Mammals exhibit ecology-related diversity in long bone morphology, revealing an ample spectrum of adaptations both within and between clades. Their occupation of unique ecological niches in postcranial morphology is thought to have occurred at different chronological phases in relation to abiotic factors such as climate and biotic interactions amongst major clades. Mammalian morphologies rapidly evolved throughout the Cenozoic, with several orders following different paths in locomotory adaptations. We assessed morphological variation in limb proportions for a rich sample of extant and fossil large mammalian clades (mainly carnivores and ungulates) to test associations with ecological adaptations and to identify temporal patterns of diversification. Phylogenetic relationships among species were incorporated into the analysis of limb bone proportions, showing significant morphological changes in relation to species substrate preference. Major climatic events appeared to have no temporal impact on patterns of morphological diversification, expressed as morphological disparity, in either clades or ecological groups. Linear stochastic differential equations supported a double-wedge diversification model for limb proportions of carnivorous clades (‘Creodonta’ and Carnivora). The concomitant increase in morphological disparity throughout the Cenozoic for the orders Carnivora and Artiodactyla had a significative impact on the disparity of Perissodactyla supporting biotic interaction as the primary driver of mammalian morphological diversification. Our findings challenge the classic idea of abiotic factors as primary driving forces in the evolution of postcranial morphologies for large terrestrial mammals and propose clade competition as a key factor in temporal diversification.

Data collection

Morphological data

To compute the analyses of the study entitled  "Morphological disparity of mammalian limb bones throughout the cenozoic: the role of biotic and abiotic factors”, we assembled from the literature and published databases 793 humerus, 805 radius, 668 femur, and 663 tibia lengths belonging to 246 living and 177 fossil species. These data were combined into the brachial index (i.e., radius to humerus ratio) and the crural index (i.e., tibia to femur ratio). Also, because the body mass of fossil species is usually estimated using long bone length, to account for the allometry in the analyses, we also collected the greatest skull length (1883 values), which is a good predictor of body mass. Collected data represented 423 species from the order Acreodi (1), Artiodactyla (206), Carnivora (178), Cimolesta (4), Condylarthra (3), Hyaenodonta (5), Oxyaenodonta (4), Dinocerata (1), Litopterna (1), Perissodactyla (59) and Procreodi (2) spanning throughout the Cenozoic. On average, data for each species were assembled by merging multiple individual references and for this reason, the collected values were checked for errors. The error checking followed the protocol proposed by (Cooper & Purvis, 2009) to estimate typographical and measurement errors.

Ecological data

To identify the selective pressures affecting long bones evolution, we collected for both extant and extinct species trophic level, diet, and locomotor behaviour, while only for extant species collected habitat type.

The trophic level category was built using body mass as a reference. This category includes:

-large predators

Extant and extinct Carnivora weighting ≥21.5 kg (Carbone et al., 1999). Examples of fossil genera included in this category were amphicyonids, nimravids, barbourofelids, machairodonts, and arctocyonids.

-large prey

Extant and extinct Perissodactyla and Artiodactyla including the fossil groups of phenacodontids, merycoidodonts, agriochoerids, and uintatherids weighting from 10 kg up to 900 kg (Estes, 1974).

-small predators

Small carnivores, weighing less than 21.5 kg (Carbone et al., 1999; Carbone et al., 2007), and feed on small vertebrate and invertebrate food sources. Examples of fossil species included in this category were Desmocyon, Cormocyon, Phlaocyon, and Archaeocyon as well as Eucyon, Limocyon, Thinocyon, Palaeonictis, Buxolestes, Paleosinopa, and Chriacus.

-small prey

Small ungulates weigh less than 10 kg (Carbone et al., 1999; Carbone et al., 2007).

-mega-herbivores

Large ungulates weighing more than 900 kg (Sunquist & Sunquist, 1989; Owen‐Smith & Mills, 2008). Among the fossil genera included in this category were Toxodon, Palaeosyops, Parvicornus, Coelodonta, Stephanorhinus, Chilotherium, Teleoceras, Paraceratherium, Amynodon, Chalicotherium, and Moropus.

-non-specialized predators

This category included species belonging to the family Ursidae which did not show particular predator morphology. Ursids consume less than 10% of ungulates meat (Penteriani & Melletti, 2020). Fossil species belonging to this category were representative of the cave bear group. The only exception was the short-faced bear Arctodus simus which is the largest omnivorous that ever lived (Figueirido et al., 2010; Meloro, 2011; Figueirido et al., 2017). The red panda (Ailurus fulgens) was also included in this category, even though it is not an ursid, as it is well-adapted to a mostly folivorous diet, similar to that of the unrelated giant panda Ailuropoda melanoleuca (Roka et al., 2021).

The diet category included: insectivore, carnivorous, omnivore, piscivore, frugivore, folivore, browser, grazer, and mixed feeders.

- Insectivores

Species feeding on insects such as herpestids, the hyena Proteles, the canid Otocyon, and the extinct genus Hemihegetotherium.  

- carnivorous

Carnivorans feed on vertebrate food, such as most of the Carnivora, Hyaenodonta, Oxyaenodonta, and the genus Pachyaena.

- omnivore

This category included generalist species such as suids and bears.

- piscivore

This category included Pantolestidae and otters.

- frugivore

Species feed mainly on fruits, such as the ring-tailed cat and the genus Bassariscus.

- folivores

Non-ungulates herbivores as for example the genera Coryphodon, Ailurus, Ailuropoda, and Tremarctos.

-browser

Ungulates feeding on leaves. Examples of species belonging to this category are Eurygenium, Uintatherium, the basal perissodactyls Hyracotherium, most rhinoceroses, and camelids.

-grazer

Species feeding on grass as for example the genera Equus, Cervus, Bos, and Bison.

-mixed feeders

This category included species that feed on both leaves and grass.

The extinct Mericoidodontinae that lived before the early Miocene, were classified as browsers, while those that lived from the middle Miocene forward were considered grazers (Mihlbachler & Solounias, 2006).

The locomotor behaviour category included amphibious, arboreal, terrestrial, scansorial, and semi-fossorial categories.

-arboreal

This category included species living on trees. Viverrids, procyonids, small felids, and the genus Miacis are examples of species in this category.

-terrestrial

This category included most of the carnivoran and ungulate species.

-amphibious

This category included species living close to bodies of water such as extant polar bears, otters, the hippopotamus, some Tapirus, and the extinct Teleoceras, Buxolestes, Palaeosinopa, and Metamynodon.

-scansorial

This category included species capable of climbing. Most felids, Mephitidae, Mustelidae, Procyonidae, and the extinct Agriochoerus antiquus and Chriacus were in this category.

-semifossorial

This category included Meles meles, Mellivora capensis, and Taxidea taxus.

For living species only we collected also habitat preferences by merging the IUCN Red List of Threatened Species (www.iucnredlist.org), the primary source, and Animal Diversity Web (animaldiversity.org) information. This category included open, mixed, closed, and wetland habitat types.

Phylogenetic tree

The phylogenetic position of both living and fossil species and node ages were collected from published articles. These data were used to assemble and calibrate the phylogenetic tree used to take the non-independence of the species into account when computing in the statistical analyses.

Time bins

In our work entitled "Morphological disparity of mammalian limb bones throughout the Cenozoic: the role of biotic and abiotic factors” we computed morphological disparity of long bones functional indices over the Cenozoic. To bin both species and the estimated ancestral character state according to the Cenozoic stages, we collected the Geological dates from the International Chronostratigraphic Chart 2021 (https://stratigraphy.org/ICSchart/ChronostratChart2021-07.pdf (Cohen et al., 2021). The species first and last appearance in the fossil record were collected from Fossilworks (http://fossilworks.org/), Paleobiology database (https://paleobiodb.org/#/), and Now database (https://nowdatabase.org/).

Analyses

For the aim of this work, we first investigated the allometric effect studying the relationships between phenotypic traits such as radius to humerus and tibia to femur ratios with the greatest skull length using Phylogenetic Generalised Least Squares regression (PGLS; Harris & Steudel, 1997). The PGLS method incorporates phylogeny as an error term by estimating the lambda parameter, which accounts for different degrees of phylogenetic signal.

To determine the functional morphospace taking into account both the effect of the allometry and phylogeny, we computed the size and phylogenetic corrected PCA (Phy-PCA; Revell, 2009). The function phyl.resid (R package phytools; Revell, 2012) was used to remove the allometric effect from the functional ratios, and the function phyl.pca (R package phytools; Revell, 2012) was used to obtain evolutionary-independent Principal Component (PCs). In this study, we also analysed the effect of trophic level, diet, substrate preferences, and habitat on limb proportions using the permutation ANOVA (Collyer & Adams, 2018).

In addition, we estimated the ancestral characters at tree nodes to correct morphological disparity computation accounting for sample incompleteness within Cenozoic time bins. The best mode of evolution was assessed for each PC vector using fitContinuous function (R package geiger; Pennell et al., 2014) which allowed testing evolutionary models alternative to Brownian Motion. The Akaike Information Criterion (AIC) was computed to compare the fitted models, with the best models selected based on the lowest AICc and a ΔAIC of less than 2. The best model of evolution for the categorical variables was investigated using fitDiscrete function (R package geiger; Pennell et al., 2014), and selected by comparing AICs and ΔAICs. The best model parameters were used to estimate the ancestral character states for both continuous and discrete variables.

Disparity through time, which is a metric that represents morphological variance among species within a specific time interval, was computed by accounting or not for the reconstructed ancestral state. The Partial Disparity Analysis (PDA) was also computed to explore how ecological groups and orders contributed to the changes in morphological disparity.

Because our study aimed to investigate the effect of temperature and primary productivity on morphological disparity in mammalian limb proportions throughout the Cenozoic, we collected climatic data from Zachos et al. (2008) and used the linear stochastic differential equation (SDE) approach to test the hypothesis that environmental changes and/or clade interactions affected morphological variation.

Finally, we used the swapOne function (R package RRphylo; (Raia et al., 2019) to validate the results in light of phylogenetic uncertainty.

References

• Carbone, C., Mace, G. M., Roberts, S. C., & Macdonald, D. W. (1999). Energetic constraints on the diet of terrestrial carnivores. Nature, 402(6759), 286-288. https://doi.org/10.1038/46266
• Carbone, C., Teacher, A., & Rowcliffe, J. M. (2007). The costs of carnivory. PLoS biology, 5(2), e22. https://doi.org/10.1371/journal.pbio.0050022
• Cohen, K. M., Finney, S. C., Gibbard, P. L., & Fan, J.-X. (2021). The ICS International Chronostratigraphic Chart. Episodes, 36, 199-204.
• Collyer, M. L., & Adams, D. C. (2018). RRPP: RRPP: An R package for fitting linear models to high-dimensional data using residual randomization. Methods in Ecology and Evolution, 9(2), 1772-1779.
• Cooper, N., & Purvis, A. (2009). What factors shape rates of phenotypic evolution? A comparative study of the cranial morphology of four mammalian clades. Journal of Evolutionary Biology, 22(5), 1024–1035.
• Estes, R. D. (1974). Social organization of the African Bovidae. In V. Geist & E. Walther (Eds.), The behaviour of ungulates and its relation to management (Vol. 1, pp. 166-205),. International Union for Conservation of Nature and Natural Resources. Morges, Switzerland.
• Figueirido, B., Pérez-Claros, J. A., Torregrosa, V., Martín-Serra, A., & Palmqvist, P. (2010). Demythologizing Arctodus simus, the'short-faced’long-legged and predaceous bear that never was. Journal of Vertebrate Paleontology, 30(1), 262-275.
• Figueirido, B., Pérez-Ramos, A., Schubert, B. W., Serrano, F., Farrell, A. B., Pastor, F. J., Neves, A. A., & Romero, A. (2017). Dental caries in the fossil record: a window to the evolution of dietary plasticity in an extinct bear. Scientific reports, 7(1), 17813.
• Harris, M. A., & Steudel, K. (1997). Ecological correlates of hind‐limb length in the Carnivora. Journal of Zoology, 241(2), 381-408. https://doi.org/https://doi.org/10.1111/j.1469-7998.1997.tb01966.x
• Meloro, C. (2011). Feeding habits of Plio-Pleistocene large carnivores as revealed by the mandibular geometry. Journal of Vertebrate Paleontology, 31(2), 428-446. https://doi.org/https://doi.org/10.1080/02724634.2011.550357
• Mihlbachler, M. C., & Solounias, N. (2006). Coevolution of tooth crown height and diet in oreodonts (Merycoidodontidae, Artiodactyla) examined with phylogenetically independent contrasts. Journal of Mammalian Evolution, 13(1), 11-36. https://doi.org/https://doi.org/10.1007/s10914-005-9001-3
• Owen‐Smith, N., & Mills, M. G. (2008). Predator–prey size relationships in an African large‐mammal food web. Journal of Animal Ecology, 77(1), 173-183. https://doi.org/10.1111/j.1365-2656.2007.01314.x
• Pennell, M. W., Eastman, J. M., Slater, G. J., Brown, J. W., Uyeda, J. C., FitzJohn, R. G., Alfaro, M. E., & Harmon, L. J. (2014). geiger v2. 0: an expanded suite of methods for fitting macroevolutionary models to phylogenetic trees. Bioinformatics, 30(15), 2216-2218. https://doi.org/10.1093/bioinformatics/btu181
• Penteriani, V., & Melletti, M. (2020). Bears of the world: Ecology, conservation and management. Cambridge University Press.
• Raia, P., Castiglione, S., Serio, C., Mondanaro, A., Melchionna, M., & Di Febbraro, M. (2019). RRphylo: Phylogenetic Ridge Regression Methods for Comparative Studies. R package version, 2(0).
• Revell, L. J. (2009). Size‐correction and principal components for interspecific comparative studies. Evolution: International Journal of Organic Evolution, 63(12), 3258-3268. https://doi.org/10.1111/j.1558-5646.2009.00804.x
• Revell, L. J. (2012). phytools: an R package for phylogenetic comparative biology (and other things). Methods in Ecology and Evolution, 3(2), 217-223. https://doi.org/doi:10.1111/j.2041-210X.2011.00169.x
• Roka, B., Jha, A. K., & Chhetri, D. R. (2021). A study on plant preferences of red panda (Ailurus fulgens) in the wild habitat: foundation for the conservation of the species. Acta Biologica Sibirica, 7, 425. https://doi.org/ https://doi.org/10.3897/abs.7.e71816
• Sunquist, M. E., & Sunquist, F. C. (1989). Ecological constraints on predation by large felids. In J. L. Gittleman (Ed.), Carnivore behavior, ecology, and evolution (pp. 283-301). Springer, Boston, MA. https://doi.org/https://doi.org/10.1007/978-1-4757-4716-4_11
• Zachos, J. C., Dickens, G. R., & Zeebe, R. E. (2008). An early Cenozoic perspective on greenhouse warming and carbon-cycle dynamics. Nature, 451(7176), 279-283. https://doi.org/10.1038/nature06588