwater turtles, hinting at coexistence with early human inhabitants in the Amazon.