No evidence for parallel evolution of cursorial limb adaptations among Neogene South American native ungulates (SANUs).

During the Neogene, many North American ungulates evolved longer limbs. Presumably, this allowed them to move more efficiently or quickly in open habitats, which became more common during this interval. Evidence suggests that open habitats appeared even earlier in South America, but no study to date has investigated whether the ungulate-like mammals of South America (South American native ungulates or SANUs) evolved similar limb adaptations. We analyzed limb elongation in the two predominant SANU groups, notoungulates and litopterns, by compiling genus-level occurrences from the late Oligocene to the Pleistocene and calculating metatarsal/femur ratio (Mt:F). None of the groups or subgroups we analyzed show a pronounced increase in Mt:F across this interval, with the possible exception of proterotheriid litopterns. Proterotheriids are thought to have inhabited forested environments rather than open ones, which raises questions about the selective forces responsible for limb elongation in ungulates. Conversely, notoungulates, which are traditionally thought to have lived in open habitats, show no strong trend of increasing Mt:F across this interval. Our study suggests that the macroevolutionary trend of limb elongation in ungulate-like mammals is not universal and is highly influenced by the evolutionary affinities of the groups being analyzed.

Introduction

The broad pattern of global climate during the Cenozoic is well-established [1]. The climate was generally warm and humid through the Paleocene and early Eocene, with an abrupt temperature peak at the Paleocene-Eocene boundary [2]; temperatures declined through the middle and late Eocene and then dropped precipitously at the Eocene-Oligocene transition 34 million years ago [3]. Temperatures fluctuated but were relatively steady during the Oligocene and early Miocene and reached a peak around the early-middle Miocene, ca. 17–15 million years ago [4]. Following this climatic optimum, temperatures gradually declined, and aridity increased.

In North America, changes in vegetational structure and mammal ecology associated with these climatic changes have been studied extensively [e.g., 5–10]. During the latter half of the Cenozoic, ungulates (hoofed herbivores) show two conspicuous trends: increased tooth crown height (hypsodonty) and pronounced ‘cursorial’ limb modifications (i.e., long, slender, stable limbs). Both of these trends appear to be adaptive responses to increasing proportions of open (i.e., relatively tree-free) habitats. Hypsodonty maintains the functional lifespan of the dentition in open habitats, which are characterized by high levels of exogenous abrasives (grit) as well as abundant phytolith-rich grasses [11-18]. Limbs that are relatively long distally and bear fewer lateral digits are biomechanically more efficient for moving quadrupedally in open habitats since they are lighter, permit more proximal muscle attachments, and result in longer stride lengths, among other factors [19-25]. Hypsodonty increased in many North American ungulate lineages around the early-middle Miocene transition [14], slightly post-dating the appearance of grass-dominated open habitats [26, 27]. An increase in cursoriality (as measured by distal limb elongation) took place some 10 million years earlier than the increase in hypsodonty, during the Oligocene [28; see also S1 File]. Similar patterns have been documented for the Neogene of Eurasia [29].

In contrast to North America, South America was separated from most other continents by large water barriers for nearly the entire Cenozoic. Consequently, few Northern Hemisphere mammal groups were able to disperse to South America during this interval. In their absence, ancient South America ecosystems were filled with mammals exclusive to (or virtually exclusive to) that continent [30-32]. These endemic groups included several clades of ungulate-like herbivores collectively referred to as South American native ungulates (SANUs), as well as xenarthrans (sloths, armadillos, and anteaters), metatherians (marsupials and relatives), and non-therian mammals, among others. The most diverse, abundant, and long-lived SANUs were notoungulates (order Notoungulata) and litopterns (order Litopterna), both of which survived until the late Pleistocene [33-37].

South American notoungulates evolved hypsodont dentitions much earlier than North American ungulates: by the late Eocene in some species and by the late Oligocene in many others [30, 38–41]. Open-habitat grasses were not major components of South American ecosystems prior to the Miocene [42, 43], but plant phytolith ecomorphology indicates that arid conditions and open habitats were present in southern South America as early as the middle Eocene [44, 45]. Thus, although these Eocene open habitats may have resembled grasslands or open shrublands in their vegetational structure, their taxonomic composition was distinct. Mammal ecomorphological data from the early Oligocene of central Chile also indicate that open habitats appeared in South America some 10 million years earlier than in North America [41, 46].

Despite longstanding interest in the so-called ‘precocious’ hypsodonty of South American notoungulates [17], no study has systematically assessed the evolution of SANU limbs during the middle to late Cenozoic. Shockey and Flynn [47], in their analysis of postcranial remains of Eocene ‘isotemnid’ notoungulates, noted that the proportion of notoungulates with a superior astragalar foramen (S1 Fig) decreased between the middle Eocene and the late Oligocene, coincident with the appearance of open habitats in South America. Although the function of this foramen (and its associated canal) is unknown both in extinct and extant mammals, it has been proposed that it could have limited full extension at the ankle joint if it transmitted neurovascular tissue, since it is located at the posterior end of the astragalar trochlea (S1 Fig). If so, its loss would permit greater range of motion at the tibioastragalar joint and could represent an adaptation for longer stride length and faster or more efficient locomotion in more open habitats [47]. However, this foramen has been lost in most species of all extant orders of mammals, independent of locomotor habit or cursorial tendencies; it only remains in a few digitigrade species (e.g., the aardvark, Orycteropus afer, and the maned wolf, Chrysocyon brachyurus), aquatic species (e.g., the neotropical otter, Lontra longicaudis and the crabeater seal, Lobodon carcinophagus), and some bears (ursids) [48-51]. This casts doubt on a tight correlation between the loss of this foramen and limb posture or locomotor performance.

In this study, we quantitatively assess limb evolution in SANUs during the late Oligocene to Pleistocene interval using the metatarsal-femur ratio (Mt:F, also known as pes length index), a commonly-used measure of distal limb elongation. Lengthening the distal limb (autopod) relative to the proximal limb (stylopod) is biomechanically advantageous because it increases stride length, which is more efficient [24, 52] and may also enable greater speed [20]. As a result, this variable has been used to study the evolution of cursoriality in a wide variety of mammals [28, 53–55]. Although both ancestry (phylogenetic relationships) and body mass can complicate precise comparisons of Mt:F among groups that are distantly related to one another and/or that differ greatly in size [56], this ratio can be a useful tool for assessing within-group trends and making general comparisons among mammals of roughly similar size [28]. Among SANUs, we focus on notoungulates and litopterns, the most diverse and abundant clades and the only ones with many species of medium to large size (ca. 10–1,000 kg) during this interval [37]. The precise evolutionary relationships of these groups are still uncertain [37], but molecular data from Pleistocene representatives of Litopterna and Notoungulata suggest they are most closely related to Perissodactyla among extant mammals [57-59].

Litoptern relationships are based on McGrath et al. [60, 61], with families Macraucheniidae (Ma) and Proterotheriidae (Pr) indicated. Toxodontian relationships are based on Shockey et al. [62], Bonini et al. [63], and Armella et al. [64]. Typothere relationships are based on Seoane and Cerdeño [65] and Croft and Anaya [66], with Hemihegetotherium sp. assumed to be closely related to Hemihegetotherium achataleptum. Typothere clades include Hegetotheriidae (He), Hegetotheriinae (H), Interatheriidae (Int), Mesotheriidae (Me), and Pachyrukhinae (P).

Material and methods

We obtained metatarsal and femur data for 31 genera and 36 species of South American native ungulates pertaining to three major clades: Litopterna, Toxodontia, and Typotheria (S1 Table and Fig 1). Museum specimens were measured to the nearest 0.1 mm using a digital or analog calipers, depending on the size of the specimen. Other measurements were taken directly from the scientific literature or measured from a photo using ImageJ [67]. The sources of each measurement are listed in S1 Table. In most cases (86% of species), measurements were from the same specimen/individual. In the remaining cases, data from different specimens/individuals from the same fossil locality were used to maximize taxon sampling. Femur length was measured parallel to the shaft from the head to the distalmost point, and metatarsal length was measured as greatest length parallel to the long axis of the bone. Each species was assigned a single age based on the specimen that was measured, with ages rounded to the nearest megannum (S1 Table). Typothere notoungulates, toxodont notoungulates, and litopterns were analyzed separately, as were smaller groups in some cases. Statistical analyses and data visualization were conducted using JMP Pro® 14.2 for Mac [68]. Reported slopes are rounded to two significant digits.

Fig 1

Phylogenetic relationships of SANUs analyzed in this study, grouped by major clade.

Results

Mean Mt:F is highest for litopterns (0.46; N = 11), followed by typothere notoungulates (0.37; N = 18), and toxodont notoungulates (0.24; N = 7) (p < 0.05, Wilcoxon test). None of these three clades shows a pronounced trend in Mt:F across the study interval as indicated by ordinary least-squares regression (Fig 2). Mt:F decreases slightly in toxodonts (slope = -0.0039; r2 = 0.25) and very slightly in typotheres (slope = -0.00088; r2 = 0.013), and it increases slightly in litopterns (slope = 0.0048; r2 = 0.13). As a whole (N = 36), SANUs show a slight upward trend from the late Oligocene to the Pleistocene (slope = 0.0011; r2 = 0.0075).

Fig 2

Metatarsal-femur ratio (Mt:F) in extinct North American ungulates and South American native ungulates.

North American ungulate data (N = 73) are from Janis and Wilhelm [28]. Data for South American native ungulates (N = 36) are provided in S1 and S2 Tables. Regression lines are ordinary least-squared regressions.

Among typotheres, mesotheriids (N = 4) show a slight downward trend in Mt:F (slope = -0.0028; r2 = 0.46), hegetotheriids (N = 8) show a negligible upward trend (slope = 0.00092; r2 = 0.20), as do interatheriids (N = 6; slope = 0.00012; r2 = 0.0001). Among litopterns, macraucheniids (N = 5) show a downward trend (slope = -0.0023; r2 = 0.17), whereas proterotheriids (N = 6) show an upward trend (slope = 0.016; r2 = 0.68).

Only a single SANU, the proterotheriid litoptern Eoauchenia primitiva, has Mt:F > 0.65 (Fig 2 and S1 Table), the lower limit among modern ungulates that are generally classified as cursorial (camelids, pecoran ruminants, equids). Most litopterns and some typothere notoungulates have Mt:F between 0.38 and 0.65, below that of extant cursorial ungulates but within the range of extant cursorial (and semiaquatic) carnivorans and rodents. Most notoungulates have Mt:F below 0.38, below that of cursorial carnivorans and rodents.

Discussion

As noted previously, North American ungulates show a marked increase in Mt:F around the Oligocene-Miocene transition [28]. A similar pattern has been documented in Eurasia [69, 70]. Our analysis demonstrates that this did not occur in South America, at least from the late Oligocene onwards. Older SANU postcranial remains are rare, but the Mt:F value of 0.33 for the middle Eocene oldfieldthomasiid notoungulate Allalmeia atalaensis [71, 72] suggests that no major change in limb proportions occurred among notoungulates during the late Paleogene. (No comparable data are available for litopterns).

It may be that the pattern we observe is the result of biases in our dataset. Although this possibility cannot be excluded entirely, we are aware of no systematic biases in the data that would significantly affect overall patterns. The taxa sampled come from throughout the continent and are relatively evenly distributed among the families sampled; no major families were unsampled. The taxa studied derive from a variety of habitats, nearly all of which have been reconstructed as at least partially open and suitable for more cursorial species. Our dataset does not sample the majority of extinct notoungulate and litoptern species that have been recorded from this interval, but it appears to be broadly representative.

Another possibility is that ordinal and subordinal patterns are obscuring different trends at the family level that do track climate, habitat, or other variables. This seems to be true to some extent, as the SANU families we analyzed show trends that vary relative to their broader taxonomic group, though such trends seem to reflect ecological specialization and body mass more than habitat per se. It is important to note that most of these more restricted clades are represented by only a few species (generally 4–8; Fig 1 and S1 Table), and interpretations of trends should be considered provisional. Nevertheless, they merit further discussion.

Family-level patterns

The overall trend in litopterns combines a downward trend among macraucheniids and an upward trend among proterotheriids. The former may relate to body mass (BM), which increases in macraucheniids during the late Cenozoic [61] and has an inverse relationship with Mt:F in our sample; the highest Mt:F pertains to diminutive Llullataruca shockeyi [60], while the lowest Mt:F corresponds to one-ton Macrauchenia patachonica. An ordinary least-squares regression of Mt:F on log(BM) for macraucheniids has a slope of -0.088 and r2 of 0.707, supporting an inverse relationship between Mt:F and BM in this family. The much weaker relationship between Mt:F and BM in proterotheriids (slope = -0.229; r2 = 0.131) suggests that increasing Mt:F is more closely related to other ecological factors.

Among toxodont notoungulates, the slight decrease in Mt:F through time may also be related to increasing average body mass, since smaller toxodonts (< 100 kg) have higher Mt:F values than larger ones (slope = - 0.088; r2 = 0.538) and the proportion of smaller toxodonts decreases through time (Fig 2 and S3 Table). However, the six species included in our sample pertain to four families (Notohippidae, Leontiniidae, Homalodotheriidae, and Toxodontidae) with varied ecological characteristics, making it unlikely that the trend can be attributed to any single factor.

Among typothere notoungulates, Mt:F decreases through time in mesotheriids. This could reflect increasing body mass [73], though there is only a very weak relationship between Mt:F and BM among the members of the family we analyzed (slope = -0.0300; r2 = 0.025). Specialization for fossoriality [74] is another potential explanation. Hegetotheriids show a very slight upward trend overall, but hegetotheriines show a slight decrease in Mt:F (slope = -0.0028) whereas pachyrukhines show a slight increase (slope = 0.0035). The former trend may relate to increasing body mass, though the relationship between Mt:F and BM is only modest in hegetotheriines (slope = -0.109; r2 = 0.384). The slight increase in pachyrukhines may correlate with cursorial specialization given low body mass variation in this clade (S3 Table).

Interatheres show no trend through time in Mt:F, a surprising result considering that the geologically youngest (middle Miocene) interathere in our analysis, Miocochilius anomopodus, has relatively smaller lateral digits than any other notoungulate (Fig 3) and probably further differed in having a subunguligrade stance [75]. Nevertheless, its Mt:F (0.36) is lower than that of some interatheres with less-specialized digits (and presumably a digitigrade stance) such as early Miocene Protypotherium australe (0.41; Fig 3). This demonstrates that lateral digit reduction and distal limb elongation were decoupled in notoungulates. To our knowledge, this phenomenon has not been reported in any other group of mammals, though a similar pattern characterized proterotheriid litopterns; the Mt:F of three-toed Eoauchenia primitiva from the Pliocene is slightly greater than that of Thoatherium minusculum from the early Miocene, a species with lateral digits proportionately smaller than those of modern horses [76, 77].

Fig 3

The pes of representative South American native ungulates.

Taxa include: Thoatherium minusculum (Proterotheriidae), YPPM-VPPU 15719 (reversed; from Scott 1910, pl. XIII, Fig 13 [78]); Diadiaphorus majusculus (Proterotheriidae), AMNH 9196 (modified from Scott 1910, pl. V, Fig 2 [78]); Theosodon lydekkeri (Macraucheniidae), AMNH 9269 (from Scott 1910, pl. XX, Fig 7 [78]); Nesodon imbricatus (Toxodontidae), YPPM-VPPU 15968, (from Scott 1912, pl. XXV, Fig 9 [79]); Scarrittia canquelensis (Leontiniidae), AMNH 29585 (from Chaffee 1952, pl. 11, Fig 2 [80]); Protypotherium australe (Interatheriidae), AMNH 9149 (from Sinclair 1909, pl. V, Fig 1 [81]); Miocochilius anomopodus (Interatheriidae), UCMP 38091 (from Stirton 1953, pl. 27 [75]); Eutypotherium lehmannnitchei (Mesotheriidae), MLP 12–1701 (reversed); Pachyrukhos moyani (Hegetotheriidae), AMNH 9481 (reversed; from Sinclair 1909, pl. X, Fig 15 [81]). Scale bars equal 5 cm in upper row (litopterns, toxodonts) and 3 cm in lower row (typotheres).

Relationship to diet and habitat

Other apparently contradictory evolutionary patterns are revealed when presumed diet and habitat preferences are considered. The SANUs with the most elongate distal limbs, proterotheriid litopterns, were most likely frugivores and/or browsers that lived in forested or partially forested habitats [76, 82–84]. Assuming such dietary interpretations are correct, limb evolution in proterotheriids was not driven by a preference for open habitats, as has been suggested for North American ungulates.

Unlike litopterns, the vast majority of notoungulates in our study are characterized by very hypsodont to hypselodont dentitions [85]. As a result, they are typically reconstructed as open-habitat grazers or mixed feeders [82, 86, 87], though some studies have questioned this interpretation [88-90]. If traditional interpretations of notoungulate paleobiology are correct, it is unclear why they did not develop cursorial specializations to a greater degree, though it may be related to the specific ecological niches they occupied (discussed further below). It is also paradoxical that the sole subunguligrade notoungulate presently known, Miocochilius anomopodus, lived in a moist tropical forest [91, 92] rather than an open, arid habitat.

South American predators

Taken as a whole, there is no evidence that SANUs evolved increasingly cursorial adaptations during the Cenozoic. Although it may be tempting to attribute the lack of such a trend to the unusual predator guild of South America, which was mainly composed of metatherian rather than placental mammals [93-97], Janis and Wilhelm [28] convincingly demonstrated that cursorial predators likely played no role in the early Miocene evolution of cursoriality in North American ungulates. Therefore, the absence of placental predators in South America cannot explain the paucity of cursorial adaptations among its native ungulates. More importantly, pursuit predators actually were present in South America during much of the Cenozoic, but they were flightless birds rather than mammals. These so-called terror birds (phorusrhacids) ranged in height from ca. 60 cm to nearly 3 m [98] and may have been capable of reaching speeds of 65 kph [99], making them large and swift enough to prey on smaller notoungulates and litopterns [100, 101]. If escape from predators were a major factor in the evolution of cursoriality in ungulates, phorusrhacids would have exerted substantial selective pressure on at least small to medium-sized SANUs. The absence of clear cursorial adaptations supports the interpretation that the evolution of mammalian pursuit predators is a consequence rather than a cause of cursoriality in ungulates.

Litopterns

How can our results be reconciled with current ideas about the paleobiology of SANUs and the evolution of cursoriality? A first step lies in distinguishing between litopterns and notoungulates when analyzing evolutionary patterns and processes. Although both taxa (along with other clades) are referred to as ungulates and lumped together as SANUs, these general terms obscure many fundamental differences between them.

Many litopterns are broadly similar to modern ungulates in limb morphology and, to a large extent, craniodental morphology [32, 37]. Unsurprisingly, they also evolved the most cursorial adaptations among SANUs. Even though litoptern Mt:F values are generally lower than those of North American ungulates (Fig 2), this may reflect differences in phylogeny and/or physiology rather than locomotor performance [cf. ref 23]. For example, Janis and Wilhelm [28] noted that the equids in their study tended to have lower Mt:F values than contemporaneous camelid and ruminant artiodactyls and suggested that this could be due to differences in digestive physiology and energy budgets in hindgut-fermenting perissodactyls and foregut-fermenting artiodactyls. The digestive physiologies of litopterns and notoungulates are not known [cf. ref 102], but hindgut fermentation is more likely considering that it is more broadly distributed among mammals [103]. Additionally, molecular data have suggested that SANUs are more closely related to perissodactyls—all of which are hindgut fermenters—than to artiodactyls [57-59]. The relatively low diversity of medium-sized (100s of kg) litopterns and notoungulates compared to small and large ones during the middle to late Cenozoic (S2 and S3 Tables) is similar to that documented for nonruminants in modern East Africa [104] and may constitute circumstantial evidence supporting hindgut fermentation in these groups.

Lower Mt:F values in litopterns could also be related to the prevalence of tridactyly (as opposed to monodactyly) in the group, which itself could be an adaptation for living in closed habitats. Since tridactyly is developed to different degrees in macraucheniids (which have three subequal digits) and proterotheriids (which are functionally monodactyl), it is likely that selective forces differed significantly between the two groups. Nevertheless, there are similarities in other aspects of their limb anatomy. In proterotheriid litopterns, the proximal third phalanx has proportions more similar to those of extinct equids with a “spring foot”—an important adaptation in the evolution of equine unguligrady [105]—than equids that retained three functional digits. The length/width ratio of this element in early Miocene Thoatherium and Diadiaphorus is 3.05 and 2.95, respectively [78]; these values are much higher than those of the early stem equine anchitheres analyzed by O’Sullivan [values ≤ 2.41; 106], indicating relatively longer, thinner proximal phalanges in these proterotheriids. This is also true for the early Miocene macraucheniid Theosodon, whose slightly lower value of 2.58 [78] is still beyond the range of the equids analyzed by O’Sullivan [106]. In contrast to the proximal phalanx, litoptern metatarsals are relatively robust, at least compared to those of artiodactyls. They are classified as broad according to the slenderness index of Morales-García [107], a category otherwise occupied only by the much larger cape buffalo (Syncerus caffer) among the 42 extant and extinct artiodactyls they analyzed.

Scott [78] noted that early Miocene proterotheriid litopterns have shorter third metapodials but longer phalanges compared to equids, and Cifelli and Villarroel [108, p. 286] suggested that this might be an alternative strategy for lengthening the distal limb. To test this proposal, they summed the lengths of the tibia, third metatarsal, and proximal phalanx in three proterotheriids (Thoatherium, Diadiaphorus, and middle Miocene Megadolodus) and compared the values to those from 22 extant artiodactyls and perissodactyls. They concluded that all three proterotheriids had very short hind limbs, comparable to extant pigs and peccaries (Suidae and Tayassuidae) as well as tapirs (Tapirus sp.). We know of no published data to which the entire proterotheriid hind limb might be compared (i.e., data that include the femur and the entire pes), but Clifford [25] analyzed forelimb proportions in 55 extant and extinct artiodactyls in addition to a limited number of perissodactyls (and carnivorans). Estimations of litoptern stylopod (humerus), zeugopod (radius), and autopod (manus) proportions based on Scott [78] result in values of 31.0%, 27.8%, and 41.2%, respectively, for Thoatherium and 31.1%, 30.0%, 38.9% for Diadiaphorus (though the former is based on data from two individuals with no elements in common and thus should be considered tentative). These proportions are most similar to those of tayassuids among extant mammals of similar size, congruent with the conclusion of Cifelli and Villarroel [108] that proterotheriids had suoid-like limbs, but more specific in suggesting greater resemblance to cursorially-adapted tayassuids than suids. Interestingly, Tayassuidae is the same extant group to which Stirton [75] qualitatively compared the limbs of the subunguligrade notoungulate Miocochilius, suggesting a parallel “plateau” of limb evolution in litopterns and notoungulates.

The selective forces favoring the evolution of cursoriality in litopterns (especially proterotheriid litopterns) remain obscure, as cursorial adaptations seem to have evolved in closed habitats rather than open ones. This observation apparently contradicts the hypothesis that cursorial modifications in mammals are an adaptive response to open habitats. It may be that such adaptations tend to be favored by natural selection in ungulate-like mammals of medium to large size regardless of habitat in the absence of countervailing selection (e.g., for semifossorial or semiaquatic habits). Another possibility, not mutually exclusive, is that an evolutionary ratchet [45, 109] operates on the postcranium, with cursorial adaptations being highly favored during certain intervals (e.g., dry or seasonal ones that result in a greater proportion of open habitat) or in mixed habitats where browsers (such as litopterns) must traverse open areas between forest patches. Extant white-lipped and collard peccaries (Dicotyles pecari and Pecari tajacu) are found in a wide variety of both forested and non-forested habitats but are generally restricted to tropical to subtropical regions [110, 111]. In this regard, they may also be reasonable models for proterotheriid litopterns, whose distributions have also been linked to warm climates, though generally ones with humid conditions [112]. The bunodont to bunolophodont dentition of the proterotheriid Megadolodus indicates it may have filled an ecological niche very much like that of a modern tayassuid or suid, whereas the lophodont dentitions of other proterotheriids suggest dietary habits closer to those of more folivorous ungulates [32, 108].

Notoungulates

Most notoungulates, in contrast to litopterns, bear little resemblance to modern ungulates; their postcranial and craniodental morphologies are reminiscent of rodents, rabbits, carnivorans, and even some marsupials [32, 36, 113–116]. This lack of correspondence may explain the absence of cursorial adaptations such as high Mt:F values and fewer functional digits, even though notoungulates may have been living in open habitats. Most notoungulates had an unspecialized ankle joint, consistent with plantigrady or digitigrady (as opposed to unguligrady) and lacked adaptations for locking limb joints like those present in extant ungulates and litopterns [117].

Interatheriid and hegetotheriid typotheres may have been ‘functionally cursorial’ despite lacking classic cursorial adaptations, as is true of many extant small mammals [54], many of which use a leaping gallop [rather than the horse gallop; 118]. Small size may also explain the scarcity of unguligrady among typotheres, as this feature may be most advantageous in large mammals [56, 119]. Small heteropod footprints (four forefoot digits and three hind foot digits) from the Miocene of La Rioja, Argentina have been considered as probably belonging to a hegetotheriid [120]. The prints are quite similar to those of modern caviids, suggesting a plantigrade to digitigrade gait. Larger hegetotheriids and interatheriids (i.e., ~10 kg; S3 Table) have Mt:F values closer to those of similarly-sized felid and mustelid carnivorans than canids [S1 Table; 28], indicating relatively short distal limbs even compared to other digitigrade and plantigrade mammals. In terms of limb proportions, suoids may also be reasonable postcranial analogs for typotheres of this size [cf. ref 75]. Smaller hegetotheriids have been compared to rodents and interpreted as cursorial with fossorial capabilities [121].

For mesotheriid notoungulates, selective pressures related to a fossorial or semifossorial lifestyle [74, 122] were likely greater than those favoring efficient or rapid locomotion in open habitats. Larger Neogene typotheres such as mesotheres may have had limb proportions similar to some extinct North American oreodonts (Merycoidodontidae), which were not included in the analysis of Janis and Wilhelm [28] due to their digitigrade stance. Based on published measurements [123], merycoidodontids had Mt:F values of 0.32–0.44, similar to those of many typotheres analyzed here.

The lack of cursorial adaptations in some toxodonts is unsurprising based on other lines of evidence. Homalodotheriids and leontiniids have both been reconstructed as browsers, and neither persisted past the Miocene [62, 80, 124, 125]. Toxodontids had ever-growing dentitions and have been reconstructed as open-habitat grazers [82, 86, 126], though some may have been browsers [90] or have had broad dietary preferences [88]. Toxodontids were tridactyl by the late early Miocene (Fig 3), but their limb proportions did not change appreciably during the remainder of the Cenozoic, at least among larger species (Fig 2). In fact, the geologically youngest species, Toxodon platensis, is notably short-legged [127]. If toxodontids occupied ecological niches like large, non-cursorial modern ungulates such as hippos and/or rhinos, it may explain their unusual, robust postcranial morphology and lack of pronounced cursorial adaptations. Testing this hypothesis requires detailed morphofunctional studies of the toxodontid postcranium, which are presently lacking.

Conclusion

The rich Cenozoic fossil record of the Northern Hemisphere has provided the foundation for much of our understanding about how mammals have responded to changing climates and habitats over the past 66 million years. Nevertheless, such studies are largely based on mammal clades that flourished in the late Cenozoic and dominate modern ecosystems. The fossil record of South America holds great potential for studies of based on different groups of mammals that lack modern representatives or have relatively modest extant diversity. As such, it provides an opportunity to evaluate the degree to which evolutionary hypotheses based on Northern Hemisphere mammals depend on the phylogenetic affinities of the groups being analyzed. This study and others have demonstrated that some South American mammal groups followed very different evolutionary trajectories than their Northern Hemisphere counterparts, resulting in mammal communities that differed in many respects from those of modern South America as well as the Northern Hemisphere, both past and present [85, 91, 97]. Going forward, it is essential to incorporate data from South America and other Southern Hemisphere continents to achieve an integrated understanding of how modern mammal faunas developed and how they might change in the future.

Line drawing of a notoungulate right astragalus illustrating the dorsal astragalar foramen.

The specimen (MLP 75-II-1-9) is from Loma Verde, Argentina, and pertains to a large isotemnid, perhaps Thomashuxleya. (A) dorsal view, anterior toward bottom; (B) posterior view, with inferior surface toward top of page. Abbreviations: af: Superior astragalar foramen; at: Astragalar trochlea; fs: Flexor sulcus.

(PDF)

Click here for additional data file.

Femur and third metatarsal (MT3) data for South American native ungulates analyzed in this study.

Species are grouped by family within one of three larger clades: Litopterns (Litop), toxodont notoungulates (N:Tox), and typothere notoungulates (N:Typ). Institutional abbreviations: ACM, Beneski Museum of Natural History, Amherst College, USA; AMNH FM, fossil mammal collection, American Museum of Natural History, New York, USA; FMNH PM, Fossil Mammals collection, The Field Museum, Chicago, USA; IGM, Instituto de Geociencias y Minería, Bogotá, Colombia; MACN, Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina; MCNAM-PV, vertebrate paleontology collections, Museo de Ciencias Naturales y Antropológicas “J. C. Moyano”, Mendoza, Argentina; MLP, Museo de La Plata, Argentina; MNHN-BOL-V, vertebrate paleontology collections, Museo Nacional de Historia Natural, La Paz, Bolivia; UATF-V, vertebrate paleontology collections, Universidad Autónoma “Tomás Frías”, Potosí, Bolivia; UCMP, University of California Museum of Paleontology, Berkeley, USA; UF, Florida Museum of Natural History, University of Florida, Gainesville, USA; YPM VPPU, Princeton University Collection, Yale Peabody Museum, New Haven, USA.

(PDF)

Click here for additional data file.

Approximate age of South American native ungulates analyzed in this study.

Ages (to nearest million years, my) are based on the provenance of the specimen(s) from which femur and metatarsal data were collected. The age listed is the value used for calculating regression equations and represents an approximate midpoint of the age for the formation or locality. Species are grouped by family within one of three larger clades: Litopterns (Litop), toxodont notoungulates (N:Tox), and typothere notoungulates (N:Typ).

(PDF)

Click here for additional data file.

Estimated body mass (BM) of South American native ungulates analyzed in this study.

Species are grouped by family within one of three larger clades: Litopterns (Litop), toxodont notoungulates (N:Tox), and typothere notoungulates (N:Typ).

(PDF)

Click here for additional data file.

Graph of metatarsal-femur ratio (Mt:F) in Eocene through Pleistocene North American ungulates and large carnivores (> 7 kg).

Note the independent increase in Mt:F in several families of ungulates from the Eocene to the early Miocene. From Janis and Wilhelm [17].

(PDF)

Click here for additional data file.

29 Jun 2021

PONE-D-21-18278

No evidence for parallel evolution of cursorial limb adaptations among Neogene South American native ungulates (SANUs)

PLOS ONE

Dear Dr. Croft,

Dear Darin,

Two reviewers have accepted to comment on your manuscript. Both of them are very positive. They clearly appreciated the manuscript and regard it as an important contribution. They made several comments and suggestions to help improve it.

In my opinion, most of the issues are minor and do not require much change. The most important issues might be:

1. extending a little the discussion about the diversity of limbs in ungulates and other mammals such as cursorial carnivores and hopper mammals (reviewer 1).

2. the comment about the increase in the Mt:F values from the Oligocene to the Miocene North American ungulates that is not so evident based on Fig. 1 (reviewer 2).

Reviewer 1, Christine Janis, revealed her identity. So, please consider mentioning her name in the acknowledgements if you feel comfortable with this possibility.

Please submit your revised manuscript by Aug 13 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Thierry

Thierry Smith, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is a well-written and nicely composed manuscript that furthers our knowledge about mammalian anatomy and adaptations over the course of the Cenozoic. The approach is traditional, and there is no attempt at complex statistical analysis – rather the data are portrayed as a figure where the morphological trends over time in the different lineages can clearly be seen. This approach is entirely appropriate for the type of data obtained and the questions being addressed in this MS, and I do hope that the other reviewer(s) don’t insist on some sort of phylogenetic multivariate analysis which would serve to confabulate, rather than illuminate, the issues here.

The authors address the issue of the commonly used “index of cursoriality” of metatarsal to femur (Mt/F) ratio employed for some modern ungulate groups (primarily equids and ruminant + camelid artiodactyls) and whether it applies to the South American native ungulates (SANUs). These animals certainly appear to be short-legged in comparison with the modern groups, even if they parallel them in some other regards (e.g., loss of digits), and I’m delighted that the authors have actually done this study (for which obtaining data from South American museums was key). They document not only the relatively shorter legs of the SANUs, but also that there is no overall trend for increasing the ratio over time (again in contrast with the modern ungulate lineages, where in North America a ‘step-up’ increase has been documented, correlated with an opening of the habitat in the Oligocene). The conclusions notes that trends noted in modern ungulate lineages should not necessarily be assumed to apply to other groups, and that data from southern continents should be added to our studies in order to better understand evolutionary trends and trajectories.

The figures are clear, and the SI provides comprehensive information about the specimens studied and the original measurements.

I don’t really have any major issues with this MS, which I think sets out what it plans to do very well, and is a good contribution to our knowledge. However, I do think that it could benefit from a little more consideration about the issue of what the Mt/F ratio actually means, and its distribution among extant ungulates.

Basically, a very high ratio (> 0.7) appears to be primarily a feature of equids (at least anchitheres and equines), and ruminant and camelid artiodactyls. I’d like to say that all of these taxa are unguligrade, but -------: secondarily digitigrade camelids don’t count, but most anchitheres have metatarsals as long (or longer than) modern Equus, and they supposedly were sub-unguligrade with a pad foot. Then you have suoids, which are unguligrade but have a low MT/F ratio, so maybe part of this relates to lateral digit reduction (although again here anchitheres upset the apple cart!). (A useful paper to consider here might be Clifford, JVP 2010, although unfortunately it only considers artiodactyls.) But lots of other ungulates (and paenungulates) from diversity of habitats (ceratomorphs, proboscideans, hyraxes) don’t have/never had (as far as we know for hyraxes) long metatarsals. So, a little more discussion about the diversity of today’s ungulates and their limb anatomy could be a good context for why one might expect, or not expect, SANUs to follow the so-called ‘established pattern’. Somewhere here you could also note that more cursorial carnivores also tend to have a higher ratio, even if they never get to be as high as the ungulates.

Another thing to think about a bit more is what a high Mt/F ratio actually means in terms of function and biomechanics. You mention that it relates to the lengthening the distal limb relative to the proximal limb, but perhaps you could be a bit more explicit about relative femur length being relatively constant in ungulates and the “distal limb” usually meaning from the tibia on down. The phalanges are not usually lengthened (but see comment below). Longer limbs do give you longer strides, but in theory lengthening the femur could give you a longer limb (which is what humans and elephants both do). So what is the biomechanics of this anatomy, and why do things that way? This is probably getting into much more than is relevant to this MS, but there’s some stuff that Hildebrand has to say about this that could be added here.

What a longer limb does is give you a longer stride, which isn’t necessarily about speed (see comments below about correlation with habitat, etc.). There may be other reasons to what long limbs or a longer stride, or to *not* want them. If you’re a mediportal mammal, without an elastic energy spring ligament system of recovery in your limbs, it may not benefit you much at all, and then if you’re a hopper you lengthen the tibia more than the metatarsals. All too much for this particular MS, but a little bit of additional discussion could be useful!

I also note that some litopterns, especially Thoatherium, have extremely long proximal phalanges, much more so than unguligrade ungulates (see brief discussion of this in Janis & Bernor, Frontiers in Ecology and Evolution, 2019). Is this somehow compensating for the shorter metapodials? What if one did a calculation of MT+PP/F ratio for all ungulates? (I’m not suggesting that you do that, but maybe mention that they appear to be doing something with their limb morphology quite unlike the equids they’re so often compared to.)

Some small issues:

p. 4: it would be nice to have a brief explanation of why the loss of the superior astragalar foramen was thought to be correlated with “cursoriality”. Also, maybe rephrase a little (especially the final sentence in the paragraph): the correlation of anatomy with habitat is a secondary inference, the primary inference is the correlation of the anatomy with a particular type of locomotor performance (e.g., longer legs enable longer strides, but correlation with open habitat [or even speed] is a secondary inference).

p. 4. Following on from the above, maybe mention that longer strides are actually correlated with locomotor efficiency at all gaits, although they may also enable greater speed. Also, Mt/F ratio declines over a more limited range of sizes than ‘several orders of magnitude’: as can be seen from Table II in Janis & Wilhelm 1993, a fox has a greater Mt/F ratio than a wolf, and a cat a greater one than a lion. (See also Fig 1 in Lovegrove & Mowoe, 2014).

p. 5. Materials and methods. How were these measurements obtained? From the real specimens, by calipers? From the photos, from ImageJ? Also, to the uninitiated it may not be clear here what you mean by ‘endemic ungulate’ as previously you’ve been calling them SANUs.

p. 7. Although Mt/F ratios have only been documented for North American ungulates, note that similar ratios are seen in Old World ruminants and equids.

p. 14. Line 3, should the reference be #35 rather than #6

p. 14. Note that the oreodonts were digitigrade, which was part of the reason that they were not included. Tayassuids and ceratomorphs were also not included, for similar reasons.

p. 14. 7 lines from the bottom ---- the only typo I’ve spotted so far! Should be “broad dietary preferences” (not “board”). Another minor error is in the SI: a couple of times you’ve put “Toxodon sp.” (sp. in italics) rather than “Toxodon sp.” (sp. in Roman type).

Fig. 1. At the expense of overloading the reader with information, perhaps there should be a little more information about who’s who in the North American ungulates (you mention on p. 12 that J&W note lower Mt/F ratios in the equids). I suggest identifying somehow the equids, the protoceratids + traguloids, and then the camelids + pecorans.

S3 Table. You say that body mass estimates are based on femur length. What is the reference group you are using to relate femur length to body mass. Do you have a reference for that?

Reviewer #2: Comments to the author

The spread of grasslands and the evolutionary response of ungulate mammals to the changing environments from closed to open landscapes is well-established in North America. Grasslands spread during the Neogene, and as a response ungulate mammals evolved teeth high higher crowns and longer limbs. However, it is unknown if the same evolutionary trends can be observed in other continents. South America is an excellent study case to evaluate if different evolutionary lineages of herbivorous mammals show similar adaptations to open environments. South America was isolated from other continents during most of the Cenozoic and a large diversity of endemic mammals evolved in relative isolation, including native ungulates (SANUs). This work is the first study to evaluate if SANUs showed the same pattern of limb elongation as North American ungulates in response to the spread of open landscapes. The finding that SANUs do not show a trend of limb elongation in response to the spread of open landscapes is somehow surprising although not completely unexpected, as SANUs have shown different patterns in the evolution of ecomorphological traits in comparison with North American ungulates (for example the early evolution of high-crown teeth in SANUs).

I consider that the manuscript is a valuable contribution to mammal palaeobiology, as it represents the first work to study patterns of limb elongation in South American ungulates across a broad geographical and temporal scales, providing the basis of comparison with evolutionary patterns seen in other continents. Therefore, I consider the work is suitable for publication, although I have several comments that I provide below and should be addressed first.

General comments

The authors discuss how the Mt:F values could be related with body size in some groups of SANUs. I think given these observations it would be appropriate to examine in more detail the possible allometric relationships between the Mt:F and body size.

Does any allometric pattern emerge for Notoungulates and/or Litopterns in a plot of log (Mt:F) vs log (body size)?

Title

I suggest avoiding the use of the acronym SANUs in the title.

Abstract

“None of the groups or subgroups we analyzed show a pronounced increase in Mt:F across this interval”

It would be helpful for the readers to clarify in the abstract if an increase in Mt:F is associated with limb elongation.

Introduction

• “Temperatures fluctuated but were relatively steady during the Oligocene and early Miocene and reached a peak around the early-middle Miocene, 18-16 million years ago”

Please add a reference for the dates of the Miocene Climatic Optimum. Notice that a recent review paper (Steinthorsdottir et al. 2021) on the Miocene climate uses dates of ca. 16.9 – 14.7 Ma for the climatic optimum

Steinthorsdottir, M., et al. (2021). The Miocene: The future of the past. Paleoceanography and Paleoclimatology, 36, e2020PA004037.

• “plant phytolith ecomorphology indicates that arid conditions and open habitats

were present in southern South America as early as the middle Eocene [23, 24].”

I consider is important to clarify (as discussed in the references cited by the authors) that the open habitats in southern South America during the Eocene were shrublands with abundant palms, a flora that was nonanalog to modern open savannas for example. Therefore, even though open habitats might have appeared in southern South America since the Eocene, the flora of these habitats was different from the open habitats that appeared in North America later during the Neogene.

• “Despite longstanding interest in the so-called ‘precocious’ hypsodonty of South American notoungulates”

Here it would be relevant to cite Madden (2015), which in chapters 1 and 2 provides a detailed account on the issue of precocious hypsodonty in South American native herbivores, for the interested reader that may wish to know more about the ‘precocious’ hypsodonty in South America.

Madden (2015). Hypsodonty in Mammals. Evolution, Geomorphology, and the Role of Earth Surface Processes. Cambridge University Press.

Materials and methods

• Please consider adding a figure of a cladogram showing the hypothesis of phylogenetic relationships of the genera of South American native ungulates sampled, showing the subclades that were analysed separately.

• Please clarify if the least square regressions are Ordinary least square (OLS) or generalized least squares (GLS). Given that closely related species might tend to have more similar Mt:F values, the residuals might be correlated and in that case it would be preferable to use GLS over OLS.

Discussion

- “As noted previously, North American ungulates show a marked increase in Mt:F

around the Oligocene-Miocene transition [14].”

Please see the comment below for Fig. 1. Although the trend of increase in Mt:F is reported in the literature, this trend is not clearly visible in the Fig. 1 as shown by the authors. Would a regression (as done for South American taxa) show this trend?

• “The taxa studied derive from a variety of habitats, nearly all of which have been reconstructed as at least partially open and suitable for more cursorial species.”

Please see the comments for Fig. 1, I think the information of habitat (open vs closed) could be indicated by the shape of the points, while the colours correspond to the subclades.

• In addition to the relation of the femur and metatarsal III, the gear ratio of the calcaneus has been also used as a proxy of locomotor habit and posture in carnivorans (e.g. Polly 2010; Polly and Head, 2015). Although the gear ratio has not been studied in detail in ungulates, it could be a relevant metric to explore in future studies. It would be interesting to assess if the gear ratio shows similar patterns than the Mt:F in ungulates from North America and the SANUs.

Polly. 2010. Tiptoeing through the trophics: Geographic variation in carnivoran locomotor ecomorphology in relation to environment. In A. Goswami and A. Friscia (eds.), Carnivoran Evolution: New Views on Phylogeny, Form, and Function

Polly and Head. 2015. MEASURING EARTH-LIFE TRANSITIONS: ECOMETRIC ANALYSIS OF FUNCTIONAL TRAITS FROM LATE CENOZOIC VERTEBRATES. In: Earth-Life Transitions: Paleobiology in the Context of Earth System Evolution. The Paleontological Society Papers, Volume 21,

Figure 1.

The data of Mt:F for North American ungulates showed in Fig. 1 clearly shows that these taxa had a higher Mt:F than South American native ungulates. However, it is not evident that there was an increase in the Mt:F values from the Oligocene to the Miocene. Would a regression line show this trend? (It does not seem it would). Is it necessary to add data from Paleogene North American ungulates to show clearly an increase in Mt:F since the Miocene? Please clarify.

Since the colours of the points and lines differentiate the clades of SANUs (namely litopterns, toxodont notoungulates and typothere notoungulates), the information of whether a taxon has been hypothesized to have lived in a closed or open environment could be indicated by the shape of the points, for example squares and circles).

Figure 2.

Please add a scale to the figure.