Raw Data and Detailed Analytical Methods
The tree topology and branch lengths
The tree includes 72 extinct, worldwide distributed species, belonging to the Laurasiatheria and Proboscidea
clades within the crown group Mammalia. Most intrafamilial relationships are not known and are therefore left as
polytomies. This is known to have minor impact on variance-covariance matrix estimation (Finarelli and Flynn 2006,
Meloro et al. 2008, Raia et al. 2010). Species first appearance records in the NOW and Paleodb databases were used
to calculate branch lengths (Finarelli and Flynn 2006, Meloro et al. 2008, Raia et al. 2010). Ages for internal nodes,
corresponding to genera, families, orders, and other inclusive nodes were taken directly from the fossil record. For
further details see Meloro et al. (2008), Meloro et al. (2010), Carotenuto et al. (2010), Raia et al. (2010), Raia (2010),
and Raia et al. (2011). Ungulates fall within the clade Laurasiatheria, which includes hedgehogs + shrews + moles, bats,
whales + artiodactyls, carnivorans, pangolins and perissodactyls (Waddell et al. 1999).
The tree was produced in Mesquite (Maddison and Maddison 2010). As for ruminants, we used the
phylogenetic topology in Hernandez-Fernandez and Vrba (2005) and Decker et al. (2009) for living species and
interfamilial relationships, and the works of specialists for placing extinct species. The higher-level relationships
between
ruminants
are
now
reaching
a
consensus
and
support
a
(camels,(chevrotains,(pronghorns,(giraffes,(deer,bovids))))) topology (Decker et al. 2009, Spaulding et al. 2009).
Within Pecorans, the phylogeny of antilocapridae follows Semprebon and Rivals (2007). Paleomerycids were allied to
cervoids (Gentry et al. 1999). Within bovinae, tragelaphini and bovini are sister groups, with nilgai and its fossil
relatives (boselaphini) sister to both. Pliocervini were allied to Cervini (Petronio et al. 2007).
The phylogeny of rhinos follows the comprehensive cladistic analysis in Cerdeño (1995), and taxonomic
attributions in Lacombat (2003), for some of the species. Phylogeny and taxonomy of tapiromorph perissodactyls was
depicted after Holbrook (1999). The phylogeny and taxonomy of equids follows Strömberg (2006), and Maguire and
Stigall (2008).
Higher-level tree topology and split ages within Carnivora follows Wesley-Hunt and Flynn (2005) and Finarelli
and Flynn [2006, see also Finarelli (2008), Meloro and Raia 2010]. Canid subfamilies ages and relationships were taken
from Wang (1994), Wang et al. (1999), Finarelli and Flynn (2006), and Tedford et al. (2010). The phylogeny of
machairodont felids follows Slater and Van Valkenburgh (2008). The current phylogeny of Felidae is robust but
molecular estimates of time of divergence are not (Johnson et al. 2006), and were thence taken directly from the fossil
record. Taxonomy and phylogeny of hyenas follow Werdelin and Solounias (1991) and Turner et al. (2008).
For proboscideans, we followed the phylogenetic descriptions and age estimates in Shoshani and Tassy
(2005), and Thomas et al. (2000).
Literature cited
1. Carotenuto F., Barbera C., and Raia P., (2010). Occupancy, range size and phylogeny in Eurasian
Pliocene to Recent large mammals. Paleobiology 36, 399-414.
2. Cerdeño, E. (1995) Cladistic analysis of the family Rhinocerotidae (Perissodactyla). Am. Mus. Novit.,
1–25.
3. Decker JE et al. (2009) Resolving the evolution of extant and extinct ruminants with high throughput
phylogenomics. PNAS 106: 18644–18649.
4. Finarelli JA, Flynn JJ (2006) Ancestral state reconstruction of body size in the Caniformia Carnivora,
Mammalia: the effects of incorporating data from the fossil record. Syst. Biol. 55:301–313.
5. Finarelli, J. A.
(2008). A Total Evidence Phylogeny of the Arctoidea (Carnivora: Mammalia):
Relationships Among Basal Taxa. J Mammal Evol 15, 231–259.
6. Flynn JJ, Finarelli JA, Zehr S, Hsu J, Nedbal MA. Molecular phylogeny of the carnivora (mammalia):
assessing the impact of increased sampling on resolving enigmatic relationships. Syst. Biol. 2005,
317-337.
7. Fortelius, M. (coordinator) [2010]. Neogene of the Old World Database of Fossil Mammals (NOW).
University of Helsinki. http://www.helsinki.fi/science/now/.
8. Gentry AW, Rössner GE, Heizmann, EPJ (1999): Suborder Ruminantia. – In: Rössner GE, Heissig K
(editors): The Miocene Land Mammals of Europe: pp. 225-258. München (Verlag Dr. Friedrich Pfeil).
9. Hernandez-Fernandez, M., and Vrba, E.S., (2005) A complete estimate of the phylogenetic
relationships in Ruminantia, a dated species-level supertree of the extant ruminants: Biological
Reviews, v. 80, p. :269–302.
10. Holbrook, L. T. (1999) The Phylogeny and Classification of Tapiromorph Perissodactyls (Mammalia).
Cladistics 15, 331–350.
11. Johnson WE, Eizirik E, Pecon-Slattery J, Murphy WJ, Antunes A, Teeling E, O’Brien SJ. 2006. The late
Miocene radiation of modern Felidae: a genetic assessment. Science 311, 73-77.
12. Maddison, W. P. and D.R. Maddison. 2010. Mesquite: a modular system for evolutionary analysis.
Version 2.73 http://mesquiteproject.org
13. Maguire, K. C., and A. L. Stigall. (2008). Paleobiogeography of Miocene Equinae of North America: a
phylogenetic biogeographic analysis of the relative roles of climate, vicariance, and dispersal.
Palaeogeography, Palaeoclimatology, Palaeoecology 267:175–184
14. Meloro C., Raia P. (2010). Cats and dogs down the tree: the tempo and mode of evolution in the
lower carnassial of fossil and living Carnivora. Evolutionary Biology 37, 177-186.
15. Meloro, C., P. Raia P, P. Piras P, C. Barbera C, and P. O’Higgins. P (2008. ) The shape of the
mandibular corpus in large fissiped carnivores: allometry, function and phylogeny. Zool. J. Linn. Soc.
154:832–845.
16. Petronio C, Krakhmalnaya T, Bellucci L, Di Stefano G (2007) Remarks on some Eurasian pliocervines:
Characteristics, evolution, and relationships with the tribe Cervini. Geobios 40:113–130.
17. Raia P., (2010) Phylogenetic community assembly over time in Eurasian Plio-Pleistocene mammals.
Palaios 25, 327–338.
18. Raia P., Carotenuto F., Eronen J.T., Fortelius M. (in press) Longer in the tooth, shorter in the record?
The evolutionary correlates of hypsodonty in Neogene ruminants. Proc. Royal Soc Lon B
19. Raia P., Carotenuto F., Meloro C., Piras P., and D. Pushkina (2010). The shape of contention.
Adaptation, history and contingency in ungulate mandibles. Evolution 64 (5), 1489 -1503.
20. Semprebon, G. M. and Rivals, F. (2007) Was grass more prevalent in the pronghorn past? An
assessment of the dietary adaptations of Miocene to Recent Antilocapridae (Mammalia:
Artiodactyla). Palaeogeography, Palaeoclimatology, Palaeoecology 253, 332–347.
21. Shoshani, J., And Tassy, P., (2005), Advances in proboscidean taxonomy and classification, anatomy
and physiology, and ecology and behaviour: Quaternary International, v. 126–128, p. 5–20.
22. Slater, G.J., And Van Valkenburgh, B. (2008) Long in the tooth: Evolution of sabertooth cat cranial
shape: Paleobiology, v. 34, p. 403–419.
23. Spaulding M., O’Leary M.A., Gatesy J. (2009). Relationships of Cetacea (Artiodactyla) Among
Mammals: Increased Taxon Sampling Alters Interpretations of Key Fossils and Character Evolution.
PLoS ONE 4: e7062.
24. Strömberg, C. A. E. (2006). The evolution of hypsodonty in equids: testing a hypothesis of adaptation.
Paleobiology 32, 236–258.
25. Tedford RH, Wang X, Taylor BE. 2010. Phylogenetic systematics of the North American fossil Caninae
(Carnivora: Canidae). Bullettin of the American Museum of Natural History n. 325.
26. The Paleobiology Database. (http://www.paleodb.org/cgi-bin/bridge.pl). Occurrence data were
retrieved for Proboscideans and Laurasiatherians on December 2010.
27. Thomas, M.G., Hagelberg, E., Jones, H.B., Yang, Z., And Lister, A.M., (2000) Molecular and
morphological evidence on the phylogeny of the Elephantidae: Proceedings of the Royal Society
London B, v. 267, p. 2493–2500.
28. Turner A, Antón M, Werdelin L. (2008) Taxonomy and evolutionary patterns in the fossil Hyaenidae
of Europe. Geobios 41, 677–687.
29. Waddell, P. J., Okada, N., and M. Hasegawa. 1999. Towards resolving the interordinal relationships of
placental mammals. Systematic Biology 48: 1-5.
30. Wang X. (1994). Phylogenetic systematics of the Hesperocyoninae (Carnivora: Canidae). Bullettin of
the American Museum of Natural History n. 221.
31. Wang, X., R. H. Tedford, and B. E. Taylor. 1999. Phylogenetic systematics of the Borophaginae
(Carnivora: Canidae). Bulletin of the American Museum of Natural History 243, 1–391.
32. Werdelin, L., Solounias, N., 1991. The Hyaenidae: taxonomy, systematics and evolution. Fossils and
Strata 30, 1–104.
33. Wesley-Hunt GD, Flynn JJ. (2005). Phylogeny of the Carnivora: basal relationships among the
carnivoramorphans, and assessment of the position of ‘Miacoidea’ relative to Carnivora. Journal of
Systematic Palaeontology 3, 1–28.
The tree in Newick format
((((((Listriodon_splendens:13.3,Bunolistriodon_lockharti:8.8):7.8,(Hyotherium_soemmeringi:11.28,(Chleuastochoerus
_stehlini:16.39,(Microstonyx_major:11.61,(Metridiochoerus_andrewsi:13.37,Propotamochoerus_palaeochoerus:5.74)
:2.87):5.59):3.88):4.32):32.45,((Dorcatherium_guntianum:11.45,Dorcatherium_naui:16.95):26.97,(((Helladotherium_d
uvernoyi:14.34,Bohlinia_attica:13.54)Sivatheriinae:5.67,Palaeotragus_coelophrys:19.22):24.84,((Walangania_africanu
s:29.92,(Micromeryx_flourensianus:27.09,(((Lagomeryx_parvulus:7.5,Procervulus_dichotomus:6.5):2,Euprox_furcatus
:14)Muntiacinae:1.5,(Cervavitus_novorossiae:15.4,Croizetoceros_ramosus:19.1)Cervinae:5.5):11.59):6.73)Cervoids:3.
36,(((Tragoportax_amalthea:6.8,Tragoportax_gaudryi:6.8,Tragoportax_rugosifrons:6.8):7.95,(Leptobos_etruscus:14.1
5,Protragelaphus_skouzesi:10.46):3.49)Bovinae:2.45,((Kobus_sigmoidalis:16.06,(Antidorcas_recki:8.22,((Gazella_borb
onica:5.19,Gazella_capricornis:2.29,Gazella_deperdita:0.99,Gazella_dorcadoides:3.39,Gazella_sinensis:3.34):0.61,Gaz
ellospira_torticornis:8.5):0.61):7.91):2.31,Palaeoryx_pallasi:13.73):3.48)Bovidae:24.29):3.36):3.36):9.62):6.42,((((((Ste
phanorhinus_etruscus:17.7,(Ceratotherium_neumayri:6.45,Ceratotherium_praecox:10.85):6.45):6.61,Lartetotherium_
sansaniensis:12.51):1.93,Dihoplus_schleiermacheri:18.44):9.2,(Aceratherium_incisivum:22.7,(Acerorhinus_zernowi:22
.66,(Alicornops_simorrensis:18.03,((Chilotherium_habereri:7.4,Chilotherium_schlosseri:8.7):11.4,Brachypotherium_br
achypus:15.78):2.24):1.12):2.24)Aceratheriinae:4.95):20.37,(Ancylotherium_pentelicum:38.24,Chalicotherium_grande
:31.24):12.77):3.81,((Cremohipparion_matthewi:10.89,Cremohipparion_mediterraneum:10.89):10.09,Hippotherium_
primigenium:17.98,(Hipparion_dietrichi:8.14,Hipparion_platyodus:9.24):12.85)Hipparionini:33.83)Perissodactyla:9.43)
:21.99,((Ursus_etruscus:45.86,(Canis_etruscus:22.6,Vulpes_alopecoides:22.6):23.28):4.54,((((Pachycrocuta_brevirostri
s:8.88,Adcrocuta_eximia:3.27):8.76,Ictitherium_viverrinum:12.02):2.58,Dinocrocuta_gigantea:12.45)Hyaenidae:20.53,
(((Homotherium_crenatidens:8.76,Machairodus_aphanistus:1.74):2.32,(Paramachairodus_orientalis:5.72,Megantereo
n_cultridens:9.48):1.36):19.73,((Acinonyx_pardinensis:18.12,Lynx_issiodorensis:16.52):9.52,Panthera_gombaszoegens
is:26.44)Felini:4.76)Felidae:8.34):9.67)Carnivora:41.44):15.57,((Mammut_borsoni:50.33,(Mammuthus_meridionalis:4
7.36,(Gomphotherium_angustidens:13.09,(Anancus_arvernensis:26.06,Tetralophodon_longirostris:18.86):6.83):13.08
):4.36):4.36,(Deinotherium_giganteum:21.75,Prodeinotherium_bavaricum:21.75):21.75)Proboscidea:51.3):10;
Species data
We compiled a database of fossil occurrences of mammals as provided by Paleodb (www.paledb.org) and
NOW (http://www.helsinki.fi/science/now/) databases. Our data span phylogenetically over western Palearctic
Laurasiatherians and Proboscideans, exclusive of small mammals. The reason for excluding small species is that they
have lower preservation potential by a sheer size effect (Damuth 1982), they are much less diverse as fossils, and
there is very little consensus about the higher-level phylogenies of many small mammals fossil groups. Our body mass
data span from 4kg to 11.2 metric tons. The smallest living ungulate is the lesser mouse-deer Tragulus kanchil at some
2Kg in body weight, the largest “large land mammal” alive is the African elephant Loxodonta africana, at some 5
metric tons. Thus, our data are entirely comparable, hence truly representative, of the mammalian size spectrum. We
concentrated on Western Palearctic mammal faunas that we are more familiar with, both in terms of taxonomy and
phylogeny. The Neogene mammalian record in the Western Palearctic is continuous and intensively studied (Bernor et
al. 2006).
The age span of the dataset covers the entire Neogene (i.e. ca. 23 Ma). Species are distributed between continents
and epochs as follows: 58 Miocene species (42 European (E), 14 Asian (As), and 2 African (Af)), 38 Pliocene species
(24E, 10As, 4Af), and 23 Pleistocene species (18E, 1As, 4Af). Given the Miocene is much longer (some 17.7 Ma) than
the other two epochs (Pliocene: ca. 2.8 Ma, Pleistocene: ca. 2.6 Ma), and given that living species were excluded (they
should belong to Pleistocene otherwise), our sampling intensity is quite even.
The record was divided in 1 My long time bins according to the fossil localities age estimates, as they appear
in the reference databases. To perform the geostatistic analyses we had to restrict our investigation to those species
with more than 2 occurrences for each time bin. This limits the number of species suitable for analyses but allows
using high-quality, continuous data.
In order to determine the movement of a species during its life span we had to identify its position in
successive, equal-time intervals. As we partitioned the record according to localities age estimates, some fossil sites
were probably put in the wrong bin because age estimates are themselves uncertain to a degree, and they are
necessarily imprecise even when estimates are obtained from absolute dating methods (which give an age with
confidence interval). Although this introduce some inaccuracy in our partitioning, it must be minor when age
estimates are direct. Furthermore, since misplacement errors are random, they could not provide any systematic
trajectory in habitat tracking. Finally, we removed from the record all of the localities with uncertain aging a priori.
The use of artificial, equal time intervals can be misleading when computing time series (Kirchner and Weil 1998). Yet,
it is appropriate to standardize the data when dealing with sampling issues. In the particular case of our study,
unequal time intervals could have misestimate the extent of species movement from one long stratigraphic interval to
a short, successive one, and the other way around because of biased sampling. Whereas wrong estimation also
applies when computing movements between a poor-sampled and a well-sampled time bin, at least one potential bias
(unequal duration) is reduced by using artificial, equal time intervals. We nonetheless performed a specific test for the
influence of sampling inequality between time bins (see below).
To compute species movements between successive intervals, we first detected the actual position of all
localities by using their paleocoordinates. Paleodb database provides the correct position of a specific fossil locality
related to its measured age (hence its paleocoordinates). For the remaining localities we computed the paleolatitude
and the paleolongitude by using the PointTracker software (www.scotese.com). The position a species occupied on
the Earth in a given time bin was identified by the Central Feature (CF) of its geographical distribution (Carotenuto et
al. 2010). We computed the CF by using Esri ArcGis 9.3. The CF identifies the weighted center of the geographic
distribution of fossil localities where a species is present. It represents that single fossil locality that minimizes the
summed distance to all other localities. As such, it is bound to correspond to a locality placed in the highest fossil sites’
density territory for a species. Identifying the movement of a species in different time bins via the crude computation
of distances between the CFs of successive time bins is not advisable because the distribution of the fossil localities is
itself uneven in time and space (Raia et al. 2009, Carotenuto et al 2010). To overcome this problem, for each time bin
we took the geometric centre (GC) of all the fossil localities occurring in that time bin (regardless of which set of
species is present in each locality) as to represent the centre of a virtual reference system, according to which the CFGC distance was computed for each species (Figure 1). The successive step was to geometrically translate all of the GC-‐CF
vectors of the same species to a single GC (Figure 1) (any GC could be used as a common reference since translation would be
geometrically equivalent) and then to compute the distances between the translated CFs (Figure 1) . For geodesic distance
computation we used the Vincenty Inverse Formula (Vincenty 1975) by using the package SDMTools in R. Distances
were obtained to an accuracy of 0.5 mm, and were calculated considering the WGS-84 ellipsoid model for the Earth,
according to the original reference system of localities coordinates of our database. These distances were explicitly
considered the outcome of habitat tracking by species. We urge the reader to consider that the “distance” could not
be zero even if the species’ occupied habitat did not change in location, since the GCs are calculated on the
distribution of fossil localities, which is clearly uneven. In addition, some bias in the distribution of fossil localities
occupied by a species in a given time interval could be, in principle, introduced by purely taphonomic factors. Rather
than focusing on the distance magnitude per se, we were interested in the distances for comparative purposes.
The species stratigraphic duration was computed as the difference in million years between the species first
and last occurrence in the fossil record.
Extinct species body sizes in the collected database were either taken from literature or estimated regressing
remains’ measures vs known body size (Damuth and MacFadden 1990).
We regressed the total distance covered by each species over its existence versus its stratigraphic duration,
both using raw data and under a phylogenetic correction. Regression was performed under phylogenetic generalized
least square (PGLS), assuming the Brownian motion model of evolution, and then correcting branch lengths by using
Pagel’s transform. The Brownian motion model assumes a phenotypic trait to evolve according to a constantvariance random walk. Differences in trait values are thus expected to be proportional to the variance/covariance
matrix derived from the tree, which includes the divergence times among species as off-diagonal elements, and their
distance to the tree root on the diagonal. Pagel’s  is a multiplier of the off-diagonal elements of the variancecovariance matrix that provides the best fit of the Brownian motion model to the trait data by means of a maximum
likelihood approach (Freckleton et al. 2002). Pagel’s  is appropriate to use if a given trait evolves under a nonBrownian motion. Stratigraphic duration (the Y variable) is not a phenotypic trait in the strict sense. Yet, a
phylogenetic correction is advisable if closely related species tend to inherit similar durations, which is clearly the case
in our data. We calculated the phylogenetic signal in duration to be significant (K = 0.308, p = 0.003; see Blomberg et
al. 2003 for a description of the K metric). This means that closely related species inherit a phenotypic trait (or a suite
of traits) having a shared effect on stratigraphic duration. PGLS regressions were performed by using the R package
ape (Paradis et al. 2004).
One potential problem with these regression models is that the relationship between duration and total distance
might be inflated by chance: the longer a species lives, the greater the distance it could move over. To this aim, we
also tested whether the maximum distance covered by a species over a single interval (that is between two successive
time bins) does correspond to the last time bin when the species lived. For each species we computed the prior
probability that the last interval would correspond to the longest distance traveled as 1/# of intervals covered. Then,
we calculated the combined probability over all of the species, and assessed if the total number of instances where
the last interval corresponds to the longest distance covered departs significantly from chance by means of the
binomial distribution. The rationale is that if habitat tracking really prolongs survival, species may have actively tried
to cover larger distances under the worst conditions, which presumably occur just before they went extinct. This is
expected assuming that extinction is not fast enough that species cannot track their preferred habitat, whether or not
they are doomed, or if environmental variables change gradually (Brett et al. 2007).
On potential problem with this procedure is that unequal sampling between successive time bins may artificially
increase the distance between CFs. This is particularly relevant before species extinction, since in the last interval the
commonness (hence the number of localities where a species occur) is expected to be low. If sampling affects distance
calculation, one could argue that distances computed between evenly sampled intervals should be smaller or less
variable than between unevenly sampled intervals. We know that distances are more variable between any two
successive intervals than between the two latest, which tend to be uniformly high. Thus, our preoccupation was to
verify that distances are not artificially large between unevenly sampled intervals. To this aim, we devised a statistical
test. We computed the distance per million years between any pair of species CFs, regardless of the fact the intervals
were successive or not. This produces a matrix of dissimilarity. Then, for the same set of paired comparison, we
produced another matrix of dissimilarity in sampling intensity between intervals, by computing the ratio of the
between the two intervals number of localities, standardized by the most locality-rich interval (in practice, consider
two intervals, one having half the localities of the other: their dissimilarity will be 0.5, two intervals with the same
number of localities will have dissimilarity = 0). These two dissimilarity matrices were tested for correlation to each
other by means of a Mantel test with randomizations. The test revealed no sensible effect of sampling inequality on
distance calculation. Of the 72 matrix correlations we produced, only 4 were positive and significant (meaning that
sampling inequality and distance are positively related). Among them, only one regards a species which covered the
largest distance just prior to extinction. More importantly, there were 7 significant negative matrix correlations. A
negative correlation is expected if tracking implies a shift in the species preferred geographic location without a shift
in range position, with no loss of commonness. This can be verified if and only if CFs are good representation of what
we call “preferred geographic position” and sampling is good enough not to introduce artificial differences in
commonness (hence the number of localities per time interval). Thus, If anything, this test confirmed the quality of the
fossil record we used here. The results of the test are now presented in the ESM table S3.
To test for the relationship between morphological stasis and species duration, we retrieved from the NOW
database information on species molar shape, including the number and shape of molar teeth cusps, and their relative
crown height (= hypsodonty) which are attributes indicative of the type of food consumed, and therefore of the
habitat exploited. NOW cheek teeth data include information on crown type (i.e. whether the tooth occlusal surface is
selenodont, lophodont, and so on); and crown relative height (brachydont, mesodont, or hypsodont). In particular, we
used the NOW variables i molar tooth shape, which describes the general shape of the cusp pattern (e.g. whether
cups are selenodont, lophodont, ectolophodont and so on, see Jernvall and Fortelius 2002); ii tooth crown type, which
is a categorical variable describing the molar cusp pattern at a much finer scale than the variable i; and relative tooth
crown height, which is an ordinal variable describing the degree of hypsodonty (the relative height of the molars
crown). Hypsodonty is of fundamental importance to achieve teeth durability and exploit grasses (Janis and Fortelius
1988). It is important to emphasize that variable i and ii and are different from each other. For instance, although
most of the rhinos here used are ectolophodont (meaning that cusps are united in forming ridges perpendicular to the
jaw, of which the external one is more pronounced) the genus Ancylotherium belongs to a crown type category on his
own. Data on carnivores were implemented by us. We ascribed each carnivore species to either one of the categories
“generalized carnivore”, “bone-cracker”, or “hypercarnivore”, depending on their carnassials and lower premolars
shape. Carnivores whose lower carnassials include a crusching area (i.e. the talonid) were considered to be
“generalized carnivore”. Strict meat-eaters lost their talonid iteratively several times in carnivores’ natural history
(Van Valkenburgh 1991). They were ascribed to the category “hypercarnivore”. Finally, some hyenas, percrocutids,
amphicyonids, and some borophagine canids show adaptations to bone crushing, including enlarged premolars and
deeper mandibles (Meloro et al. 2008), and were consequently ascribed to the category “bone cracker”.
Teeth shape variables were reduced by principal component analysis, by using the function dudi.mix in the
R package ade4 (Dray and Durfur 2007). dudi.mix is specifically meant to perform ordination with mixed variables
(in our case we have both categorical and ordinal variable types). The two first principal components (PCs) were
retained for analyses. We computed morphological disparity for all clades in the tree by using the PC scores applying
the function tip.disparity in the package Geiger in R, computing the average euclidean distance among all
species in the clades. These disparity values were regressed against the average of the total distances covered by all
species included in the clade. A significant negative relationship would indicate that morphological disparity is lower in
clades composed of long-distance travelling species. Since the tree includes polytomous clades, we resolved
multichotomies at random 100 times and re-ran the correlation each time, thereby originating a family of correlation
statistics to be inspected for significance. The same analysis was repeated after including body size (besides teeth
shape variables) to calculate PC scores. The results are qualitatively very similar whether or not body size is
considered. As such, herein we present only the results of the purely morphological analysis (that is excluding size
data).
Literature cited
1. Bernor, R.L., V. Fahlbuch and H.W. Mittmann, Editors, The Evolution of Western Eurasian Neogene
Mammal Faunas, Wiley, New York (1996).
2. Blomberg, S. P., T. Garland, and A. R. Ives. 2003. Testing for phylogenetic signal in comparative data:
behavioral traits are more labile. Evolution 57:717–745.
3. Brett, C.E., Hendy, A.W., Bartholomew, A.J., Bonelli, J., And Mclaughlin, P.I., 2007, Response of
shallow marine biotas to sea-level fluctuations: A review of faunal replacement and the process of
habitat tracking: PALAIOS, v. 22, p. 228–244.
4. Carotenuto F., Barbera C., and Raia P., (2010). Occupancy, range size and phylogeny in Eurasian
Pliocene to Recent large mammals. Paleobiology 36, 399-414.
5. Damuth, J. & Mcfadden, B. J. 1990 Body size in mammalian paleobiology: estimation and biological
implications. New York, NY: Cambridge University Press.
6. Dray, S. and Dufour, A.B. (2007): The ade4 package: implementing the
duality diagram for
ecologists. Journal of Statistical Software. 22(4): 1-20.
7. Freckleton, R. P., P. H. Harvey, and M. Pagel. 2002. Phylogenetic analysis and comparative data: a
test and review of evidence. Am. Nat. 160:712–726.
8. Kirchner JW, and A. Weil (1998). No fractals in fossil extinction statistics. Nature 395, 337-338.
9. Meloro, C., P. Raia, P. Piras, C. Barbera, and P. O’Higgins. 2008. The shape of the mandibular corpus
in large fissiped carnivores: allometry, function and phylogeny. Zool. J. Linn. Soc. 154:832–845.
10. Paradis E., Claude J. & Strimmer K. 2004. APE: analyses of phylogenetics and evolution in R language.
Bioinformatics 20: 289-290.
11. Raia, P., Carotenuto, F., Meloro, C., Piras, P., Barbera, C. & Kotsakis, T. 2009 More than three million
years of community evolution. The temporal and geographical resolution of the Plio-Pleistocene
Western Eurasia mammal faunas. Palaeog. Palaeoclim. Palaeoecol. 276, 15–23.
12. Van Valkenburgh, B. (1991) - Iterative evolution of hypercarnivory in canids (Mammalia: Carnivora):
evolutionary interactions among sympatric predators. Paleobiology, 17, 340–362.
13. Vincenty, T. (1975). "Direct and Inverse Solutions of Geodesics on the Ellipsoid with application of
nested equations". Survey Review XXIII (misprinted as XXII) (176): 88–93.
14. Janis, C. M., and M. Fortelius. 1988. On the means whereby mammals achieve increased functional
durability of their dentitions, with especial reference to limiting factors. Biol. Rev. 63:197–230.
15. Jernvall, J., and M. Fortelius. 2002. Common mammals drive the evolutionary increase of hypsodonty
in the Neogene. Nature 417:538–540.
Supplementary Figure 1 – The phylogeny used in this study. Terminal branches in green identify species which covered
the largest distance over their last temporal interval.
Table S1 - Raw Data and Principal Component Scores. max.dist = longest distance covered over a single time bin (in
km). durations = stratigraphic duration in million years. dis.tot = total distance covered over the species existence (in
km). mass = log10 of the body size in grams. FA = first appearance datum, LA = last appearance datum.
Species
Aceratherium incisivum
Acerorhinus zernowi
Acinonyx pardinensis
Adcrocuta eximia
Alicornops simorrensis
Anancus arvernensis
Ancylotherium pentelicum
Antidorcas recki
Bohlinia attica
Brachypotherium brachypus
Bunolistriodon lockharti
Canis etruscus
Ceratotherium neumayri
Ceratotherium praecox
Cervavitus novorossiae
Chalicotherium grande
Chilotherium habereri
Chilotherium schlosseri
Chleuastochoerus stehlini
Cremohipparion matthewi
Cremohipparion mediterraneum
Croizetoceros ramosus
Deinotherium giganteum
Dihoplus schleiermacheri
Dinocrocuta gigantea
Dorcatherium guntianum
Dorcatherium naui
Euprox furcatus
Gazella borbonica
Gazella capricornis
Gazella deperdita
Gazella dorcadoides
Gazella sinensis
Gazellospira torticornis
Gomphotherium angustidens
Helladotherium duvernoyi
Hipparion dietrichi
Hipparion platyodus
Hippotherium primigenium
max.dist
2773.869
1234.745
1501.238
2583.009
1682.699
1811.455
1322.014
2219.4
1062.294
702.3691
3682.738
1808.274
1370.173
487.2853
1241.384
657.1003
5389.536
1424.034
1667.711
1441.874
1486.253
1649.005
2740.265
2586.474
5912.629
905.329
1306.317
1164.175
845.5416
1323.722
1511.169
1037.981
2198.733
672.9007
3187.784
1196.197
908.2502
1577.687
2112.389
dist.tot durations FA
LA
mass PC1
PC2
3761.87
9.14 20.42
5.30 6.041 0.336 -0.338
1934.2
6.20 11.60
1.81 5.954 0.336 -0.338
2613.38
4.30
3.59
0.78 4.699 0.311 1.408
3296.3
4.15
2.58
0.78 4.845
0.54 1.887
2366.42
8.25
3.59
0.78 5.813 -0.05 -0.344
2643.37
11.40 11.59
5.33 6.544 1.607 -1.469
1669.83
4.15
6.99
5.30 6.444 0.511 -1.057
2973.01
2.64 13.59
5.30 4.447 -0.902 -1.496
1812.79
2.95 13.64 11.61 5.813 -0.82 -0.714
1073.23
10.07
7.74
5.30 6.204 0.589 -0.516
7003.88
6.35 11.19
5.30 5.072 1.978 -0.188
4700.94
4.30
8.69
7.75 4.477 -0.153 1.637
3288.49
7.67
8.89
7.00 6.079 -0.085 1.085
810.42
10.62
8.19
7.00 6.322 -0.352 0.716
2482.77
4.80
8.99
5.33 4.903 -0.82 -0.714
1180.69
7.50 15.19
5.30 5.973 1.282 -1.044
6807.58
8.91
2.49
0.78 5.845 -0.352 0.716
2941.83
7.31
7.99
5.30 6.022 -0.167 0.303
2740.41
4.10
7.74
7.00 4.652 1.978 -0.188
1836.59
3.05 11.19
5.30 5.021 -0.484 0.184
1758.59
3.05
8.99
7.25 5.272 -0.751 -0.185
2140.78
3.06
3.19
0.13 4.903 -1.086 -1.083
9948.52
14.10 11.19
4.20 7.051 1.534 0.216
6707.59
10.35
8.99
5.30 6.079 -0.218 0.901
6182.45
4.15
2.58
0.13
5.58 -0.231 1.874
1302.02
3.20 15.19
5.30
4 0.484 -1.545
2420.19
5.80 11.59
9.00 4.556 0.484 -1.545
1586.74
8.70 11.09
2.59 4.699 -0.82 -0.714
1407.82
4.71
8.99
7.00
4.38 -1.456 -0.258
2486.68
3.23
8.69
7.75 4.342 -1.508
0.34
3221.06
6.53 11.59
5.30 4.362 -0.953 -0.899
2075.96
9.90 15.96
9.00 4.362 -1.641 0.155
3997.4
5.68 11.19
4.20 4.362 -1.255 0.162
829.12
2.50 15.99
5.33 5.167 -0.516 -1.49
6467.74
15.46 12.49
4.20 6.533 1.607 -1.469
2781.58
3.23
9.49
4.20
6 -1.374 0.524
1309.7
2.40
4.89
0.78 5.279 -0.484 0.184
3155.37
6.43
3.39
0.13 5.167 -0.484 0.184
6330.47
10.61
3.19
0.78 5.438 -0.751 -0.185
Homotherium crenatidens
Hyotherium soemmeringi
Ictitherium viverrinum
Kobus sigmoidalis
Lagomeryx parvulus
Lartetotherium sansaniensis
Leptobos etruscus
Listriodon splendens
Lynx issiodorensis
Machairodus aphanistus
Mammut borsoni
Mammuthus meridionalis
Megantereon cultridens
Metridiochoerus andrewsi
Micromeryx flourensianus
Microstonyx major
Pachycrocuta brevirostris
Palaeoryx pallasi
Palaeotragus coelophrys
Panthera gombaszoegensis
Paramachairodus orientalis
Procervulus dichotomus
Prodeinotherium bavaricum
Propotamochoerus palaeochoerus
Protragelaphus skouzesi
Stephanorhinus etruscus
Tetralophodon longirostris
Tragoportax amalthea
Tragoportax gaudryi
Tragoportax rugosifrons
Ursus etruscus
Vulpes alopecoides
Walangania africanus
4444.469
901.4213
1840.515
2133.872
978.1225
1108.771
1540.787
2723.964
2675.284
1280.424
1024.645
2059.229
3819.161
2153.522
691.9995
1070.261
6750.95
1069.497
3360.689
1389.143
3060.444
2836.732
658.955
746.371
1095.906
1712.501
2549.22
1088.392
2299.561
900.3368
1829.236
1438.742
3083.018
7123.41
1615.26
2680.76
4984.69
1209.03
2878.33
3289.06
4753.22
5208.44
1999.79
1786.93
4405.17
4200.96
2862.95
1685.71
2361.12
8799.5
1632.51
4901.18
1751.78
4174.21
3023.07
1636.34
1165.11
1702.39
2665
6751.94
2156.48
4789.51
1214.92
3145.23
2259.81
5721.79
5.75
5.35
7.65
2.68
6.60
8.07
3.86
11.91
4.30
7.46
13.75
4.30
3.86
2.68
8.00
11.35
4.30
4.10
4.24
3.25
3.60
4.75
10.90
7.10
2.44
4.30
14.92
3.60
4.70
2.40
2.69
4.30
12.80
2.58
5.29
5.32
17.99
16.89
2.58
2.49
2.58
11.60
16.89
15.96
2.49
16.89
2.49
2.58
8.99
4.89
15.96
5.32
4.89
9.49
20.49
11.60
11.60
15.96
16.89
11.59
16.89
16.89
16.89
15.18
12.74
16.89
0.13
0.13
1.81
9.00
5.33
0.13
0.78
0.13
1.81
5.33
13.70
0.13
13.70
0.78
0.13
5.30
1.81
11.10
0.13
2.50
3.60
7.75
7.25
7.75
3.60
11.60
3.60
11.61
11.10
13.70
15.16
11.61
13.65
5.364
4.799
4.255
5.053
3.699
5.766
5.602
4.959
4.477
5.342
6.855
6.797
4.799
5.176
3.602
5.519
5.079
5.301
5.602
4.954
4.929
4.477
6.643
5.079
4.813
6.146
6.652
5.104
4.903
5.134
4.954
4.041
4.256
0.74
1.978
0.575
-0.902
-0.82
0.336
-1.056
1.148
-0.119
0.74
1.708
0.561
-0.032
-0.423
-0.82
1.978
0.54
-1.374
-0.953
-0.02
0.74
-0.82
1.267
1.978
-1.241
-0.05
1.607
-0.953
-0.82
-0.686
0.379
-0.153
-0.902
1.234
-0.188
1.83
-1.496
-0.714
-0.338
0.296
0.209
1.581
1.234
-0.504
-0.282
1.221
1.973
-0.714
-0.188
1.887
0.524
-0.899
0.66
1.234
-0.714
-0.154
-0.188
0.709
-0.344
-1.469
-0.899
-0.714
-0.53
0.911
1.637
-1.496
Table S2 – References for body size data.
Species
Reference
Aceratherium incisivum
NOW Database
Acerorhinus zernowi
NOW Database
Acinonyx pardinensis
NOW Database
Adcrocuta eximia
NOW Database
Alicornops simorrensis
NOW Database
Anancus arvernensis
Average of estimates for best equations in Christiansen (2004, Table 7)
in Meloro C. et al. - Effect of predation on prey abundanceand survival
in Plio-Pleistocene mammalian communities. Evolutionary Ecology
Research, 9: 1–21 (2007)
Ancylotherium pentelicum
NOW Database
Antidorcas recki
Kappelman J. et al. - Bovids as indicators of Plio-Pleistocene
paleoenvironments in East Africa. Journal of Human Evolution 32,
229–256 (1997)
Bohlinia attica
NOW Database
Brachypotherium brachypus
NOW Database
Bunolistriodon lockharti
NOW Database
Canis etruscus
NOW Database
Ceratotherium neumayri
NOW Database
Ceratotherium praecox
NOW Database
Cervavitus novorossiae
NOW Database
Chalicotherium grande
NOW Database
Chilotherium habereri
NOW Database
Chilotherium schlosseri
NOW Database
Chleuastochoerus stehlini
Deng T. - Late Cenozoic environmental changes in the Linxia Basin
(Gansu, China) as indicated by cenograms of fossil mammals.
Vertebrata PalAsiatica pp. 282-298 (2009)
Cremohipparion matthewi
NOW Database
Cremohipparion mediterraneum
NOW Database
Croizetoceros ramosus
NOW Database
Deinotherium giganteum
NOW Database
Dihoplus schleiermacheri
NOW Database
Dinocrocuta gigantea
Deng T. - Late Cenozoic environmental changes in the Linxia Basin
(Gansu, China) as indicated by cenograms of fossil mammals.
Vertebrata PalAsiatica pp. 282-298 (2009)
Dorcatherium guntianum
NOW Database
Dorcatherium naui
NOW Database
Euprox furcatus
NOW Database
Gazella borbonica
NOW Database
Gazella capricornis
NOW Database
Gazella deperdita
NOW Database
Gazella dorcadoides
NOW Database
Gazella sinensis
NOW Database
Gazellospira torticornis
NOW Database
Gomphotherium angustidens
NOW Database
Helladotherium duvernoyi
NOW Database
Hipparion dietrichi
NOW Database
Hipparion platyodus
NOW Database
Hippotherium primigenium
NOW Database
Homotherium crenatidens
Deng T. - Late Cenozoic environmental changes in the Linxia Basin
(Gansu, China) as indicated by cenograms of fossil mammals.
Vertebrata PalAsiatica pp. 282-298 (2009)
Hyotherium soemmeringi
NOW Database
Ictitherium viverrinum
NOW Database
Kobus sigmoidalis
Kappelman J. et al. - Bovids as indicators of Plio-Pleistocene
paleoenvironments in East Africa. Journal of Human Evolution 32,
229–256 (1997)
Lagomeryx parvulus
NOW Database
Lartetotherium sansaniensis
NOW Database
Leptobos etruscus
Body size estimates with for bovids TLML equation in Janis (1990) by
m3 lenght in Masini, F. - I bovini villafranchiani dell’Italia. PhD
dissertation, Università di Firenze (1988)
Listriodon splendens
NOW Database
Lynx issiodorensis
NOW Database
Machairodus aphanistus
NOW Database
Mammut borsoni
Average of estimates for best equations in Christiansen (2004, Table 7)
in Meloro C. et al. - Effect of predation on prey abundanceand survival
in Plio-Pleistocene mammalian communities.Evolutionary Ecology
Research, 9: 1–21 (2007)
Mammuthus meridionalis
Christiansen P. - Body size in proboscideans, with notes on elephant
metabolism. Zoological Journal of the Linnean Society 140, 523–549,
(2004)
Megantereon cultridens
Body size estimates with all felids M1 length equation in Van
Valkenburgh (1990) by M1 length in Sardella, R. - Sistematica e
distribuzione stratigrafica dei macairodontini dal Miocene superiore al
Pleistocene. PhD dissertation, Università ‘La Sapienza’ Roma (1993)
Metridiochoerus andrewsi
Raia P. et al. - One size does not fit all: no evidence foran optimal body
size on islands. Global Ecology and Biogeography 19, 475–484 (2010)
Micromeryx flourensianus
NOW Database
Microstonyx major
NOW Database
Pachycrocuta brevirostris
NOW Database
Palaeoryx pallasi
NOW Database
Palaeotragus coelophrys
NOW Database
Panthera gombaszoegensis
Body size estimates with all felids M1 lenght equation in Van
Valkenburgh (1990) by M1 lenght in Schaub, S. - Revision de quelques
Carnassiers villafranchiens du Niveau des Etouaires
(Montagne de Perrier, Puy-de-Dôme). Eclog. Geol. Helvetiae, 42: 492–
506 (1949)
Paramachairodus orientalis
NOW Database
Procervulus dichotomus
NOW Database
Prodeinotherium bavaricum
NOW Database
Propotamochoerus
palaeochoerus
NOW Database
Protragelaphus skouzesi
NOW Database
Stephanorhinus etruscus
NOW Database
Tetralophodon longirostris
NOW Database
Tragoportax amalthea
NOW Database
Tragoportax gaudryi
NOW Database
Tragoportax rugosifrons
NOW Database
Ursus etruscus
NOW Database
Vulpes alopecoides
NOW Database
Walangania africanus
Body size estimates with all ruminants FLML equation in Janis (1990)
by m1 lenght in Barry J.C. et al. - Oligocene and early Miocene
Ruminants. (Mammalia, Artiodactyla) form Pakistan and Uganda.
Palaeontologia Electronica Article Number: 8.1.22A (2005)
Table S3- Correlation between the inter-CF distances and the ratios between the number of localities per time bin
were a species was present. The table shows the matrix correlation test results, performed by using a randomized
version of Mantel’s test. § = significant correlation. * = species covering the longest distance just prior to extinction.
Species
R
p
Aceratherium incisivum
-0.068
0.700 *
Acerorhinus zernowi§
-0.947
1.000 *
Acinonyx pardinensis
-0.127
0.560 *
Adcrocuta eximia
-0.506
0.850 *
Alicornops simorrensis
0.464
0.200 *
Anancus arvernensis§
0.803
0.020
Ancylotherium pentelicum
-0.615
0.820 *
Antidorcas recki
-0.349
0.780 *
Bohlinia attica
0.999
0.140 *
Brachypotherium brachypus
0.387
0.210
Bunolistriodon lockharti
0.763
0.210 *
Canis etruscus
Ceratotherium neumayri
-0.419
0.820
0.142
0.350
Ceratotherium praecox
-0.354
0.940 *
Cervavitus novorossiae
-0.528
0.900 *
Chalicotherium grande
-0.273
0.780 *
Chilotherium habereri
-0.154
0.460
Chilotherium schlosseri
-0.061
0.570 *
Chleuastochoerus stehlini
Cremohipparion matthewi§
0.877
0.160
-0.816
1.000 *
Cremohipparion mediterraneum
0.741
0.260 *
Croizetoceros ramosus§
0.999
0.010 *
Deinotherium giganteum
0.110
0.230
Dihoplus schleiermacheri
-0.101
0.580
Dinocrocuta gigantea§
-0.989
1.000
0.889
0.050
Dorcatherium naui
-0.050
0.490
Euprox furcatus§
-0.730
0.960 *
Dorcatherium guntianum
Gazella borbonica§
0.773
Gazella capricornis
-0.211
0.500 *
Gazella deperdita
0.674
0.240 *
Gazella dorcadoides
-0.370
0.630 *
Gazella sinensis
-0.215
0.690
Gazellospira torticornis
-0.853
0.840
0.530
0.020
Helladotherium duvernoyi
-0.194
0.810
Hipparion dietrichi
-0.510
0.800 *
Hipparion platyodus
0.513
0.160 *
Hippotherium primigenium
0.320
0.110
Homotherium crenatidens§
-0.480
Gomphotherium angustidens§
Hyotherium soemmeringi
Ictitherium viverrinum
0.974
0.010
1.000 *
0.080
-0.496
0.740 *
Kobus sigmoidalis
0.007
0.390 *
Lagomeryx parvulus
0.423
0.260
Lartetotherium sansaniensis§
-0.286
0.970
Leptobos etruscus
-0.576
0.860
0.193
0.200
Listriodon splendens
Lynx issiodorensis
-0.254
0.830 *
Machairodus aphanistus§
-0.834
1.000 *
Mammut borsoni
-0.162
0.540
Mammuthus meridionalis
-0.020
0.320 *
Megantereon cultridens
-0.469
0.730
Metridiochoerus andrewsi
-0.248
0.620 *
Micromeryx flourensianus
-0.129
0.700 *
Microstonyx major
-0.045
0.690
0.922
0.070
Pachycrocuta brevirostris
Palaeoryx pallasi
-0.535
0.820 *
Palaeotragus coelophrys
-0.119
0.690 *
Panthera gombaszoegensis
0.136
0.310 *
Paramachairodus orientalis
-0.247
0.670 *
Procervulus dichotomus
0.439
0.050
Prodeinotherium bavaricum
0.205
0.170 *
-0.775
0.890 *
Protragelaphus skouzesi
0.991
0.130 *
Stephanorhinus etruscus
0.126
0.600 *
Propotamochoerus
palaeochoerus
Tetralophodon longirostris
-0.067
0.660
Tragoportax amalthea
-0.077
0.530
Tragoportax gaudryi
0.371
0.290
Tragoportax rugosifrons
0.972
0.160 *
-0.191
0.680 *
Vulpes alopecoides
0.992
0.110 *
Walangania africanus
0.645
0.330
Ursus etruscus