Cats in the forest: predicting habitat adaptations from humerus morphometry in extant and () Author(s): Carlo Meloro, Sarah Elton, Julien Louys, Laura C. Bishop, and Peter Ditchfield Source: Paleobiology, 39(3):323-344. 2013. Published By: The Paleontological Society DOI: http://dx.doi.org/10.1666/12001 URL: http://www.bioone.org/doi/full/10.1666/12001

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/ terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. Paleobiology, 39(3), 2013, pp. 323–344 DOI: 10.1666/12001

Cats in the forest: predicting habitat adaptations from humerus morphometry in extant and fossil Felidae (Carnivora)

Carlo Meloro, Sarah Elton, Julien Louys, Laura C. Bishop, and Peter Ditchfield

Abstract.—Mammalian carnivores are rarely incorporated in paleoenvironmental reconstructions, largely because of their rarity within the fossil record. However, multivariate statistical modeling can be successfully used to quantify specific anatomical features as environmental predictors. Here we explore morphological variability of the humerus in a closely related group of predators (Felidae) to investigate the relationship between morphometric descriptors and habitat categories. We analyze linear measurements of the humerus in three different morphometric combinations (log-transformed, size- free, and ratio), and explore four distinct ways of categorizing habitat adaptations. Open, Mixed, and Closed categories are defined according to criteria based on traditional descriptions of species, distributions, and biome occupancy. Extensive exploratory work is presented using linear discriminant analyses and several are included to provide paleoecological reconstructions. We found no significant differences in the predictive power of distinct morphometric descriptors or habitat criteria, although sample splitting into small and large cat guilds greatly improves the stability of the models. Significant insights emerge for three long-canine cats: populator, orientalis, and sp. from Olduvai Gorge (East ). S. populator and P. orientalis are both predicted to have been closed-habitat adapted taxa. The false ‘‘sabertooth’’ Dinofelis sp. from Olduvai Gorge is predicted to be adapted to mixed habitat. The application of felid humerus ecomorphology to the carnivoran record of Olduvai Gorge shows that the older stratigraphic levels (Bed I, 1.99–1.79 Ma) included a broader range of environments than Beds II or V, where there is an abundance of cats adapted to open environments.

Carlo Meloro* and Sarah Elton.** Hull York Medical School, University of Hull, Loxley Building, Cottingham Road Hull HU6 7RX, UK. *Corresponding author. Present address: Dipartimento di Scienze della Terra, Universita´ degli Studi di Napoli, Federico II, Naples, Italy. E-mail: [email protected]. **Present address: Department of Anthropology, Durham University, Durham, U.K. Julien Louys† and Laura C. Bishop. Research Centre in Evolutionary Anthropology and Palaeoecology, School of Natural Sciences and Psychology, Liverpool John Moores University, Byrom Street, Liverpool L3 3AF, U.K. †Present address: School of Earth Sciences, University of Queensland, Brisbane, Australia Peter Ditchfield. Research Laboratory for Archaeology and the History of Art, School of Archaeology, University of Oxford, Dyson Perrins Building, South Parks Road, Oxford OX1 3QY, U.K.

Accepted: 8 January 2013 Published online: 18 March 2013 Supplementary materials deposited at Dryad: 10.5061/dryad.s587t

Introduction DeGusta and Vrba 2003, 2005a,b; Kovarovic Adaptations in the functional morphology and Andrews 2007; Plummer et al. 2008), have of the postcranial skeleton can be powerful informed paleohabitat reconstruction. Al- indicators of locomotion and habitat exploita- though other taxa, such as primates (Elton tion. For fossil species, whose behavior cannot 2001, 2002, 2006), marsupials (Bassarova et al. be observed directly, identifying such adapta- 2009), and suids (Bishop 1999), have been tions and linking them to habitat are impor- subject to similar analyses, terrestrial carnivo- tant aspects of paleobiological reconstruction. rans (fissiped Carnivora) are generally under- This approach also informs paleoecological represented in ecomorphic studies (Gonyea and paleoenvironmental reconstruction, with 1976; Lewis 1997). Conventionally, it is as- these ‘‘ecomorphological’’ methods shedding sumed that morphological diversity in the light not only on the themselves but carnivorans reflects adaptations to specific also on the environments they inhabited. functions (e.g., foraging and feeding, posture) Numerous ecomorphic studies, focused most- more than the environment they occupy (Ewer ly on bovids from Plio- African 1973; Van Valkenburgh 1985, 1987, 1988, 1989, paleontological sites (Kappelman 1988; Plum- 1999, 2007; Biknevicius and Van Valkenburgh mer and Bishop 1994; Kappelman et al. 1997; 1996; Anyonge 1996; Janis and Wilhem 1993;

Ó 2013 The Paleontological Society. All rights reserved. 0094-8373/13/3903-0001/$1.00 324 CARLO MELORO ET AL.

Garland and Janis 1993; Carrano 1999; Farlow Werdelin 2003; Andersson 2004; Schutz and and Pianka 2002; Wroe et al. 2005; Meloro Guralnick 2007; Polly 2008; Polly and Mac- 2011a,b; Samuels et al. 2012; Walmsley et al. Leod 2008; Meachen-Samuels and Van Val- 2012). Because they occupy large geographic kenburgh 2009; Lewis and Lague 2010; ranges and high trophic levels, carnivorans Walmsley et al. 2012). tend to be more eurybiomic—able to exploit Here, we explore the relationships between numerous habitats and biomes—than other the functional morphology of the carnivoran mammalian clades (Herna´ndez Ferna´ndez postcranial skeleton and habitat preferences, and Vrba 2006). This reinforces the largely focusing on a single family of fissiped carniv- unexplored notion that carnivorans are ‘‘gen- orans, the Felidae. We develop models based eralists’’ in their skeletal adaptations to habitat on modern species and apply these to fossil and hence have limited value when included felids. The felids, or cats, are a speciose and in studies that aim to reconstruct paleohabitat. widespread family of ‘‘hypercarnivores’’ (Ew- Neglecting carnivoran fauna when under- er 1973; Gittleman 1985; Martin 1989; Kitch- taking ecomorphic-based paleoenvironmental ener 1991; Turner and Anton´ 1997; Sunquist reconstructions may exclude important infor- and Sunquist 2002). Although this hyper- mation about local and regional habitats and carnivory has resulted in relative dental how these are exploited by different members homogeneity within the family (Holliday of mammalian communities. It is thus impor- and Steppan 2004; Meloro and Raia 2010), tant to address whether the carnivoran skel- other features do differ according to the prey eton can yield sufficiently detailed on which they specialize (Christiansen 2008; information about functional morphological Slater and Van Valkenburgh 2008; Meachen- adaptations related to particular habitats. Samuels and Van Valkenburgh 2009; Meloro Variations in habitat preferences within the 2011a). The most exceptional skull and post- felids suggest the utility of this group. Within cranial morphologies, seen in the extinct the , there are obvious differ- saber- and dirk-tooth cats (distinguished by ences in habitat exploitation between the extremely long canine teeth), are possibly a extant (Panthera leo), which tends to hunt result of extreme adaptations to a specialized in an open, savanna environment, and the hunting technique (Christiansen 2008; Slater tiger (Panthera tigris), which is more restricted and Van Valkenburgh 2008; Anton´ et al. 2004, to tropical and temperate forested areas. This 2005; Palmqvist et al. 2007; McHenry et al. indicates that even though carnivorans are 2007; Meachen-Samuels and Van Valkenburgh eurytopic and eurybiomic, niche differentia- 2010; Meachen-Samuels 2012). Felids vary tion does occur. It can further be inferred that greatly in size, with the smallest members if this differentiation has reasonably deep (such as the black-footed cat, Felis nigripes) evolutionary roots, there may be morpholog- having body masses under 2 kg and the ical adaptations, even if subtle, to these largest extant forms (such as the tiger, Panthera different habitats. Indeed, distinct skeletal tigris) weighing as much as 300 kg (Kitchener metrics that correlate with habitat exploitation et al. 2010). Some extinct taxa, such as the in large carnivorous predators have already dirk-tooth Smilodon populator, were likely to been identified and used to explore adaptation have been even larger than this, exceeding 400 to habitat in both extant and extinct taxa kg (Christiansen and Harris 2005). This (Lewis 1997; Meloro 2011b). These have rarely diversity in body mass is reflected in locomo- been followed up with morphometric surveys tion, with smaller taxa generally being much that specifically examine the correlations more arboreal than bigger forms (Gittleman between skeletal morphology and habitat 1985; Kitchener 1991; Turner and Anton´ 1997; adaptations in terrestrial carnivorans (but see Sunquist and Sunquist 2002; Kitchner et al. Polly 2010), but several studies have indicated 2010; Samuels et al. 2012). Felids also exploit the strong relationship between appendicular an array of habitats, commensurate with their skeleton morphometry and locomotion or near-cosmopolitan distribution. Many species behavior (Anyonge 1996; Andersson and are adapted to a broad range of habitats (e.g., MORPHOMETRIC HABITAT-ADAPTATION 325 the , Panthera pardus,isatypical ology of fossil specimens must take this into habitatgeneralistthatcanbefoundin account. Most ecomorphic studies focus on woodlands as well as deserts), whereas others single bones (see Polly 2010), and some on occur only in association with specific envi- epiphyses only, given that these are the long ronmental conditions (e.g., the Andean cat, bone segments most likely to be preserved Leopardus jacobita, occurs only in association (e.g., Elton 2001, 2002, 2006). We therefore with rocky outcrops in the arid zones of the present a broad range of statistical analyses high Andes, typically above 4200 m) (Mac- designed to improve the resolution of existing donald et al. 2010). This diversity of habitats methods. seems to be reflected in a variety of adapta- We also aim to identify an objective habitat tions in the appendicular skeleton, suggesting classification for use in ecomorphic analyses. that the latter evolved in response to the It is usual for three or four habitat categories former (Kitchener et al. 2010). to be defined a priori (Kappelman et al. 1997; We restrict our analyses to a single bone, the Bishop 1999; Elton 2001, 2002; DeGusta and humerus. The humerus has been shown in Vrba 2003, 2005a,b; Plummer et al. 2008; primates to be highly informative about Louys et al. 2012), although some studies locomotor adaptations and habitat preferences have used as many as seven (Kovarovic and (Elton 2001, 2002, 2006), and forelimb bone Andrews 2007). Notwithstanding the relative- proportions (radius/humerus length) have ly large number of studies correlating habitat been used in previous studies of felids and adaptations and long bone morphology, there carnivorans to distinguish adaptation to dif- is no consensus about how to categorize large ferent habitats (Gonyea 1978; Lewis 1997; in discrete and distinct habitats Meachen-Samuels and Van Valkenburgh objectively. Most studies rely on reviews of 2009; Meloro 2011b; Samuels et al. 2012). The biology and ethology to categorize the most humerus is one of the three long bones of the common environments exploited by different forelimb, and articulates proximally with the species. These are often defined as Open (for scapula, providing information about shoul- example, grassland), Mixed (mixture of grass- der function including rotation, extension and land and tree cover), and Closed (forest). In flexion, and distally with the radius and ulna, her study of the Plio-Pleistocene East African reflecting elbow flexion and extension. The carnivore guild, Lewis (1997) assigned carniv- shape of the distal portion of the humerus oran species to these three habitat types, discriminates cursorial from non-cursorial defining Mixed as having around 20% canopy Carnivora, and Andersson and Werdelin cover, Open as having less than this, and (2003)arguedthatthisfeaturemightbe Closed as having more than 20%. An alterna- indicative of adaptations to open/closed tive advocated by some authors (e.g., Herna´n- habitats. Cursorial carnivores are generally dez Ferna´ndez 2001; Herna´ndez Ferna´ndez expected to be better adapted to open envi- and Pela´ez-Campomanes 2003) is to catego- ronments (Janis and Wilhem 1993; Samuels et rize environmental preferences by number of al. 2012) and this generalization applies also to biomes occurring in a species’ geographical Felidae (Walmsley et al. 2012). Thus, consid- range. However, few published studies have ering the ecological diversity of the Felidae addressed the issue of habitat categorization (Macdonald et al. 2010) we expect the mor- in detail. Here, we explore how to define phometry of the humerus to correlate with habitat categories, quantifying presence/ab- habitat adaptations. sence in particular environments and using Our focus on a single bone does not imply number of biomes overlapped by the geo- that only this bone may be informative, but graphical range of a taxon to examine a rather aims to identify its potential for species’ environmental preferences. paleobiological and paleoenvironmental re- Early ecomorphic studies (e.g., Kappelman construction. Associated skeletons and skele- 1988) used ratios as a means of size-correcting tal regions are rare in the fossil record, so any morphometric data. Later studies questioned method that aims to reconstruct the paleobi- this approach (DeGusta and Vrba 2005a,b; 326 CARLO MELORO ET AL.

Kovaric and Andrews 2007; Louys et al. 2012), ern specimen, taxonomy was reassessed arguing that simple linear measurements were following species accounts in the IUCN Red equally informative as ratios in the discrimi- List (IUCN 2009). When accurate geographic nant analyses that form the basis of ecomor- information was available, modern specimens phic studies. Given that previous studies of belonging to species with large geographic carnivorans similar to ours used only ratios ranges—the wild cat (Felis silvestris), lion (Van Valkenburgh 1987; Lewis 1997; Samuels (Panthera leo), leopard (Panthera pardus), and et al. 2012), here we examine the utility of tiger (Panthera tigris)—were assigned to sub- simple raw measurements versus ratios and species (Table 1). A total of 111 extant residuals (another common way of generating specimens across 11 genera were included in ‘‘size free’’ data) as predictors of habitat the analyses (Supplementary Table 1). Sample preference. size was not equally distributed across taxa. In short, we seek to assess whether the Inevitably, most of the extant sample was humeral morphometry of one family of biased in favor of trophy-hunted species (e.g., carnivorans, the felids, allows the recovery of and ). To get maximum taxo- useful information about habitat exploitation. nomic and hence environmental coverage, We assess methods surrounding ecomorphic non-pathological captive specimens were in- reconstruction, in order to examine particular cluded (13% of the sample). Several of these types of scaling methods and modern com- specimens derive from captive breeding cen- parative data sets and to determine whether ters where general conditions for the the way in which habitat is categorized for approximate their natural environment (A. taxa in a modern comparative sample influ- Kitchener personal communication 2009). For ences analytic results. In addition, we inves- this study, captivity was determined to be a tigate whether it is possible to recover accurate negligible source of morphometric variation information from highly fragmentary materi- (Supplementary Table 2, but see O’Regan and al, using data from the epiphyses of modern Kitchener 2005 for caveats). Thirty-one percent specimens as a proxy for the data collected of the specimens either had no locality from incomplete fossils. Finally, we use the recorded or were only located to a continent; methods we develop to reconstruct the habitat the rest of the sample being wild-caught with preferences of three extinct felids, Paramachair- good locality data. Approximately half the odus orientalis, Smilodon populator, and Dinofelis extant sample could not be assigned to sex; sp., as well as those of taxa represented by within the rest of the sample, males and fragmentary fossil material from different females were equally distributed. Sexual stratigraphic intervals of the hominin East dimorphism is generally high in felids, but African fossil site Olduvai Gorge, Tanzania. because it is uncorrelated with habitat adap- This approach provides an example of how tation at interspecific level (Gittleman and Van felid humerus ecomorphology can be used to Valkenburgh 1997), we assume it to be a inform paleoenvironmental reconstruction. negligible source of morphometric variation. Both sexes were pooled in the analyses. Materials and Methods Two complete specimens (one fossil, one cast) of cats of the subfamily Machairodonti- Sample Size nae plus five humeri from Olduvai Gorge Complete and incomplete humeri belonging were analyzed. The sabertooth Paramachairo- to both extant and extinct members of the dus orientalis (BMNH M8960) is represented by Felidae, housed in the Natural History Muse- a complete but slightly deformed humerus um London (BMNH) London; Royal Museum from Pikermi, Greece (a late fossil for Central Africa (RMCA), Tervuren; Nation- site); the specimen of Smilodon populator, the al Museum of Scotland (NMS) Edinburgh; and biggest Pleistocene dirk-tooth cat from South National Museum (KNM), Nairobi, America, was a cast from a complete skeleton were included in the ecomorphological anal- housed in the Natural History Museum, yses (Supplementary Table 1). For each mod- London (BMNH). The material from Olduvai MORPHOMETRIC HABITAT-ADAPTATION 327

TABLE 1. Habitat categories and basic ecological data for felid species analyzed (body weight averaged for males plus females, from IUCN 2009; locomotion as in Meachen-Samuels and Van Valkenburgh 2009). Habitat C is summarized as percentage of specimens recorded in Forest or Grassland.

Habitat C % specimens Body weight Species (kg) Locomotion Habitat A Habitat B Forest Grass Habitat D Acinonyx jubatus 40.917 Terrestrial Open Open 50 50 Open Caracal aurata 6.200 Terrestrial Closed Closed 100 0 Closed Caracal caracal 11.500 Scansorial Open Open 50 50 Mixed Felis chaus 5.150 Terrestrial Mixed Mixed 100 0 Closed Felis margarita 2.500 Terrestrial Open Open 0 100 Open Felis nigripes 1.525 Terrestrial Open Open 0 100 Mixed Felis silvestris grampia 4.167 Scansorial Closed Mixed 100 0 Closed Felis silvestris lybica 4.833 Scansorial Open Open 33 67 Mixed Leopardus geoffroyi 4.350 Terrestrial Mixed Mixed 100 0 Open Leopardus guigna 2.200 Unknown Closed Closed 100 0 Closed Leopardus pardalis 10.131 Scansorial Closed Closed 0 100 Closed Leopardus wiedii 3.200 Arboreal Closed Closed 100 0 Closed Leptailurus serval 12.250 Terrestrial Open Open 0 100 Mixed Lynx canadensis 10.025 Terrestrial Mixed Closed 100 0 Mixed Lynx lynx 20.100 Scansorial Mixed Mixed 67 33 Closed Lynx pardinus 11.050 Terrestrial Mixed Mixed 100 0 Closed Lynx rufus 9.300 Scansorial Open Mixed 0 100 Open Neofelis nebulosa 15.500 Arboreal Closed Closed 100 0 Closed Panthera leo 150.529 Terrestrial Open Open 20 80 Open Panthera leo persica 147.500 Terrestrial Open Open 100 0 Open Panthera onca 79.167 Scansorial Closed Closed 100 0 Closed Panthera pardus 35.042 Scansorial Mixed Mixed 50 50 Mixed Panthera pardus fusca 49.667 Scansorial Mixed Mixed 100 0 Mixed Panthera tigris 169.375 Terrestrial Closed Closed 100 0 Closed Panthera tigris altaica 243.000 Terrestrial Closed Closed 100 0 Closed Panthera uncia 42.188 Scansorial Closed Open 0 100 Open Pardofelis badia 1.950 Unknown Closed Closed 100 0 Closed Pardofelis marmorata 4.000 Arboreal Closed Closed 100 0 Closed Pardofelis temminckii 11.750 Scansorial Mixed Mixed 100 0 Closed Prionailurus bengalensis 5.050 Scansorial Closed Closed 100 0 Closed Prionailurus planiceps 2.000 Terrestrial Closed Closed 100 0 Closed Prionailurus rubiginosus 1.350 Unknown Mixed Mixed 100 0 Closed Prionailurus viverrinus 9.625 Terrestrial Closed Closed 100 0 Closed Puma concolor 57.125 Scansorial Mixed Mixed 100 0 Closed Puma yagouaroundi 5.150 Scansorial Closed Mixed 100 0 Closed

Gorge includes two complete humeri belong- observer (C.M.) using an osteometric board ing to Dinofelis sp. indet D (OLD 74/54, OLD (for greatest bone length), spreading calipers 74/348 [Werdelin and Lewis 2001]) from Bed I (for physiological length), or Sylvac digital measured in the Kenya National Museum calipers interfaced to a laptop computer. Most (Nairobi), and three distal fragments housed measurements were taken on the left humer- at the Natural History Museum, London: us. If that was not available, we substituted M20240, recorded from DKI 25 IV 35 and the right, assuming that because asymmetry tentatively assigned to Panthera sp. from Bed I; was fluctuating and not directional, no sys- M14676, belonging to Panthera leo from Bed II tematic bias would be introduced. (cf. Leakey 1965); and M14677, classified as Measurement error was calculated by mea- Panthera leo from Bed V (Upper Pleistocene). suring the same specimen of serval (Leptailu- Only one humerus from Bed I (OLD 5067 FLK rus serval BMNH 1981.988) three times on NI I 4, Panthera pardus) was too incomplete to be included here. separate occasions (cf. DeGusta and Vrba 2003, 2005a,b). Overall, the mean error was Linear Measurements and Error Estimation less than 5%, consistent with that seen in other Forty linear measurements of the humerus studies. The mean error estimate for each (Table 2) were taken to 0.5 mm by a single measurement is given in Table 2. Given that 328 CARLO MELORO ET AL.

TABLE 2. Measurement error and description for 40 humerus linear measurements. Only absolute values of differences in multiple measurement comparison are reported. Abbreviations: L, length; ML, mediolateral; AP, anterior-posterior.

ID Description Error L Length 0.12% PhL Physiological L—from central tip of epiphyses 0.62% DtL Deltopectoral crest max L 1.32% DtPhL Deltopectoral crest physiological L 2.57% Mds_ML Midshaft ML 0.10% Mds_AP Midshaft AP 1.77% APmH Head max AP 0.14% APartH Head articular surface AP 5.07% APsH Head shaft AP 1.05% ML_H Head max ML 0.66% ML_artH Head articular surface ML 1.69% H_H Head surface height 0.42% BcG_W Bicipital groove width 4.48% BcG_D Bicipital groove depth 1.45% GT_AP Greater tubercle max AP 1.47% GT_ML Greater tubercle max ML 0.00% Sb_ML Subspinosus scar ML 2.89% Sb_AP Subspinosus scar AP 3.32% LT_AP Lesser tubercle max AP 0.16% LT_ML Lesser tubercle max ML 0.00% Dst_ML Distal epiphysis maximum ML 0.14% Dst_AP1 Distal epiphysis medial articular surface AP 5.15% Dst_AP2 Distal epiphysis lateral articular surface AP 1.27% TrL Trochlea max L 1.52% CpL Capitulum max L 2.90% TrAP Trochlea AP at the midpoint 0.11% Cd_L1 Trochlea superior-inferior maximum L 1.15% Cd_L2 Trochlea superior-inferior medium L 1.85% Cd_L3 Trochlea superior-inferior minimum L 5.24% Pj_Tr Projection of trochlea in vertical plane 2.13% Dst_art_ML Distal articular surface ML 2.38% Of_ML Olecranon fossa ML 0.72% Of_H Olecranon fossa height 8.22% Of_Pr Olecranon fossa projection 0.64% PrTb_L Pronator tubercle L 3.16% UmF_L Ulnar medial fossa L 1.00% UlF_ML Ulnar lateral fossa ML 5.78% UlF_AP Ulnar lateral fossa AP 12.23% UlF_pj Ulnar lateral fossa depth 5.20% ExC_L Extensor carpii scar L 1.14% the sample represents cats having a wide Caro 1996) to assign species to one of three range of body mass (1–200 kg), the measure- categories (Open, Mixed, Closed); (B) descrip- ments were log-transformed for statistical tions from the IUCN Cat Specialist Group analyses. This enables an assumption of (IUCN 2009) that were then used to assign normality to be met and also scaled the data species to the above categories; (C) a GIS- (cf. Kovarovic and Andrews 2007). Initial based approach to assign each specimen to examinations showed that when data are log grassland or forest biome; (D) a similar GIS- transformed for discriminant analyses, the based method assigning specimens to open or percentages of correctly classified cases were closed biome. always higher than when raw data were used. Presence or Absence in Particular Biomes (A).—We used data from Ortolani and Caro Habitat Categorization (1996), who recorded the presence or absence Four different ways of determining habitat of each carnivoran species in a series of broad categories were examined and compared biomes, to assign each felid species to one of (Table 1): (A) presence or absence in particular three categories (Open, Mixed, or Closed). The biomes (based on raw data from Ortolani and biomes used by Ortolani and Caro (1996) were MORPHOMETRIC HABITAT-ADAPTATION 329

temperate forest, tropical forest, grassland, 5.2.0.2. We used open-source shape files arctic, riparian, and desert. Species were first showing species-specific range maps for the given a habitat score, by recording presence in Carnivora, taken from Greneyer et al. (2006), tropical or temperate forest as þ1 and presence to check for potential outliers. We also used in grassland, arctic, or desert as 1; riparian these maps to identify species centroid coor- was not considered because it is generally dinates (where the geographic range was associated with semiaquatic species like otters continuous and not fragmented), assumed to and no felids occurred exclusively in this be a representative locality for captive speci- category. Absence was recorded as 0. These mens or for those with no recorded locality. By scores were summed, with positive values overlaying the WWF world ecoregion polygon used to indicate preference for Closed envi- (Olson et al. 2001), which describes 14 ronments (presence in forests), zero values ecoregions (plus two categories: rock and ice, preference to Mixed environments (balanced and lake), on the range maps, we were able to between forest and open environments), and extract an ecoregion (biome) for each speci- negative scores preference for Open environ- men locality. The ecoregions extracted for the ments. For example, the lion is recorded in sampled localities were ascribed to either grassland (1) and desert (1), which sum to ‘‘forest’’ (e.g., tropical and subtropical moist, 2 and is thus categorized as Open. The dry broadleaf forest, temperate coniferous European lynx (Lynx lynx) occurs in temperate forest) or ‘‘grassland/shrubland’’ (e.g., mon- forest (þ1) and grassland (1), summing to tane grassland and shrubland, tropical and zero, and thus is assigned to Mixed. For subtropical grassland, savanna and shrub- several taxa we did not possess the Ortolani lands, deserts and xeric shrublands). In this and Caro (1996) categorizations, and conse- way each individual extant specimen could be quently we based their categorization on the confidently ascribed to ‘‘forest’’ or ‘‘grass- IUCN species description (see B). land/shrubland.’’ Descriptions from the IUCN Specialist Group GIS-based Approach: Open versus Closed (B).—We used descriptions of habitat prefer- (D).—The fourth method builds on the prin- ences from the IUCN Cat Specialist Group ciples of method (C), but instead of simply (IUCN 2009) to help assign each as Open, extracting a single habitat category for each Mixed, or Closed. For instance, the lion, specimen, we used the Xtools of ESRI software described as preferring ‘‘open woodland–thick ArcGIS 3.2 (ESRI, Redlands, Calif.) to assess bush, scrub, grass complexes’’ was scored as the interaction between a species’ geographic Open. The European lynx, described as range and ecoregion polygons to quantify the preferring ‘‘forested areas’’ as well as being relative proportion of each biome occupied by present in ‘‘more open, thinly wooded, thick an individual taxon. In order to assign species scrub woodland and barren, rocky areas to either Open, Mixed, or Closed, we classified above the treeline, alpine tundra,’’ is scored the 14 biomes as either ‘‘forest’’ or ‘‘grassland/ as Mixed. In many cases, categorization using shrubland’’ as in method C. For each species, this method correlates with results of method the relative percentages of occurrence in forest (A) but some differences exist; e.g., the or grassland biomes were summarized in a Scottish wild cat (Felis silvestris grampia)is ratio that is equal to % occurrence grassland considered as Mixed (coniferous forest þ divided by % occurrence forest. Species with a Mediterranean shrubland) using this method, ratio between 0 and 0.9 were classified as rather than Open, as was the case using Closed, between 0.9 to 1.1 Mixed, and above method (A). 1.1 Open. Seventy-two percent of the geo- GIS-based Approach: Grassland versus Forest graphic range of the lion was found in (C).—The third method used a specimen- grassland/shrubland biomes and 25% in specific rather than a broad species-based forest biomes (3% is in the biome ‘‘lake’’ or approach. Each individual skeleton with ‘‘rocks & ice,’’ which are not considered here), accurate locality data (longitude and latitude) giving a ratio of 2.88. Consequently, all was plotted using DIVA-GIS software, version skeletal specimens of lion are classified as 330 CARLO MELORO ET AL.

‘‘open.’’ The Eurasian lynx (Lynx lynx), in the percentage of correctly classified speci- contrast, has a ratio of 0.63 (38% in grassland mens after cross-validation. In addition, we and 61% in forest biomes) and is considered ran LDA models on a subset of variables that Closed. could simulate bias introduced by analyses of fossils. In these cases, we selected a subsample Scaling Measurements of variables from the proximal or distal Three separate data sets, derived from epiphyses of specific bones and re-ran the different measurement scaling/size correction discriminant analyses on that subset (cf. Elton methods, were analyzed to examine their 2001, 2002). relative efficacy. Data set 1 contained log- transformed linear measurements. Data set 2 Nested Ecological Analyses was obtained by applying univariate regres- Meloro (2011a) recently demonstrated that sion models of log maximum length versus all LDA model performance can be improved if the other logged variables, then using the data sets are split according to ecologically unstandardized residuals as ‘‘size-free’’ vari- meaningful groups. Because the purpose of ables. Only one of these residuals, log Phys- our study is to predict habitat adaptation of iological Length (n ¼ 112, rs ¼ 0.230, p ¼ 0.012) large fossil species, we performed a nested was correlated with log Maximum Length. ecological analysis splitting our felid sample at Data set 3 used log-transformed linear mea- a body mass threshold of 7 kg, based on the surements (Table 2) combined in 27 functional findings of studies by Van Valkenburgh (1985, ratios (after Bishop 1999; Elton 2001, 2002). Six 1988, 1989) and Meloro and O’Higgins (2011). ratios exhibited no correlation with log Max- The majority of small cats (,7 kg) tend to hunt imum Length but the majority were negative- prey smaller than themselves and are capable ly influenced by humerus length (see of an arboreal life style (Meachen-Samuels and Supplementary Table 3). All the ratios were Van Valkenburgh 2009: Table 1). Consequent- retained for statistical analyses. ly, the morphology of their humerus mor- We used the software PASW Statistic 18 to phometry is expected to correlate to specific perform linear discriminant analysis (LDA) habitat adaptations differently than in larger with a forward stepwise procedure (with an F species. Lewis and Lague (2010) have also entry probability of p ¼ 0.05). This option demonstrated that long bone allometry of allows for the selection of the most appropri- felids (including the humerus) is better de- ate morphometric variables for discriminating scribed by a second polynomial regression, preassigned categories (Hair et al. 1998). suggesting that allometric differences occur Although the more relaxed criterion of F ¼ between small and larger taxa. 0.15 used by several previous studies (Kappel- Because the number of ‘‘small’’ cat (,7 kg) man et al. 1997; Bishop 1999; Elton 2001, 2002; specimens is relatively low (n ¼ 29), the large Plummer et al. 2008) generally leads to the number of variables relative to the sample size inclusion of more variables in the models, the precluded any meaningful LDAs. However, more stringent criterion used here should the data set of large taxa (.7 kg, n ¼ 82 select a smaller number of the most powerful specimens) was reanalyzed separately to test discriminatory variables, which could be an whether LDA on these taxa improved when advantage when applying the methods to compared to the overall data set. Because the fragmentary fossils. To assign fossil specimens, big cat LDA models performed so much better which lack an a priori habitat classification to a on data set 1 only (log-transformed data set), category, we used the modern specimens of we restricted our analyses to this data set. known habitat as a ‘‘training set.’’ Twelve separate LDAs were therefore un- Sensitivity Analyses dertaken, combining each scaled data set and In order to validate the efficacy of our LDA each habitat categorization method. For all models in making predictions irrespective of models, we interpreted the validity of LDA on unequal taxonomic sample size (Kovarovic et the basis of the Wilks’ lambda statistics and al. 2011), we performed two kinds of sensitivity MORPHOMETRIC HABITAT-ADAPTATION 331 analyses. First, we repeated the most accurate those from the distal humerus (Fig 1). In the LDA after removing all the specimens belong- analyses using ratios, anteroposterior maxi- ing to a particularly abundant taxon from the mum head diameter/mediolateral head articu- original sample. We repeated the LDA by lar surface width, the bicipital groove ratio, the excluding first Felis silvestris grampia (n ¼ 9, olecranon fossa ratio, and three of the trochlea the most abundant small felid), then Panthera ratios are all frequently selected. (Fig. 2). Both pardus (n ¼ 12, representative medium-size proximal and distal data are equally selected in felid of Mixed habitat), and finally Panthera leo ratio-based models (Fig. 2). (n ¼ 17, the most abundant large felid). Table 4 shows percentages of correctly A second sensitivity analysis was conducted classified specimens for each of the 12 models to test for the effect of sample size (number of using ‘‘leave-one-out’’ classification in the specimens) or body mass (in grams, log LDA. A combination of logged linear mea- transformed) on percentage of correctly clas- surements with habitat categorization method sified specimens for the 32 extant species A gives the most consistent and accurate sampled. On the basis of the results from all classifications (see also Fig. 3). Residuals and the LDA models, nonparametric Spearman ratios are less consistent and effective overall. correlation was applied to identify positive or Analyses using habitat categorization meth- negative significant correlations. ods C and D are less accurate than those using the other two methods. Results Proximal and Distal Epiphyses Whole Humerus The proximal and distal humerus models For the whole humerus data set, all 12 data were derived from logged linear measure- combinations yielded statistically significant ments and ratios; residuals were not used (p , 0.0001) Wilks’ lambda values in the linear because corrections were based on whole discriminant analysis (Table 3). humerus length, which would not be available The stepwise procedure reduced the number for fragmented fossil specimens. Sixteen linear of variables selected in the models. For models discriminant models (using each of the four using habitat categorization methods A, B, or D habitat categorization methods, and two data (using Open/Mixed/Closed categories), six to sets divided into proximal and distal ele- nine variables were selected. For models using ments) were therefore derived. All but one of method C (which uses only two habitat the discriminant functions were significant (p categories—forest and grassland) the number , 0.001). The exception was the model using of variables decreased to four, three, or two habitat categorization method B (IUCN cate- depending on the measurement-scaling data set gorization) and ratios for the distal humerus, used. Certain variables were selected in most of in which the second extracted function was the LDA models regardless of habitat categori- nonsignificant (p ¼ 0.32). Again, the stepwise zation or measurement-scaling data set (Fig. 1, procedure considerably reduced the number section ‘‘logged data’’). Mediolateral head of variables selected (Figs. 1, 2). For logged articular surface, bicipital groove depth, head linear measurements, six to nine variables surface height, and mediolateral subspinosus were selected for proximal models (with scar were commonly selected for the proximal greater tubercle mediolateral length selected end, whereas distal epiphysis maximum me- most often) and three to six for distal (with diolateral, olecranon fossa projection, trochlea distal mediolateral width and trochlea superi- superior-inferior medial border, and extensor or-inferior medial border being frequently carpi scar length were commonly selected selected) (Fig. 1); for ratios the number of measurements from the distal portion (Fig. 1). variables selected ranged from one—antero- These variables also tended to be selected in the posterior maximum head diameter/mediolat- models of the ‘‘size-free’’ data set (see Fig. 1, eral head articular surface width—and six section ‘‘size-free’’). Data from the proximal variables for proximal models, and up to five humerus were selected more frequently than for distal models (Fig. 2). 332 CARLO MELORO ET AL.

TABLE 3. Wilks’ lambda for 12 discriminant analyses based on overall measurements. For all values, p , 0.0001.

Habitat A Habitat B Habitat C Habitat D Logged (1) 0.241 0.299 0.794 0.271 Size-free (2) 0.297 0.349 0.859 0.385 Ratios (3) 0.368 0.557 0.86 0.415

The percentages of correctly classified spec- categorization method C. The stepwise proce- imens differed markedly between the models dure reduced the number of variables to as (Table 5). For the proximal humerus, the few as one (for proximal humerus only, logged model with the best classification rate used greater tubercle mediolateral width for habitat habitat categorization method A (Ortolani and categorization method D) to as many as ten Caro habitat) and data set 1 (logged linear (logged humerus measurements on the whole measurements). For the distal humerus, hab- humerus for habitat categorization method D). itat categorization methods C (GIS sample Proximal humerus measurements were select- based) and D (geographic range based) along ed more frequently than distal ones (Fig. 4), with data set 3 (ratios) gave the best results with bicipital groove descriptors being one of (Table 5). In general, the proximal humerus the most informative in the whole humerus models gave better classification results than and proximal humerus analyses (see Fig. 4). those for the distal humerus. Mediolateral width at the distal end is the most frequently selected variable in the distal Big Cats models. The highest percentage of correctly Discriminant functions were significant at p classified specimens is in models for the whole , 0.05, except for the distal humerus model humerus using logged linear measurements, using logged linear measurements and habitat with habitat categorization methods A and D

FIGURE 1. Variable selection by each categorization scheme (¼Meth) is indicated in gray. Results from different data sets are reported, including the whole bone LDA models (¼logged data), the ‘‘size-free’’ LDA model after regressing out bone length, and the epiphyses logged data (proximal and distal separated by line). Abbreviations: Meth A, presence or absence in particular biomes; Meth B, descriptions from IUCN specialist group; Meth C, GIS-based approach, grassland versus forest; Meth D, GIS-based approach, open versus closed; L, length; ML, mediolateral; AP, anterior-posterior. MORPHOMETRIC HABITAT-ADAPTATION 333

FIGURE 2. Variable selection by each categorization scheme (¼ Meth) is indicated in gray. Results from the ratio data sets (whole bone or epiphyses only) are shown. Abbreviations as in Figure 1.

2 giving the best results (Table 6) and method C v (9) ¼ 22.393, p ¼ 0.008), with models based on markedly worse. Accuracy declines with the data set 1 (logged linear measurements) use of proximal and distal elements on their having the highest percentage of correctly own, with distal models being the least classified cases (Fig. 5A). To compare the effective classifiers (Table 6). predictive accuracy of distinct habitat catego- rizations, we pooled percentage of correctly Summary of Models classified cases from different data sets. There When the percentage of correctly classified are no statistically significant differences be- cases is compared across different data sets tween reclassification rates for the different without taking different habitat methodolo- habitat categorization methods (all K-W p . gies into account there are no differences in 0.1, Fig. 5B). In general, the models using predictive power for the categories Open habitat criterion D obtain the highest correct 2 (Kruskal-Wallis v (9) ¼ 14.820, p ¼ 0.096) and prediction rate for the category Open but not 2 Mixed (K-W v (6) ¼ 16.075, p ¼ 0.065). The for Closed (Fig. 5B). When only big cats are predictive power for the Closed category considered, the category Closed is predicted at changes according to the data set used (K-W approximately 90% accuracy when habitat

TABLE 4. Percentage of correctly classified cases after leave-one-out procedure, for 12 discriminant models based on the overall data set.

Habitat A Habitat B Habitat C Habitat D Logged (1) Open 74.4 71.4 70.6 69.7 Mixed 75.9 69.4 69.0 Closed 78.6 71.9 61.9 75.0 Size-free (2) Open 82.1 69.1 64.7 60.6 Mixed 62.1 58.3 62.1 Closed 66.7 75.0 40.5 70.8 Ratios (3) Open 71.8 71.4 67.7 69.7 Mixed 65.5 44.4 75.9 Closed 64.3 59.4 64.3 60.4 334 CARLO MELORO ET AL.

three taxa close to the centroids for the different habitat categories (Open, Mixed, Closed). Open-adapted specimens exhibit on average a larger subspinosus scar, a higher head, a larger distal epiphysis and a higher trochlea relative to the Closed-adapted taxa.

Sensitivity Analyses The LDA models excluding specific taxa were applied to discriminate habitat categori- zation method A (Ortolani and Caro) for data set 1 (logged data). Excluding Felis silvestris grampia, the LDA yields two significant functions (at p , 0.0001) associated with ten measurements. The percentage of correctly classified cases is highest for Closed (81.8%), followed by Open (79.5%) and then Mixed (69.0%). The exclusion of the leopard also FIGURE 3. Plot of the discriminant functions extracted for shows little effect on the LDA model (signif- Habitat A (Ortolani and Caro 1996) when using logged linear measurements (number of variables selected ¼ icant after a selection of eight variables) with eight). percentage of correctly classified cases im- proving for Closed (92.7%) and Mixed (70.6%) criterion A is used; this result is never but not for Open (66.7%). On the other hand, achieved in any other LDA models. Conse- the exclusion of all lion specimens rendered quently, the Ortolani and Caro (1996) criterion the LDA model nonsignificant (significant only after adding at least seven lion speci- is probably the best for fitting ecomorphology mens). data of big cats (Fig. 5B). Species show distinct percentages of correct- We used UPGMA cluster analysis to sum- ly classified cases depending on the analyses marize differences and similarities in LDA (Table 7). A positive correlation is recorded model performance when different data sets between number of specimens and log body (Fig. 5C) and methods (Fig. 5D) were used. weight (rspearman ¼ 0.547, p , 0.0001). Number Figure 6 illustrates the major differences of specimens also correlates positively with between the proximal and distal epiphyses of percentage of correct cases when habitat A

TABLE 5. Percentage of correctly classified cases after leave-one-out procedure, for 16 discriminant models using proximal or distal humerus region.

Habitat A Habitat B Habitat C Habitat D Proximal log (1) Open 66.7 64.3 73.8 81.8 Mixed 65.5 63.9 44.8 Closed 69 65.6 64.7 68.8 Proximal ratio (3) Open 66.7 64.3 52.4 75.8 Mixed 72.4 47.2 72.4 Closed 57.1 59.4 66.2 43.8 Distal raw (1) Open 66.7 71.4 57.1 66.7 Mixed 41.4 44.4 62.1 Closed 67.4 30.3 69.6 57.1 Distal ratio (3) Open 69.2 78.6 76.2 81.8 Mixed 58.6 19.4 34.5 Closed 53.5 39.4 58 49 MORPHOMETRIC HABITAT-ADAPTATION 335

FIGURE 4. Variable selection by each categorization scheme (¼ Meth) is indicated in gray. Results from the whole log data set and epiphyses log data set of a subsample of large cats are reported. Abbreviations as in Figure 1.

(Ortolani and Caro 1996) is applied to the body mass had no influence on the other LDA logged data set (rspearman ¼ 0.362, p , 0.004) models. and to the ‘‘size-free’’ data set (rspearman ¼ 0.522, p , 0.002). For the latter data set there is also a Application to Fossil Specimens positive correlation between log body weight Habitat prediction of fossil specimens varies and percentage of correct cases (rspearman ¼ according to the method and data set used 0.357, p , 0.044). No other significant correla- (Fig. 7). The sabertooth/dirk-tooth cats Para- tions emerged, suggesting that sample size and orientalis and Smilodon populator

TABLE 6. Percentage of correctly classified cases after leave-one-out procedure, for 11 discriminant models based on subsample of large felids.

Habitat A Habitat B Habitat C Habitat D Log (1) Open 81.2 85.7 67.6 85.7 Mixed 91.7 90.5 97.5 Closed 95.5 72.7 61.4 80.8 Proximal log (1) Open 75 71.4 73.5 78.6 Mixed 79.2 61.9 41.7 Closed 63.6 72.7 68.2 19.2 Distal log (1) Open 75 78.6 n.s 74.3 Mixed 58.3 70.8 57.1 Closed 54.5 57.7 n.s 54.5 336 CARLO MELORO ET AL.

FIGURE 5. A, B, Box and whisker plots of percentage of correctly classified cases summarized in 40 LDA models. Central bar indicates the mean value, top and bottom of box indicate the 25% and 75% quartiles, whiskers indicate maximum and minimum values. C, D, UPGMA tree based on absolute differences in percentage of correctly classified cases. Cophenetic correlation is 0.8549 for data sets (C) and 0.9626 for methods (D). Abbreviations: DistalBig, data set of distal logged data for only big cats; DistalRatio, data set of distal ratios; DistalRaw, data set of distal logged data; ProximalBig, data set of proximal logged data for only big cats; ProxRatio, data set of proximal ratios; ProxRaw, data set of proximal logged data; Ratio, data set of ratios; Raw, data set of logged data; RawBig, data set of logged data for big cats only; Residual, data set ‘‘size-free.’’ Habitat definitions as in Figure 1. are generally predicted to be adapted to by similar proportions, as Open, Mixed, or Closed habitat. Dinofelis sp., on the other Closed. hand, exhibits a broader range of adaptations: Because all fossil specimens are large felids, depending on the analysis used, it is classified, the models based on the big cats data set 1 MORPHOMETRIC HABITAT-ADAPTATION 337

standing the cosmopolitan and eurybiomic nature of the family (Kitchener 1991; Turner and Anton´ 1997; Sunquist and Sunquist 2002; Kitchener et al. 2010). Single bones, even if fragmentary, can be ecologically informative. Previous work (Gonyea 1976; Anyonge 1996; Lewis 1997; Meloro 2011b) has demonstrated that comparative long bone indices, such as intermembral index, can be used to recon- struct habitat preference and locomotor strat- egy in a broad range of large carnivores. However, the probability of fossilization for FIGURE 6. Schematic representation of proximal and distal humerus epiphyses for three taxa that closely resemble in mammalian carnivores is generally low (Dam- the discriminant function scores the centroid of each uth 1982) and even lower for particularly large habitat categorization. Abbreviations: Sb_ML, subspino- felids (Gittleman and Harvey 1982; Gittleman sus scar mediolateral; H_H, head surface height; Dst_ML, distal epiphysis maximum ML; Cd_L3, trochlea superior- 1985; Turner and Anton´ 1997), so few rela- inferior minimum length. tively complete and associated skeletons are recovered. Developing accurate models based (logged linear measurements) using habitat on single bones and bone elements is thus method A are, as expected, the most accurate important. The resubstitution rates for the for the complete specimens. With this classi- felid humerus in this study are similar to those fication scheme, both Paramachairodus orienta- observed in discriminant analyses from other lis and Smilodon populator are classified as studies of large mammals, including bovids, Closed, and Dinofelis sp. (OLD 74/01) is suids, and primates (Kappelman et al. 1997; classified as Mixed. For the analyses of Bishop 1999; Elton 2001, 2002; DeGusta and proximal humerus we chose data set 1/ Vrba 2003, 2005a,b; Kovarovic and Andrews method A for big cats, which yielded the best 2007; Plummer et al. 2008). This indicates that rate of cross-validation accuracy (average 72.6 carnivorans, important components of past %) when compared with the other methods. and present biotas, can be as paleoecologically Again, P. orientalis and S. populator are informative as their prey (Herna´ndez Ferna´n- classified as Closed, and both Dinofelis sp. dez 2001; Herna´ndez Ferna´ndez and Pela´e- from Olduvai Bed I as Mixed. Data set 1/ Campomanes 2003; Herna´ndez Ferna´ndez et method B (IUCN classification scheme) had al. 2006; Herna´ndez Ferna´ndez and Vrba the best rate of classification for the distal 2006). humerus of big cats (average 69.03%) and it One promising avenue of research seeks to validates adaptation to Closed habitat in P. combine ecomorphology-based reconstruc- orientalis and S. populator;bothDinofelis tions of past habitats for different mammalian specimens from Bed I are classified as Open. groups likely to be sympatric and contempo- The distal fragment Panthera sp. (M 20240) raneous to construct a more holistic picture of from Bed I is predicted as Closed whereas the the environmental context of ecological com- two distal fragments of lions from Bed II and munities. This approach was recently used by Bed V are predicted as Open. The fossil Polly (2010), who examined calcaneum eco- humeri of cats from Olduvai Gorge show the morphology in different North American breadth of habitats present at Bed I relative to carnivoran communities and found strong Bed II and Bed V (Upper Pleistocene). correlations between community ‘‘ecometrics’’ and environmental variables. Taking this Discussion further, using multiple carnivorans and prey Our results clearly indicate that accurate species from the same locality may provide a information about habitat exploitation can be wealth of information about biome and recovered from the felid humerus, notwith- paleoenvironments that cannot be recovered 338 CARLO MELORO ET AL.

TABLE 7. Percentage of correctly classified cases (numbers are in decimals, 1.00 ¼ 100%) for each species based on a selection of LDA analyses performed on the whole humerus measurement data set. Meth ¼ habitat categorization. All of the 100% correct cases are in bold.

Logged data Size-free data Ratios Meth Meth Meth Meth Meth Meth Meth Meth Meth Meth Meth Meth Species A B C D A B C D A B C D Acinonyx jubatus 1.00 1.00 0.67 1.00 1.00 1.00 0.67 1.00 1.00 1.00 0.33 0.83 Caracal aurata 0.50 0.50 0.50 1.00 0.00 0.50 1.00 1.00 0.00 0.00 0.50 0.00 Caracal caracal 1.00 1.00 1.00 1.00 1.00 1.00 0.00 0.50 0.50 1.00 0.50 1.00 Felis chaus 0.00 0.00 0.00 0.50 0.00 0.00 0.00 0.00 0.50 0.50 0.50 0.00 Felis margarita 0.00 0.00 0.50 0.50 1.00 1.00 0.50 1.00 0.00 0.00 0.50 0.00 Felis marmorata 1.00 0.00 1.00 1.00 0.00 0.00 1.00 1.00 1.00 1.00 1.00 0.00 Felis nigripes 0.50 0.50 0.00 0.00 1.00 1.00 1.00 0.00 0.00 0.50 0.00 0.00 Felis silvestris grampia 1.00 1.00 1.00 0.89 0.78 0.67 0.78 0.56 1.00 0.67 1.00 0.89 Felis silvestris lybica 0.00 0.33 0.33 0.33 1.00 1.00 0.67 0.00 0.67 0.33 0.33 0.67 Leopardus geoffroyi 0.50 1.00 0.00 0.00 0.50 0.00 0.50 0.50 0.50 0.50 0.50 0.00 Leopardus guigna 0.00 1.00 1.00 1.00 0.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 Leopardus pardalis 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 Leopardus wiedii 0.00 0.00 1.00 1.00 0.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00 Leptailurus serval 0.83 0.50 0.50 1.00 0.67 0.83 0.83 0.83 0.50 0.33 0.67 0.83 Lynx canadensis 1.00 1.00 1.00 1.00 1.00 0.75 0.25 1.00 1.00 0.75 1.00 0.75 Lynx lynx 1.00 0.67 0.67 0.67 1.00 1.00 0.67 0.33 1.00 1.00 0.33 0.00 Lynx pardinus 1.00 0.50 1.00 1.00 1.00 0.50 0.50 1.00 1.00 0.00 1.00 1.00 Lynx rufus 0.00 1.00 0.00 0.00 0.00 1.00 0.00 1.00 1.00 1.00 1.00 0.00 Neofelis nebulosa 1.00 0.67 0.00 0.33 0.67 0.67 0.33 0.67 0.67 0.33 0.67 0.67 Panthera leo 0.88 0.94 0.47 0.88 0.76 0.53 0.18 0.53 0.88 0.82 0.65 0.94 Panthera onca 1.00 0.67 0.67 1.00 1.00 1.00 1.00 1.00 0.00 0.67 1.00 0.33 Panthera pardus 0.92 0.67 0.67 0.58 0.50 0.75 0.75 0.67 0.50 0.17 0.42 0.83 Panthera tigris 0.00 0.25 1.00 0.25 0.75 0.50 1.00 0.75 0.00 0.25 0.75 0.25 Panthera uncia 0.75 0.25 1.00 0.25 0.50 0.00 1.00 0.25 0.75 1.00 1.00 0.50 Pardofelis badia 1.00 1.00 1.00 1.00 0.00 1.00 1.00 0.00 0.00 0.00 1.00 1.00 Pardofelis temminckii 0.00 0.00 0.00 1.00 1.00 0.00 0.00 1.00 1.00 0.00 0.00 1.00 Prionailurus bengalensis 1.00 0.67 1.00 0.33 1.00 0.67 0.67 0.33 1.00 0.67 0.67 1.00 Prionailurus planiceps 1.00 1.00 1.00 1.00 0.00 1.00 0.00 1.00 1.00 1.00 1.00 1.00 Prionailurus rubiginosus 0.00 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.00 1.00 1.00 Prionailurus viverrinus 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.75 0.75 0.75 0.75 Puma concolor 0.50 0.00 1.00 0.50 0.50 0.50 1.00 1.00 0.50 0.50 0.50 0.00 Puma yagouaroundi 0.00 1.00 1.00 0.00 0.00 0.00 1.00 1.00 0.00 0.00 1.00 0.00 by focusing on specimens from a single dressed (Navarro et al. 2004). Using logged species, genus, or even family. linear measurements retains a significant size The analyses we present here highlight signal. For ecomorphic reconstruction using several methodological issues. Attempts to felids, it is likely that size, known to be hugely correct the data using residuals or ratios to take size into account did not increase the accuracy of the statistical models, and indeed in many cases yielded resubstitution rates that were lower than the logged linear measure- ments. These results therefore lend support to the use of minimally manipulated data in ecomorphic analysis of large mammals (sensu DeGusta and Vrba 2003). Investigations on the teeth of much smaller mammals—voles— suggested similar conclusions, indicating that residual or ratio-based scaling of morphomet- ric data is not always justified, must be FIGURE 7. Percentage of predicted cases for fossil speci- validated through experimentation, and must mens. Stratigraphy of specimens from Olduvai is also be appropriate to the question being ad- indicated. MORPHOMETRIC HABITAT-ADAPTATION 339 influential both biologically and ecologically as seen in Table 1, but it does highlight a grade in carnivorans (Gittleman 1985; Carbone et al. shift in the felids. Separation of these two 1996, 2007) as well as in mammals generally grades results in more accurate discrimination (Damuth and McFadden 1990), is an impor- for modern specimens, so a similar separation tant explanatory and discriminatory variable method (using estimates of mass based on the when considering habitat adaptation. This has size of the bone) should be used for fossil felid also been noted for primates (Elton 2001, 2002) specimens, as supported by the more consis- and bovids (Plummer et al. 2008; Louys et al. tent allocation of fossils to habitat types in our 2012). study when using the big cat modern training One challenge when retaining a size signal set rather than the full modern sample. by using logged linear measurements is There were no statistically significant differ- accounting for allometric effects, especially if ences in the discriminatory power of GIS- scaling factors differ between taxonomic based habitat categories (methods C and D) groups or adaptive grades. In the discriminant versus more traditional ways of assigning taxa analyses, resubstitution rates improved on the to groups (methods A and B) based on species’ whole when the modern sample was divided biology (see Fig. 5B). However, in the discrim- according to size, with 7 kg the threshold inant analyses, some habitat categorization between ‘‘small’’ and ‘‘big’’ cats. UPGMA methods worked better than others, and this cluster analysis suggests that data set 1 (log was also evident in the UPGMA cluster linear measurement) for big cats clearly analysis (Fig. 5D). For the whole humerus outperforms the others and is placed outside and proximal humerus, method A gave the all other clusters (Fig. 5C). This clearly points most accurate and consistent resubstitution to scaling differences within the felids, as results when the whole modern training set noted in other studies (Bertram and Biewener was used. When we used the big cats training 1990; Christiansen 1999; Lewis and Lague set, methods A and D yielded the best results. 2010). There is no hard-and-fast rule about The GIS-based methods (C and D) developed where the threshold should be drawn, how- here, which use specimen-specific information ever, with one study on interspecific scaling of based on geographic location, may have a the carnivoran postcranium pegging the size significant drawback in that their accuracy threshold at 100 kg (Bertram and Biewener relies on the sample’s having a representative 1990) and another at 50 kg (Christiansen geographical distribution. This is often prob- 1999). Another potential threshold is at 25 lematic with museum collections, as speci- kg, based on metabolism and hunting behav- mens tend to be drawn from a relatively small ior, because carnivorans bigger than this tend number of localities, often chosen for ease of to kill prey larger than themselves (Carbone et access (Cardini et al. 2007). Smaller or forest al. 1999, 2007). The choice of threshold should specimens may thus be more poorly repre- be appropriate to the question being ad- sented than the larger, charismatic open dressed. In our case, 7 kg is the most landscape animals, and that certainly seems meaningful threshold, as shown in previous to be the case for our sample. The overall research on locomotor behavior (Van Valken- habitat classification may thus be skewed. burgh 1985, 1987). Specifically, arboreality is Cluster analysis demonstrates that methods C primarily dictated by body mass (Van Val- and D (both GIS-based) stand apart from the kenburgh 1987); in Felidae, most taxa with others, in having a lower percentage correctly mean body masses ,7 kg tend to use arboreal classified. On the other hand, methods A and substrates frequently, whereas bigger forms B perform in a similar way to each other (Fig. such as caracal, serval, lynxes, and most of the 5D). pantherine cats tend to be more terrestrial, On balance, therefore, method A, based on albeit with many taxa being scansorial as well data from Ortolani and Caro (1996), who arboreal (Meachen-Samuels and Van Valken- recorded the presence or absence of carnivo- burgh 2009). This does not preclude the ran species, rather than individual specimens exploitation of closed habitats in taxa .7 kg, in broad biomes, seems to provide the best 340 CARLO MELORO ET AL. habitat categorization for extant species of big extensor carpii tuberosity to a lesser extent. cats. The error is relatively low for these measure- Unsurprisingly, the jackknifed classification ments. Both distal mediolateral width and rates were higher for models using the whole trochlea superior-inferior length provide in- humerus than for those employing either the formation about elbow functional morpholo- proximal or distal epiphyses separately. The gy. The elbow joint has been identified as proximal humerus was the most functionally functionally informative in other studies of informative in our felid sample, demonstrated Carnivora as a whole (Andersson and Werde- both by the dominance of proximal humerus lin 2003; Andersson 2004) because of its role in measurements in the stepwise LDA and by the supination as well as its high correlation with better resubstitution results from the analyses body mass. The extensor carpii muscle influ- using only the proximal region compared with ences movements of the forepaw that are distal. The proximal humerus is also highly needed in prey grappling and climbing informative in cercopithecid primates, al- (Barone 1980). though the distal humerus is also a good One major purpose of ecomorphic discrim- discriminator (Elton 2001). In felids, variables inant analysis is to reconstruct the habitat such as greater tubercle mediolateral width, preferences of extinct taxa. The results from bicipital groove depth and width, head artic- our study give useful insights into the ular surface height and width of subspinosus paleoecology of the saber/dirk toothed cats tuberosity (Fig. 6) are constantly selected in Smilodon populator and Paramachairodus orien- most models, and generally have a low degree talis and the enigmatic false sabertooth Dino- of measurement error. Similar measurements, felis sp. The most robust combination of albeit contained in ratios, were selected in logged linear measurements, habitat categori- primate models (Elton 2001), suggesting that zation method A and the big cat data set, they may be functionally informative across a supported strongly the assignment of both number of mammalian groups. The greater saber cats to the Closed category while tubercle is the attachment site for the supra predicting Dinofelis sp. as ‘‘mixed.’’ This is in and subspinosus muscle complex, part of the line with previous research that suggested that rotator cuff of the shoulder. The bicipital the large, specialist stalker S. populator needed groove (or sulcus intertubercularis) is the to exploit environments with extensive forest passage for the tendon of the biceps muscle cover (Gonyea 1976; Kurten´ and Werdelin (Reighard and Jennings 1901; Barone 1980). 1990; Christiansen and Harris 2005). Paleoeco- These muscles facilitate movement between logical data on P. orientalis are generally the scapula and the humerus in the vertical scanty, although its European counterpart and lateral planes. They are implicated in both (Pristinosmilus ogygia) has been suggested to prey capture and climbing, activities that can be a better climber than the leopard (Salesa et be linked to habitat adaptations in the Felidae al. 2005, 2006, 2009). Polly and MacLeod (e.g., arboreal taxa occur more in forested (2008) also predicted semidigitigrade locomo- biomes [Ewer 1973; Nowell and Jackson 1996; tor behavior for this big cat, indicating Kitchener et al. 2010]). climbing ability and possibly suggesting an In felids, models based on the distal adaptation to mixed-closed environments. humerus were less consistent and accurate, Lewis (1997) and Werdelin and Lewis (2001) and cluster analysis indicated a clear partition reported adaptations in Dinofelis sp. to tree between proximal (better performance) and climbing for carcass transport and possibly to distal humerus fractions (Fig. 5C). Relatively ‘‘mixed/closed’’ habitat exploitation. Our few measurements were selected in the step- analyses suggest that this taxon was probably wise procedure for the whole humerus. more eurybiomic and capable of adapting to a Measurements that most frequently emerged range of environments including open grass- were maximum mediolateral width of the land. These findings are supported by the distal humerus and trochlea superior-inferior allometric investigation of Lewis and Lague medial length (Fig. 6), as well as the size of the (2010), who recognized that the Dinofelis MORPHOMETRIC HABITAT-ADAPTATION 341 humerus had similarities with medium-sized Conclusion Mixed cats such as the leopard and the puma. The results presented in this paper show Predictions for the fragmentary material of that the use of carnivorans has great promise Panthera sp. of Olduvai Gorge Bed I are more in aiding paleoecological reconstructions. For enigmatic and possibly suggest that the felids, the best linear discriminant function specimen does not belong to P. leo but to a model, which predicted habitat categories at a large cat with adaptations to Closed habitats. very high rate of accuracy (over 90% for The fossil lions included in our analyses are Mixed and Closed categories in jack-knifed consistently assigned to the Open category, the classifications), used logged linear measure- same category as the modern lion. This ments in a training set comprising big (.7 kg) suggests that the habitat preferences of Afri- species. This suggests that the use of logged can Pleistocene lions were broadly similar to linear measurements could be preferable to those of extant forms. residual or ratio-based scaled data, although Overall, these results strongly encourage the this should be verified on a study-by-study inclusion of felids in paleoenvironmental recon- basis and with reference to the question being structions. It clearly emerges that a broad range addressed. It also indicates that a narrower, of habitats existed at Olduvai Bed I, with grade-based modern comparative sample possibly more forested conditions than in later might be more appropriate than one compris- intervals of Bed II and Bed V (Upper Pleisto- ing a larger number of species. The choice of cene). This conclusion is supported by other studies (Fernandez-Jalvo` et al. 1998, Plummer habitat categorization method is less straight- and Bishop 1994; Plummer et al. 2009), and if the forward, but a broad, biome-based classifica- ecomorphologies of a broader range of mam- tion using robust and comparable data from malian species were examined more robust the literature may be the preferable option. results might emerge. Louys et al. (2011) recently The proximal humerus appears to be func- demonstrated that broad ecological categories tionally informative, in line with earlier within mammalian communities correlate with ecomorphic research (Elton 2001). The high percentage of vegetation cover in extant tropical reclassification results for the proximal hu- ecosystems; thus future ‘‘taxon-free’’ studies merus on its own indicates that even isolated, have the potential to predict multidimensional fragmentary bones can yield useful informa- environmental variables. tion about habitat preferences in fossils. The Inevitably, a small number of inconsisten- fossil sabertooths Smilodon populator and Para- cies emerged in the multiple models used to machairodus orientalis show adaptations to assign the fossil specimens to habitat category, Closed environments whereas Dinofelis sp. with some resubstitutions into the Mixed from Olduvai Bed I clearly belongs to the category for both sabertooth cats and lions. Mixed category. This interpretation, together This may reflect measurement error or statis- with results emerging from other fragmentary tical inaccuracy. Alternatively, it may hint at fossil cats, validates previous paleoenviron- adaptations in extinct animals to environ- mental reconstructions of Olduvai Gorge, with ments that are unknown in the modern world. older stratigraphic intervals (Bed I) being Fossils may inhabit functional categories of characterized by higher abundance of forest- their own (Albrecht 1992), which may lead to adapted taxa. unexpected results in discriminant analysis. Acknowledgments Similarly, habitats in the past may not resem- ble those seen today. This has been suggested We are grateful to museum curators and by faunal evidence for Olduvai Bed I, which staff of The Natural History Museum of had a more species-rich woodland than that London, National Museum of Scotland, Royal seen today. It is therefore possible that Museum for Central Africa (Tervuren, Bel- significant differences exist between the func- gium) and National Museums of Kenya tion and structure of past and present ecosys- (Nairobi) for providing access to museum tems (Fernandez-Jalvo` et al. 1998). specimens. In particular, we would like to 342 CARLO MELORO ET AL. thank P. Jenkins, L. Tomsett, R. Portela- Biknevicius, A. R., and B. Van Valkenburgh. 1996. Design for killing: craniodental adaptations of predators. Pp. 393–428 in Miguez, A. Salvador, D. Hills, J. J. Hooker, P. Gittleman 1989. Brewer, and A. Currant (The Natural History Bishop, L. C. 1999. Suid paleoecology and habitat preference at Museum, London); A. Kitchener and J. Her- African and Pleistocene hominid localities. Pp. 216–225 man (National Museum of Scotland, Edin- in T. G. Bromage and F. Schrenk, eds. African biogeography, climate change and human evolution. Oxford University Press, burgh); E. Gilissen and W. Wendelen (Royal Oxford. Museum for Central Africa, Tervuren); E. Carbone, C., G. M. Mace, S. C. Roberts, and D. W. Macdonald. Mbua, M. Mungu, F. Nderitu and O. Mwebi 1999. Energetic constraints on the diet of terrestrial carnivores. Nature 402:286–288. (Kenya National Museum, Nairobi). A visit to Carbone, C., A. Teacher, and J. M. Rowcliffe. 2007. The costs of the Royal Museum of Central Africa was carnivory. PLoS Biology 5(2):e22. supported by the Synthesys grant ‘‘Ecomor- Carrano, M. T. 1999. What, if anything, is a cursor? Categories vs. continua for determining locomotor habit in mammals and phology of extant African carnivores’’ (BE- dinosaurs. Journal of Zoology 247:29–42. TAF 4901) to C. Meloro. We are grateful to the Cardini, A., A.-U. Jansson, and S. Elton. 2007. Ecomorphology of Governments of Kenya and Tanzania for vervet monkeys: a geometric morphometric approach to the study of clinal variation. Journal of Biogeography 34:1663–1678. kindly providing permission to study Olduvai Christiansen, P. 1999. Scaling of the limb long bones to bodymass and Koobi Fora fossil material. in terrestrial mammals. Journal of Morphology 239:167–190. C. P. Egeland and M. Dom´ınguez-Rodrigo ———. 2008. Evolution of skull and mandible shape in cats. PLoS ONE 3(7):e2807. kindly provided assistance in verifying strati- Christiansen, P., and M. Harris. 2005. Body size of Smilodon graphic interval for Olduvai Gorge fossil (Mammalia: Felidae). Journal of Morphology 266:369–384. material. The associate editor and several Damuth, J. 1982. Analysis of the preservation of community anonymous reviewers greatly improved the structure in assemblages of fossil mammals. Paleobiology 8:434–446. quality of our manuscript. Our research was Damuth, J., and B. J. MacFadden. 1990. Body size in mammalian generously supported by the Leverhulme paleobiology. Cambridge University Press, Cambridge. Trust project ‘‘Taxon-Free Palaeontological DeGusta, D., and E. S. Vrba. 2003. A method for inferring paleohabitats from the functional morphology of bovid astrag- Methods for Reconstructing Environmental ali. Journal of Archaeological Science 30:1009–1022. Change’’ (F/00 754/C). ———. 2005a. Methods for inferring paleohabitats from discrete traits of the bovid postcranial skeleton. Journal of Archaeolog- Literature Cited ical Science 32:1115–1123. ———. 2005b. Methods for inferring paleohabitats from the Albrecht, G. H. 1992. Assessing the affinities of fossils using functional morphology of bovid phalanges. Journal of Archae- canonical variates and generalized distances. Journal of Human ological Science 32:1099–1113. Evolution 7:49–69. Elton, S. 2001. Locomotor and habitat classification of cercopith- Andersson, K. 2004. Elbow-joint morphology as a guide to forearm function and foraging behaviour in mammalian carnivores. ecoid postcranial material from Sterkfontein Member4, Bolt’s Zoological Journal of the Linnean Society 142:91–104. Farm and Members 1 and 2, . Palae- Andersson, K., and L. Werdelin. 2003. The evolution of cursorial ontologia Africana 37:115–126. carnivores in the Tertiary: implications of elbow-joint morphol- ———. 2002. A reappraisal of the locomotion and habitat ogy. Proceeding of the Royal Society of London B 270:S163–S165. preference of Theropithecus oswaldi. Folia Primatologica 73:252– Anton,´ M., M. J. Salesa, J. F. Pastor, I. M. Sa´nchez, S. Fraile, and J. 280. Morales. 2004. Implications of the mastoid anatomy of larger ———. 2006. 40 years on and still going strong: the use of the extant felids for the evolution and predatory behaviour of hominin-cercopithecid comparison in human evolution. Journal sabertoothed cats (Mammalia, Carnivora, Felidae). Zoological of the Royal Anthropological Institute 12:19–38. Journal of the Linnean Society 140:207–221. Ewer, R. F. 1973. The carnivores. Cornell University Press, Ithaca, Anton,´ M., A. Galobart, and A. Turner. 2005. Co-existence of N.Y. scimitar-toothed cats, lions and hominins in the European Farlow, J. O., and E. R. Pianka. 2002. Body size overlap, habitat Pleistocene: implications of the post-cranial anatomy of Homo- partitioning and living space requirements of terrestrial verte- therium latidens (Owen) for comparative paleoecology. Quater- brate predators: implications for the paleoecology of large nary Science Review 24:1287–1301. theropod dinosaurs. Historical Biology 16:21–40. Anyonge, W. 1996. Locomotor behaviour in Plio-Pleistocene saber- Fernandez-Jalvo,` Y., C. Denys, P. Andrews, T. Williams, Y. tooth cats: a biomechanical analysis. Journal of Zoology Dauphin, and L. Humphrey. 1998. Taphonomy and palae- 238:395–413. oecology of Olduvai Bed-I (Pleistocene, Tanzania). Journal of Barone, R. 1980. Anatomia Comparata dei Mammiferi Domestici, Human Evolution 34:137–172. Vol. 1. Osteologia. Edagricole, Bologna. Garland, T., Jr., and C. M. Janis. 1993. Does metatarsal/femur ratio Bassarova, M., C. M. Janis, and M. Archer. 2009. The calcaneum— predict the maximal running speed in cursorial mammals? on the heels of marsupial locomotion. Journal of Mammalian Journal of Zoology 229:133–151. Evolution 16:1–23. Gittleman, J. L. 1985. Carnivore body size: ecological and Bertram, J. E. A., and A. A. Biewener. 1990. Differential scaling of taxonomic correlates. Oecologia 67:540–554. the long bones in the terrestrial Carnivora and other mammals. ———, ed. 1989. Carnivore behavior, ecology, and evolution, Vol. Journal of Morphology 204:157–169. 1. Cornell University Press, Ithaca, N.Y. MORPHOMETRIC HABITAT-ADAPTATION 343

Gittleman, J. L., and P. H. Harvey. 1982. Carnivore home-range implications for community convergence. Global Ecology and size, metabolic needs and ecology. Behavioral Ecology and Biogeography 20:717–729. Sociobiology 10:52–63. Louys, J., Montanari, S., Plummer, T., Hertel, F., and L. C. Bishop, Gittleman, J. L., and B. Van Valkenburgh. 1997. Sexual size 2012. Evolutionary divergence and convergence in shape and dimorphism in the canines and skulls of carnivores: effects of size within African antelope proximal phalanges. Journal of size, phylogeny, and behavioural ecology. Journal of Zoology Mammalian Evolution. doi: 10.1007/s10914-012-9211-4. 242:97–117. Macdonald, D. W., A. J. Loveridge, and K. Nowell. 2010. Dramatis Gonyea, W. J. 1976. Behavioral implications of saber-toothed felid personae: an introduction to the wild felids. Pp. 3–58 in D. W. morphology. Paleobiology 2:332–342. Macdonald and A. J. Loveridge, eds. Biology and conservation Hair, J. F., R. E. Anderson, R. L. Tatham, and W. C. Black. 1998. of wild felids. Oxford University Press, Oxford. Multivariate data analysis, 5th ed. Prentice Hall, Upper Saddle Martin, L. D. 1989. Fossil history of terrestrial Carnivora. Pp. 536– River, N.J. 568 in Gittleman 1989. Herna´ndez Ferna´ndez, M. 2001. Bioclimatic discriminant capacity McHenry, C. R., S. Wroe, P. D. Clausen, K. Moreno, and E. of terrestrial faunas. Global Ecology and Biogeography Cunningham. 2007. Supermodeled sabercat, predatory behavior 10:189–204. in Smilodon fatalis revealed by high-resolution 3D computer Herna´ndez Ferna´ndez, M., and P. Pela´ez-Campomanes. 2003. The simulation. Proceedings of the National Academy of Sciences bioclimatic model: a method of palaeoclimatic qualitative USA 104:16010–16015. inference based on mammal associations. Global Ecology and Meachen-Samuels, J. 2012. Morphological convergence of the Biogeography 12:507–517. prey-killing arsenal of sabertooth predators. Paleobiology 38:1– Herna´ndez Ferna´ndez, M., and E. Vrba. 2006. Plio-Pleistocene 14. climate change in the Turkana Basin (East Africa): evidence from Meachen-Samuels, J., and B. Van Valkenburgh. 2009. Forelimb large mammal faunas. Journal of Human Evolution 50:595–626. indicators of prey-size preference in the Felidae. Journal of Herna´ndez Ferna´ndez, M., M. Alberdi, B. Azanza, P. Montoya, J. Morphology 270:729–744. Morales, M. Nieto, and P.Pela´ez-Campomanes. 2006. Identification ———. 2010. Radiographs reveal exceptional forelimb strength in problems of arid environments in the Neogene-Quaternary the sabertooth cat, Smilodon fatalis. PLoS ONE 5:e11412. mammal record of Spain. Journal of Arid Environments 66:585–608. Meloro, C. 2011a. Feeding habits of Plio-Pleistocene large Holliday, J. A., and S. J. Steppan. 2004. Evolution of hyper- carnivores as revealed by their mandibular geometry. Journal carnivory: the effect of specialization on morphological and of Vertebrate Paleontology 31:428–446. taxonomic diversity. Paleobiology 30:108–128. ———. 2011b. Locomotor adaptations in Plio-Pleistocene large IUCN 2009. The IUCN red list of threatened species, Version carnivores from the Italian Peninsula: palaeoecological implica- 2009.1. tions. Current Zoology 57:269–283. Janis, C. M., and P. B. Wilhem. 1993. Were there mammalian Meloro, C., and P. O’Higgins. 2011. Ecological adaptations of pursuit predators in the Tertiary? Dances with wolf avatars. mandibular form in Fissiped Carnivora. Journal of Mammalian Journal of Mammalian Evolution 1:103–125. Evolution 18:185–200. Kappelman, J. 1988. Morphology and locomotor adaptations of Meloro, C., and P. Raia. 2010. Cats and dogs down the tree: the the bovid femur in relation to habitat. Journal Morphology tempo and mode of evolution in the lower carnassial of fossil 198:119–130. and living Carnivora. Evolutionary Biology 37:177–186. Kappelman, J., T. Plummer, L. Bishop, A. Duncan, and S. Appleton. Navarro, N., X. Zatarain, and S. Mountuire. 2004. Effects of 1997. Bovids as indicators of Plio-Pleistocene paleoenvironments morphometric descriptor changes on statistical classification in East Africa. Journal Human Evolution 32:229–256. and morphospaces. Biological Journal of the Linnean Society Kitchener, A. C. 1991. The natural history of the wild cats. 83:243–260. Comstock Associates, Ithaca, N.Y. Nowell, K., and P. Jackson. 2006. Wild cats: status survey and Kitchener, A. C., B. Van Valkenburgh, and N. Yamaguchi. 2010. conservation action plan. IUCN Produced Services Unit, Cam- Felid form and function. Pp. 83–106 in D. W. Macdonald and A. bridge, U.K. J. Loveridge, eds. Biology and conservation of wild felids. Olson, D., E. Dinerstein, E. D. Wikramanayake, N. D. Burgess, G. Oxford University Press, Oxford. V. N. Powell, E. C. Underwood, J. A. D’amico, I. Itoua, H. E. Kovarovic, K. M., and P. Andrews. 2007. Bovid postcranial Strand, J. C. Morrison, C. J. Loucks, T. F. Allnutt, T. H. Ricketts, ecomorphological survey of the Laetoli paleoenvironment. Y. Kura, J. F. Lamoreux, W. W. Wettengel, P. Hedao, and K. R. Journal of Human Evolution 52:663–680. Kassem. 2001. Terrestrial ecoregions of the world. Bioscience Kovarovic, K., Aiello, L. C., Cardini, A., and C. A. Lockwood. 51:933–938. 2011. Discriminant function analyses in archaeology: are O’Regan, H. J., and A. Kitchener. 2005. The effects of captivity on classification rates too good to be true? Journal of Archaeolog- the morphology of captive, domesticated and feral mammals. ical Science 38:3006–3018. Mammal Review 34:215–230. Kurten,´ B., and L. Werdelin. 1990. Relationships between North Ortolani, A., and T. M. Caro. 1996. The adaptive significance of and South American Smilodon. Journal of Vertebrate Paleontol- color patterns in carnivores: Phylogenetic tests of classic ogy 10:158–169. hypotheses. Pp. 132–188 in J. L. Gittleman, ed. Carnivore Lewis, M. E. 1997. Carnivoran paleoguilds of Africa: implications behavior, ecology, and evolution, Vol. 2. Comstock, Ithaca, N.Y. for hominid food procurement strategies. Journal of Human Palmqvist, P., V. Torregrosa, J. A. Perez-Claros,´ B. Mart´ınez- Evolution 32:257–288. Navarro, and A. Turner. 2007. A re-evaluation of the diversity of Lewis, M. E., and M. R. Lague. 2010. Interpreting sabretooth cat (Mammalia, Carnivora, ) and (Carnivora; Felidae; Machairodontinae) postcranial morphology the problem of species identification in extinct carnivores. in light of scaling patterns. Pp. 411–465 in A. Goswami and A. R. Journal of Vertebrate Paleontology 27:160–175. Friscia, eds. Carnivoran evolution: new views on phylogeny, Plummer, T. W., and L. C. Bishop. 1994. Hominid palaeoecology at form and function. Cambridge University Press, Cambridge. Olduvai Gorge, Tanzania as indicated by antelope remains. Louys, J., Meloro, C., Elton, S., Ditchfield, P., and L. C. Bishop. Journal of Human Evolution 27:47–75. 2011. Mammal community structure correlates with arboreal Plummer, T. W., L. C. Bishop, and F. Hertel. 2008. Habitat heterogeneity in faunally and geographically diverse habitats: preference of extant African bovids based on astragalus 344 CARLO MELORO ET AL.

morphology: operationalizing ecomorphology for palaeoenvi- Schutz, H., and R. P. Guralnik. 2007. Postcranial element shape ronmental reconstruction. Journal of Archaeological Science and function: assessing locomotor mode in extant and extinct 35:3016–3027. mustelid carnivorans. Zoological Journal of the Linnean Society Plummer, T. W., P. W. Ditchfield, L. C. Bishop, J. D. Kingston, J. V. 150:895–914. Ferraro, D. R. Braun, F. Hertel, and R. Potts. 2009. Oldest Slater, J. G., and B. Van Valkenburgh. 2008. Long in the tooth: evidence of toolmaking hominins in a grassland-dominated evolution of sabertooth cat cranial shape. Paleobiology 34:403– ecosystem. PLoS ONE 4:e7199. 419. Polly, P. D., and N. Macleod. 2008. Locomotion in fossil Carnivora: Sunquist, M., and F. Sunquist. 2002. Wild cats of the world. an application of eigensurface analysis for morphometric University of Chicago Press, Chicago. comparison of 3D surfaces. Palaeontologia Electronica 11.2.8A. Turner, A., and M. Anton.´ 1997. The big cats and their fossil Polly, P. D. 2008. Adaptive zones and the pinniped ankle: a 3D relatives. Columbia University Press, New York. quantitative analysis of carnivoran tarsal evolution. Pp. 165–194 Van Valkenburgh, B. 1985. Locomotor diversity between past and in E. Sargis and M. Dagosto, eds. Mammalian evolutionary present guilds of large predatory mammals. Paleobiology morphology: a tribute to Frederick S. Szalay. Springer, Dordrecht. 11:406–428. ———. 2010. Tiptoeing through the trophics: geographic variation ———. 1987. Skeletal indicators of locomotor behavior in living in carnivoran locomotor ecomorphology in relation to environ- and extinct carnivores. Journal of Vertebrate Paleontology 7:62– ment. Pp. 374–410 in A. Goswami and A. Friscia, eds. 182. Carnivoran evolution: new views on phylogeny, form and ———. 1988. Trophic diversity in past and present guilds of large function. Cambridge University Press, Cambridge. predatory mammals. Paleobiology 14:155–173. Reighard, J., and H. S. Jennings. 1901. Anatomy of the cat. Henry ———. 1989. Carnivore dental adaptations and diet: a study of Holt, New York. trophic diversity within guilds. Pp. 410–436 in Gittleman 1989. Salesa, M. J., M. Anton,´ A. Turner, and J. Morales. 2005. Aspects of ———. 1999. Major patterns in the history of carnivorous the functional morphology in the cranial and cervical skeleton of the sabre-toothed cat Paramachairodus ogygia (Kaup, 1832) mammals. Annual Review of Earth and Planetary Sciences (Felidae, Machairodontinae) from the Late Miocene of Spain: 27:463–93. implications for the origins of the machairodont killing bite. ———. 2007. Dej´ a` vu: the evolution of feeding morphologies in Zoological Journal of the Linnean Society 144:363–377. the Carnivora. Integrative and Comparative Biology 47:147–163. ———. 2006. Inferred behaviour and ecology of the primitive Walmsley, A., S. Elton, J. Louys, L. C. Bishop, and C. Meloro. 2012. sabretoothed cat Paramachairodus ogygia (Felidae, Machair- Humeral epiphyseal shape in the Felidae: the influence of odontinae) from the Late Miocene of Spain. Journal of Zoology phylogeny, allometry, and locomotion. Journal of Morphology 268:243–254. 273:1424–1438. ———. 2009. Functional anatomy of forelimb in Pristinosmilus Werdelin, L., and M. E. Lewis. 2001. A revision of the genus ogygia (Felidae, Machairodontinae, ) from the Late Dinofelis (Mammalia, Felidae). Zoological Journal of the Linnean Miocene of Spain and the origins of the sabre-toothed felid Society 132:147–258. model. Journal of Anatomy 216:381–396. Wroe, S., C. McHenry, and J. Thomason. 2005. Bite club: Samuels, J. X., J. A. Meachen, and S. A. Sakai. 2012. Postcranial comparative bite force in big biting mammals and the prediction morphology and the locomotor habits of living and extinct of predatory behaviour in fossil taxa. Proceeding of the Royal carnivorans. Journal of Morphology 274:121–146. Society of London B 272:619–625.