Abstract
In this study, we aimed to achieve three objectives: (1) to precisely characterize the body plans of Elephantidae and other large herbivorous mammals; (2) based on this analysis, to determine whether the body plans of the extinct woolly mammoth (Mammuthus primigenius) and steppe mammoth (M. trogontherii) differ from those of modern‐day Elephantidae: the Asian elephant (Elephas maximus), the African bush (Loxodonta africana), and forest (L. cyclotis) elephants; (3) to analyze how the body plans have changed in extant perissodactyls and proboscideans compared with their Paleogene ancestors. To accomplish this, we studied mammoth skeletons from the collections of Russian museums and compared this data with a large number of skeletons of extant elephantids, odd‐toed, and even‐toed ungulates, as well as their extinct relatives. We showed that three genera of Elephantidae are characterized by a homogeneous body plan, which is markedly different from other large herbivores. Elephantids break the interrelationship, that exists in artiodactyls and perissodactyls, between the total length of the head and neck on one side and the limb's segments on the other. Their limbs are very tall (inferior in this regard among large ungulates only to the giraffe), and, contrary to the other large herbivorous mammals, elongated due to the length of the proximal segments. This allows them to effectively utilize the principle of inverted pendulum (straight‐legged walking) in locomotion. The biggest differences in the body plan of mammoths compared with extant elephants are a markedly larger pelvis, elongated fore‐ and hindlimbs (due to the increased relative length of their proximal segments), and different proportions of the skull. The body plans of plesiomorphic Paleogene proboscideans and perissodactyls differed markedly from their descendants in every body part; these differences are related, on the one hand, to the allometric growth, and on the other hand, to the advancement of the locomotor apparatus in the course of their evolution. The most notable difference in the body plan between Paleogene proboscidean Moeritherium and extant Elephantidae is the ~2‐fold increase in relative limb height.
Keywords: Artiodactyla, Late Pleistocene, locomotion, Mammuthus, Mammuthus, Perissodactyla, Proboscidea, Yakutia
In this study, we examine the differences in the body proportions of Elephantidae and other large herbivorous mammals, identify similarities and unique features, and associate these features with the benefits and constraints imposed by their body plan. We also analyze how body plans have changed in extant perissodactyls and proboscideans compared to their Paleogene ancestors and offer biomechanical explanations for these reinforcements.
1. INTRODUCTION
Elephants have been a subject of intense interest for researchers in anatomy, morphology, and biomechanics for a long time. These huge animals have a unique combination of robustness and low mobility in their intervertebral and limb joints, as well as the ability to move fast and overcome obstacles. Elephants can swim and ascend mountains, but they cannot jump at all, and even during the fastest locomotion they do not have aerial stages leaving at least one foot on the ground to support their huge body. Despite the inability to run, elephants can reach surprisingly high speeds during locomotion, with the highest speed recorded in experiments with Indian elephants being 25 km/h (Hutchinson et al., 2003). The linear sizes and body mass of various extinct proboscideans have received extensive attention (Christiansen, 2004; Huang et al., 2023; Larramendi, 2016; Larramendi et al., 2017). Elephant myology has long attracted the attention of researchers (Drujinin, 1941a, 1941b; Сuviеr & Laurillard, 1850); the musculature of the fore‐ and hindlimbs of the African and Asian elephants was most extensively examined in a large‐scale study by Gambaryan (1974; Gambaryan & Ruhkyan, 1974), and then the subject was rediscovered in a recent study of the antebrachium and manus muscles and fascial elements of the African elephant (Nagel et al., 2018). Research of elephant locomotion has a long history, covering the study of gaits, kinematics of limb joint movement, speed, mechanical stresses, the morphology of digital cushion, and pressure during support. Muybridge (1887), Gregory (1912), Howell (1944), Gambaryan (1974), Gambaryan and Ruhkyan 1974), Alexander et al. (1979), Hutchinson et al. (2003, 2006, 2011), Ren et al. (2008), Genin et al. (2010), Panagiotopoulou et al. (2012, 2016), and Warner et al. (2013, 2014) have all contributed to this field.
Studies on the anatomy of the woolly mammoth (Mammuthus primigenius) have taken more than two centuries (Tilesio, 1815). Different authors have studied the body size of various woolly mammoth individuals (Averianov et al., 1992; Boeskorov, 2021; Boeskorov et al., 2009; Garutt, 1964; Vereshchagin & Tikhonov, 1990), with individual parameters such as height at the shoulders and body length being considered. Several studies have attempted to calculate the body mass of an adult mammoth (Boeskorov et al., 2007; Larramendi, 2016). According to Larramendi (2016), Siberian M. primigenius was close in height at the shoulders, and apparently, in body mass to the Indian elephant Elephas maximus. Boeskorov (2021) analyzed new finds of skeletons and mummy carcasses of woolly mammoths from the north of Eastern Siberia, noting, that mammoths, on average, have longer body lengths than Asian elephants.
In this study, we aim to explore the anatomy of elephants and other large herbivorous mammals from a different perspective. Our goal is to examine the differences in their body proportions, identify similarities and unique features of body ratios, and associate these features with the benefits and constraints imposed by their body plan. Our research has several objectives. First, we will describe the body plan of extant elephantids, perissodactyls, and artiodactyls; and characterize in detail how much the former differs from the latter. Second, we will compare the proportions of woolly and steppe (M. trogontherii) mammoths' skeletal elements with those of modern‐day elephants, and answer the question: How different is the body plan between the mammoths and extant elephants? Finally, using proboscideans and perissodactyls as examples, we will illustrate how the body plans of these animals have changed compared with their Paleogene ancestors and offer biomechanical explanations for these reinforcements.
2. MATERIAL
In this study, we utilized osteological material from museum collections, namely, the Azov Museum of History, Archaeology and Paleontology (AMZ), Azov, Russia; the P. A. Lazarev's Mammoth Museum, Yakutsk, Russia; the Royal Museum for Central Africa (RMCA), Tervuren, Belgium; Zoological Institute of the Russian Academy of Sciences (ZIN), Saint Petersburg, Russia; and the Zoological Museum of the Lomonosov Moscow State University (ZMMU), Moscow, Russia. The studied material includes five complete skeletons of the woolly mammoth. These are the skeletons of the following individuals.
Adams mammoth. The skeleton with mummified skin and muscles was found near the Lena River delta on the Bykovsky Peninsula in 1799 (Figure 1). The age of the find, according to radiocarbon dating, is 34,450 ± 2500 and 35,800 ± 1200 years BP, the time of the Karginian Interstadial of the Late Pleistocene (Heintz & Garutt, 1964). It was a very old male, about 65–70 years old (last generation of teeth, M3/m3; Garutt, 1964; Vereshchagin & Tikhonov, 1990).
FIGURE 1.
Studied skeletons of the woolly mammoth (Mammuthus primigenius) individuals in the left lateral view. Photographs of Taimyr and Chekurovsky mammoths were mirrored to match photographs of Berezovsky and Adams mammoths.
Taimyr mammoth. The skeleton of an adult mammoth with remains of muscles, skin, and fur preserved on it was discovered in 1948 on the Mamontovaya River (a tributary of the Schrenck River) on the Taimyr Peninsula (Figure 1). The age of the find, according to radiocarbon dating, is 11,450 ± 250 yBP, the time of the Sartanian Glaciation final, Late Pleistocene (Heintz & Garutt, 1964). The skeleton belongs to an old male, about 50 years old (last generation of teeth, M3/m3; Garutt, 1964; Vereshchagin & Tikhonov, 1990).
Berezovsky (Berezovka) mammoth. The carcass of this mammoth was found on the bank of the Berezovka River, a right tributary of the Kolyma River, in August 1900 (Figure 1). Later radiocarbon analysis showed two different dates of the Berezovsky mammoth: 31,750 ± 2500 and 44,000 ± 3500 yBP, time of Karginian Interstadial of the Late Pleistocene (Heintz & Garutt, 1964). It was an adult male mammoth of middle age, about 40 years old (changing between the penultimate and the ultimate generation of teeth: M2/M3, m2/m3; Garutt, 1964; Vereshchagin & Tikhonov, 1990).
Allaikhovsky mammoth. An almost complete skeleton of an adult male mammoth (M3) was found in 1973 at Achchygyi Allaikha River (Averianov et al., 1992; Lazarev, 2008). Achchagyi‐Allaikha mammoth bones mass accumulation where the Allaikhovsky mammoth skeleton was found, was radiocarbon dated 12,490 ± 80 to 12,400 ± 60 yBP, time of the Boelling warming during the final of Sartanian Glaciation of the Late Pleistocene (Nikolskiy et al., 2010).
Chekurovsky mammoth. The skeleton of this mammoth with some remains of soft tissues was found in 1960 in the lower reaches of the Lena River near the village of Chekurovka (Figure 1). The skeleton belongs to an adult female, about 35 years old (changing between the penultimate and the ultimate generation of teeth: M2/M3, m2/m3). Radiocarbon dating of this skeleton showed the age of 22,760 ± 80, 22,940 ± 100, and 23,220 ± 110 yBP, time of Sartanian Glaciation of the Late Pleistocene (Boeskorov et al., 2009).
For comparison, we have studied other representatives of the Elephantidae: five specimens of the Asian elephant (Elephas maximus), five specimens of the African bush elephant (Loxodonta africana), and one specimen of the African forest elephant (L. cyclotis). The body plan was also studied in 12 species (25 specimens) of perissodactyls and 15 species (22 specimens) of artiodactyls (Table S1). The juvenile specimens were excluded from the analysis, as they display distinct differences in body plan compared with adults. For comparison, we used the previously published datasets of the vertebral column properties in artiodactyls, perissodactyls, and carnivorans (Belyaev et al., 2021, 2023, 2024). Finally, to compare the body plan in extant and extinct large herbivores, we used additional reference data on proboscideans Moeritherium, Numidotherium koholense, Mammut americanum (Buesching mastodon ISM 71.3.261; Owosso mastodon UMMP VP 23498), Megabelodon lulli (DMNH 1261), Mammuthus trogontherii (AMZ KP‐28689), perissodactyls Arenahippus grangeri (UM 115547), Hyrachyus eximius, and the embrithopod Arsinoitherium zitteli (Andrews, 1906; Baygusheva et al., 2012; Cook, 1928; Cope, 1873a; Court, 1994; Osborn, 1936; Wood et al., 2011).
3. METHOD
3.1. Measurements
To determine the differences in body plan of studied large herbivores, we measured all body parts of these animals. Our measurement system was inspired by the ideas of Petr Gambaryan (1964, 1974) and is also similar in limb measurements to the one used by Larramendi (2016). In the case of all mammoth skeletons and some other elephantids (and other ungulates) studied, we had to take measurements from mounted skeletons. As a consequence, in the case of some measurements, we had to replace the optimal ones with those that were possible to take while working in constrained conditions. The utilized measurement system is presented in Figure 2.
FIGURE 2.
Scheme of the measurements used. Left lateral view of the (a) vertebral column; (b) forelimb; (c) hindlimb; (d) ventral view of the pelvis; (e) left lateral and (f) frontal view of the skull; (g) rib; (h) frontal view of the mandible of the Elephas maximus (ZMMU S‐182211). Textual explanations of the measurements are given in Section 3.1.
Measurements of the skull and mandible (1–3). Skull measurements are the most complicated when working with mounted elephantid skeletons. Consequently, we were forced to replace one of the optimal measurements (the greatest width of the skull) with two more easily made measurements – the greatest width of the skull in the eye socket area and the greatest width of the mandible. To partially compensate for this shortcoming, we used published data for mammoths (Averianov, 1996; Averianov et al., 1992; Tikhonov, 1992) and measured the maximum occipital width in some elephantids studied.
Skull length (1). The greatest distance between the occipital condyle (most convex point) and the most rostral point of the premaxillary bones.
Skull width (2). The greatest width of the skull is in the eye socket area.
Mandible width (3). The greatest width of the mandible. Measured at the widest point of the mandible; for example, between condyles or between the widest part of the left and right ramus.
Dimensions of the vertebral regions (4–7) were measured on the articulated vertebral columns, along the ventral sagittal line (along the curvature of the backbone in case of mounted skeletons; Figure 2a). The vertebrae were tightly articulated before measurement. If the mounted skeletons studied had large gaps between vertebrae, we reduced this space from the “osteological length.” These “osteological lengths” underestimate the actual length by the total length of the intervertebral disks in the respective region.
Cervical region (4). The length of the cervical region was measured from the anterior edge of C1 (atlas) to the posterior edge of C7.
Thoracic region (5). The length of the thoracic region was measured from the anterior edge of T1 to the posterior edge of the last thoracic vertebra.
Lumbar region (6). The length of the lumbar region was measured from the anterior edge of L1 to the posterior edge of the last lumbar vertebra.
Sacral region (7). The length of the sacral region was measured from the anterior edge of S1 to the posterior edge of the last sacral vertebra.
Body width (8–10) was assessed through the following measurements:
Longest rib length (8). The length of the rib was measured using a measuring tape along its curvature on the outer side.
Pelvis width (9). Distance between the most distant points of the right and left wings of the ilium.
Pelvis minor width (10). Distance between the closest points of the right and left acetabulum.
Measurements of fore‐ and hindlimb segments (11–19). Working with limb segments, we measured the length of the segments between the articular surfaces of the bones instead of the greatest length of the bones in general (Figure 2b,c). Based on this, the sum of the lengths of the limb segments can legitimately be considered as the height of the fore‐ and hindlimb.
Scapula length (11). The distance from the upper point of the cranial border of the scapula to the most concave point of the glenoid socket.
Scapula width (12). Distance between the most distant points of the cranial and caudal edges of the scapula.
Humerus (13). The distance from the most proximal (most convex) point of the head of the humerus to the most distant from it (most convex) point of the trochlea.
Antebrachium (14). The smallest distance between the proximal and distal articular surface (between concave points) of the larger antebrachium bone. In the case of elephantids, the measurement was made on the ulna and in the case of the odd‐toed and even‐toed ungulates on the radius.
Manus (15). Distance from the most proximal (most convex) point of the upper row of carpals to the distal point of the last phalanx. On disarticulated skeletons, the measurement was made along the straightened segments, and on mounted skeletons along the anterior surface of the manus bones using a measuring tape.
Ilium (16). The shortest distance between the upper point of the tuber sacrale and the upper point of the acetabulum. Unlike other limb segment measurements, the length of the ilium segment does not lie in the parasagittal plane.
Femur (17). The distance from the most proximal (most convex) point of the head of the femur to the most distant from it (most convex) point of the medial condyle.
Crus (18). The smallest distance between the medial condyle articular surface and distal articular surface (between concave points) of the tibia.
Pes (19). Distance from the most proximal (most convex) point of the astragalus to the distal point of the last phalanx. On disarticulated skeletons, the measurement was made along the straightened segments, and on mounted skeletons along the anterior surface of the pes bones using a measuring tape.
3.2. Relative size
To switch from the study and comparison of linear dimensions of the body parts to the study and comparison of body ratios, we expressed all studied variables in relation to the trunk length. Trunk length is estimated as the sum of thoracic, lumbar, and sacral lengths. Even though the number of trunk vertebrae can vary considerably among large herbivorous mammals (elephantids and perissodactyls have +5–6 trunk vertebrae compared with the artiodactyls) and the anteroposterior length of vertebral centra can be elongated or shortened, we believe that trunk length, in general, is an appropriate and consistent parameter. Our opinion is based on the fact that, regardless of the vertebral formula and characteristics of vertebral centra, the trunk characterizes the part of the body in which the majority of internal organs of the organism are located and it can be viewed as a structural basis of the body.
3.3. Data analysis
The data analysis was performed using IBM SPSS Statistics 23. Before conducting the analysis, the variables were tested for normality using the Kolmogorov–Smirnov (K–S) test. If the distribution was found to be normal, parametric statistics were employed for further analysis. Student's t‐test (independent samples) was used to compare different variables in elephantids vs. ungulates. On the other hand, if the K–S test indicated a non‐normal distribution, non‐parametric statistics were employed. The Mann–Whitney U test was used to compare different variables in elephantids vs. ungulates.
For the principal component analysis (PCA), R v4.3.1 (R Core Team, 2023) was used. We used PCA to visualize differences in body plans between various large herbivorous mammals. Initially, we performed PCA (covariance matrix) using the absolute values of the skeletal measurements, which resulted in only one PC that explained ~90% of the variance, primarily related to size. Therefore, we performed PCA twice using the 19 relative variables described in Sections 3.1–3.2 for the entire sample (1) and specifically for the elephantids studied (2). In the whole sample (1), the first principal component (PC1) explained 58.5% of the variance and was mostly related to a longer humerus and femur, wider scapula and pelvis, shorter cervical region, and shorter manus and pes. The second principal component (PC2) explained 25.8% of the variance and was mostly related to the longer cervical region, antebrachium, manus, and pes. In the elephantid‐only sample (2), the PC1 explained 61.9% of the variance and was mostly related to the wider pelvis, and longer and wider skull. The PC2 explained 14.6% of the variance and was mostly related to the longer ribs, longer humerus and femur, shorter lumbar region, and narrower pelvis. PCA loadings between PC1/PC2 and 19 variables are presented in Tables S2.1 and S2.2.
4. RESULTS
In general, the body plan of Elephantidae is unique and has no close analogs among modern‐day large ungulates (Figure 3a). It remains rather homogeneous in representatives of three studied genera (Elephas, Loxodonta, and Mammuthus) and is characterized by rather minor variations in the proportions of different body parts (Figure 3b).
FIGURE 3.
PCA biplots. Nineteen body parts ratios in a studied sample of large herbivorous mammals (a); the same set of variables in a sample of Elephantidae studied (b). The circle colors indicate the taxonomic group: Equidae (orange), Tapiridae (yellow), Rhinocerotidae (lilac), Hippopotamidae (green), Bovinae (pale blue), Ovibos moschatus (white/pale blue), Camelus bactrianus (grey), Giraffidae (white), Hyracoidea (black), Elephas maximus (red), Loxodonta (dark red), Mammuthus primigenius (pink). Note that data on the Paleogene proboscidean Moeritherium lyonsi is rough approximations based on Andrews's (1906) measurements and depictions of different individuals of this species and limb segment ratios of the other Paleogene proboscidean Numidotherium koholense (Court, 1994).
4.1. Head and neck
Head. The head in elephantids is massive; it is very long, wide, and robust. The relative length (hereinafter in relation to the trunk length) of the skull in elephantids is long (mean from 44.6% to 52.0%). Among artiodactyls and perissodactyls, these ratios are very similar to the equids, musk ox, and giraffe (mean 43.8%, 46.7%, and 50.9%, respectively) and longer than in other studied large ungulates (Table 1, Figure 4). The relative width of the skull and mandible in elephantids is almost unparalleled among other large herbivorous mammals. Only hippos are close to the elephantids in the relative mandible width (mean = 26.9%) due to the unique specialization of the angular region of their jaw. Among studied elephantids, the skull length in the woolly mammoth (mean 52%) is markedly longer than those in the Asian and African elephants (mean ~45%). The relative width of the skull in the eye socket area in the woolly mammoth (mean = 36.5%) is markedly wider than those in the Asian and African elephants (mean ~31%). However, the relative width of the skull in the occiput area, on average, is similar in the E. maximus (n = 5, mean ~35%), M. primigenius, and L. africana (n = 4 in both species, mean ~34%), and narrower in the L. cyclotis (n = 1, mean ~30%). The relative mandible width in studied Elephantidae is very close on average (Table 1).
TABLE 1.
Characteristics of the skull, mandible, and vertebral column. Dimensions are given as a percentage of the cumulative length of the trunk (thoracic + lumbar + sacral regions).
| Taxa | n | Skull L′ (%) | Eye socket W (%) | Occiput W (%) | Mandible W (%) | Vertebral formula | Vertebral region length (%) | Costa (%) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| T | L | S | C | T | L | S | ||||||||
| T + L + S = 100% | ||||||||||||||
| Proboscidea | Elephas maximus | 5 | 44.6 | 30.5 | 35.1 | 26.7 | 19/20 | 3/4 | 4 | 22.8 | 71.5 | 13.3 | 15.1 | 56.5 |
| Loxodonta africana | 5 | 44.8 | 31.6 | 34.4 | 25.4 | 20/21 | 2/3 | 4 | 22.5 | 72.1 | 11.3 | 16.6 | 59.3 | |
| Loxodonta cyclotis | 1 | 44.7 | 29.6 | 30.3 | 27.0 | 21 | 2 | N/A | N/A | 72.0 | 11.2 | 16.8 | 58.6 | |
| Mammuthus primigenius† | 5 | 52.0 | 36.5 | 34.4 | 27.4 | 18–20 | 3/4 | 4 | 20.7 | 69.4 | 15.6 | 15.0 | 59.4 | |
| Adams† | >44.9 | ≤38.5 | N/A | 27.8 | 18 | 4 | 4 | 21.0 | 67.8 | 16.6 | 15.6 | 56.1 | ||
| Taimyr† | 49.0 | 35.2 | 33.4 | 26.5 | 19 | 4 | 4 | 20.2 | 70.1 | 15.6 | 14.3 | 51.1 | ||
| Berezovsky† | 53.8 | 41.0 | 37.3 a | 31.8 | 19 | 4 | 4 | 20.8 | 68.8 | 15.3 | 15.9 | 62.7 | ||
| Allaikhovsky† | 54.1 | 34.1 | 33.1 | 27.1 | 20 | 4 | 4 | 21.6 | 70.7 | 15.8 | 13.5 | 62.2 | ||
| Chekurovsky† | 51.2 | 35.9 | 33.5 | 23.8 | 19 | 3 | 4 | 20.0 | 69.4 | 14.7 | 15.9 | 56.5 | ||
| Mammut americanum† | 2 | 46.2 | 30.9 | 34.6 | 27.3 | 20/21 | 2/3 | 4/5 | 20.3 | 73.7 | 9.8 | 17.0 | 55.6 | |
| Perissodactyla | Rhinocerotidae | 7 | 38.9 | 16.6 | N/A | 20.7 | 18–20 | 3/4 | 4/5 | 29.1 | 75.1 | 12.4 | 12.5 | 56.7 |
| Tapiridae | 3 | 38.1 | 10.8 | N/A | 17.9 | 18/19 | 4/5 | 6/7 | 23.7 | 62.0 | 16.6 | 21.4 | 44.5 | |
| Equidae | 13 | 43.9 | 18.9 | N/A | 16.5 | 18/19 | 5/6 | 4–7 | 45.8 | 61.3 | 22.1 | 16.5 | 44.4 | |
| Arenahippus grangeri† | 1 | 29.7 | N/A | N/A | N/A | 17 | 7 | 5 | 21.6 | 50.4 | 30.7 | 18.9 | 28.8 | |
| Hyrachyus eximius† | 1 | N/A | N/A | N/A | N/A | 17 | 7 | N/A | 19.7 | 47.3 | 33.6 | 19.1 | N/A | |
| Artiodactyla | Hippopotamidae | 5 | 40.1 | 23.4 | N/A | 26.9 | 15 | 4 | 5 | 26.4 | 62.4 | 20.3 | 17.3 | 49.5 |
| Camelus | 1 | 33.8 | 16.3 | N/A | 12.5 | 12 | 7 | 5 | 64.4 | 55.0 | 31.9 | 13.1 | 38.1 | |
| Cervidae | 2 | 39.8 | 18.0 | 13.2 | 13.1 | 13 | 6 | N/A | 42.9 | 53.6 | 29.4 | 17 | 43.5 | |
| Giraffa | ≤4 | 50.9 | 22.4 | N/A | 16.7 | 14/15 | 4/5 | 3/4 | 114.4 | 64.3 | 21.0 | 14.6 | 56.4 | |
| Okapia | 1 | 42.6 | 17.0 | N/A | 12.5 | 14 | 5 | 4 | 55.5 | 59.0 | 24.5 | 17.0 | 53.5 | |
| Bovinae | 8 | 36.3 | 19.5 | N/A | 12.5 | 13/14 | 5/6 | 4/5 | 34.0 | 56.6 | 26.6 | 16.7 | 46.5 | |
| Ovibos | 1 | 46.7 | 27.3 | N/A | 11.0 | 13 | 6 | 5 | 30.8 | 52.4 | 29.5 | 18.1 | 49.3 | |
| K–S test p‐value | 0.200 | 0.001 | N/A | 0.038 | 0.004 | 0.065 | 0.200 | 0.200 | 0.024 | |||||
FIGURE 4.
Head and neck dimensions in relation to trunk length in various large herbivorous mammals and their extinct ancestors. The illustration of M. lyonsi and Arsinoitherium zitteli is based on Andrews's (1906) and Osborn's (1936) depictions. The illustration of the Paleogene equoid Arenahippus grangeri is based on a 3D skeleton from the UMORF website (https://umorf.ummp.lsa.umich.edu/wp/specimen‐data/?Model_ID=1675) and Wood et al. (2011) figures.
Neck. The relative length of the neck in elephantids (mean from 20.7% to 22.8%) is shorter than those in other large herbivorous mammals (U = 10, p < 0.001, mean diff = −14.7%). These ratios are shorter than those of short‐necked ungulates, such as tapirs, rhinos, and hippos (mean from 23.7% to 29.1%). Among studied elephantids, the woolly mammoth is characterized by a slightly shorter neck (mean = 20.7%) compared with the Asian and African elephants (mean 22.5% and 22.8%, respectively).
4.2. Trunk
The vertebral formula in studied elephantids consists of 18–21 thoracic, 2–4 lumbar, and 4 sacral vertebrae. The African elephants of genus Loxodonta are characterized by a slightly higher number of the thoracic (20/21 vs. 18–20) and a slightly lower number of the lumbar (2/3 vs. 3/4) vertebrae.
Backbone. The ratio of relative lengths of the thoracic–lumbar–sacral regions of the vertebral column in the studied elephantids is quite similar. The biggest difference in the backbone ratios is the relatively shorter lumbar (~11%) and longer sacral (~17%) region in the African elephants. The relative length of the thoracic region in elephantids (mean from 69.4% to 72.1%; Table 1, Figure 5) is smaller only in relation to rhinoceroses (mean = 75.1%) and noticeably larger than those in odd‐toed (~61%) and even‐toed ungulates (mean from 52.4% to 64.3%). The relative length of the lumbar region in elephants (mean from 11.2% to 15.6%) is very short; it is similar to the rhinoceroses and tapirs (mean 12.4% and 16.6%, respectively) and noticeably shorter than those in hippopotamids (~20%), giraffe (~21%), equids (~22%), and other even‐toed ungulates (mean from 24.5% to 31.9%). The sacral region's relative length in elephantids lies in the variability range of the other studied taxa.
FIGURE 5.
Ratios of the vertebral column in various mammals: (a) Diceros bicornis (ZMMU S‐93020); (b) Elephas maximus (ZMMU S‐182210); (c) Moeritherium lyonsi; (d) Equus ferus caballus (ZMMU S‐102019); (e) Arenahippus grangeri (UM 115547); (f) Camelus bactrianus (ZIN 11173); (g) Canis lupus (ZMMU S‐107021); (h) Acinonyx jubatus (ZMMU S‐171619).
Thoracic cage. The relative rib length in elephantids (mean from 56.5% to 59.4%) is significantly longer than in other large herbivorous mammals (U = 46, p < 0.001, mean diff = 10.5%). In studied ungulates, it is matched only in okapi (53.5%), giraffe (56.4%), and rhinocerotids (mean 56.7%). Among the Elephantidae studied, the longer relative rib length is characteristic for the African elephant and woolly mammoth (mean ~59%); in the Asian elephant, it is somewhat shorter (mean = 56.5%).
Pelvis. The relative width of the pelvis in elephantids (mean from 55.6% to 71.9%) is, on average, almost two times wider than in any other large herbivorous mammals. In other megaherbivores it is also markedly wider (mean 42.9% and 48.0% in hippos and rhinos, respectively) than in other studied large ungulates (mean from 31.3% to 37.4%). Among studied elephantids, the narrowest relative pelvis width is characteristic of the African elephants (mean ~56%); it is markedly wider in the Asian elephant (mean = 62.5%) and especially in the woolly mammoth (mean = 71.9%).
4.3. Limbs
The body plan of elephantids is characterized by a very tall appearance (Figure 6). The height of limbs relative to trunk length in elephantids (130–150%) is not only markedly higher than in other megaherbivores (~85%–95% in Rhinocerotidae and Hippopotamidae) but also significantly higher than in such “tall” ungulates as equids and camels (~110%–120%). Among the Elephantidae studied, the woolly mammoth is characterized, on average, by the tallest limbs relative to the length of the trunk (Tables 2 and 3).
FIGURE 6.
Forelimb height in relation to the trunk length in various large herbivorous mammals and their extinct ancestors: (a) Giraffa camelopardalis (ZMMU); (b) Moeritherium lyonsi; (c) Elephas maximus (ZMMU S‐182211); (d) Arsinoitherium zitteli; (e) Camelus bactrianus (ZMMU F 2749); (f) Arenahippus grangeri (UM 115547); (g) Equus hemionus (ZMMU); (h) Bison bonasus (ZMMU F 2751); (i) Rhinoceros unicornis (ZMMU F 2757); (j) Hippopotamus amphibius (ZMMU F 2755). The illustration of the M. lyonsi (b) is based on Andrews's (1906) depictions of this species and Court's (1994) photographs of the Paleogene proboscidean N. koholense.
TABLE 2.
Characteristics of the forelimb.
| Taxa | Scapula W (%) | Scapula L′ (%) | Humerus (%) | Antebr. (%) | Manus (%) | Forelimb sum (%) | |
|---|---|---|---|---|---|---|---|
| Proboscidea | Elephas maximus | 36.4 | 38.8 | 42.9 | 34.3 | 18.2 | 134.1 |
| Loxodonta africana | 33.4 | 38.4 | 46 | 35.1 | 19.7 | 139.1 | |
| Loxodonta cyclotis | 33.6 | 35.9 | 43.8 | 32.6 | 22.4 | 134.5 | |
| Mammuthus primigenius† | 37.0 | 39.9 | 47.7 | 37.2 | 19.7 | 144.5 | |
| Adams† | 35.6 | 42.2 | 48.3 | 37.8 | ≤20.0 | ≤148.3 | |
| Taimyr† | 34.2 | 37.2 | 43.1 | 33.7 | 17.4 | 131.4 | |
| Berezovsky† | 39.8 | 43.4 | 52.6 | 38.5 | 21.4 | 156.0 | |
| Allaikhovsky† | 40.1 | 39.6 | 50.1 | 37.3 | 21.1 | 148.1 | |
| Chekurovsky† | 35.3 | 37.1 | 44.1 | 38.8 | 18.8 | 138.8 | |
| Mammut americanum† | 39.6 | 37.3 | 38.6 | 27.7 | 18.4 | 122.1 | |
| Perissodactyla | Rhinocerotidae | 14.7 | 30.1 | 23.1 | 20.2 | 19.5 | 92.8 |
| Tapiridae | 14.4 | 28.3 | 22.9 | 20.0 | 23.4 | 94.6 | |
| Equidae | 15.5 | 26.3 | 21.3 | 26.3 | 35.9 | 109.8 | |
| Arenahippus grangeri† | 13.7 | 21 | 23.4 | 21.1 | 18.9 | 84.4 | |
| Hyrachyus eximius† | 14.6 | ≤24.2 | ≤30.4 | ≤22.5 | ≤21.1 | ≤98.2 | |
| Artiodactyla | Hippopotamidae | 18.1 | 26.1 | 23.1 | 16.3 | 20.8 | 86.3 |
| Camelus | 16.6 | 29.7 | 25.0 | 30.0 | 36.3 | 120.9 | |
| Cervidae | 18.1 | 27.7 | 26.9 | 29.3 | 43.2 | 127.1 | |
| Giraffa | 16.7 | 37.3 | 33.8 | 52 | 77.5 | 200.5 | |
| Okapia | 18.0 | 35.0 | 29.0 | 35.5 | 49.5 | 148.3 | |
| Bovinae | 18.2 | 32.3 | 22.4 | 22.6 | 28.7 | 106.1 | |
| Ovibos | 20.7 | 33.7 | 28.1 | 27.8 | 31.7 | 121.3 | |
| K–S test p‐value | <0.001 | 0.009 | <0.001 | 0.184 | 0.013 |
Abbreviation: †, fossil.
TABLE 3.
Characteristics of the hindlimb.
| Taxa | Pelvis W max (%) | Pelvis W min (%) | Ilium (%) | Femur (%) | Crus (%) | Pes (%) | Hindlimb sum (%) | |
|---|---|---|---|---|---|---|---|---|
| Proboscidea | Elephas maximus | 62.5 | 19.4 | 31.3 | 52.7 | 30.1 | 17.6 | 131.7 |
| Loxodonta africana | 56.2 | 16.8 | 30.2 | 55.1 | 31.8 | 18.5 | 135.5 | |
| Loxodonta cyclotis | 55.6 | 19.1 | 31.3 | 54.9 | 28.5 | 20.4 | 135.1 | |
| Mammuthus primigenius † | 71.9 | 20.6 | 35.0 | 56.0 | 30.8 | 19.0 | 140.7 | |
| Adams † | 66.8 | 19.5 | 32.7 | 56.1 | 31.7 | ≤19.0 | ≤139.5 | |
| Taimyr † | 64.3 | 19.9 | 33.4 | 52.6 | 27.0 | 17.6 | 130.6 | |
| Berezovsky † | 82.0 | 23.9 | 38.2 | 59.9 | 33.3 | 19.3 | 150.8 | |
| Allaikhovsky † | 72.2 | 21.6 | 34.1 | 57.6 | 32.1 | 19.6 | 143.4 | |
| Chekurovsky † | 74.1 | 18.2 | 36.5 | 53.5 | 30.0 | 19.4 | 139.4 | |
| Mammut americanum † | 76.9 | 22.4 | 33.4 | 47.3 | 27.1 | 16.9 | 124.8 | |
| Perissodactyla | Rhinocerotidae | 48.0 | 14.4 | 22.5 | 29.4 | 18.9 | 21.3 | 92.1 |
| Tapiridae | 32.3 | 13.0 | 18.9 | 30.4 | 24.5 | 25.5 | 99.3 | |
| Equidae | 34.4 | 11.4 | 21.1 | 29.8 | 26.8 | 42.3 | 119.9 | |
| Arenahippus grangeri † | 22.2 | 10.8 | 14.4 | 30.7 | 29.7 | 27.8 | 102.6 | |
| Hyrachyus eximius † | N/A | N/A | ≤14.6 | ≤32.1 | ≤27.5 | ≤32.2 | ≤106.4 | |
| Artiodactyla | Hippopotamidae | 42.3 | 16.4 | 22.1 | 30.3 | 19.7 | 23.8 | 95.9 |
| Camelus | 31.3 | 11.3 | 16.9 | 34.4 | 26.9 | 38.1 | 116.3 | |
| Cervidae | 26.7 | 12.0 | 15.7 | 31.5 | 34.6 | 50.9 | 132.7 | |
| Giraffa | 33.8 | 13.7 | 20.6 | 40.2 | 41.2 | 80.4 | 182.4 | |
| Okapia | 34.5 | 14.5 | 25.9 | 32.8 | 32.3 | 50.8 | 141.8 | |
| Bovinae | 36.5 | 12.6 | 19.1 | 28.4 | 25.7 | 34.5 | 107.7 | |
| Ovibos | 32.2 | 13.2 | 18.9 | 31.1 | 29.3 | 34.9 | 114.1 | |
| K–S test p‐value | <0.001 | 0.002 | <0.001 | <0.001 | 0.200 | 0.017 |
Abbreviation: †, fossil.
The difference between the height of the fore‐ and hindlimbs (estimated as the sum of the lengths of the respective segments) can vary considerably in large ungulates. Thus, in our sample, the height of the hindlimb in Equidae is on average 10% taller than the height of the forelimb (and opposite, the height of the forelimb is 18% taller than that of the hindlimb in giraffe). In elephants, the height of the forelimb is slightly taller than the height of the hindlimb. On average, this difference in the woolly mammoth and African bush elephant (+3.7%) is slightly more pronounced than in the Asian elephant (+2.4%).
4.3.1. Forelimbs
Scapula. The scapula in elephantids is unparalleled among ungulates in size and shape (Figure 6); it is characterized by one of the greatest relative lengths (U = 14, p < 0.001, mean diff = 9.8%) and significantly exceeds all other ungulates in relative width (U = 0, p < 0.001, mean diff = 19.1%). The relative length of the scapula in elephantids (~39%) exceeds it in other large herbivorous mammals by 1.2–1.5 times; only giraffe (37.3%), okapi (35%), and representatives of the genus Bison (35–38.9%) occurred within the range of variability of elephants. The width of the scapula in elephantids is on average ~ 2 times wider than in other studied large herbivores. Moreover, the minimum relative width of the scapula in elephantids is more than 1.5 times wider than the maximum among non‐elephantids (21.8% in Bubalus bubalis).
The length of the scapula in the three Elephantidae genera studied is almost identical. The width of the scapula is almost identical in the woolly mammoth and Asian elephant and slightly narrower in the genus Loxodonta.
Humerus. The humerus relative length in elephantids (mean from 42.9% to 47.7%) is very long; it is on average ~ 2 times longer than those in odd‐toed and even‐toed ungulates (21%–29%) and ~1.3 times longer than in giraffe (33.8%). The humerus is the longest segment in the elephantid forelimb, which is not characteristic of any modern‐day ungulates. Moreover, in the majority of taxa studied (equids, bovines, camels, and giraffe), the humerus is quite the opposite—the shortest segment of the forelimb (Figure 6). Among the Elephantidae studied, the shortest relative humerus length is characteristic for the Asian elephants (~43%) and somewhat longer for the African bush elephant (~46%) and the woolly mammoth (~48%).
Antebrachium. The relatively longest antebrachium among studied ungulates belongs to the giraffe (52%). The antebrachium relative length in elephantids (mean from 34.3% to 37.2%) is significantly longer (mean diff = 11.2%, t = 9.997, p < 0.001, 95% CI: 9%–13.5%) than in the dataset of the ungulates we studied. Besides the extreme antebrachium elongation in giraffe, the closest to the elephants' antebrachium relative length is characteristic of okapi (35.5%), moose, camel, and some equids (~30%). In megaherbivores, relative antebrachium length is ~2 times shorter than in elephants (16.3% in Hippopotamidae and 20.2% in Rhinocerotidae). The relative length of the antebrachium in the Asian and African elephants (mean 34.3% and 35.1%) is almost identical; in the woolly mammoth, it is somewhat longer (mean 37.2%).
Manus. Contrary to the more proximal segments of the forelimb, the manus relative length in elephantids (mean from 18.2% to 19.7%) is significantly shorter (U = 25, p < 0.001, mean diff = −13%) than in other large herbivorous mammals; it is slightly shorter than in other megaherbivores (~20% in Hippopotamidae and Rhinocerotidae). Relative manus length is significantly longer in bovines (~29%) and especially in camels and equids (~36%). The relatively longest manus among studied ungulates belongs to the giraffe (77.5%), followed by okapi (49.5%). Among the Elephantidae, the relative manus length is slightly shorter in the Asian elephant (18.2%), and somewhat longer in the woolly mammoth and African elephant (19.7%).
4.3.2. Hindlimbs
Ilium. The ilium relative length (between the dorsal edge of acetabulum and the most distant part of the tuber sacrale) in elephantids (mean from 30.2% to 35.0%) is significantly longer (U = 0, p < 0.001, mean diff = 11.9%) than in other large herbivorous mammals we studied. Among the latter it is shorter in cervids, camel, bovids, and tapirids (~15%–19%) and longer in equids, hippos, and rhinos (~21%–23%). The relative length of the ilium in the African and Asian elephants (mean 30.2% and 31.3%) is almost identical; in the woolly mammoth, it is somewhat longer (mean 35%).
Femur. The femur relative length in elephantids (mean from 52.7% to 56.0%) is very long; it is on average almost 2 times longer than in odd‐toed and even‐toed ungulates (mean 29%–34%) and ~1.3–1.4 times longer than in giraffe (40.2%). The femur is the longest segment in the elephantid hindlimb; this is also characteristic of other megaherbivores, as well as tapirs (mean ~30% for the femur vs. 22%–26% for the pes). In other large herbivorous mammals we have studied the longest hindlimb segment is pes (29%–34% femur vs. 35%–51% pes). The relative length of the femur in the African elephant and woolly mammoth (mean ~55%–56%) is almost identical; in the Asian elephant, it is somewhat shorter (mean ~ 53%).
Crus. The crus relative length in elephantids (mean from 30.1% to 31.8%) is significantly longer (mean diff = 5.5%, t = 5.587, p < 0.001, 95% CI: 3.5%–7.5%) than in the studied sample of artiodactyls and perissodactyls. However, contrary to the previously described limb segments, this difference is substantial only between elephantids and other megaherbivores (mean ~19%–20% in hippos and rhinos). In other large ungulates we studied, crus relative length is, on average, only slightly inferior to the elephantids (mean from 24.5% to 34.6%). The relatively longest crus among studied ungulates belongs to the giraffe (41.2%). The relative length of the crus in the woolly mammoth and the Asian elephant (mean 30.8% and 30.1%) is almost identical; in the African bush elephant, it is somewhat longer (mean 31.8%).
Pes. Contrary to the more proximal segments of the hindlimb, the pes relative length in elephantids (mean from 17.6% to 19%) is significantly shorter than in other large herbivores (U = 0, p < 0.001, mean diff = −19.0%); it is slightly shorter than in other megaherbivores (mean ~21%–24% in hippos and rhinos). Relative pes length is significantly longer in bovines and musk ox (~35%) and especially in camels and equids (mean ~38% and ~42%, respectively). The relatively longest pes among studied ungulates belongs to the giraffe (80.4%), followed by moose (54.9%) and okapi (50.8%). Among the Elephantidae studied, the relative pes length is almost identical on average (mean ~18%–19%).
5. DISCUSSION
5.1. Head and neck
In previous studies, it was observed that the woolly mammoth had a relatively longer head compared with extant elephants. The length of the Mammuthus primigenius head, from the rostrum to the occiput was ~50% of the height at the shoulders, while that of the African elephant is about 40% (Vereshchagin & Tikhonov, 1990; Zalensky, 1909). This is probably associated with the development of larger tusks and as a result longer premaxillaries. Adult and old male woolly mammoths had tusks that were 250–350 cm long and weighed 50–75 kg (Boeskorov, 2021). The record‐size tusk of M. primigenius measured up to 430 cm in length and weighed up to 125 kg (Boeskorov et al., 2020). Among extant Elephantidae, larger tusks are characteristic of the savanna elephant, with the record‐size tusk measuring up to 349 cm in length and the heaviest up to 107 kg in mass (Larramendi, 2023). The tusks of male Asian elephants are considerably smaller, with the average length being 150–160 cm and weight being 20–25 kg; the record‐size tusk measuring up to 326 cm in length and the heaviest up to 73 kg in mass (Larramendi, 2023; Shoshani & Eisenberg, 1982; Walker et al., 1968). Female E. maximus are mostly tuskless (Wilson & Mittermeier, 2011).
Our data illustrate different proportions between the skulls of woolly mammoths and extant elephantids. The skull of M. primigenius is relatively longer and wider at the eye sockets than those of the Asian (on average by +7.4% and +6% of the trunk length, respectively) and African (by ~+7% and ~+5%–7%, respectively) elephants. However, the maximal occipital width in the studied elephantids has a large variability and overlap: 31%–40% in the Asian elephant, 32%–40% in the savanna elephant, ~33% in the Taimyr, Allaikhovsky, and Chekurovsky mammoths, and 37% in the Berezovsky mammoth (44% based on Averianov, 1996). The relative width of the mandible, on average, differs only slightly by ~1%. A relatively longer skull is characteristic of other Mammuthus species as well, so, based on our measurements and Baygusheva et al. (2012), the female specimen of the steppe mammoth (AMZ KP‐28689) has a skull ~50% of the trunk length (~5% longer than in extant elephantids). The occipital width of this specimen is ~36%. Based on the measurements of 3D models of the skeletal bones of the Buesching mastodon (ISM 71.3.261) from the UMORF website (https://umorf.ummp.lsa.umich.edu/wp/mammutidae2/), its skull is relatively long (~49% of the trunk length) and wide (~34% at the eye sockets; 38.5% at the occiput). The female Owosso mastodon (https://umorf.ummp.lsa.umich.edu/wp/specimen‐data/?Model_ID=1335) skull is notably smaller in length and width than those of Buesching mastodon and elephantids (~43% length; ~28% width at the eye sockets; 31% width at the occiput). The relative height of the mastodon skull at the occiput is similar to those of elephantids (21.5% and 23% vs. 22%–29%), however, the height from the occlusal surface of the teeth is notably lower (~29% vs. 40%–46%; Figure 7, Table 1).
FIGURE 7.
Head and neck dimensions in relation to the trunk length in various proboscideans. The illustration of the Mammut americanum is based on a 3D skeleton of the Buesching mastodon from the UMORF website (https://umorf.ummp.lsa.umich.edu/wp/mammutidae2/). The illustration of the Megabelodon lulli is based on Cook (1928).
The paleontological data on Paleogene perissodactyls and proboscideans show that relative cranial sizes in both phylogenetic lineages have substantially increased over time. The fossil remains of Arenahippus grangeri (Wood et al., 2011) reveal that plesiomorphic perissodactyls had substantially smaller skull sizes than any of the extant Tapiridae, Rhinocerotidae, and Equidae (Figure 4). The skull of this equoid is about 1.2–1.4 times shorter (relative to the trunk length) than the skulls of extant tapirs and rhinoceroses (~30% vs. 35.4%–43.2%) and 1.4–1.6 times shorter than their equid descendants (40.5%–48%; Table 1). The relative skull size of Paleogene proboscidean Moeritherium can only be roughly estimated. Based on Andrews (1906) data on presacral vertebrae count, measurements/figures of the skulls, thoracic, lumbar vertebrae, and sacrum we estimated the skull length of this animal to be ~31% of the trunk length. This is significantly shorter than the skulls of the studied Elephantidae (mean from 44.6% to 52%). The height and robustness of the skull of extant elephantids is also significantly greater than the notably lower skulls of their Paleogene ancestors. Thus, the relative height of the skull at the occiput of Moeritherium is more than twice lower than those of elephantids (~10%), and the height from the occlusal surface of the teeth is almost 4 times lower (~12%; Figures 4 and 7).
The significant increase in relative head dimensions in large herbivorous mammals is probably associated with a variety of factors, including among others: (1) an increase in the size of dental crowns and the development of mesodont and hypsodont cheek teeth; (2) the use of the head as a tournament weapon; and (3) allometric growth. The evolution of high‐crowned molars has been well studied among equids (Mihlbachler et al., 2011), elephantids (Lister, 2013; Lister & Sher, 2001; Maglio, 1973; Ranjan et al., 2023), and ungulates in general (Fortelius et al., 2006). In addition, the evolution of a relatively taller skull in elephantids compared with other proboscideans is associated with a transition from side‐to‐side to fore‐aft chewing, accompanied by a rearrangement and forward shift of the masticatory musculature (Maglio, 1973). The use of the head for tournament fighting is a characteristic of artiodactyls, perissodactyls, and proboscideans. In all of these groups, it is accompanied by the development of various weapons, including long canines, keratinous horns, bony antlers, ossicones, bone core/keratin surrounding horns, and tusks (Wilson & Mittermeier, 2011). Lastly, since the body mass of the animal grows proportionally to the linear dimensions (lengths) cubed and the strengths of muscles grow proportionally to their cross‐section (lengths squared), larger animals require a disproportional increase in the strength of their neck muscles and, hence, an increase in the area for their attachment to the skull. Further discussion of this point for elephantids is available in Larramendi et al. (2017).
In contrast to their huge heads, large herbivorous mammals generally have relatively short necks (with obvious exceptions; Figures 4 and 7). The mechanical advantage of a shorter neck is manifested by the reduced moment of force acting on vertebral flexion at the withers, where the head and neck are connected to the trunk by a nuchal ligament (Gambaryan, 1974; Sokolov, 1979). Finally, the presence of a proboscis allows elephantids to break the interrelationship that exists between the total length of the head and neck on one side and the forelimbs on the other. Among other megaherbivores, the relative length of the head and neck is similar to the length of the sum of the limb segments (~65% vs. ~85%–95%). However, elephants have a twofold difference in these two characteristics (~70% vs. ~140%).
5.2. Trunk
The ratios of the vertebral column regions in the trunk of the Elephantidae are very similar to the other megaherbivores, as well as some other odd‐toed and even‐toed ungulates (Figure 5). The thoracic region dominates the length of the elephantid's trunk, accounting for ~70%. Among extant ungulates, only rhinoceroses have, on average, a relatively longer thoracic region than elephants (~75%). The lumbar region in elephantids, on average, is ~5 times shorter than the thoracic region; in some perissodactyls and artiodactyls this ratio is quite similar: in rhinoceroses, the thoracic region, on average, is ~6 times longer than the lumbar region, in tapirs, it is ~4 times longer, in equids, giraffes, and hippos it is ~3 times longer. In smaller ungulates like tragulids and antilopins, the thoracic region is only ~1.5 times longer than the lumbar region, and in some carnivorans, these modules are approximately equal in length (Figure 5g,h; Belyaev et al., 2021, Belyaev et al., 2024). The reason for such widely differing ratios of the vertebral column regions is rooted in significant differences in the engagement of sagittal flexibility of the backbone (mostly localized in the lumbar region) during locomotion between dorsostable and dorsomobile mammals.
As was shown by Alexander et al. (1985), the storage and recoil of elastic energy in the aponeurosis of the m. longissimus thoracis et lumborum allow mammals to reduce metabolic costs of running utilizing gallop and makes the gallop the most energy‐efficient gait at high speeds. However, sagittal mobility (as well as storage and recoil of elastic energy) in the lumbar region is actively engaged by mammals only during asymmetrical gaits (canter, gallop) and is practically not engaged during symmetrical ones (Haussler et al., 2001).
Allometric growth and adaptation of the locomotor apparatus to gigantism lead large herbivorous mammals to develop more dorsostable modes of locomotion. For artiodactyls, the upper mass limit for dorsomobile modes of gallop (saltatorial and saltatorial‐cursorial running forms in terms of Gambaryan, 1974) is ~200–400 kg (Belyaev et al., 2022). When their body mass increases beyond this margin, artiodactyls tend to switch to a running form that is more suitable for that size. For instance, the musk ox adopts the mediportal running form,1 which is uncharacteristic of other Caprinae (Gambaryan, 1974). Among the three groups of megaherbivores, we observe three solutions to the same problem. Rhinoceroses are the largest extant animals that have retained the ability to gallop (Alexander & Pond, 1992); however, their gallop is characterized by an almost immobile backbone and the lowest engagement of the sagittal flexibility across all of the odd‐toed and even‐toed ungulates (Belyaev et al., 2023). The common hippo employs gallop to run underwater; however, it employs the trot (symmetrical gait) for fast terrestrial running (Coughlin & Fish, 2009). Finally, the mode of fast movement in elephants is exclusively fast walk (symmetrical gait) with no unsupported phases (Gambaryan, 1974; Gambaryan & Ruhkyan, 1974; Hutchinson et al., 2003). Thus, it is not surprising that the relative length of the lumbar region, which is actively engaged while galloping, is significantly shortened in evolutionary lineages of odd‐toed and even‐toed ungulates transitioning to a more dorsostable running mode (Belyaev et al., 2022, 2023). Among studied large ungulates, the relative length of the thoracic region is the least different from the length of the lumbar region in the Bactrian camel (the thoracic region is ~1.7 times longer; Figure 5f). Since a longer lumbar region is strongly correlated with higher sagittal flexibility of the backbone (Belyaev et al., 2021, 2022), the camel is, as expected, the most dorsomobile species studied here.
The fossil remains indicate that the proportions of the vertebral column in proboscideans changed rather little during their evolution (Figure 5b,c). The Paleogene Moeritherium, like modern‐day elephants, has 23 presacral vertebrae (19 T + 4 L; Andrews, 1906). Based on Andrews's (1906) measurements, we estimate that the relative length of the thoracic region in Moeritherium is slightly shorter in comparison to extant Elephantidae (~66% vs. 69%–72%), while the lumbar region is slightly longer (~19% vs. 11%–16%). This similarity probably suggests that dorsostability appeared early in the evolution of proboscideans. Despite the similarities in the proportions of the vertebral regions, the morphology of vertebrae has changed significantly between the Paleogene and modern‐day proboscideans. This is particularly evident in the taller and wider appearance of the vertebrae centra, as well as a marked elongation of the spinous processes (Figure 5b,c).
In contrast to the proboscideans, perissodactyls underwent significant changes in the proportions of their vertebral column during evolution. The Paleogene perissodactyls have a similar (or slightly higher in the case of rhinocerotoids) number of trunk vertebrae as extant representatives (Cope, 1873b; Scott et al., 1941; Wood et al., 2011), but their vertebral formula consisted of a higher number of lumbar vertebrae (7), and the relative length of their lumbar region was markedly longer (Figure 5a,d,e). Thus, the relative length of the lumbar region of the equoid A. grangeri is one‐third longer than those of the extant equids, and the lumbar region of the rhinocerotoid Hyrachyus eximius is almost three times longer than those of modern‐day rhinoceroses (Table 1). Such rearrangements are associated with the development of a more dorsostable locomotion mode in odd‐toed ungulates; a detailed discussion of this issue is presented in Belyaev et al. (2023). The neck region of Paleogene perissodactyls was slightly shorter than in modern‐day tapirs and rhinoceroses and more than twice shorter than in extant equids (Table 1; Figure 4). The increase in the relative length of the cervical region in equids was accompanied by a significant increase in the flexibility of the intervertebral joints in the sagittal and frontal planes (Belyaev et al., 2023; Clayton & Townsend, 1989). Finally, extant equids have larger rib and pelvic sizes than A. grangeri by an average of ~1.5 times. This difference between Paleogene and modern‐day rhinoceroses is probably even higher (~2 times). Such a difference is a by‐product of the allometric growth.
Finally, elephantids have a very massive thorax. Their ribs, on average, are relatively longer compared with other megaherbivores and markedly longer compared with other large artiodactyls and perissodactyls we studied (Table 1). The pelvic dimensions of elephantids are markedly larger than those of any artiodactyls and perissodactyls (Table 3). Among the proboscideans studied, the width of the pelvis of extant elephants is found to be the narrowest, which is consistent with Larramendi et al. (2017). Thus, the relative width of the pelvis in the Megabelodon lulli (~69%) is, on average, 6% wider than those of the Asian elephants and ~13% wider than those of the African elephants. In the woolly mammoth (average ~72%), steppe mammoth (~79% in AMZ KP‐28689), and American mastodon (~72% in UMMP VP 23498 and ~80% in ISM 71.3.261), the pelvis is even wider. A wider pelvis may be attributed to larger body volume and the sizes of the intestines, particularly the digestive system. Evidence of body mass suggests that there are relationships between it and the environment (diet) for large Pleistocene herbivorous mammals (Saarinen et al., 2016, 2021). It can be hypothesized that the presence of hindgut chambers for bacterial fermentation of cellulose in perissodactyls and proboscideans (Alexander, 1993), may be linked to the unique dorsostability of their backbone (Belyaev et al., 2023; Gambaryan, 1974).
5.3. Limbs
Before we begin the discussion of the limb proportions, it should be noted that the ilium segment (measurement 16; Figure 2) in large herbivorous mammals should not be considered a limb segment in the biomechanical sense. Since the sacroiliac joints in large ungulates are immobile or almost immobile, and the lumbosacral joint, as well as the lumbar region in general, are rather low‐mobile (Belyaev et al., 2022), the pelvic segment is almost static relative to the backbone during the contact and swing phases. However, we will consider the pelvic segment as part of the hindlimb in the context of the body plan, as its length determines the height of the animal at the rump, similar to how the scapula determines the height of the animal at the shoulders.
In the studied herbivores, we observe two strategies of limb elongation during evolution. The first strategy was realized in elephantids with limbs elongated due to the elongation of the proximal segments (Figures 6b,c and 8). The relative length of the scapula, humerus, and antebrachium in the forelimb and the ilium and femur in the hindlimb in Elephantidae is superior to all studied artiodactyls and perissodactyls (with the only exception in some segments in giraffe). In the case of humerus and femur, this difference is close to twofold. Overall, elephants have a very tall appearance (Figure 6c). Their fore‐ and hindlimb height is ~130%–150% of their trunk length, which is about 50% taller than those of rhinoceroses and hippopotamuses, and about 20% taller than traditionally considered “tall” equids and camels. This difference in the relative limb height is further enhanced by the nearly straight joints of elephantid limbs (Figure 6). In contrast, the relative length of the manus in elephantids is almost identical to the other megaherbivores (Table 2; Figure 6c,h,i), while the relative length of the pes is slightly shorter (by ~5%, Table 3). The very short manus and pes of elephantids' limbs are similar to those of many large plesiomorphic Paleogene mammals: pantodonts, taeniodonts, uintatheres, embrithopods, etc. (Cope, 1873b, Carroll, 1993; see Arsinoitherium at Figures 3a and 6d).
FIGURE 8.
Forelimb height in relation to the trunk length in various proboscideans. In the American mastodon, Asian elephant, and mammoth, the height of the forelimb corresponds to the average in the species. The illustration of the M. americanum is based on the Buesching mastodon; M. primigenius is based on the Berezovsky mammoth (Figure 1).
The figures in Court (1994) allow us to estimate the proportions of the segments of the forelimb in Paleogene proboscidean Numidotherium koholense. The scapula and humerus, which are approximately the same length, are the longest segments of the forelimb. The antebrachium segment is ~1.5 times shorter, and the manus is ~2 times shorter than the two proximal segments. In Elephantidae, the ancestral proportion scapula ≈ brachium > antebrachium > manus has deviated toward scapula < brachium > antebrachium > manus. It is worth noting that among extant large artiodactyls and perissodactyls, the length of the scapula is always longer than the length of the humerus. The length of the humerus exceeding the length of the antebrachium is characteristic only of rhinoceroses, tapirs, and hippos (Table 2). However, even in these taxa, the length of the manus is either equal to (rhinos) or exceeds the length of the antebrachium (tapirs and hippos), rather than being shorter than the antebrachium. A rough estimate of the relative length of the humerus and femur in M. lyonsi shows that the corresponding segments in studied Elephantidae elongated more than twofold in the course of evolution: humerus 39%–53% vs. ~21% in M. lyonsi; femur 48%–60% vs. ~23% (18% and 20%, respectively, in the restoration of Larramendi (2016)). If we apply this value to the proportions of the forelimb of Numidotherium, the total height of the limbs of Moeritherium was ~65%–70% of the trunk length (Figure 6b,c). This is considerably shorter than the shortest relative limb length in extant large ungulates noted in hippos and rhinos and twice as short as the limbs of elephantids (Tables 2 and 3). The body plan of Moeritherium is so different from elephantids (Figure 3a) that it may be a by‐product of adaptation to a semi‐aquatic lifestyle (Liu et al., 2008), making these animals possibly the most aberrant taxa of plesiomorphic proboscideans. However, it should be noted that unlike the other ratios presented in this article, these values are rough approximations and should be taken with caution.
The most pronounced difference between the limbs of the woolly mammoth and modern‐day elephants is that the limbs of the mammoth were relatively taller (on average +6%–11% in the forelimb and +5%–9% in the hindlimb). In the forelimb, this is due to small positive differences in the lengths of all segments and it becomes more noticeable in the length of the humerus (1.7%–4.8%) and antebrachium (2.2%–3%). In the hindlimb, it is due to a slightly longer ilium (3.7%–4.8%) and femur (0.9%–3.3%). The steppe mammoth probably had even relatively taller limbs, as seen in the AMZ KP‐28689 specimen (Baygusheva et al., 2012), where the forelimb was ~155% of the trunk length (on average +17%–20% compared with the extant elephantids). The mammutid M. americanum had shorter limbs compared with extant elephantids (~115% of the trunk length in UMMP VP 23498 and ~130% in ISM 71.3.261; Figure 8). The ISM 71.3.261 male has long proximal segments, with the scapula at ~40% vs. ~36%–39% and ilium at ~34.5% vs. ~30%–31% in extant Elephantidae. The scapula width in Buesching mastodon is widest in our sample (44.4%; Figure 8). The American mastodon's larger scapula and pelvis are associated with a significantly heavier body mass than the extant elephants (Larramendi, 2016). The smaller female Owosso mastodon had noticeably smaller scapula (34.4% in length and 34.8% in width) and similar to the male length of ilium (34.2%). However, the humerus (35.6% and 41.5% vs. ~43%–46%), antebrachium (26.3% and 29% vs. ~33%–35%), femur (44.6% and 50% vs. ~53%–55%), and crus length (25.8% and 28.5% vs. ~28.5%–32%) in mastodons are notably shorter than those in extant Elephantidae. Measurements of the Warren mastodon (Osborn, 1936; Warren, 1852) indicate that the relative length of its limbs was significantly shorter than those of the Buesching mastodon (~110% vs. ~130%). This is probably associated with an overestimated thoraco‐lumbo‐sacral length, which resulted in our calculations in significantly shorter limbs, ribs, and a smaller and narrower skull and pelvis. However, this assumption requires further verification. The forelimb height of the amebelodontid M. lulli was ~123% of the trunk length, shorter than the mastodon and extant elephantids (Figure 8). Their scapula is the narrowest and one of the longest in studied proboscideans (~28% and ~41%, respectively). The other segments are notably shorter than those of extant elephantids, especially the humerus (~36%) and femur (~48%).
What is the biomechanical benefit of the elephantid limb structure? Specialized limbs that are very tall, almost straight at the joints, and elongated due to proximal segments allow elephantids to utilize the principle of inverted pendulum (straight‐legged walking) in locomotion (Cavagna et al., 1976; Ren & Hutchinson, 2007). The same principle is characteristic of human locomotion, but in the case of elephantids, there is an additional layer of adaptations associated with the size‐related reinforcements. The main biomechanical profit of the tall limbs in elephantids is the ability to walk relatively fast by increasing stride length, despite the secondary loss of the ability to use asymmetrical gaits and running in general (Gambaryan, 1974; Gambaryan & Ruhkyan, 1974; Hutchinson et al., 2003). Despite this benefit, a tall limb is very dangerous to its owner, especially when that owner is a graviportal animal. Many of the unique features of elephantid limbs are attributed to the necessity to ensure static and dynamic safety.
(1) First, in contrast to the odd‐toed and even‐toed ungulates, elephantids have ligamentous instead of muscular connection between the scapula and the trunk. The dorsal angle of the scapula is connected to the third thoracic vertebra by a ligament up to 5–6 cm thick, as well as a few smaller ligaments up to 1 cm thick between the scapula, 2nd and 4th ribs, and vertebrae (Gambaryan, 1974; Gambaryan & Ruhkyan, 1974). These ligaments occur due to collagenization of the dorsal muscular fibers of the m. serratus ventralis. One of the benefits of ligamentous connection between the scapula and the trunk is a reduction of the muscle energy expending to maintain body weight. Finally, the ligamentous connection in the forelimb of elephants creates a structure similar to the hip joint in the hindlimb (Kuznetsov, 2020).
(2) Despite very tall limbs, the vertical fluctuations of the dorsal angle of scapula and hip joint during elephant locomotion are very low. While fast walking, these fluctuations do not exceed 10–15 cm or 3%–4% of the animal's height (Gambaryan & Ruhkyan, 1974; Hutchinson et al., 2006). In comparison, the tapir's gallop is characterized by vertical fluctuations of the withers by 14% of the animal's height (Gambaryan, 1964). This results in the elephant's center of mass moving almost horizontally. Moreover, the vertical fluctuations of the elephants' center of mass decreases with increasing movement speed (Genin et al., 2010).
(3) Since the apex of the inverted pendulum of the forelimb corresponds to the ligamentous connection of the scapula to the trunk, it is noticeably taller than the pendulum of the hindlimb, the apex of which is the hip joint. This difference corresponds to the ~30%–40% of the trunk length or from ~50 cm in L. cyclotis to ~110 cm in M. trogontherii in linear dimensions. It can be hypothesized that a markedly taller forelimb is related to the increase in the size of the head and tusks during the proboscidean evolution, which shifts the body's center of mass forward. This is expressed by the peak ground reaction forces of the forelimb during the contact phase: in trotting horses they are 1.23 times higher than those of the hindlimb (Parmentier et al., 2023), in the galloping African buffalo (Syncerus caffer) 1.51 times higher, and in fast walking African elephant they are 1.79 times higher (Alexander et al., 1979). The length of the forelimb (including the scapula) is 1.18 times longer than that of the hindlimb at the hip joint in the domestic horse (on average 1.11 times taller in the equid sample), 1.25 times taller in the African buffalo, and 1.31 times taller in the African elephant. The shorter inverted pendulum of the hindlimb is compensated for by the higher pace angle (i.e., the angle by which the limb as a whole turns around the point of its contact with the ground) compared with the forelimb (50° vs. 36°; Gambaryan & Ruhkyan, 1974). The trade‐off of a higher pace angle is a slightly higher vertical fluctuation at the hip joint compared with the dorsal angle of scapula (Gambaryan, 1974; Gambaryan & Ruhkyan, 1974).
(4) The pillar‐shaped limbs of elephants have nearly straight joints. The more erect postures allow large mammals to reduce the joint moment arms of the ground reaction force (Biewener, 2005; Gambaryan, 1974). During locomotion, the angles between bones at the elephant's shoulder, elbow, wrist, hip, and knee joints rarely fall below 120° (Gambaryan, 1974; Ren et al., 2008). Moreover, the most straightened position of the elephants' joints was observed in the contact phase, and the bent position is characteristic of the limb in the swing phase. Joints that function in an almost straightened position are vulnerable to unintended hyperextension and therefore require reliable static and dynamic muscular control – especially for the ball‐and‐socket type joints, such as the shoulder and the hip.
(5) The enormous scapula and pelvis of elephantids are associated with relative and absolute strengthening of the shoulder/elbow and hip/knee joint muscles. So, the relative mass of the hip and elbow joint extensors in elephants corresponds to the maximum values among the artiodactyls and perissodactyls (>60% of the total mass of the hindlimb and >20% of the total mass of the forelimb, respectively; Gambaryan, 1974; Gambaryan & Ruhkyan, 1974). In addition to increasing mass and functional cross‐section, some elephant muscles reappear as pinnate muscles. Thus, the main extensor of the knee joint m. vastus lateralis in elephants is not only heavier than in any artiodactyl and perissodactyl (10.3%–11.8% of the hindlimb muscles mass vs. 4.5%–9%; Gambaryan, 1974) but is also five‐pinnate (Gambaryan, 1957). As a result, this muscle is not only very efficient for low‐amplitude motions but also incredibly powerful. Some other muscles, in contrast, reappear into pinnate‐ligament muscles, becoming relatively lighter and performing mostly static functions. This is the case with the m. infraspinatus, which is responsible for abduction and counteracting hyperadduction in the shoulder joint (Drujinin, 1941a, 1941b).
(6) The shape of the limb bones also fulfills the function of effective muscle control. The scapula of an elephant is exceptionally long and wide, its caudal angle expanding backward, while its robust scapular spine is accompanied by an almost boomerang‐shaped acromion/metacromion process (Figure 6). This creates better levers for controlling unintended shoulder movements. No wonder the muscles that are responsible for preventing unintended hyperextension in the shoulder joint are attached to the caudal angle of the scapula (m. anconeus longus, m. teres major, m. spinodeltoideus) and acromion/metacromion processes (m. acromiodeltoideus; Gambaryan, 1974; Gambaryan & Ruhkyan, 1974), maximizing their distance from the joint creating optimal leverage for control. All these muscles are relatively heavier in elephants than in any odd‐toed and even‐toed ungulate (Gambaryan, 1974; Gambaryan & Ruhkyan, 1974).
(7) Elephants have a remarkable adaptation at the most distal point of their inverted pendulum as well. Their manus and pes anatomy feature huge digital cushions that markedly increase the foot's support area (Gambaryan, 1974; Gambaryan & Ruhkyan, 1974). In lighter ungulates (e.g., tapirs), most loads are concentrated directly on the hooves, which are in contact with the ground (Michilsens et al., 2009). The increased size of the digital cushion in megaherbivores allows to offset the increase in body mass, resulting in foot loads in large ungulates virtually identical to their smaller relatives (Warner et al., 2013). The short muscles of the manus and pes in elephants are well developed, together with the ligamentous apparatus, plantar aponeurosis, expanded sesamoids (that form massive “predigits”), and fat pad that support the skeletal elements (Gambaryan & Ruhkyan, 1974; Hutchinson et al., 2011). These adaptations allow them to lower peak foot pressures that occur at the beginning of the contact phase, distributing the load from the individual bones of the manus and pes to the digital cushion.
(8) Another feature that we want to focus on is the ulna as the main support bone of the antebrachium in elephantids in contrast to the radius in even‐toed and odd‐toed ungulates. Remains of the Paleogene N. koholense (Court, 1994) indicate that the ulna became the main support bone of the antebrachium in proboscideans at the early stages of their evolution. A possible explanation for this phenomenon is the distribution of the load on the foot during locomotion. In elephants, the load on the lateral part of the foot is slightly higher than on the medial part of the foot, while in contrast, in horses and rhinoceroses, the higher load is either on the medial part of the foot or evenly distributed (Panagiotopoulou et al., 2012, 2019). This slight difference in load distribution along the forelimb may result in the primarily radial or ulnar support in the antebrachium. Another example of the ulna as a main support bone in antebrachium in a large herbivorous mammal is presented in the afrotherian Arsinoitherium (Andrews, 1906; Figure 6).
To summarize, how effective is the inverted pendulum as a limb type in elephantids? Elephants are capable of rapid acceleration over short distances, reaching speeds of 25 km/h under experimental conditions (Hutchinson et al., 2003, 2006) and unconfirmed reports of speeds up to 40–50 km/h (Gambaryan & Ruhkyan, 1974; Howell, 1944). In contrast to other megaherbivores, the elephants are capable of long seasonal migrations (annual migration up to 500 km in Gourma elephants) and daytime travel over long distances (up to 60 km per day; Wilson & Mittermeier, 2011). These animals are capable of living in a variety of environments, moving up in the highlands of the Himalayas (up to 3000 m), and overcoming water obstacles, including the ability to swim between islands (Choudhury, 1999; Wilson & Mittermeier, 2011). Their extinct ancestors occupied all continents except Australia and Antarctica, thrived in different environments including the mammoth steppe, and left behind an unprecedented fossil record (Osborn, 1936, 1942).
The relatively taller limbs of mammoths, with enlarged proximal segments compared with modern‐day elephants may indicate a further increase in mass and strength of proximal joints musculature. This reinforcement may be related to the peculiarities of a cool climate with pronounced seasonality. First, the warm and cold seasons of the year may have required mammoths to migrate even more distance in search of seasonal pastures, making a relatively taller pendulum more beneficial. Second, the snow cover and ice crust in winter, mud, and soft ground in spring and fall could have led to additional muscular strengthening and stabilization of proximal limb joints (primarily shoulder and hip) against unintended hyperextension.
The second strategy of the limbs elongation was realized in odd‐toed and even‐toed ungulates with limbs elongated mostly due to the distal segments (Gambaryan, 1974; Gregory, 1912). Thus, based on measurements from Wood et al. (2011) and a 3D model of the skeleton of the small Paleogene equoid A. grangeri from UMORF website (https://umorf.ummp.lsa.umich.edu/wp/specimen‐data/?Model_ID=1675), and measurements of the small Paleogene rhinocerotoid H. eximius from Cope (1873a), we can follow the changes in body plan in perissodactyls. Modern‐day tapirs and rhinoceroses are similar to Paleogene perissodactyls in their short‐necked and short‐legged appearance (Tables 2 and 3; Figures 3a and 6e,h). However, in contrast to extant representatives of the order, the relative length of the head of the Paleogene perissodactyls was ~1.5 times shorter, the ribs ~1.5–2 times shorter, and the pelvis was notably smaller in every measured dimension. The relative height of the hindlimb of the plesiomorphic odd‐toed ungulates notably exceeded the height of the forelimb (by ~8% in H. eximius and by ~18% in A. grangeri), which is consistent with the more dorsomobile mode of locomotion (Belyaev et al., 2023; Preuschoft & Franzen, 2012). In the relative height, compared with the extant tapirs and rhinoceroses, the Paleogene perissodactyls are characterized by a notably shorter scapula (by 5%–10%, Table 2) and ilium (by 3%–10%, Table 3), and notably longer crus and pes (~2%–10%, Table 3). In addition to differences in limb segment ratios, Paleogene odd‐toed ungulates showed higher diversity in the morphology of virtually all parts of the limb skeleton and were noticeably different in their appearance from modern‐day tapirs (MacLaren & Nauwelaerts, 2020). The most apomorphic limb skeleton among the odd‐toed ungulates is characteristic of the equids; in addition to such well‐known morphological features as a reduction in the number of digits in the manus and pes (Kovalevsky, 1873; Osborn, 1910), these animals are characterized by a significant limb's elongation (Gambaryan, 1974). This occurs primarily due to an increase in the relative length of the most distal limb segments—the manus (~2 times) and pes (~1.5 times)—compared to the Paleogene A. grangeri (Tables 2 and 3; Figure 6e,f).
Before considering the question of the biomechanical benefit of distal limb segment elongation, it is important to note two points: (1) unlike in proboscideans, where the elongation of limb segments occurs alongside a drastic increase in body size, the elongated distal limb segments in modern‐day large even‐toed ungulates and equids are inherited from their smaller Paleogene and Neogene ancestors; (2) the distal limb segments in larger ungulates are usually relatively shorter than in smaller ones (Gambaryan, 1974; Grеgоrу, 1912). Elongating the distal segments while maintaining the length (or even shortening) of the proximal segments allows the shift of the limb's joints and muscles controlling their movements proximally. This results in a heavier upper half of the limb and a lighter lower half. Compared with a limb that is elongated saving proportions of segments and respective mass distribution, a limb elongated due to the distal segments allows for a lower moment of inertia of the pendulum (in this case simple pendulum, not inverted) during the swing phase and makes the natural period of oscillation shorter. This allows to keep low energy expenditure during the swing phase and to maintain or even increase stride frequency, enabling faster running speed. However, comparing segment inputs to the overall leg length, the shortening of the femur on the proximal end in ungulates does not go as far as the shortening of the pes on the distal end in proboscideans. Also, even in the heaviest ungulates, the legs do not approach the shape of a straight column as closely as they do in proboscideans. Thus, the legs of ungulates never deviate as much towards the inverted pendulum from the basic therian structure characterized by three‐segment legs crouched in a Z‐like zigzag. The three segments are the scapula, humerus and antebrachium+manus in the forelimb, and the femur, crus and pes in the hindlimb (Fischer, 1998; Inuzuka, 1992; Kuznetsov, 1985, 2020). The benefit of the three‐segment limbs is the ability to optimize joint kinematics and thus reduce the mechanical work of muscles against each other within the leg in the contact phase, the advantage being most substantial when running with asymmetric gaits, such as the gallop (Kuznetsov, 1995, 1999). Overall, the locomotor strategy of ungulates (and carnivorans), keeping the Z‐like leg structure for economic galloping, is quite opposite to the locomotor strategy of proboscideans (and humans), whose legs adopted the principle of inverted‐pendulum for economic walking.
6. CONCLUSION
In this study, we characterized the unique body plan of the Elephantidae, which is markedly different from any artiodactyls and perissodactyls. Elephants have a very tall appearance, second only to giraffes. Their limbs are elongated due to proximal segments, almost straight at the joints. This allows them to effectively utilize the principle of inverted pendulum (straight‐legged walking) in locomotion. Due to the proboscis, elephantids break the interrelationship observed in ungulates between the total length of the head and neck on one side and the limb's segments on the other.
The body plan of the woolly and steppe mammoths differed in some aspects from their extant relatives. The skull proportions vary markedly between mammoths and extant elephantids. Mammoths' skulls are relatively longer and wider at the eye sockets, while the width at the occiput in elephantids has a large variability and overlap. The longer skull in mammoths is associated with the longer and heavier tusks. The limbs of Mammuthus primigenius were relatively taller due to the increased relative length of their proximal segments. One notable difference in body plan among the studied elephantids is the relatively larger size of the mammoth's pelvis. So, the relative maximal width of the pelvis in the woolly and steppe mammoths is 1.3–1.4 times wider than those of the extant elephants. Among other proboscideans, the American mastodon was characterized by a surprisingly similar to the elephantids' exterior. However, compared with extant elephants their skull is notably lower, limbs are shorter, and pelvis is 1.3‐1.4 times wider. The amebelodontid Megabelodon lulli has a long skull and relatively lower limbs.
The body plan of plesiomorphic Paleogene proboscideans and perissodactyls differed from their extant descendants in every body part. These differences are related to the allometric growth and the advancement of the locomotor apparatus. Thus, the relative length of the lumbar region of the equoid Arenahippus grangeri was one‐third longer than those of the extant equids, and the lumbar region of the rhinocerotoid Hyrachyus eximius was almost three times longer than those of modern‐day rhinoceroses. The cervical region of Paleogene perissodactyls was slightly shorter compared with the modern‐day tapirs and rhinoceroses, and more than twice shorter than in extant equids. Counterintuitively, the body plan of elephants is even more different from the Paleogene proboscidean Moeritherium than it is from modern‐day artiodactyls and perissodactyls. Compared with Moeritherium, Elephantidae has limbs approximately two times taller (relative to their trunk length), and markedly larger skulls. In contrast, the proportions of the vertebral regions between Paleogene and extant proboscideans are almost similar. This indicates that dorsostability appeared early in the proboscidean's evolution.
AUTHOR CONTRIBUTIONS
R.B., G.B., M.R., and N.P. worked with the material, R.B. and G.B. conceptualized the study and wrote the initial manuscript, R.B. analyzed the data, prepared tables and figures. All authors edited, read, and approved the final version of the manuscript.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Supporting information
Table S1. Studied material.
Table S2. PCA loadings between PC1/PC2 and nineteen body parts in a studied sample.
ACKNOWLEDGMENTS
We deeply appreciate Ekaterina Petrova, Gennady Baryshnikov, and Anton Rezvyi (ZIN) for helping us measure the mammoth skeletons; Sergey Sukhovey for support in measurements of the African elephants. We are thankful to Andrey Lissovsky (IEE), Ekaterina Obolenskaya, Natalia Spasskaya, Sergey Kruskop, Anna Tikhomirova (ZMMU), Emmanuel Gilissen (RMCA), Galina Timonina (AMZ), and Vadim Titov for access to osteological collections. Finally, we also acknowledge two anonymous reviewers for the valuable advice and recommendation of additional material, which greatly improved this study. The study was supported by the Russian Science Foundation (grant no. 23‐24‐00569, https://rscf.ru/project/23‐24‐00569/).
Belyaev, R.I. , Boeskorov, G.G. , Kuznetsov, A.N. , Rotonda, M. & Prilepskaya, N.E. (2025) Comparative study of the body proportions in Elephantidae and other large herbivorous mammals. Journal of Anatomy, 246, 63–85. Available from: 10.1111/joa.14143
Footnotes
Mediportal form of running is typical for large and heavy ungulates (e.g., bovines, rhinocerotids). The vertical fluctuations of the center of gravity of the body are reduced compared with cursorial running form. The limb contact phase is characterized by decreased limb joint amplitudes and decreased pace angle (Gambaryan, 1974). Sagittal flexibility in the lumbar region is low, backbone significantly less involved as an active link during gallop (Belyaev et al., 2021).
Contributor Information
Ruslan I. Belyaev, Email: belyaev.ruslan@gmail.com.
Alexander N. Kuznetsov, Email: alexanderk@technion.ac.il.
DATA AVAILABILITY STATEMENT
The data that supports the findings of this study are available in the supplementary material of this article.
REFERENCES
- Alexander, R.M. (1993) The relative merits of foregut and hindgut fermentation. Journal of Zoology (London), 231, 391–401. Available from: 10.1111/j.1469-7998.1993.tb01927.x [DOI] [Google Scholar]
- Alexander, R.M. , Dimery, N.J. & Ker, R.F. (1985) Elastic structures in the back and their role in galloping in some mammals. Journal of Zoology (London), 207, 467–482. [Google Scholar]
- Alexander, R.M. , Maloiy, G.M.O. , Hunter, B. , Jayes, A.S. & Nturibi, J. (1979) Mechanical stresses in fast locomotion of buffalo (Syncerus caffer) and elephant (Loxodonta africana). Journal of Zoology, 189, 135–144. Available from: 10.1111/j.1469-7998.1979.tb03956.x [DOI] [Google Scholar]
- Alexander, R.M. & Pond, C.M. (1992) Locomotion and bone strength of the white rhinoceros, Ceratotherium simum . Journal of Zoology, 227, 63–69. Available from: 10.1111/j.1469-7998.1992.tb04344.x [DOI] [Google Scholar]
- Andrews, C.W. (1906) A Descriptive Catalogue of the Tertiary Vertebrata of the Fayûm, Egypt. Based on the Collection of the Egyptian Government in the Geological Museum, Cairo, and on the Collection in the British Museum (Natural History). London: Natural History Museum, London, p. 324. [Google Scholar]
- Averianov, A.O. (1996) Sexual dimorphism in the mammoth skull, teeth, and long bones. In: Shoshani, J. & Tassu, P. (Eds.) Proboscidea. Oxford: Oxford University Press, pp. 260–267. [Google Scholar]
- Averianov, A.O. , Baryshnikov, G.F. , Garutt, V.E. , Garutt, N.V. & Fomocheva, N.L. (1992) The Volga mammalian fauna in the geological‐mineralogical museum of Kazan university. Kazan City: Kazan university Publ. house, p. 161 [in Russian]. [Google Scholar]
- Baygusheva, V.S. , Titov, V.V. & Timonina, G.I. (2012) Two skeletons of Mammuthus trogontherii from the sea of Azov region. Quaternary International, 276–277, 242–252. Available from: 10.1016/j.quaint.2011.05.045 [DOI] [Google Scholar]
- Belyaev, R.I. , Kuznetsov, A.N. & Prilepskaya, N.E. (2021) How the even‐toed ungulate vertebral column works: comparison of intervertebral mobility in 33 genera. Journal of Anatomy, 239, 1370–1399. Available from: 10.1111/joa.13521 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Belyaev, R.I. , Kuznetsov, A.N. & Prilepskaya, N.E. (2022) From dorsomobility to dorsostability: a study of lumbosacral joint range of motion in artiodactyls. Journal of Anatomy, 241, 420–436. Available from: 10.1111/joa.13688 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Belyaev, R.I. , Kuznetsov, A.N. & Prilepskaya, N.E. (2023) Truly dorsostable runners: vertebral mobility in rhinoceroses, tapirs, and horses. Journal of Anatomy, 242, 568–591. Available from: 10.1111/joa.13799 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Belyaev, R.I. , Nikolskaia, P. , Bushuev, A.V. , Panyutina, A.A. , Kozhanova, D.A. & Prilepskaya, N.E. (2024) Running, jumping, hunting, and scavenging: functional analysis of vertebral mobility and backbone properties in carnivorans. Journal of Anatomy, 244(2), 205–231. Available from: 10.1111/joa.13955 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Biewener, A.A. (2005) Biomechanical consequences of scaling. Journal of Experimental Biology, 208(9), 1665–1676. [DOI] [PubMed] [Google Scholar]
- Boeskorov, G. , Tikhonov, A. & Suzuki, N. (2007) Yukagir Mammoth. Saint Petersburg: Saint‐Petersburg University Publishing House. [Google Scholar]
- Boeskorov, G.G. (2021) Body dimensions of the woolly mammoth Mammuthus primigenius (Blumenbach) of the second half of the late Pleistocene in the north of eastern Siberia. Natural Resources of the Arctic and Subarctic, 26(1), 38–48. [Google Scholar]
- Boeskorov, G.G. , Cherkashina, A.P. , Belolyubsky, I.N. & Zaitsev, A.I. (2009) Features of the morphology and paleoecology of the Chekurovsky mammoth. Russian Geology, 5, 84–90. [Google Scholar]
- Boeskorov, G.G. , Tikhonov, A.N. , Shchelchkova, M.V. , Ballard, J.P. & Mol, D. (2020) Big tuskers: maximum sizes of tusks in woolly mammoths – Mammuthus primigenius (Blumenbach) – from East Siberia. Quaternary International, 537, 88–96. Available from: 10.1016/j.quaint.2019.12.023 [DOI] [Google Scholar]
- Carroll, R.L. (1993) Vertebrate Paleontology and Evolution.
- Cavagna, G.A. , Thys, H. & Zamboni, A. (1976) The sources of external work in level walking and running. The Journal of Physiology, 262(3), 639–657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choudhury, A. (1999) Status and conservation of the Asian elephant Elephas maximus in north‐eastern India. Mammal Review, 29, 141–174. Available from: 10.1046/j.1365-2907.1999.00045.x [DOI] [Google Scholar]
- Christiansen, P. (2004) Body size in proboscideans, with notes on elephant metabolism. Zoological Journal of the Linnean Society, 140, 523–549. [Google Scholar]
- Clayton, H.M. & Townsend, H.G.G. (1989) Cervical spinal kinematics: a comparison between foals and adult horses. Equine Veterinary Journal, 21(3), 193–195. [DOI] [PubMed] [Google Scholar]
- Cook, H.J. (1928) A new longirostrine elephant from Nebraska Trilophodon phippsi, sp.nov. Proceedings of the Colorado Museum of Natural History, 8(4), 37–45. [Google Scholar]
- Cope, E.D. (1873a) On the osteology of the extinct tapiroid Hyrachyus . Proceedings of the American Philosophical Society, 13(90), 212–224. [Google Scholar]
- Cope, E.D. (1873b) On the short footed Ungulata of the Eocene of Wyoming. Proceedings of the American Philosophical Society, 13(90), 38–74. [Google Scholar]
- Coughlin, B.L. & Fish, F.E. (2009) Hippopotamus underwater locomotion: reduced‐gravity movements for a massive mammal. Journal of Mammalogy, 90(3), 675–679. Available from: 10.1644/08-MAMM-A-279R.1 [DOI] [Google Scholar]
- Court, N. (1994) Limb posture and gait in Numidotherium koholense, a primitive proboscidean from the Eocene of Algeria. Zoological Journal of the Linnean Society, 111(4), 297–338. Available from: 10.1006/zjls.1994.1026 [DOI] [Google Scholar]
- Сuviеr, G. & Laurillard, C.L. (1850) Anatomie comparée, Vol. 4. Paris: Imprimerie d‘E. Duverger. [Google Scholar]
- Drujinin, A.N. (1941a) Morphological‐functional analysis of the muscles of the fore‐girdle and limbs in the Indian elephant. In: Schmalhausen, I.I. (Ed.) Proceedings in memory of a N. Severtsov, Vol. 2. Moscow: Publishing house of the Russian Academy of Sciences, pp. 209–277. [Google Scholar]
- Drujinin, A.N. (1941b) Morphological‐functional analysis of the muscles of the fore‐girdle and limbs in the Indian elephant. In: Schmalhausen, I.I. (Ed.) Proceedings in memory of a N. Severtsov, Vol. 2. Moscow: Publishing house of the Russian Academy of Sciences, pp. 279–298. [Google Scholar]
- Fischer, M.S. (1998) Die lokomotion von Procavia capensis (Mammalia, Hyracoidea): zur evolution des bewegungssystems bei saugetieren. Abhandlungen des Naturwissenschaftlichen Vereins in Hamburg, 33, 1–188. [Google Scholar]
- Fortelius, M. , Eronen, J. , Liu, L. , Pushkina, D. , Tesakov, A. , Vislobokova, I. et al. (2006) Late Miocene and Pliocene large land mammals and climatic changes in Eurasia. Palaeogeography Palaeoclimatology Palaeoecology, 238(1–4), 219–227. [Google Scholar]
- Gambaryan, P.P. (1957) On the function of the pinnate muscles. Proceedings of the Armenian SSR Academy of Sciences, 25, 87–91. [Google Scholar]
- Gambaryan, P.P. (1964) Morphofunctional analysis of the musculature of extremities of Tapirus americanus . Zoological Papers of the Armenian SSR Academy of Sciences, 13, 5–50. [Google Scholar]
- Gambaryan, P.P. (1974) How animals run: anatomical adaptations. New York: John Wiley. [Google Scholar]
- Gambaryan, P.P. & Ruhkyan, R.G. (1974) Morpho‐functional analysis of the muscles of the limbs of elephants. Proceedings of the Zoological Institute of the USSR Academy of Sciences, 54, 190–265. [Google Scholar]
- Garutt, V.E. (1964) Das Mammut Mammuthus primigenius (Blumenbach). Zeimsen Verlag. Wittenberg Lutherstadt, 1964, 140. [Google Scholar]
- Genin, J.J. , Willems, P.A. , Cavagna, G.A. , Lair, R. & Heglund, N.C. (2010) Biomechanics of locomotion in Asian elephants. Journal of Experimental Biology, 213(5), 694–706. Available from: 10.1242/jeb.035436 [DOI] [PubMed] [Google Scholar]
- Grеgоrу, W.К. (1912) Notes on the principles of quadrupedal locomotion and on the mechanism of the limbs in hoofed animals. Annals of the New York Academy of Sciences, 22, 267–294. [Google Scholar]
- Haussler, K.K. , Bertram, J.E.A. , Gellman, K. & Hermanson, J.W. (2001) Segmental in vivo vertebral kinematics at the walk, trot and canter: a preliminary study. Equine Veterinary Journal, 33, 160–164. [DOI] [PubMed] [Google Scholar]
- Howell, A.B. (1944) Speed in animals: their specialization for running and leaping. Chicago: University of Chicago Press. [Google Scholar]
- Huang, S. , Eyres, A. , Fritz, S. , Eronen, J. & Saarinen, J. (2023) Chapter 6 ‐ environmental change and body size evolution in Neogene large mammals. In: Casanovas‐Vilar, I. , van den Hoek‐Ostende, L. , Janis, C. & Saarinen, J. (Eds.) Evolution of Cenozoic land mammal faunas and ecosystems—25 years of the NOW database of fossil mammals. Cham: Springer. [Google Scholar]
- Hutchinson, J.R. , Delmer, C. , Miller, C.E. & Boyde, A. (2011) From flat foot to fat foot: structure, ontogeny, function, and evolution of elephant “sixth toes”. Science, 334, 1699–1703. [DOI] [PubMed] [Google Scholar]
- Hutchinson, J.R. , Famini, D. , Lair, R. & Kram, R. (2003) Are fast‐moving elephants really running? Nature, 422(6931), 493–494. [DOI] [PubMed] [Google Scholar]
- Hutchinson, J.R. , Schwerda, D. , Famini, D.J. , Dale, R.H. , Fischer, M.S. & Kram, R. (2006) The locomotor kinematics of Asian and African elephants: changes with speed and size. The Journal of Experimental Biology, 209(19), 3812–3827. Available from: 10.1242/jeb.02443 [DOI] [PubMed] [Google Scholar]
- Inuzuka, N. (1992) Evolution of the shoulder girdle with special reference to the problems of the clavicle. Journal of the Anthropological Society of Nippon, 100(4), 391–404. Available from: 10.1537/ase1911.100.391 [DOI] [Google Scholar]
- Kovalevsky, V.O. (1873) Sur l'Anchitherium aurelianense Cuv. et sur l'histoire paleontologique des chevaux. In: Memoires de l'Academie des sciencer de St‐Petersbourg. Saint‐Petersburg: Typography of Imperial Academy of Sciences. [Google Scholar]
- Kuznetsov, A.N. (1985) Comparative functional analysis of the fore and hind limbs in mammals. Zoologicheskiĭ Zhurnal, 64(12), 1862–1867. [Google Scholar]
- Kuznetsov, A.N. (1995) Energetical profit of the third segment in parasagittal legs. Journal of Theoretical Biology, 172(1), 95–105. Available from: 10.1006/jtbi.1995.0007 [DOI] [Google Scholar]
- Kuznetsov, A.N. (1999) Structural types of limbs and evolution of running technique in tetrapods. Zoologicheskie Issledovania, 3, 1–94 [in Russian]. [Google Scholar]
- Kuznetsov, A.N. (2020) Reconstructing serial homology with a special emphasis on the nature of limbs in vertebrates and with references to Cruveilhier, Wyman, Belogolowy, and Balinsky. Journal of Morphology, 281(12), 1617–1627. Available from: 10.1002/jmor.21274 [DOI] [PubMed] [Google Scholar]
- Larramendi, A. (2016) Shoulder height, body mass, and shape of proboscideans. Acta Palaeontologica Polonica, 61(3), 537–574. [Google Scholar]
- Larramendi, A. (2023) Estimating tusk masses in proboscideans: a comprehensive analysis and predictive model. Historical Biology, 1–14, 1–14. Available from: 10.1080/08912963.2023.2286272 [DOI] [Google Scholar]
- Larramendi, A. , Palombo, M.R. & Marano, F. (2017) Reconstructing the life appearance of a Pleistocene giant: size, shape, sexual dimorphism and ontogeny of Palaeoloxodon antiquus (Proboscidea: Elephantidae) from Neumark‐Nord 1 (Germany). Bollettino Della Societa Paleontologica Italiana, 56(3), 299–317. [Google Scholar]
- Lazarev, P.A. (2008) Large mammals of the anthropogene of Yakutia. Novosibirsk: Nauka, p. 2160. [Google Scholar]
- Lister, A.M. (2013) The role of behaviour in adaptive morphological evolution of African proboscideans. Nature, 500, 331–334. [DOI] [PubMed] [Google Scholar]
- Lister, A.M. & Sher, A.V. (2001) The origin and evolution of the woolly mammoth. Science, 294(5544), 1094–1097. Available from: 10.1126/science.1056370 [DOI] [PubMed] [Google Scholar]
- Liu, A.G. , Seiffert, E.R. & Simons, E.L. (2008) Stable isotope evidence for an amphibious phase in early proboscidean evolution. PNAS, 105(15), 5786–5791. Available from: 10.1073/pnas.0800884105 [DOI] [PMC free article] [PubMed] [Google Scholar]
- MacLaren, J.A. & Nauwelaerts, S. (2020) Modern tapirs as Morphofunctional analogues for locomotion in endemic Eocene European Perissodactyls. Journal of Mammalian Evolution, 27, 245–263. Available from: 10.1007/s10914-019-09460-1 [DOI] [Google Scholar]
- Maglio, V.J. (1973) Origin and evolution of the Elephantidae. Transactions of the American Philosophical Society, 63, 1–149. [Google Scholar]
- Michilsens, F. , Aerts, P. , Van Damme, R. & D'Août, K. (2009) Scaling of plantar pressures in mammals. Journal of Zoology, 279(3), 236–242. [Google Scholar]
- Mihlbachler, M.C. , Rivals, F. , Solounias, N. & Semprebon, G.M. (2011) Dietary change and evolution of horses in North America. Science, 331(6021), 1178–1181. Available from: 10.1126/science.1196166 [DOI] [PubMed] [Google Scholar]
- Muybridge, E. (1887) Animal locomotion: an electrophotographic investigation of consecutive phases of animal movements. Philadelphia: University of Pennsylvania. [Google Scholar]
- Nagel, R.M. , Forstenpointner, G. , Soley, J.T. & Weissengruber, G.E. (2018) Muscles and fascial elements of the antebrachium and manus of the African elephant (Loxodonta africana, Blumenbach 1797): starring comparative and functional considerations. Anatomia, Histologia, Embryologia, 47(3), 195–205. Available from: 10.1111/ahe.12340 [DOI] [PubMed] [Google Scholar]
- Nikolskiy, P.A. , Basilyan, A.E. , Sulerzhitsky, L.D. & Pitulko, V.V. (2010) Prelude to the extinction: revision of the Achchagyi–Allaikha and Berelyokh mass accumulations of mammoth. Quaternary International, 219(1), 16–25. [Google Scholar]
- Osborn, H.F. (1910) The age of mammals in Europe, Asia and North America. New York: The Macmillan Company, p. 635. [Google Scholar]
- Osborn, H.F. (1936) Proboscidea: a monograph on the discovery, evolution, migration and extinction of the Mastodonts and elephants of the world, I: Moeritherioidea, Deinotherioidea, Mastodontoidea. New York: The American Museum Press, p. 802. [Google Scholar]
- Osborn, H.F. (1942) Proboscidea: a monograph on the discovery, evolution, migration and extinction of the Mastodonts and elephants of the world. In: Stegodontoidea and Elephantoidea, Vol. II. New York: The American Museum Press, p. 870. [Google Scholar]
- Panagiotopoulou, O. , Pataky, T.C. , Day, M. , Hensman, M.C. , Hensman, S. , Hutchinson, J.R. et al. (2016) Foot pressure distributions during walking in African elephants (Loxodonta africana). Royal Society Open Science, 3, 160203. Available from: 10.1098/rsos.160203 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Panagiotopoulou, O. , Pataky, T.C. , Hill, Z. & Hutchinson, J.R. (2012) Statistical parametric mapping of the regional distribution and ontogenetic scaling of foot pressures during walking in Asian elephants (Elephas maximus). Journal of Experimental Biology, 215, 1584–1593. [DOI] [PubMed] [Google Scholar]
- Panagiotopoulou, O. , Pataky, T.C. & Hutchinson, J.R. (2019) Foot pressure distribution in white rhinoceroses (Ceratotherium simum) during walking. PeerJ, 7, e6881. Available from: 10.7717/peerj.6881 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Parmentier, J.I.M. , Bosch, S. , van der Zwaag, B.J. , Weishaupt, M.A. , Gmel, A.I. , Havinga, P.J.M. et al. (2023) Prediction of continuous and discrete kinetic parameters in horses from inertial measurement units data using recurrent artificial neural networks. Scientific Reports, 13, 740. Available from: 10.1038/s41598-023-27899-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Preuschoft, H. & Franzen, J.L. (2012) Locomotion and biomechanics in Eocene mammals from Messel. Palaeobiodiversity and Palaeoenvironments, 92, 459–476. Available from: 10.1007/s12549-012-0103-7 [DOI] [Google Scholar]
- R Core Team . (2023) R: a language and environment for statistical computing. R Core Team: R Foundation for Statistical Computing, Vienna, Austria, p. 2023. [Google Scholar]
- Ranjan, P.B. , Titov, V.V. , Patnaik, R. & Tesakov, A.S. (2023) Progress in comparison of Elephantidae Fauna based on dental study between the Plio‐Pleistocene of the south of Eastern Europe and the upper Siwaliks of India. Stratigraphy and Geological Correlation, 31, 721–736. Available from: 10.1134/S0869593823060102 [DOI] [Google Scholar]
- Ren, L. , Butler, M. , Miller, C. , Paxton, H. , Schwerda, D. , Fischer, M.S. et al. (2008) The movements of limb segments and joints during locomotion in African and Asian elephants. The Journal of Experimental Biology, 211(17), 2735–2751. Available from: 10.1242/jeb.018820 [DOI] [PubMed] [Google Scholar]
- Ren, L. & Hutchinson, J.R. (2007) The three‐dimensional locomotor dynamics of African (Loxodonta africana) and Asian (Elephas maximus) elephants reveal a smooth gait transition at moderate speed. Journal of the Royal Society Interface, 5(19), 195–211. Available from: 10.1098/rsif.2007.1095 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Saarinen, J. , Cirilli, O. , Strani, F. , Meshida, K. & Bernor, R. (2021) Testing equid body mass estimate equations on modern zebras—with implications to understanding the relationship of body size, diet and habitats of Equus in the Pleistocene of Europe. Frontiers in Ecology and Evolution, 9, 622412. Available from: 10.3389/fevo.2021.622412 [DOI] [Google Scholar]
- Saarinen, J. , Eronen, J. , Fortelius, M. , Seppä, H. & Lister, A.M. (2016) Patterns of diet and body mass of large ungulates from the Pleistocene of Western Europe, and their relation to vegetation. Palaeontologia Electronica, 19(2), 1–58. Available from: 10.26879/443 [DOI] [Google Scholar]
- Scott, W.B. , Jepsen, G.L. & Wood, A.E. (1941) The mammalian Fauna of the White River Oligocene: part V. Perissodactyla . Transactions of the American Philosophical Society, New Series, 28(5), 747–980. [Google Scholar]
- Shoshani, J. & Eisenberg, J.F. (1982) Elephas maximus. Mammalian Species, No, 182, 1–8. [Google Scholar]
- Sokolov, V.E. (1979) European bison: morphology, systematics, evolution, ecology. Moscow: Nauka, p. 496. [Google Scholar]
- Tikhonov, A.N. (1992) Mammoth, Mammuthus primigenius (Elephantidae, Proboscidae) from Bolshoy Lyakhovsky Island, eastern Siberia. Proceedings of the Zoological Institute of the Russian Academy of Sciences, 270, 167–185. [Google Scholar]
- Tilesio, W.G. (1815) De skeleto mammonteo Sibirico ad maris glacialis littora anno 1807 effosso, cui praemissae Elephantini generis specierum distinctiones. Mémoires de l'Académie Impériale des Sciences de St. Pétersbourg, 5, 406–514. [Google Scholar]
- Vereshchagin, N.K. & Tikhonov, A.N. (1990) Exterior of a mammoth (p. 40). Yakutsk: Publishing House IM SO AN USSR. [Google Scholar]
- Walker, E. , Warnick, F. , Lange, K. , Uible, H. , Hamlet, S. , Davies, M. et al. (1968) Mammals of the world, 2nd edition. Baltimore: John Hopkins Press, p. 1500. [Google Scholar]
- Warner, S.E. , Henry, V. , Roskilly, K. , Hildebrandt, T. , Panagiotopoulou, O. & Hutchinson, J.R. (2014) Regional variation in digital cushion pressure in the forefeet of horses and elephants. PeerJ, 2, e231v1. Available from: 10.7287/peerj.preprints.231v1 [DOI] [Google Scholar]
- Warner, S.E. , Pickering, P. , Panagiotopoulou, O. , Pfau, T. , Ren, L. & Hutchinson, J.R. (2013) Size related changes in foot impact mechanics in hoofed mammals. PLoS One, 8(1), e54784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Warren, J.C. (1852) The mastodon giganteus of North America. Boston: John Wilson and Son. [Google Scholar]
- Wilson, D.E. & Mittermeier, R.A. (2011) Handbook of the mammals of the world – volume 2: hoofed mammals. Barcelona: Lynx Ediciones. [Google Scholar]
- Wood, A.R. , Bebej, R.M. , Manz, C.L. , Begun, D.L. & Gingerich, P.D. (2011) Postcranial functional morphology of hyracotherium (Equidae, Perissodactyla) and locomotion in the earliest horses. Journal of Mammalian Evolution, 18, 1–32. Available from: 10.1007/s10914-010-9145-7 [DOI] [Google Scholar]
- Zalensky, V.V. (1909) Species zoological attributes of the mammoth. Scientific results of the expedition equipped by the imperatorial Academy of Sciences for the excavation of the mammoth found on the Berezovka River in 1901 . St. Petersburg, 2, 36–42. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Studied material.
Table S2. PCA loadings between PC1/PC2 and nineteen body parts in a studied sample.
Data Availability Statement
The data that supports the findings of this study are available in the supplementary material of this article.