Data and code from: Phylogenetic evidence that migratory behavior increases ungulate extinction rates
Data files
Oct 01, 2026 version files 1.27 MB
-
All_ungulate_state_data.csv
33.06 KB
-
Data_to_add_extinct_ungulates.csv
14.42 KB
-
Extinct-extant_migration_differences.csv
5.22 KB
-
Functional_diversity_declines.csv
25.13 KB
-
HiSSE_all_rates_extant_only.csv
82.46 KB
-
HiSSE_all_rates_fossil_inclusive.csv
75.11 KB
-
Migration_extinction_rates_code.R
125.82 KB
-
Rate_tests_extant_only.csv
30.30 KB
-
Rate_tests_fossil_inclusive.csv
14.60 KB
-
README.md
36.31 KB
-
Ungulate_phylogenies_fossil_inclusive.nex
826.35 KB
Abstract
Widespread instances of migratory collapse have been documented within the past several decades, indicating that migration may be a fragile strategy in the Anthropocene. Yet the fragility of migration as a strategy for navigating seasonally variable resource availability may extend deeper into evolutionary time. We used phylogenetic imputation to infer the migratory habits of 51 ungulate species that went extinct during the late Quaternary and evaluated whether migration enhances extinction rates on macroevolutionary timescales. We found that approximately 80% (41 of 51) of recently extinct species were migratory, whereas only 46% (95 of 207) of extant ungulates are migratory. Furthermore, Hidden State Speciation and Extinction modeling indicated that migratory species likely experienced higher speciation and extinction rates compared to non-migratory species, resulting in elevated lineage turnover for migratory ungulates, though diversification parameter estimates depended strongly upon whether extinct species were included in analyses. The extinctions of these ungulates led to negligible declines in global ungulate functional richness (0.8%) but more substantial declines (2.7%) across Africa. Altogether, our findings indicate that migration has complex macroevolutionary consequences, enhancing both speciation and extinction rates, which may render migration an evolutionarily fragile strategy, particularly as increasingly widespread and intensive anthropogenic activity cause gradients of resource availability to become severed (e.g., via human infrastructure) and more unpredictable (e.g., via climate change).
Dataset DOI: 10.5061/dryad.jm63xsjsp
Description of the data and file structure
The data in this repository were compiled and generated as part of the analyses described in the publication associated with this repository. Data on the body mass, dietary grass fraction, and latitudinal range centroids of extinct ungulates were collated from published literature. These extinct ungulates were grafted to a published set of mammal-wide phylogenies, with placements informed by their taxonomic relationships to extant species. Phylogenetic imputation was then used to infer the migratory habits of extinct ungulates from the relationships between migration and other characters within extant ungulates. Then, Hidden State Speciation and Extinction modeling was conducted to ascertain the effects of migration on ungulate macroevolutionary dynamics.
Files and variables
File: All_ungulate_state_data.csv
Description: All character data necessary for the analyses described in the paper associated with this repository, including placements of taxa within the phylogeny, taxonomic information, extinction statuses, body masses, dietary grass fractions, distribution data, and migratory statuses. Migratory statuses of extinct taxa were inferred via phylogenetic imputation; values for extinct taxa within this file correspond to their median migratory statuses across all 100 imputations. These data were used to generate figures 1 and S1 in the associated paper.
Variables
- Scientific.name: the scientific name (genus followed by species) of the ungulate species included in these analyses (207 extant, 51 extinct)
- Tip.label: the name of the tip associated with the particular taxa within the phylogeny set also included in this repository (Ungulate_phylogenies_fossil_inclusive.nex)
- Genus: the genus to which the ungulate species belongs
- Family: the family to which the ungulate species belongs
- Order: the order to which the ungulate species belongs
- Extinct: whether the species is extinct (1) or extant (0)
- Body.mass: the body mass of the species, in kilograms (kg)
- Dietary.grass.fraction: the proportion of the species' diet comprised of grass, ranging from a diet comprised of no grass (0) to a diet comprised entirely of grass (1).
- Latitudinal.range.centroid: the latitudinal center of the range of the species, in decimal degrees
- Latitude.minimum: the minimum latitude of the range of the species, in decimal degrees
- Latitude.maximum: the maximum latitude of the range of the species, in decimal degrees
- African: whether the species occurs in African (1) or does not (0)
- Migratory.status: whether the species exhibits migratory behavior (1) or does not (0)
File: Extinct-extant_migration_differences.csv
Description: The difference between the fraction of extinct ungulates that migrated versus the fraction extant ungulates that exhibit migratory behavior considering all ungulates or just African ungulates, as well as the statistical significance of those differences, for all 100 tree topologies. These data were used to generate figure S1 in the associated paper.
Variables
- tree: the identity of the tree used within the set of 100 ungulate phylogenies (1-100)
- difference_all: the difference between the fraction of extinct ungulates that migrated versus the fraction extant ungulates that exhibit migratory behavior, considering all ungulates
- significance_all: the statistical significance of the difference between extinct and extant ungulates when considering all ungulates, as assessed using logistic phylogenetic generalized linear modeling
- difference_African: the difference between the fraction of extinct ungulates that migrated versus the fraction extant ungulates that exhibit migratory behavior, considering just African ungulates
- significance_African: the statistical significance of the difference between extinct and extant ungulates when considering just African ungulates, as assessed using logistic phylogenetic generalized linear modeling
File: Functional_diversity_declines.csv
Description: The functional diversity of different subsets of ungulate species for all 100 tree topologies. Functional diversity was estimated from Gower distance matrices built using data on the four characters employed in this study (body mass, dietary grass fraction, latitudinal range centroid, and migration). These data were used to generate figure 3 in the associated paper.
Variables
- tree: the identity of the tree used within the set of 100 ungulate phylogenies (1-100)
- All_extant: the estimated functional diversity of all extant ungulate species for a given tree
- All_extant+extinct: the estimated functional diversity of all extant and extinct ungulate species for a given tree
- All_extinct: the estimated functional diversity of all extinct ungulate species for a given tree
- All_relative.decline: the relative decline in functional diversity when extinct ungulates were lost, when considering all ungulate species
- North.America_extant: the estimated functional diversity of extant North American ungulate species for a given tree
- North.America_extant+extinct: the estimated functional diversity of extant and extinct North American ungulate species for a given tree
- North.America_extinct: the estimated functional diversity of extinct North American ungulate species for a given tree
- North.America_relative.decline: the relative decline in functional diversity when extinct ungulates were lost, when considering just North American ungulate species
- South.America_extant: the estimated functional diversity of extant South American ungulate species for a given tree
- South.America_extant+extinct: the estimated functional diversity of extant and extinct South American ungulate species for a given tree
- South.America_extinct: the estimated functional diversity of extinct South American ungulate species for a given tree
- South.America_relative.decline: the relative decline in functional diversity when extinct ungulates were lost, when considering just South American ungulate species
- Eurasia_extant: the estimated functional diversity of extant Eurasian ungulate species for a given tree
- Eurasia_extant+extinct: the estimated functional diversity of extant and extinct Eurasian ungulate species for a given tree
- Eurasia_extinct: the estimated functional diversity of extinct Eurasian ungulate species for a given tree
- Eurasia_relative.decline: the relative decline in functional diversity when extinct ungulates were lost, when considering just Eurasian ungulate species
- Africa_extant: the estimated functional diversity of extant African ungulate species for a given tree
- Africa_extant+extinct: the estimated functional diversity of extant and extinct African ungulate species for a given tree
- Africa_extinct: the estimated functional diversity of extinct African ungulate species for a given tree
- Africa_relative.decline: the relative decline in functional diversity when extinct ungulates were lost, when considering just African ungulate species
File: HiSSE_all_rates_extant_only.csv
Description: The state-specific estimates of all macroevolutionary parameters resulting from Hidden State Speciation and Extinction modeling for all 100 tree topologies, when extant-only phylogenies were employed. These data were used to generate figure S6 in the associated paper.
Variables
- parameter_type: the type of macroevolutionary parameter (extinction rate, speciation rate, lineage turnover rate, net diversification rate, or extinction fraction)
- state: the complete state (0A, 0B, 1A, 1B) to which the parameter estimate corresponds, reflecting both the hidden state and observed state
- observed_state: the observed state (migratory or non-migratory) to which the parameter estimate corresponds
- hidden_state: the hidden state (A or B) to which the parameter estimate corresponds
- value: the estimated value of a given macroevolutionary parameter for a given state; units for macroevolutionary rates are in species per lineages per million years, while extinction fraction is unitless
File: HiSSE_all_rates_fossil_inclusive.csv
Description: The state-specific estimates of all macroevolutionary parameters resulting from Hidden State Speciation and Extinction modeling for all 100 tree topologies, when fossil-inclusive phylogenies were employed. These data were used to generate figure S5 in the associated paper.
Variables
- parameter_type: the type of macroevolutionary parameter (extinction rate, speciation rate, lineage turnover rate, net diversification rate, or extinction fraction)
- state: the complete state (0A, 0B, 1A, 1B) to which the parameter estimate corresponds, reflecting both the hidden state and observed state
- observed_state: the observed state (migratory or non-migratory) to which the parameter estimate corresponds
- hidden_state: the hidden state (A or B) to which the parameter estimate corresponds
- value: the estimated value of a given macroevolutionary parameter for a given state; units for macroevolutionary rates are in species per lineages per million years, while extinction fraction is unitless
File: Migration_extinction_rates_code.R
Description: Code to conduct all analyses and generate all figures for the paper associated with this dataset. Note that two files necessary to run the entirety of this script are housed elsewhere. The set of mammal-wide phylogenies onto which extinct taxa were grafted as part of this project are available from https://vertlife.org/phylosubsets/. We used 100 different tree configurations from the published set of birth-death node-dated completed trees here. Continental occurrence data for extant ungulates were derived from the IUCN Red List (https://www.iucnredlist.org/); continental occurrence data for extinct ungulates are present in the file 'Data to add extinct ungulates.csv', in column 'Continent'.
File: Rate_tests_extant_only.csv
Description: Within-tree averages of all macroevolutionary parameters as well as the results of phylogenetic generalized linear modeling evaluating the effects of migration on all macroevolutionary parameters for all 100 tree topologies, when extant-only phylogenies were employed. These data were used to generate figures S3 and S4 in the associated paper.
Variables
- tree_num: the identity of the tree used within the set of 100 ungulate phylogenies (1-100)
- extinction_1: the average extinction rate for migratory lineages for a given tree, in units of species per lineages per million years
- extinction_0: the average extinction rate for non-migratory lineages for a given tree, in units of species per lineages per million years
- extinction_difference: the average difference in extinction rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- extinction_effect: the modeled effect of migration on extinction rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- extinction_significance: the statistical significance of the effect of migration on extinction rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- speciation_1: the average speciation rate for migratory lineages for a given tree, in units of species per lineages per million years
- speciation_0: the average speciation rate for non-migratory lineages for a given tree, in units of species per lineages per million years
- speciation_difference: the average difference in speciation rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- speciation_effect: the modeled effect of migration on speciation rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- speciation_significance: the statistical significance of the effect of migration on speciation rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- turnover_1: the average lineage turnover rate for migratory lineages for a given tree, in units of species per lineages per million years
- turnover_0: the average lineage turnover rate for non-migratory lineages for a given tree, in units of species per lineages per million years
- turnover_difference: the average difference in lineage turnover rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- turnover_effect: the modeled effect of migration on lineage turnover rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- turnover_significance: the statistical significance of the effect of migration on lineage turnover rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- net.div_1: the average net diversification rate for migratory lineages for a given tree, in units of species per lineages per million years
- net.div_0: the average net diversification rate for migratory lineages for a given tree, in units of species per lineages per million years
- net.div_difference: the average difference in net diversification rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- net.div_effect: the modeled effect of migration on net diversification rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- net.div_significance: the statistical significance of the effect of migration on net diversification rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- ext.frac_1: the average extinction fraction for migratory lineages for a given tree, which is unitless
- ext.frac_0: the average extinction fraction for non-migratory lineages for a given tree, which is unitless
- ext.frac_difference: the average difference in extinction fraction between migratory and non-migratory lineages for a given tree, which is unitless
- ext.frac_effect: the modeled effect of migration on extinction fraction for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- ext.frac_significance: the statistical significance of the effect of migration on extinction fraction for a given tree, as assessed using logistic phylogenetic generalized linear modeling
File: Rate_tests_fossil_inclusive.csv
Description: Within-tree averages of all macroevolutionary parameters as well as the results of phylogenetic generalized linear modeling evaluating the effects of migration on all macroevolutionary parameters for all 100 tree topologies, when fossil-inclusive phylogenies were employed. These data were used to generate figures 2 and S2 in the associated paper.
Variables
- tree_num: the identity of the tree used within the set of 100 ungulate phylogenies (1-100)
- extinction_1: the average extinction rate for migratory lineages for a given tree, in units of species per lineages per million years
- extinction_0: the average extinction rate for non-migratory lineages for a given tree, in units of species per lineages per million years
- extinction_difference: the average difference in extinction rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- extinction_effect: the modeled effect of migration on extinction rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- extinction_significance: the statistical significance of the effect of migration on extinction rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- speciation_1: the average speciation rate for migratory lineages for a given tree, in units of species per lineages per million years
- speciation_0: the average speciation rate for non-migratory lineages for a given tree, in units of species per lineages per million years
- speciation_difference: the average difference in speciation rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- speciation_effect: the modeled effect of migration on speciation rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- speciation_significance: the statistical significance of the effect of migration on speciation rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- turnover_1: the average lineage turnover rate for migratory lineages for a given tree, in units of species per lineages per million years
- turnover_0: the average lineage turnover rate for non-migratory lineages for a given tree, in units of species per lineages per million years
- turnover_difference: the average difference in lineage turnover rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- turnover_effect: the modeled effect of migration on lineage turnover rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- turnover_significance: the statistical significance of the effect of migration on lineage turnover rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- net.div_1: the average net diversification rate for migratory lineages for a given tree, in units of species per lineages per million years
- net.div_0: the average net diversification rate for migratory lineages for a given tree, in units of species per lineages per million years
- net.div_difference: the average difference in net diversification rate between migratory and non-migratory lineages for a given tree, in units of species per lineages per million years
- net.div_effect: the modeled effect of migration on net diversification rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- net.div_significance: the statistical significance of the effect of migration on net diversification rate for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- ext.frac_1: the average extinction fraction for migratory lineages for a given tree, which is unitless
- ext.frac_0: the average extinction fraction for non-migratory lineages for a given tree, which is unitless
- ext.frac_difference: the average difference in extinction fraction between migratory and non-migratory lineages for a given tree, which is unitless
- ext.frac_effect: the modeled effect of migration on extinction fraction for a given tree, as assessed using logistic phylogenetic generalized linear modeling
- ext.frac_significance: the statistical significance of the effect of migration on extinction fraction for a given tree, as assessed using logistic phylogenetic generalized linear modeling
File: Data_to_add_extinct_ungulates.csv
Description: All the data necessary to add the extinct ungulate species to the phylogeny.
Variables
- Scientific name: the scientific name (genus followed by species) of the ungulate species included in these analyses (207 extant, 51 extinct)
- Tip label: the name of the tip associated with the particular taxa within the phylogeny set also included in this repository (Ungulate_phylogenies_fossil_inclusive.nex)
- Genus: the genus to which the ungulate species belongs
- Family: the family to which the ungulate species belongs
- Order: the order to which the ungulate species belongs
- mrca1: the tip name of the ungulate species to which the extinct taxa is most closely related, used to define where in the phylogeny the species should be added; NA values in this column indicate that the closest relative is another extinct taxa, and thus that taxa needed to be added first, or that the extinct taxa is already present in the tree (for taxa with NAs in this column and 'mrca2')
- mrca2: the tip name of the other ungulate species to which the extinct taxa is most closely related, used to define where in the phylogeny the species should be added; NA values in this column indicate that there is only one closely related species, or that the extinct taxa is already present in the tree (for taxa with NAs in this column and 'mrca1')
- nodeAge: the first recorded occurrence of the species in the fossil record (in millions of years), and thus the age used to anchor the node when adding the species to the phylogeny set
- tipAge: the last recorded occurrence of the species in the fossil record (in millions of years), and thus the age used to anchor the tip when adding the species to the phylogeny set
- branchLength: the difference between the first and last recorded occurrence of the species in the fossil record (in millions of years), and thus the age used to define the branch length when adding the species to the phylogeny set
- Continent: the continents on which the species occurred
- African: whether the species occurs in Africa (1) or does not (0)
- Latitude minimum: the minimum latitude of the range of the species, in decimal degrees
- Latitude maximum: the maximum latitude of the range of the species, in decimal degrees
- Latitudinal range centroid: the latitudinal center of the range of the species, in decimal degrees
- Body mass: the body mass of the species, in kilograms (kg)
- Dietary grass fraction: the proportion of the species' diet comprised of grass, ranging from a diet comprised of no grass (0) to a diet comprised entirely of grass (1)
- Taxonomy_source: the citation of the data sources from which taxonomic information was derived (see 'Access information' below)
- Mass_source: the citation of the data sources from which body mass data were derived (see 'Access information' below)
- Diet_source: the citation of the data sources from which diet data were derived (see 'Access information' below)
- Diet_notes: description of the type of data used to estimate dietary grass fraction
File: Ungulate_phylogenies_fossil_inclusive.nex
Description: The set of 100 different ungulate phylogenies employed in the analyses described in the paper associated with this repository, including 38 additional extinct ungulate species which were added to the phylogeny using their taxonomic relationships to extant species. Note that this file was generated using the R code provided in this repository.
Code/software
Analyses were conducted in R using publicly available R packages, as cited in the text of the manuscript associated with this dataset. The R code to reproduce the analyses in the associated manuscript is included in this data repository.
Access information
Other publicly accessible locations of the data:
- These data are also available in the Supporting Information of the publication associated with this repository
Data was derived from the following sources:
- Abraham, J. O., Upham, N. S., Damian-Serrano, A., and Jesmer, B. R. (2022). Evolutionary causes and consequences of ungulate migration. Nature Ecology and Evolution, 6(7), 998-1006.
- Alcover, J. A., Perez-Obiol, R., Yll, E. I., & Bover, P. (1999). The diet of Myotragus balearicus Bate 1909 (Artiodactyla: Caprinae), an extinct bovid from the Balearic Islands: evidence from coprolites. Biological Journal of the Linnean Society, 66(1), 57-74.
- Barrón-Ortiz, C. I., Avilla, L. S., Jass, C. N., Bravo-Cuevas, V. M., Machado, H., and Mothé, D. (2019). What is Equus? Reconciling taxonomy and phylogenetic analyses. Frontiers in Ecology and Evolution, 7, 343.
- Bernor, R. L., Armour-Chelu, M., Gilbert, H., Kaiser, T. M., and Schulz, E. (2010). “Equidae,” in Cenozoic mammals of Africa, eds L. Werdelin and W. L. Sanders (Berkeley, CA: University of California Press), 685–721.
- Bibi, F., and Kiessling, W. (2015). Continuous evolutionary change in Plio-Pleistocene mammals of eastern Africa. Proceedings of the National Academy of Sciences, 112(34), 10623-10628.
- Bibi, F., and Cantalapiedra, J. L. (2023). Plio-Pleistocene African megaherbivore losses associated with community biomass restructuring. Science, 380(6649), 1076-1080.
- Bocherens, H., Koch, P. L., Mariotti, A., Geraads, D., & Jaeger, J. J. (1996). Isotopic biogeochemistry (13 C, 18 O) of mammalian enamel from African Pleistocene hominid sites. Palaios, 306-318.
- Bradham, J. L., DeSantis, L. R., Jorge, M. L. S., & Keuroghlian, A. (2018). Dietary variability of extinct tayassuids and modern white-lipped peccaries (Tayassu pecari) as inferred from dental microwear and stable isotope analysis. Palaeogeography, Palaeoclimatology, Palaeoecology, 499, 93-101.
- Carrasco, T. S., Scherer, C. S., Ribeiro, A. M., and Buchmann, F. S. (2022). Paleodiet of Lamini camelids (Mammalia: Artiodactyla) from the Pleistocene of southern Brazil: insights from stable isotope analysis (δ13C, δ18O). Paleobiology, 48(3), 513-526.
- Cerling, T. E., Andanje, S. A., Blumenthal, S. A., Brown, F. H., Chritz, K. L., Harris, J. M., ... and Uno, K. T. (2015). Dietary changes of large herbivores in the Turkana Basin, Kenya from 4 to 1 Ma. Proceedings of the National Academy of Sciences, 112(37), 11467-11472.
- Chritz, K. L., Dyke, G. J., Zazzo, A., Lister, A. M., Monaghan, N. T., and Sigwart, J. D. (2009). Palaeobiology of an extinct Ice Age mammal: Stable isotope and cementum analysis of giant deer teeth. Palaeogeography, Palaeoclimatology, Palaeoecology, 282(1-4), 133-144.
- Codron, D., Bousman, C. B., Buschke, F., Clauss, M., Lewis, C., van der Merwe, R., ... and Brink, J. S. (2023). Competition drives the evolution of emergent neutrality in the dietary niches of mammalian herbivores. Quaternary International, 650, 28-39.
- Crowley, B. E., Godfrey, L. R., and Samonds, K. E. (2023). What can hippopotamus isotopes tell us about past distributions of C4 grassy biomes on Madagascar?. Plants, People, Planet, 5(6), 997-1010.
- Croitor, R. (2018). Plio-Pleistocene deer of western Palearctic: taxonomy, systematics, phylogeny. Institute of Zoology of the Academy of Sciences of Moldova.
- Croitor, R. (2021). Taxonomy, systematics and evolution of giant deer Megaloceros giganteus (Blumenbach, 1799)(Cervidae, Mammalia) from the Pleistocene of Eurasia. Quaternary, 4(4), 36.
- Davis, E. B. (2007). Family antilocapridae. The Evolution of Artiodactyls. Johns Hopkins University Press, Baltimore, Maryland, 227-240.
- Faith, J. T. (2014). Late Pleistocene and Holocene mammal extinctions on continental Africa. Earth-Science Reviews, 128, 105-121.
- Faith, J. T., Potts, R., Plummer, T. W., Bishop, L. C., Marean, C. W., and Tryon, C. A. (2012). New perspectives on middle Pleistocene change in the large mammal faunas of East Africa: Damaliscus hypsodon sp. nov.(Mammalia, Artiodactyla) from Lainyamok, Kenya. Palaeogeography, Palaeoclimatology, Palaeoecology, 361, 84-93.
- Faith, J. T., Tryon, C. A., Peppe, D. J., Beverly, E. J., and Blegen, N. (2014). Biogeographic and evolutionary implications of an extinct Late Pleistocene impala from the Lake Victoria Basin, Kenya. Journal of Mammalian Evolution, 21, 213-222.
- Faith, J. T., Rowan, J., Du, A., and Koch, P. L. (2018). Plio-Pleistocene decline of African megaherbivores: No evidence for ancient hominin impacts. Science, 362(6417), 938-941.
- Fernandez, P., Bouzouggar, A., Collina-Girard, J., and Coulon, M. (2015). The last occurrence of Megaceroides algericus Lyddekker, 1890 (Mammalia, Cervidae) during the middle Holocene in the cave of Bizmoune (Morocco, Essaouira region). Quaternary International, 374, 154-167.
- Gasparini, G. M., and Ubilla, M. (2011). Platygonus sp.(Mammalia: Tayassuidae) in Uruguay (Raigón? Formation; Pliocene–early Pleistocene), comments about its distribution and palaeoenvironmental significance in South America. Journal of Natural History, 45(45-46), 2855-2870.
- Guthrie, R. D. (1991). New paleoecological and paleoethological information on the extinct helmeted muskoxen from Alaska. In Annales Zoologici Fennici (pp. 175-186). Finnish Zoological Publishing Board, formed by the Finnish Academy of Sciences, Societas Biologica Fennica Vanamo, Societas pro Fauna et Flora Fennica, and Societas Scientiarum Fennica.
- Hayashi, S., Kubo, M. O., Sánchez-Villagra, M. R., Taruno, H., Izawa, M., Shiroma, T., ... and Fujita, M. (2023). Variation and process of life history evolution in insular dwarfism as revealed by a natural experiment. Frontiers in Earth Science, 11, 1095903.
- Hempel, E., Bibi, F., Faith, J. T., Koepfli, K. P., Klittich, A. M., Duchêne, D. A., ... and Westbury, M. V. (2022). Blue turns to gray: paleogenomic insights into the evolutionary history and extinction of the blue antelope (Hippotragus leucophaeus). Molecular biology and evolution, 39(12), msac241.
- Hendey, Q. B. (1974). Faunal dating of the late Cenozoic of Southern Africa, with special reference to the Carnivora. Quaternary Research, 4(2), 149-161.
- Holanda, E. C., & Ferrero, B. S. (2013). Reappraisal of the genus Tapirus (Perissodactyla, Tapiridae): systematics and phylogenetic affinities of the South American tapirs. Journal of Mammalian Evolution, 20, 33-44.
- Kubo, M. O., and Fujita, M. (2021). Diets of Pleistocene insular dwarf deer revealed by dental microwear texture analysis. Palaeogeography, Palaeoclimatology, Palaeoecology, 562, 110098.
- Kuzmin, Y. V., Kosintsev, P. A., Stepanov, A. D., Boeskorov, G. G., and Cruz, R. J. (2017). Chronology and faunal remains of the Khayrgas cave (Eastern Siberia, Russia). Radiocarbon, 59(2), 575-582.
- Lambert, W. D., and Holling, C. S. (1998). Causes of ecosystem transformation at the end of the Pleistocene: evidence from mammal body-mass distributions. Ecosystems, 1, 157-175.
- Liang, Q., Chen, Q., Zhang, N., Zhang, W., Ning, C., and Cai, D. (2023). Subsistence strategies in the Late Neolithic and Bronze Age in Nenjiang River Basin: A zooarchaeological and stable isotope analysis of faunal remains at Honghe site, Northeast China. International Journal of Osteoarchaeology, 33(2), 271-284.
- Lynch, S., Sánchez-Villagra, M. R., & Balcarcel, A. (2020). Description of a fossil camelid from the Pleistocene of Argentina, and a cladistic analysis of the Camelinae. Swiss journal of palaeontology, 139, 1-17.
- McFadden, B. (1992). Fossil horses. Systematics, Paleobiology, and Evolution of the Family Equidae: Cambridge University Press, New York.
- Mead, J. I., O'Rourke, M. K., and Foppe, T. M. (1986). Dung and diet of the extinct Harrington's mountain goat (Oreamnos harringtoni). Journal of Mammalogy, 67(2), 284-293.
- Nacarino-Meneses, C. (2023). Evolution of Equid Body Size. In The Equids: A Suite of Splendid Species (pp. 113-141). Cham: Springer International Publishing.
- O’Brien, H. D., Faith, J. T., Jenkins, K. E., Peppe, D. J., Plummer, T. W., Jacobs, Z. L., ... and Tryon, C. A. (2016). Unexpected convergent evolution of nasal domes between Pleistocene bovids and Cretaceous hadrosaur dinosaurs. Current Biology, 26(4), 503-508.
- Orlando, L., Metcalf, J. L., Alberdi, M. T., Telles-Antunes, M., Bonjean, D., Otte, M., ... and Cooper, A. (2009). Revising the recent evolutionary history of equids using ancient DNA. Proceedings of the National Academy of Sciences, 106(51), 21754-21759.
- Palombo, M. R., Kohler, M., Sola, S. M., and Giovinazzo, C. (2008). Brain versus body mass in endemic ruminant artiodactyls: a case studied of Myotragus balearicus and smallest Candiacervus species from Mediterranean Islands. Quaternary International, 182(1), 160-183.
- Pardi, M. I., and DeSantis, L. R. (2021). Dietary plasticity of North American herbivores: a synthesis of stable isotope data over the past 7 million years. Proceedings of the Royal Society B, 288(1948), 20210121.
- Pickford, M. (2013). Locomotion, diet, body weight, origin and geochronology of Metridiochoerus andrewsi from the Gondolin karst deposits, Gauteng, South Africa. Annals of the Ditsong National Museum of Natural History, 3(1), 33-47.
- Polling, M., ter Schure, A. T., van Geel, B., van Bokhoven, T., Boessenkool, S., MacKay, G., ... and Gravendeel, B. (2021). Multiproxy analysis of permafrost preserved faeces provides an unprecedented insight into the diets and habitats of extinct and extant megafauna. Quaternary Science Reviews, 267, 107084.
- Prado, J. L., Sánchez, B., and Alberdi, M. T. (2011). Ancient feeding ecology inferred from stable isotopic evidence from fossil horses in South America over the past 3 Ma. BMC ecology, 11, 1-13.
- Prado, J. L., and Alberdi, M. T. (2017). Fossil Horses of South America. Phylogeny, systemics and ecology. Cham: Springer.
- Prothero, D. R. (2021). The systematics of North American peccaries (Mammalia: artiodactyla: tayassuidae) (Vol. 85). New Mexico Museum of Natural History and Science.
- Saarinen, J., Eronen, J., Fortelius, M., Seppä, H., and Lister, A. M. (2016). Patterns of diet and body mass of large ungulates from the Pleistocene of Western Europe, and their relation to vegetation.
- Saysette, J. E. (1999). Postcranial estimators of body mass in pecorans with emphasis on Capromeryx (Mammalia: Artiodactyla). Colorado State University.
- Sewell, L. (2019). Using a multiproxy analysis of springbok fossils to track 2 million years of vegetation changes experienced by hominins in southern Africa (Doctoral dissertation, Bournemouth University).
- Smith, F. A., Lyons, S. K., Ernest, S. M., Jones, K. E., Kaufman, D. M., Dayan, T., ... and Haskell, J. P. (2003). Body mass of late quaternary mammals: ecological archives E084‐094. Ecology, 84(12), 3403-3403.
- Sondaar, P. Y., & Van der Geer, A. A. E. (2005). Evolution and extinction of Plio-Pleistocene island ungulates. Quaternaire, 2, 241-256.
- Spasskaya, N. N., Pavlinov, I. Y., Sharko, F. S., Boulygina, E. S., Tsygankova, S. V., Nedoluzhko, A. V., ... and Mashchenko, E. N. (2021). Morphometric and genetic analyses of diversity of the Lena horse (Equus lenensis Russanov, 1968; Mammalia: Equidae). Russian Journal of Theriology, 20(1), 82-95.
- Stefaniak, K., Stachowicz-Rybka, R., Borówka, R. K., Hrynowiecka, A., Sobczyk, A., Moskal-del Hoyo, M., ... and Kovalchuk, O. (2021). Browsers, grazers or mix-feeders? Study of the diet of extinct Pleistocene Eurasian forest rhinoceros Stephanorhinus kirchbergensis (Jäger, 1839) and woolly rhinoceros Coelodonta antiquitatis (Blumenbach, 1799). Quaternary International, 605, 192-212.
- Upham, N. S., Esselstyn, J. A., & Jetz, W. (2019). Inferring the mammal tree: species-level sets of phylogenies for questions in ecology, evolution, and conservation. PLoS biology, 17(12), e3000494.
- Welker, F., Smith, G. M., Hutson, J. M., Kindler, L., Garcia-Moreno, A., Villaluenga, A., ... and Gaudzinski-Windheuser, S. (2017). Middle Pleistocene protein sequences from the rhinoceros genus Stephanorhinus and the phylogeny of extant and extinct Middle/Late Pleistocene Rhinocerotidae. PeerJ, 5, e3033.
- Weston, E. M., and Lister, A. M. (2009). Insular dwarfism in hippos and a model for brain size reduction in Homo floresiensis. Nature, 459(7243), 85-88.
- Willerslev, E., Davison, J., Moora, M., Zobel, M., Coissac, E., Edwards, M. E., ... and Taberlet, P. (2014). Fifty thousand years of Arctic vegetation and megafaunal diet. Nature, 506(7486), 47-51.
- Wimberly, A. N. (2023). Predicting body mass in Ruminantia using postcranial measurements. Journal of Morphology, 284(10), e21636.
- Wu, D. D., Ding, X. D., Wang, S., Wójcik, J. M., Zhang, Y. I., Tokarska, M., ... and Zhang, Y. P. (2018). Pervasive introgression facilitated domestication and adaptation in the Bos species complex. Nature ecology and evolution, 2(7), 1139-1145.
- Yang, D., Pisano, A., Kolasa, J., Jashashvili, T., Kibii, J., Cano, A. R. G., ... and Souron, A. (2022). Why the long teeth? Morphometric analysis suggests different selective pressures on functional occlusal traits in Plio-Pleistocene African suids. Paleobiology, 48(4), 655-676.
We first identified extinct species suitable for inclusion in our analyses. We included any ungulate taxa (Perissodactyla and Artiodactyla, excluding Cetacea) that have persisted through the late Quaternary (late Pleistocene and Holocene) within extant families, as families without extant representatives are difficult to place phylogenetically. This limited the available species pool to taxa that arose within the Neogene (< 23 Ma, the interval during which phylogenetic evidence suggests that migratory behavior in ungulates first evolved).
Migratory behavior in extant ungulates is strongly positively associated with other characters, specifically body mass, dietary grass fraction, and latitude: these three characters together explain ~44% of variation in whether or not ungulate species are capable of migration. We therefore used the correlations between migration and these characters within extant species to infer migratory behavior in extinct taxa. Data on the migratory habits of extant ungulates, as well as their diets, body masses, and latitudinal niches, were available from Abraham et al. 2022. Migratory behavior varies substantially within species and across populations (with some populations of a species being entirely non-migratory and others almost completely migratory) but is not well-catalogued at population scale. For this reason, and because anthropogenic factors such as physical barriers and habitat alterations have modified the distribution of migratory behavior in many taxa, migratory behavior was coded as binary: a species was considered migratory if it was explicitly described as such in published literature or if any population, past or present, exhibited seasonal round-trip movements between discrete ranges. While this inclusive classification may overestimate the importance of migration for some taxa or underestimate it for data-deficient ones, it provides a consistent framework for macroevolutionary comparison across taxa and time.
We then compiled data on the dietary grass fractions, body masses, and latitudinal range centroids of extinct ungulates. We estimated latitudinal range centroids from fossil locality data in the Paleobiology Database (http://paleodb.org), refined with expert knowledge from additional published sources. Latitude provides a relative and internally consistent proxy for the magnitude of seasonality to which species are exposed, particularly given that Earth’s climate has varied markedly through geological time and fine-scale paleoclimatic reconstructions sufficient to infer migratory behavior are unavailable. Body mass estimates were derived from various sources via targeted literature searches; estimates were generally derived from lineage-specific extrapolations of body mass from the skeletal allometry of extant relatives. Dietary grass fraction data were likewise compiled from various published sources. These data were predominately estimated via stable isotope analysis of fossilized material (e.g., tooth enamel) or DNA metabarcoding of coprolites. Values range from 0 (a diet comprising no grass) to 1 (a diet comprising exclusively grass). Reliable quantitative diet data were available for a limited number of species (n = 51) and served as the biggest barrier to the inclusion of taxa.
One of the most widely used proxies for reconstructing herbivore dietary grass fraction, stable carbon isotope data, works best in the tropics and especially in tropical Africa, where nearly all grasses are C4-photosynthesizing and grasses consequently have a unique isotopic signature; outside of Africa, especially at higher latitudes, both C3- and C4-photosynthesizing grasses are prevalent, so stable carbon isotopes cannot readily differentiate grasses from non-grasses. Accordingly, dietary grass estimates were available for the majority (>75%) of extinct African ungulates that met our other inclusion criteria, such that African taxa are well-represented in this study. We therefore categorized both extinct and extant species as either African or non-African: for extinct species, we determined whether a species occurred in Africa from the aforementioned fossil locality data from the Paleobiology Database; for extant species, we determined whether a species occurred in Africa from species range maps curated by the IUCN. This allowed us to repeat all analyses on only the African taxa as a ‘best quality’ data subset.
We then incorporated the 51 extinct ungulates with all available character data into a published set of mammal-wide phylogenies. For our analyses, we used 100 different tree configurations from the published set of birth-death node-dated completed trees, in which divergence times among major lineages were estimated using fossil-calibrated backbones. These phylogenies were constructed by combining a 31-gene supermatrix with Bayesian inference to generate time-scaled backbone and patch trees. Species lacking genetic data were placed using taxonomic constraints and branch lengths inferred from local birth-death models. The full trees contained 5,911 mammal species, including 13 recently extinct ungulate species for which genetic data were available. For our analyses, we pruned each tree in the set down to only ungulates (13 extinct and 207 extant species).
The remaining 38 extinct taxa not already included in the phylogeny set were grafted onto all 100 phylogenetic trees using the ‘bind.tip’ function in the ‘phytools‘ R package, following the procedure described in Turvey et al. 2021. For these 38 extinct species, we determined their phylogenetic placements using a multi-step approach specific to each species. First, we used Linnaean taxonomic classifications as a guiding framework. If a fossil species was confidently assigned to a particular genus, it was positioned within the crown or stem of that genus accordingly. Where available, we consulted the most recent morphological, molecular, or combined-data phylogenies for each extinct species to guide more precise placements. In such cases, extinct taxa were inserted on a case-by-case basis based on their inferred evolutionary relationships. For example, the three extinct antilocaprid species Capromeryx furcifer, Stockoceros conklingi, and Tetrameryx shuleri, all members of the extinct tribe Stockoceratini, were placed as a polytomy with respect to one another, and this polytomy was grafted as sister to their closest living relative, Antilocapra americana. In cases where finer-scale phylogenetic resolution was unavailable, taxa were conservatively placed as polytomies within the broader crown group of the relevant genus. For instance, no detailed phylogenetic information could be found for the extinct species Antidorcas australis and A. bondi, and so both were added as a polytomy with the extant A. marsupialis.
Once placements were determined, we anchored each extinct taxon within the tree by defining its branch length as the difference between its first and last appearances in the fossil record: the first occurrence was used as the node age (representing the branching point from its closest known relative), and the last occurrence was treated as the tip age. This approach allowed us to consistently define branch lengths across species, though we recognize that using observed stratigraphic ranges of fossil species necessarily truncates branch lengths at both ends; despite the fact that our approach underestimates branch lengths, the error it introduces should be negligible relative to the substantial age of ungulates overall (~60 Ma crown age).
Given that many placements produced polytomous clades, we used the ‘fix.poly’ function from the ‘RRphylo’ R package to resolve these polytomies into dichotomies by assigning arbitrarily small branch lengths (0.0001) to newly created internodes. This entire process was repeated across all 100 trees so we could propagate phylogenetic uncertainty through our subsequent analyses.
