Palaeotherium

Palaeotherium is the type genus of the extinct perissodactyl family Palaeotheriidae, a Palaeogene-exclusive lineage within the superfamily Equoidea that diverged from the extant Equidae (horses and relatives) by the Palaeocene to early Eocene. The genus lived in Europe and possibly the Middle East and ranged from the middle Eocene to the early Oligocene. Fossils of Palaeotherium were first described in 1782 by the French naturalist Robert de Lamanon and then closely studied by another French naturalist Georges Cuvier since 1798. He erected the genus name in 1804 and recognized multiple species up to 1824 based on overall fossil sizes and limb bone morphologies, although several were eventually reclassified to other perissodactyl genera by other naturalists. It was the fourth fossil mammal genus to be described with official taxonomic authority and is recognized as an important milestone within the field of palaeontology due to contributing to the developing ideas of evolution, extinction, and succession and demonstrating the morphological diversity of different species within one genus.

Since Cuvier's research efforts, many other naturalists from Europe and the Americas recognized many species of Palaeotherium, some valid, some reclassified to different genera afterward, and others being eventually rendered invalid. In particular, the German palaeontologist Jens Lorenz Franzen, as part of his dissertation in 1968, modernized its taxonomy due to his recognition of many subspecies, which were subsequently accepted by other palaeontologists. Today, there are fourteen known species recognized, many of which have multiple subspecies. In 1992, the French palaeontologist Jean-Albert Remy recognized two subgenera that most species are classified to based on cranial anatomies: the specialized Palaeotherium and the more generalized Franzenitherium.

Palaeotherium is a derived member of its family with tridactyl (or three-toed) forelimbs and hindlimbs, small post-canine diastemata, and premolars that are usually developed into molar-like forms. It shares many similar anatomical traits to other perissodactyls and had a large diversity in anatomical traits by species, with some species like P. magnum, P. curtum, and P. crassum being stockier in build and P. medium being more cursorial (or adapted for running). The genus ranges in size from the small species P. lautricense, with an estimated weight of 36 kg, to the massive P. giganteum, thought to have been capable of weighing over 700 kg. P. magnum, known by two mostly complete skeletons from France, could have reached approximately 1.3 m in shoulder height and 2.52 m in length. The large-sized species were therefore amongst the largest mammals in the Eocene of Europe. Palaeotherium may have lived in herds and, as demonstrated by its dentition, was able to actively niche partition with another palaeothere Plagiolophus by specializing on softer leaves and fruit, although both were mostly leaf-eating.

Palaeotherium and other genera of the subfamily Palaeotheriinae likely descended from the earlier subfamily Pachynolophinae, which lived in both Europe and Asia as opposed to North America unlike undisputed members of the Equidae. By the time that the first species P. eocaenum appeared in the middle Eocene, western Europe was an archipelago that was isolated from the rest of Eurasia, meaning that it and subsequent species lived in an environment with various other faunas that also evolved with strong levels of endemism. The Iberian Peninsula had its own level of endemism with several species that are only known within the region, although they were replaced by more widespread species from central Europe by the late Eocene. Within both the middle and late Eocene, Palaeotherium consistently maintained a high species diversity and endured major environmental changes leading to a faunal turnover that occurred by the beginning of the late Eocene.

By the early Oligocene, most of its species went extinct along with many genera of western European mammals as part of the Grande Coupure extinction and faunal turnover event, the causes of the extinctions being attributed mainly to environmental changes from increased glaciation and seasonality, negative interactions with immigrant faunas from Asia (competition and/or predation), or some combination of the two. P. medium survived past the Grande Coupure probably due to its cursorial nature that allowed it to travel across open lands more efficiently and escape immigrant carnivores; it was the last species of its genus and went extinct not long after the faunal turnover event.

First descriptions
In 1782, the French naturalist Robert de Lamanon described a fossil skull retaining the upper and lower jaws that was collected from the quarries of Montmartre, a large hill near Paris, France, that belonged to the nobleman Philippe-Laurent de Joubert. He recognized that the morphologies of its molars and incisors were roughly akin to those of ruminants but noted that its dentition still lacked any modern analogues. As a result, he hypothesized that the animal was extinct, had amphibious behaviors, and fed on both herbs and fish.

Since 1796, the French naturalist Georges Cuvier innovated the idea of vanished worlds of extinct animals, but as his observations of fossils were mostly limited to drawings and very fragmentary fossil materials stored at the National Museum of Natural History, France, his palaeontological insight remained limited early on. The fossils of Montmartre were credited with great importance to the field of palaeontology, as the fossil taxa found near Paris were embedded in deeper and harder sediments, falling between the Pleistocene-aged mammals and the Cretaceous-aged reptiles. In 1798, he documented fossils from Montmartre, suggesting initially that they could have belonged to the canid genus Canis based on its dentition. Not long after in the same year, he changed his mind and thought that the fossil mammal instead would have been within the order of pachyderms, theorizing that it would have been closest to tapirs and that it would have had trunks like them. He also figured out that the animals of Montmartre were of multiple species with different sizes and numbers of toes.

Early taxonomy and depictions
In 1804, Georges Cuvier described the sets of fossils from the gypsum quarries of the outskirts of Paris (known as the Paris Basin). Describing the skull previously reported by de Lamanon, he confirmed that it belonged to a mammal, had a complete set of 44 total teeth, and had molar morphologies similar to those of rhinoceroses and hyraxes. The naturalist, recognizing that its separate affinities from other mammals, established the genus name Palaeotherium and established the first species name Palaeotherium medium. The genus name means "ancient beast," for which the etymology is a compound of the Greek prefix παλαιός meaning 'old' or 'ancient' and the suffix θήρ  meaning 'beast' or 'wild animal'. He debunked Lamanon's hypothesis that Palaeotherium was an omnivorous amphibian and suspected that it had trunks akin to those of tapirs.

Later, he wrote about a species that he deemed to have similar dentition to P. medium to the extent that it belonged to Palaeotherium. Cuvier observed that the species had larger-sized dentition compared to the other species based on imprints that he was provided. Thus, he established the species Palaeotherium magnum. He also gave mentions to the postcranial fossils of the genus and listed a newly recognized species named Palaeotherium minus. In a later journal of the same year, Cuvier described a mostly complete skeleton from the French commune of Pantin that he determined to have belonged to P. minus.

In 1805, Cuvier described additional postcranial fossil bones of Palaeotherium. He noted that its forefeet consisted of three short toes and that no other animal had postcranial bones that closely resembled those of the extinct genus. Nonetheless, he also gave emphasis to some of the fossil foot bones resembling those of either tapirs or rhinoceroses. Based on the metacarpal bone shapes located on the front feet, he erected the species name P. crassum, mentioning that it was a distinct species from P. medium.

In 1812, he examined more metacarpal bone material that he classified as belonging to Palaeotherium. He stated that the newer material was nearly the size of those of P. crassum but that it was shorter than even those of P. minus. Thus, he felt the need to establish another species P. curtum based on the fossils. He also listed five new additional species from surrounding areas of France that he did not further elaborate on, such as the rhinoceros-sized P. giganteum, the ox-sized P. tapiroïdes, the pig-sized P. buxovillanum plus P. aurelianense, and the sheep-sized P. occitanicum.

The naturalist also suggested palaeobiologies of the four species of Palaeotherium that he described from the gypsum quarries. He acknowledged that P. magnum had skull and limb bone material but lacked vertebra and rib fossils. Regardless, he was able to speculate based on available material that P. magnum would have resembled a tapir the size of a horse with bare amounts of hair. He also hypothesized that P. crassum would have resembled a tapir and been the size of one, which in theory would have caused people to confuse the two. P. medium, he suggested, would have also resembled a tapir but differed by higher legs and longer feet. He was able to construct a speculative skeletal reconstruction drawing of P. minus because of a previously found skeleton and hypothesized that it was smaller than a sheep and could have been cursorial with its slender legs and face. Finally, he theorized that P. curtum would have been the bulkiest species with lower legs compared to P. minus that were stocky like those of P. crassum. Cuvier also suggested that Palaeotherium as in the entire genus was tridactyl (or three-toed).

In 1822, Cuvier recognized additional species of Palaeotherium based on postcranial materials. He also depicted a drawn reconstruction of the skeleton of P. magnum, outlining that it was the size of a Javan rhinoceros, was stocky in body build, and had a massive head. Palaeotherium was also depicted in 1822 drawings by the French palaeontologist Charles Léopold Laurillard under the direction of Cuvier. In 1824, he listed most species of Palaeotherium that he previously named and described, namely P. magnum, P. medium, P. crassum, P. latum, P. curtum, P. minus, and P. aureliense. He also recognized an additional species P. isselanum, but he did not describe its fossils.

Palaeotherium magnum, Palaeotherium medium and "Plagiolophus minus" (= Plagiolophus) are notably depicted in the forms of one sculpture representing each species in the Crystal Palace Dinosaurs attraction in the Crystal Palace Park in the United Kingdom, open to the public since 1854 and constructed by English sculptor Benjamin Waterhouse Hawkins. The original large-sized P. magnum sculpture was lost at some point after 1958 and was replaced by a new replicated model in 2023. The other two palaeothere statues in the park represent the medium-sized P. medium and the small-sized "P. minus" (= Plagiolophus minor). Both the large P. magnum and P. medium were posed in standing positions whereas the smallest sitting statue was made to represent "P. minus". The models' resemblances to tapirs reflected early perceptions that the palaeothere species resembled them in body plan appearances. Despite this, the sculptures differ from living tapirs in several ways, such as shorter plus taller faces, higher eye positions, slender legs, longer tails, and the presence of three toes on the forelimbs unlike the four toes of the forelimbs of tapirs.

Of the three sculptures, P. medium most closely resembles a tapir, and it has remained mostly intact but suffered from damages that were eventually repaired. P. medium was depicted as having durable skin and a slender face with a trunk, representing archaic perceptions that it was a slow animal that lived in closed habitats. The original P. magnum sculpture was last known from a 1958 photograph of it that reveals that it was the largest sculpture of the three species and that it had a highly robust form with large and deep eyes, a proportionally large head, bulky legs, and a muscular-looking body plan. The trunk of the model appears to start from the upper section of the skull and descends down to the lower lip. The overall anatomy of the sculpture appears to have been based off of elephants compared to the other two palaeothere statues.

Palaeotherium proved to be a significant find to the field of palaeontology in multiple other aspects. For one, both the skeletal reconstruction drawing and the life restoration in Cuvier's works were incorporated into textbooks and handbooks around the world up to the 20th century. The genus was also incorporated into old orthogenesis models of the evolution of the horse theory as early as 1851 by British biologist Richard Owen and followed by other 19th century European naturalists such as Jean Albert Gaudry and Vladimir Kovalevsky.

Later 19th century taxonomy
Throughout much of the 19th century, many species were classified under Palaeotherium, some of which were eventually reclassified under different genera. For instance, "P." aurelianense was reclassified to its own genus Anchitherium by Christian Erich Hermann von Meyer in 1844. In an 1839–1864 osteography, the French naturalist Henri Marie Ducrotay de Blainville relisted "P." tapiroides, "P." buxovillanum and "P." occitanicum as species belonging to Lophiodon, but the latter two were eventually reclassified to Paralophiodon and Lophiaspis, respectively in the 20th century. In 1862, Swiss zoologist Ludwig Ruetimeyer defined the previously recognized genera Plagiolophus and Propalaeotherium as distinct from Palaeotherium and containing the species P. minor and P. isselanum, respectively.

In 1853, Pomel erected the species P. duvali based on fossil limbs that he thought to have been less stocky compared to those of P. curtum. In an 1839–1864 osteography, Blainville listed Palaeotherium species previously recognized by other taxonomists and erected P. girondicum. In 1863, the French naturalist Jean-Baptiste Noulet created the species P. castrense based on an incomplete mandible from the commune of Viviers-lès-Montagnes, where it was later studied in Castres. In 1869, Pictet and Humbert erected the species Plagiolophus siderolithicus using fossil molars from a museum collection whose form is similar to that of P. minor but differs mainly by the dimensions. The same year, German palaeontologist Oscar Fraas erected P. suevicum based on teeth that he thought to have distinct enamel. Gervais in 1875 described fossil bones and teeth from the French commune of Dampleux, noting that the particular species was smaller than others classified to Palaeotherium and that the dental measurements were similar to those of Plagiolophus minor. He assigned the fossils to the newly erected species P. eocaenum.

Palaeotherium skeletons
For much of paleontological history, Palaeotherium was not known by any complete skeleton since its initial description by Cuvier. This changed when in 1873, the French geologist Gaston Casimir Vasseur uncovered the first complete skeleton of P. magnum from a gypsum quarry in the commune of Vitry-sur-Seine. The quarry was owned by the civil engineer Fuchs, who donated the skeleton to the National Museum of Natural History, France. The skeleton was first described by Gervais in an academic journal the same year, who noted that it allows for more accurate confirmations of the species' anatomical traits. He pointed out that the skeleton had a skull measures 0.5 m long, a longer neck than previously expected, and a less stocky build compared to tapirs and rhinoceroses. The naturalist said that the extraction process was difficult but completed by multiple skillful workers. Since then, it has been displayed at the Gallery of Paleontology and Comparative Anatomy exhibit of the museum, where it had been noted as an important and famous component of the gallery to the modern day.

During the 20th century, a second complete skeleton of P. magnum was excavated from the plasters within the French commune of Mormoiron. It was sent to the geological department of the University of Lyon and described after preparation by the Austrian geologist Frédéric Roman in 1922. Roman depicted a drawing of a reconstruction of the skeleton of P. magnum based on the Mormoiron skeleton within his 1922 monography. According to Austrian palaeontologist Othenio Abel in 1924, it was the most complete skeleton of Palaeotherium to have been found and amongst the most complete of early Cenozoic mammal skeletons, missing only a few ribs and a left femur.

20th century revisions
In 1904, Swiss palaeontologist Hans Georg Stehlin first created the species name P. lautricense based on an upper jaw from a collection at the Muséum de Toulouse that originated from sandstone deposits at Castres. He also wrote about two somewhat crushed skulls of the species, producing a sketched reconstruction of it based on the first one. In his monography for palaeotheres, published the same year, Stehlin considered most species within Palaeotherium to be potentially valid and created more of them, but he noted that most taxonomists were cautious invalidating species erected by Cuvier. Stehlin also revised P. girondicum as ''P. magnum var. girondicum, or P. magnum girondicum''. He established the subspecies name ''P. curtum var. perrealense, or P. curtum perrealense'', based on jaw fragments from La Débruge. He erected P. Mühlbergi based on dental material in the Swiss municipality of Obergösgen that Ruetimeyer examined back in 1862. He also stated that the recent excavations at Mormont from Natural History Museum of Basel had yielded fossils that he classified along with a mandible identified by Pictert in 1869 under the new species name P. Renevieri. Finally, he also determined that an additional species Palaeotherium Rütimeyeri, which he described as having primitive premolars, was present in the municipality of Egerkingen.

In 1917, French palaeontologist Charles Depéret recognized two additional species of Palaeotherium called P. Euzetense and P. Stehlini. German palaeontologist Wilhelm Otto Dietrich named the German species P. Kleini in 1922, basing it off of fossils from the locality of Mähringen and mentioning that it would have been the size of P. curtum and P. Heimi.

In 1968, upcoming German palaeontologist Jens Lorenz Franzen, then a graduate student, made major revisions of Palaeotherium within his dissertation. He synonymized or rendered dubious statuses many species of Palaeotherium that were erected throughout the 19th and early 20th centuries, including those named by Cuvier. He also erected P. pomeli using fossils from a locality in Castres and reclassified "Plagiolophus" siderolithicum into Palaeotherium. Furthermore, Franzen converted some species into subspecies, namely P. magnum girondicum, P. magnum stehlini, P. medium suevicum, and P. medium euzetense. In addition, he named the following subspecies that he named in his thesis: P. castrense robiacense, P. crassum robustum, P. curtum villerealense, P. curtum frohnstettense, P. muehlbergi praecursum, and P. duvali priscum. Not all species within Palaeotherium had any recognized subspecies in them.

In 1975, Spanish palaeontologist María Lourdes Casanovas-Cladellas erected the species P. crusafonti from a left maxilla with dentition from the Spanish site of Roc de Santa. In 1980, both she and José-Vicente Santafé Llopis established the second Iberian species P. franzeni, taking into account morphological differences of the dental fossils from the Spanish municipality of Sossís. In 1985, the French palaeontologist Jean-Albert Remy named a subspecies P. muehlbergi thaleri in honor of fellow palaeontologist Louis Thaler, having documented that its fossils were from the commune of Saint-Étienne-de-l'Olm and that both the holotype and paratype each consist of a skull with a mandible.

In 1991, Casanovas-Cladellas and Santafé Llopis erected P. llamaquiquense from partial jaw material from the Spanish locality of Llamaquique in the city of Oviedo, where the name derived from. The next year in 1992, Remy proposed the creation of two subgenera of Palaeotherium based on cranial characteristics. The first subgenus he listed was Palaeotherium, which includes the type species P. magnum along with P. medium, P. crassum, P. curtum, P. castrense, P. siderolithicum, and P. muehlbergi. The second subgenus name that he proposed was Franzenitherium, which includes the type species P. lautricense plus P. duvali and was named in honor of Franzen's review of Palaeotherium. The Spanish palaeontologist Miguel Ángel Cuesta Ruiz-Colmenares established the species P. giganteum using dentition from the Mazaterón site in the Duero Basin in 1993, considering it to be the largest species of Palaeotherium known. In 1998, Casanovas-Cladellas et al. formally recognized P. crassum sossissense from a fragmented right maxilla with dentition from Sossís in Spain. They also invalidated the previously named P. franzeni due to reassigning the material to P. magnum stehlini.

Classification
Palaeotherium is the type genus of the Palaeotheriidae, largely considered to be one of two major hippomorph families in the superfamily Equoidea, the other being the Equidae. Alternatively, some authors have proposed that equids are more closely related to the Tapiromorpha than to the Palaeotheriidae. It is also usually thought to consist of two families, the Palaeotheriinae and Pachynolophinae; not all authors agree on the latter as a palaeotheriid subfamily, however. Some authors have also considered the Plagiolophinae to be a separate subfamily, while others group its genera into the Palaeotheriinae. The geographic range of the palaeotheres were in contrast to equids, which are generally thought to have been an endemic radiation in North America. Some of the most basal equoids of the European landmass are of uncertain affinities, with some genera being thought to potentially belong to the Equidae. Palaeotheriids are well-known for having lived in western Europe during much of the Palaeogene but were also present in eastern Europe, possibly the Middle East, and, in the case of pachynolophines (or pachynolophs), Asia.

The Perissodactyla makes its earliest known appearance in the European landmass in MP7 of the Mammal Palaeogene zones. During the temporal unit, many genera of basal equoids such as Hyracotherium, Pliolophus, Cymbalophus, and Hallensia made their first appearances there. A majority of the genera persisted to the MP8-MP10 units, and "pachynolophines" (probably true palaeotheres) such as Propalaeotherium and Orolophus arose by MP10. The MP13 unit saw the appearances of later pachynolophines such as Pachynolophus and Anchilophus along with definite records of the first palaeotheriines such as Palaeotherium and Paraplagiolophus. The palaeotheriine Plagiolophus has been suggested to have potentially made an appearance by MP12. It was by MP14 that the subfamily proceeded to diversify, and the plagiolophines were generally replaced but still reached the late Eocene. In addition to more widespread palaeothere genera such as Plagiolophus, Palaeotherium, and Leptolophus, some of their species reaching medium to large sizes, various other palaeothere genera that were endemic to the Iberian Peninsula, such as Cantabrotherium, Franzenium and Iberolophus, appeared by the middle Eocene.

The phylogenetic tree for several members of the family Palaeotheriidae within the order Perissodactyla (including three outgroups) as created by Remy in 2017 and followed by Remy et al. in 2019 is defined below:

As shown in the above phylogeny, the Palaeotheriidae is defined as a monophyletic clade, meaning that it did not leave any derived descendant groups in its evolutionary history. Hyracotherium sensu stricto (in a strict sense) is defined as amongst the first offshoots of the family and a member of the Pachynolophinae. "H." remyi, formerly part of the now-invalid genus Propachynolophus, is defined as a sister taxon to more derived palaeotheriids. Both Pachynolophus and Lophiotherium, defined as pachynolophines, are defined as monophyletic genera. The other pachynolophines Eurohippus and Propalaeotherium consistute a paraphyletic clade in relation to members of the derived and monophyletic subfamily Palaeotheriinae (Leptolophus, Plagiolophus, and Palaeotherium), thus making Pachynolophinae a paraphyletic subfamily clade.

List of lineages
Since 1968, many species of Palaeotherium have multiple defined subspecies due to taxonomic revisions conducted by Franzen involving new species plus subspecies erections and conversions of some species into subspecies that were accepted by subsequent authors. From his dissertation was he able to justify the subspecies by proof of various intraspecific variations. The following table defines the species and subspecies of Palaeotherium and additional information about them:

Skull
The Palaeotheriidae is diagnosed in part as generally having orbits that are wide open in the back area and are located in the middle of the skull or in a slight frontal area of it. The nasal bones are slightly extensive to very extensive in depth. Palaeotherium is characterized as having calvaria that range in base length from 150 mm to 520 mm depending on the species. The pterygoid crest, which is located on the pterygoid processes of the sphenoid bone, does not cover the optic foramen, which is separated from other cranial cavities at the temporal fossa. The zygomatic process of the squamosal bone is elongated and extends to the maxilla at a back angle of the orbit. The genus is also diagnosed by the presence of an anastomosis (anatomical connection between two passageways) roughly at the sphenoid bone and prominent temporalis muscle developments.

According to Remy, the subgenus Palaeotherium is the more specialized one of the two, characterized by the orbit being located more in front of the skull's middle length. The optic foramen is separated by a bony wall, and there are two optic canals in total. The cranium is constricted in its front area behind the postorbital processes and close to the suture for the frontal bone and parietal bone. The other subgenus Franzenitherium in comparison has more generalized skull traits, its orbit being aligned within the middle length of the head. It has a front constriction of the cranium near the postorbital processes, and the optic foramen crosses through the skull from one side to the other. Not all species are placed in any subgenus due to having inadequate skull evidence for further analyses.

The height and weight proportions of the skull of Palaeotherium are roughly equivalent with those of other taxa within the Equoidea. In comparison to other equoids where the skull's maximum width extends above the front root of the parallel zygomatic arches, those of Palaeotherium and most other palaeotheres (except Leptolophus) extend back to the joint of the squamosal bone and mandible. Unlike that of Plagiolophus, the maxillary hiatus, or an opening of the maxillary sinus, in Palaeotherium is wide, diamond-shaped, and has oblique back edges. Palaeotherium differs from most other palaeotheres by the nasal opening stretching up to the P3 tooth at minimum (noticeable in P. duvali and P. siderolithicum) or up to the front edge of the orbit above M3 in the case of P. magnum. Similar to other palaeotheres, the back process (tissue projection) of the premaxilla is reduced, but its morphology can vary. The maxilla can extend to the nasal opening but can also vary in proportions. While the shapes and proportions of the nasal bones vary by species, they extend beyond P1 in adults and sometimes even the canine like in equines. The nasomaxillary suture, which unites the maxilla and nasal bones, is short and strongly curved.

The jugal bone and lacrimal bone, both located in front of the orbit, are weak in development. The latter bone is elongated in its back and touches the posterolateral process of the nasal bones. Members of the Equoidea have relatively shortened front areas of the face. The orbits of Palaeotherium, unlike those of other equoids, are proportionally smaller and are situated somewhat in front of the skull's mid-length area; they might be more forwards in the case of P. medium. Similar to other Palaeogene equoids, the front edge of the orbit is aligned with M1 or M2 while the back area is wide. In most species, the infraorbital foramen in adults is located above P4 or M1; in P. curtum frohnstettense, it extends to above M2. Each zygomatic arch is wide, and its uneven narrowing in the front area under the orbit may be the result of either species traits or sexual differentiation. The squamosal process of the postorbital is elongated and reaches the maxilla at a back angle of the orbital floor, the roof of the maxillary sinus. The orbit is shallow within its front area, its back opening of the infraorbital canal having a small distance of 10 mm to 15 mm. The canal itself is smaller than those of most other palaeotheriines except for that of Leptolophus, and it has a tendency to shorten in later species.

The side wall of the snout is usually concave but may be interrupted by other local concavities that are normally poorly distinguished. The palatine bone has a well-developed front area, which in the case of P. curtum villerealense can rise near the orbit's top. The sphenopalatine foramen, large and oval-shaped, is above the back of M4. The optic canal is small, has a primitive form of opening in front of the pterygoid crest, and is separated from it by a thick bony wall except in the cases of P. lautricense and P. duvali. Whereas there are two optic canals in most species that are nearly parallel of each other and separate in the end, there is only one long one in P. lautricense and possibly P. duvali. The sphenoid bone's wings are well-developed in their backs, and a suture for the parietal and sphenoid bones separates the frontal bone from the squamosal bone. The postorbital processes of the frontal bone are not very elongated. The cranial vault is broad, domed, and wider than the overall skull. The maximum front narrowing of the cranium, with the exceptions of those of P. lautricense and P. duvali, is set far back to roughly where the frontoparietal suture occurs. The skull's top peaks at the far back area, although this is not observed in P. lautricense. The sagittal crest can be prominent and depends on the age and sex of the individual for development. The nuchal crest, where the neck attaches to the head, is prominent and, except in P. lautricense, extends outwards plus backwards past the occipital condyles. The temporal fossae are large but vary in proportion. In Palaeotherium and not Plagiolophus, the overall basicranium's axis is thick plus wide.

The horizontal ramus of the mandible is overall thick plus tall and has an elongated mandibular symphysis, but the width and lower area morphology vary by species. It is wide in both the front and back areas and low compared to equines. The joint for the squamosal and mandible of Palaeotherium is low compared to those of Plagiolophus and Leptolophus. The angular process, located above the angle of the mandible, is blocked from further expansion by the mandibular notch and is well-developed in its rear like in Palaeogene equids. The coronoid process of the mandible, an upper eminence, is both broad in the front plus back and stocky. The condyloid process, an upper process of the mandible, is transversely elongated and cylindrical in shape. Both the zygomatico-mandibular fossa and pterygoid fossa are prominent.

Dentition
Derived palaeotheres are generally diagnosed as having selenolophodont upper molars and selenodont lower molars that are mesodont, or medium-crowned, in height. The canines strongly protrude and are separated from the premolars by medium to long diastemata and from the incisors by short ones in both the upper and lower dentition. The other teeth are paired closely with each other in both the upper and lower rows. The dental formula of Palaeotherium is for a total of 44 teeth, consistent with the primitive dental formula for early-middle Palaeogene placental mammals. The post-canine diastemata of Palaeotherium are small. The premolars and preceding deciduous teeth both tend to have molarized forms and have newly developing hypocone cusps on them. The mesostyle cusp present in the molars thicken from M1 to M3. The lingual lobes (or divisions) in the upper molars are closely aligned with the ectolophs (crests or ridges of upper molar teeth). The ectolophs themselves are W-shaped, being made up of two articulated crescents.

The incisors are shovel-shaped and, like in modern horses, are used for chewing at right angles in relation to their longitudinal axes. They have no cutting functions but instead are used for grasping food akin to how tweezers grasp items. The canines are proportionally large-sized and are dagger-shaped. They were probably not used for cutting or chewing food given how they are oriented. Instead, they were probably used for biting functions for self-defense and sexual selection.

The decreased lengths of postcanine diastemata in Palaeotherium and the equid subfamily Anchitheriinae may be correlated with increases in body size. The trend may be due to the necessity to improve chewing performances through molarization and proportional size increases of the premolars and the enlargements of the molar row, the latter trend of which may play a role in decreasing diastamata lengths. Early species such as P. castrense have nearly absent postcanine diastemata. In later species, the postcanine diastemata can vary from shortened such as in P. crassum and P. curtum to elongated like in P. medium and P. magnum. The separation of cheek teeth from the incisors and canines attests to their independent and specific chewing functions. The distance from the canine to the second premolar is up to twenty percent (twenty-five percent in the case of P. magnum) of the total length of the second premolar to third molar dental row.

Late Eocene species of Palaeotherium tend to have more molariform premolars. The non-molarized premolars are composed of four to five cusps (one to two external, two intermediate, and one internal) while the molarized premolars and molars have six cusps (two external, two intermediate, and two internal). The upper molars are medium-crowned (shorter than those of modern equids) and have ectolophs that are about twice the height of the inner cusps and curve into a W shape. The lower molarized premolars and molars are about half as wide as their upper counterparts. The lower cheek teeth's occlusal surfaces have patterns resembling two mesiodistal crescents with an outwards convex side. M3 has a hook-shaped and curved hypoconulid cusp. The non-molarized premolars have talonids (crushing regions of cheek teeth) that are only semi-developed as elongated ridges.

The lingual side (front side in relation to the mouth) of the upper molars are at about the same heights at different stages of the teeth resulting from shifting stages in chewing. On the other hand, in regard to chewing stages, the crowns on the buccal side of the upper molars increase in height and move forward. In the lower molars, the crowns instead shift the opposite way towards the buccal side (back side). This is due to the chewing function being emphasized on the buccal side of the upper molars for shearing through food vertically and the lingual side of the lower molars for slowly chewing through food in a horizontal manner. A Palaeotherium individual would have moved its lower jaw in a circular movement, pushing forward the upper molars in a manner of occlusion during and after eruption, especially at their buccal side. The upper molars go through an abrasion process that causes their outer part of their crowns to curve. This ensures that the distance from the front cutting edge of the ectolophs to the axis of rotation remains the same.

Compared to the earlier-appearing pachynolophines, the palaeotheriines have more molarized deciduous premolars. For instance, Stehlin illustrated a Mormont fossil of P. renevieri with erupted dP1-dP4 plus an unerupted M1. dP1 appears small and triangular in shape with two buccal cusps (paracone and metacone cusps) and a smaller posterolingual cusp. dP2-dP4, in comparison, are molariform in shape and have four major cusps. Stehlin theorized that the dP1 tooth is unreplaced by any adult P1 due to the similar sizes of the milk tooth to the adult tooth. A juvenile skull of P. magnum with deciduous premolars was described by Remy in 1985, who noted their molarized forms. As is the case for the juvenile P. renevieri, the dP1 of the juvenile P. magnum is triangular in shape and has two close buccal cusps plus a smaller posterolingual cusp. It also shares the trait of molariform, four-cusped dP2-dP4. While Remy proposed that an adult P1 had already replaced its deciduous counterpart in P. magnum at an early age, there is no strong evidence to support his claim.

Vertebrae and ribs
The overall postcranial anatomy of Palaeotherium is best known from a skeleton of P. magnum uncovered from Mormoiron. Within the vertebral column are seven large-sized cervical vertebrae total for a series of C1-C7, typical of most mammals. In total, they measure 65 cm long within the individual skeleton. The atlas (C1) was trapezoidal and closely resembled those of equids, but it differed from theirs in its transverse processes having less rounded outlines. The spiny process of the atlas of Palaeotherium is elongated and thin compared to those of horses. The proceeding cervical vertebrae, C2-C4, are wider than they are long and appear roughly quadrangular in shape. Their lateral process are proportionately wider than in horses but also decrease in length gradually from the front to the back. C7 has a well-pronounced spinous process more akin to tapirs than horses. Based on the anatomy of the cervical vertebrae, the neck of P. magnum is long and muscular but not low like in tapirs.

There are also 17 thoracic vertebrae, one vertebra less than in horses and two less than in tapirs, that measure 77 cm long in total. The first ten vertebrae, T1-T10, are about equal in individual lengths with each other but are slightly shorter than the last seven, T11-T17. The first thoracic vertebrae have strong spinous processes with few variations, but their lengths decrease as seen in T5-T6. The thoracic vertebrae generally have rectangular and elongated shapes, differing from the triangular and sequentially shorter spinous processes in horses. The arrangements of the thoracic vertebrae reveal that the withers (or scapula ridges) of P. magnum are lower than a horse's but higher than a tapir's.

Like horses, P. magnum also has six lumbar vertebrae measuring 65 mm long, one more vertebra than that of donkeys, zebras, and tapirs. All the vertebrae have equal lengths and, especially L3, have strong lateral processes that furrow toward the back, widen to a third of their total lengths, slightly narrow, and then finally swell in the end. They differ from the nearly straight lateral process shapes of equids. The lumbar vertebrae portion of Palaeotherium was probably wider than that of the horse.

The sacrum, connected from its back underneath the pelvis, is triangular similar to that of the Equidae but is slightly wider in its front area. It appears to compose of six sacral vertebrae total. The final region of the vertebral column is composed of fifteen caudal vertebrae that compose the tail, although there is the possibility that one to two additional ones are hidden by the pelvis. The caudal vertebrae, especially in C5-C11, are slender due to their lengths being twice their widths. The first four, however, have roughly equal lengths and widths. The tail, in terms of length and vertebrae shape, is similar to that of equids. The skeleton's tail measures 35 cm long, the first four vertebrae being 11 cm long.

Although the Mormoiron skeleton has nineteen complete or fragmented ribs, P. magnum would have had thirty-four total based on the number of thoracic vertebrae. Like in equids, the front ribs are strong and flattened. One rib between the forelimbs, probably articulated at T10, measures 49 cm long, suggesting that the rib cage size is larger than those of horses and approximate to those of modern rhinoceroses. The back portion of the thorax would have been wider than in horses and roughly comparable to those of tapirs and rhinoceroses but never being as long as those of the latter. The ribs sharing space with the sternum do not directly connect to it but are instead separated from it. The sternum was approximately the same size as the thorax. Contrary to the 1922 reconstruction by Roman, there is no evidence that the rib cage had a fan-shaped spread.

Limbs
The Palaeotheriidae includes both the primitive members with tetradactyl (four-toed) forelimbs plus tridactyl (three-toed) hindlimbs (Propalaeotherium) and more derived members with tridactyl forelimbs plus hindlimbs (Palaeotherium, Plagiolophus). Most species of Palaeotherium, most notably P. medium, P. magnum, P. crassum, P. curtum, and P. muehlbergi, have tridactyl limbs. It is unclear as to whether or not P. eocaenum is tetradactyl based on a possible manus tentatively assigned to it. Both the manus and pes bones are short plus robust. The side digits are capable of reaching the ground and are not much more slender than the middle one. While previous studies have suggested that modern tapirs serve as analogues for European perissodactyls of the Eocene, a 2020 study by Jame A. MacLaren and Sandra Naewelaerts suggested that no one species of tapir serves as an analogue for any of the extinct species. P. medium has a more unique foot morphology compared to other Palaeotherium species due to narrower plus higher feet and stretched metapodial bones. The tridactyl foot morphology with all three digits being functional suggests digitigrade locomotion.

In P. magnum, the morphology of the scapula, which articulates with the humerus, is similar to those in modern equids, giving off a narrow appearance due to its length being twice its width. Its highest point aligns with the transverse processes of the front thoracic vertebrae. It is elongated, nearly rectangular in shape, and slightly narrow at its base. Its glenoid fossa is concave plus shallow whereas its coracoid process articulates underneath the vertebral column. Compared to equids, the scapula, while similar, is more triangular while its glenoid fossa is deeper and more rounded.

The humerus' head articulates with the glenoid cavity of a scapula while its lower end connects with the forearm's bones. The head is slightly oblique, making it more similar to those of rhinoceroses than those of tapirs. The greater tubercle of the humerus appears long and narrow as in tapirs. The deltoid tuberosity is located about halfway along the humerus' length, slightly lower than in Anchitherium and modern horses. The humerus of P. medium is more slender. The lips of its trochlea (articular surface of the elbow joint) has a larger slope. The bicipital groove of the humerus, which separates the greater tubercle and lesser tubercle, is deep. The morphology of the humerus of P. medium implies more cursorial adaptations.

In P. magnum, the forearm, consisting of the radius and ulna, is 20% larger than the humerus. In comparison, tapirs have slightly longer humeri compared to the radius and ulna, and the forearm of horses are 25% larger than their humeri. The radius, articulated with the humerus and carpal bones, is slightly arched and has a roughly circular front end. The upper part of the radius is fused with the upper ulna. Its general shape appears similar to those of tapirs but is slightly thinner and has less prominent extremities. The ulna is well-developed and lengthy, being as strong as the radius and having a less pronounced back curvature compared to tapirs and horses. Its olecranon process is enlarged but is only slightly hollow in its internal face.

The palaeotheriines Palaeotherium and Plagiolophus, despite being known as tridactyl genera, display large morphological diversity of the forelimb. The former has long plus narrow carpals, its metacarpal bones being close in length to each other plus developing into wide ungual phalanges at their ends and the middle one being slightly more robust. Palaeotherium has an exceptional amount of shape variation displayed in its third metacarpal and varies in manus dimensions by species with known limb remains. P. curtum has very robust forelimb bones including a stocky and stocky manus, which suggests that it was stocky in build. P. magnum and P. crassum are suggested to have forelimbs that to at least some level resemble those of tapirs, especially the mountain tapir (Tapirus pinchaque). The forelimb morphology of P. magnum rather than P. crassum may more closely resemble those of tapirs to the more gracile forms of the radius and metacarpals in the latter. P. medium, along with Plagiolophus spp, appear to be the most cursorial palaeotheres due to their elongated and gracile metacarpals; those of P. medium and Plagiolophus are of approximately equal proportional lengths.

In the Mormoiron skeleton, the femur is strong and stocky, its upper (or proximal) end being enlarged and having a large trochanter (or femoral tubercle) that does not extend beyond the size of the femoral head. On the external face of the bone, here is a third trochanter that is well-developed, triangular in shape, and barely curved forward. Its position is slightly more centred compared to that of P. medium. The bone narrows slightly after the third trochanter then swells quickly.

The tibia, like other limb bones, is strong and thick in build, its front crest reaching the midlength of the bone. It hosts a shallow front fossa for the patella, differing from the deeper ones of Anchitherium and Equus. Its joint end for the astragalus is somewhat oblique, and the tibia's malleolus is less developed compared to the aforementioned equids. The fibula is slender in form and fused to the tibia. It is proportionally wider in P. magnum than in P. crassum and P. medium, the latter two species of which have lesser-developed ones.

Palaeotherium has a straighter and less concave trochlea of the astragalus than in Plagiolophus. The calcaneum is semirectangular in shape and is slightly wide on its back end. The cuboid bone is high and narrow similar to that in Anchitherium. As in the metacarpal bones, the middle metatarsal bone (finger III) is larger and more well-developed than the others. Finger IV of the hind foot of P. magnum appears slightly arched and is slightly longer than finger II.

Footprints
Palaeotheriids are known from footprint tracks assigned to ichnotaxa, among them being the ichnogenus Palaeotheriipus, named by the palaeontologist Paul Ellenberger in 1980 and suggested to have belonged to Palaeotherium. The ichnogenus, described from the department of Gard in France, is defined as having been produced by large-sized perissodactyls, of which the middle toe imprint is deep while the two other outer toes are less pitted. The three toes have rounded nail ends based on the shape of the footprints. The ichonotaxon's characteristics were said by Ellenberger to correspond with those of P. medium or P. cf. crassum. The ichnogenus is diagnosed as being a very short and tridactyl footprint in which the outer digits II and IV are flat and less pitted than digit III. It differs from another palaeothere footprint ichnogenus Plagiolophustipus, suggested to have been made by Plagiolophus, by the presence of smaller and broader digits. Lophiopus, likely produced by Lophiodon, differs by smaller digit sizes and divergent outer digit imprints while Rhinoceripeda, attributed to the Rhinocerotidae, is oval-shaped and has three or five digits. Palaeotheriipus is known from both France and Iran whereas Plagiolophustipus is currently known from Spain.

The type ichnospecies is Palaeotheriipus similimedius from the lacustrine limestone of the French commune of Garrigues-Sainte-Eulalie in the department of Gard. The footprint in question measures 115 mm long and 140 mm wide; it is therefore wider than it is long. The three fingers diverge widely from each other in angles of at least 50°. The hoof of finger III appears to be wider than those of the outer toes. Ellenberger suggested that the ichnospecies most closely corresponds with either P. medium euzetense or P. medium perrealense due to the more flexible fingers. Fingers II and IV measure approximately 50 mm long, and the length of finger IV is 65 mm.

A second ichnospecies from Iran attributed to Palaeotheriipus brings the possibility that palaeotheres could have extended in geographical range to the region by the middle to late Eocene. It was named P. sarjeanti after William A. S. Sarjeant and was found in eastern Iran. The ichnospecies is diagnosed as being large-sized perissodactyl footprints each with a finger III that is relatively round, lesser-developed, and is broader and longer than outer fingers II and IV, which nearly share the same size and shape. The manus is less elongated than the pes. The footprint for the pes measures 295 mm in length and 240 mm in width, with the lengths of finger II measuring 25 mm, digit III being 65 mm, and finger IV measuring 35 mm. In comparison, the footprint for the manus measures 200 mm long and 240 mm wide, the only clearly discernable digit, finger III, measuring 25 mm long.

Additional footprints from the d’Apt-Forcalquier basin in France, dated to the middle Eocene and described first by G. Bessonat et al. in 1969, are recorded to be larger than the footprints of P. similimedius, measuring 170 mm long and 160 mm wide. Finger II measures 80 mm long, finger III is 75 mm in length, and that of finger IV is 85 mm. They are referred to the species P. magnum.

Size
Palaeotherium is characterized by the inclusion of small to large-sized species, the skull base length ranging from 150 mm to 520 mm depending on the species. The length of the P2 to M3 dental row ranges from 64 mm long in the smallest species P. lautricense to 217 mm long in the largest species P. giganteum. That of P. magnum is not far behind the latter with the one dental set measuring 208.6 mm in length, and it was previously considered the largest species of Palaeotherium. P. medium is estimated to be the size of a subadult South American tapir (Tapirus terrestris), larger than the roe deer-sized Plagiolophus minor. The P. magnum Mormoiron skeleton demonstrates that individuals could have reached approximately 1.3 m in shoulder height and 2.52 m in length. The head and neck together measure 1.04 m, and its forelimb (humerus to hoof) also measures 1.04 m in length.

In 2015, Remy calculated the body mass of several Eocene European perissodactyl species based on a predictive body mass formula that was originally devised by Christine M. Janis in 1990. He estimated that the small species P. lautricense could have weighed just as comparatively little as 36 kg. P. siderolithicum could have had an average weight of around 61 kg. P. aff. ruetimeyeri had a larger body mass estimate of 196 kg while that of P. pomeli is 206 kg. P. castrense robiacense had a comparatively massive size estimate of 447 kg. According to Piere Perales-Gogenola et al. in 2022, the largest species P. giganteum could have had a body weight as massive as over 700 kg. Alternatively, MacLaren and Naewelaerts gave the large-sized P. magnum a lower weight estimate of 240.3 kg.

Palaeobiology
Palaeotherium has a large range of species that range in morphology, from small-sized to large-sized and from bulky but slow (P. magnum, P. curtum, P. crassum) to light but cursorial (P. medium). The evolutionary history of the palaeotheres might have had emphasized macrosmatic (derived smell) traits rather than sight or hearing, evident by the smaller orbits and a seeming lack of a derived auditory system. The macrosmatic trait could have allowed palaeotheres to keep track of their herds, implying gregarious behaviours. The wide diversity of palaeothere forelimb morphologies attests to different levels of mobility by species. They generally had smaller hindlimbs compared to forelimbs, suggesting less tendencies towards cursoriality due to being adapted towards closed and stable environments. In 2000, Giuseppe Santi proposed that that Palaeotherium could have been able to stand on its hind legs to reach high plants. P. magnum in particular may have been able to browse on plants at over 2 m tall. If it were capable of being bipedal, it could have reached 3 m tall, 3.5 m if it were more vertically bipedal. However, Jerry J. Hooker argued that there is no evidence for facultative bipedalism in P. magnum unlike in the contemporary artiodactyl Anoplotherium. The long neck of P. magnum brings the possibilities of it browsing on higher plants and/or drinking water from below. Fossil evidence points to Palaeotherium being amongst the largest mammals to inhabit Europe during the middle to late Eocene, with only a few contemporary mammalian groups such as lophiodonts, anoplotheriids, and other palaeotheres reaching similar to larger body sizes.

Both Palaeotherium and Plagiolophus have dentitions that are both capable of chewing through harder items such as fruits without wearing their teeth down quickly compared to their pachynolophine predecessors (i.e. Hyracotherium and Propalaeotherium). The shifts in dietary capabilities were the result of changes in the efficiencies of the mastication processes. The two derived genera have brachyodont dentition, the hypsodonty index suggesting that both genera were mostly folivorous (leaf-eating) and did not have especially frugivorous (fruit-eating) tendencies because of the reduced proportions of rounded cusps. While both genera may have incorporated some fruit into their diets, the higher lingual tooth wear in Plagiolophus indicates it ate more fruit than Palaeotherium. Because of their likely tendencies to browse on higher plants, evident by their long necks and the woodland environments that they inhabited, it is unlikely that ground minerals, usually consumed from grazing on ground plants, significantly affected the tooth wear of either of the genera. The tooth wear in both genera could have been the result of scratches from chewing on fruit seeds. It is likely that Palaeotherium ate softer food such as younger leaves and fleshy fruit that may have had hard seeds while Plagiolophus leaned towards consuming tough food such as older leaves and harder fruit.

Middle Eocene
For much of the Eocene, a hothouse climate with humid, tropical environments with consistently high precipitations prevailed. Modern mammalian orders including the Perissodactyla, Artiodactyla, and Primates (or the suborder Euprimates) appeared already by the early Eocene, diversifying rapidly and developing dentitions specialized for folivory. The omnivorous forms mostly either switched to folivorous diets or went extinct by the middle Eocene (47–37 Ma) along with the archaic "condylarths". By the late Eocene (approx. 37–33 Ma), most of the ungulate form dentitions shifted from bunodont cusps to cutting ridges (i.e. lophs) for folivorous diets.

Land-based connections to the north of the developing Atlantic Ocean were interrupted around 53 Ma, meaning that North America and Greenland were no longer well-connected to western Europe. From the early Eocene up until the Grande Coupure extinction event (56 Ma - 33.9 Ma), the western Eurasian continent was separated into three landmasses, the former two of which were isolated by seaways: western Europe (an archipelago), Balkanatolia, and eastern Eurasia (Balkanatolia was in between the Paratethys Sea of the north and the Neotethys Ocean of the south). The Holarctic mammalian faunas of western Europe were therefore mostly isolated from other continents including Greenland, Africa, and eastern Eurasia, allowing for endemism to occur within western Europe. The European mammals of the late Eocene (MP17 - MP20 of the Mammal Palaeogene zones) were mostly descendants of endemic middle Eocene groups as a result.

Palaeotherium made its first appearance from its earliest-known representative P. eocaenum in the MP13 unit. By then, it would have coexisted with perissodactyls (Palaeotheriidae, Lophiodontidae, and Hyrachyidae), non-endemic artiodactyls (Dichobunidae and Tapirulidae), endemic European artiodactyls (Choeropotamidae (possibly polyphyletic, however), Cebochoeridae, and Anoplotheriidae), and primates (Adapidae). Both the Amphimerycidae and Xiphodontidae made their first appearances by the level MP14. The stratigraphic ranges of the early species of Palaeotherium also overlapped with metatherians (Herpetotheriidae), cimolestans (Pantolestidae, Paroxyclaenidae), rodents (Ischyromyidae, Theridomyoidea, Gliridae), eulipotyphlans, bats, apatotherians, carnivoraformes (Miacidae), and hyaenodonts (Hyainailourinae, Proviverrinae). Other MP13-MP14 sites have also yielded fossils of turtles and crocodylomorphs, and MP13 sites are stratigraphically the latest to have yielded remains of the bird clades Gastornithidae and Palaeognathae.

The Egerkingen α + β locality, dating to MP14, records fossils of P. eocaenum, P. ruetimeyeri, and P. castrense castrense. Other mammal genera recorded within the locality include the herpetotheriid Amphiperatherium, ischyromyids Ailuravus and Plesiarctomys, pseudosciurid Treposciurus, omomyid Necrolemur, adapid Leptadapis, proviverrine Proviverra, palaeotheres (Propalaeotherium, Anchilophus, Lophiotherium, Plagiolophus), hyrachyid Chasmotherium, lophiodont Lophiodon, dichobunids Hyperdichobune and Mouillacitherium, choeropotamid Rhagatherium, anoplotheriid Catodontherium, amphimerycid Pseudamphimeryx, cebochoerid Cebochoerus, tapirulid Tapirulus, mixtotheriid Mixtotherium, and the xiphodonts Dichodon and Haplomeryx.

MP16 marks the first appearances of several species of Palaeotherium in the Central European region, namely P. castrense robiacense, P. pomeli, P. siderolithicum, and P. lautricense, some of which are exclusive to the unit (P. pomeli and P. lautricense) and one of which makes its final temporal appearance (P. castrense). The locality of Robiac in France records the likes of ''Palaeotherium aff. ruetimeyeri and all the aforementioned species from the region in MP16 along with the herpetotheriids Amphiperatherium and Peratherium, apatemyid Heterohyus, nyctithere Saturninia, omomyids (Necrolemur, Pseudoloris, and Microchoerus), adapid Adapis, ischyromyid Ailuravus, glirid Glamys, pseudosciurid Sciuroides, theridomyids Elfomys and Pseudoltinomys, hyaenodonts (Paracynohyaenodon, Paroxyaena, and Cynohyaenodon), carnivoraformes (Simamphicyon, Quercygale, and Paramiacis), cebochoerids Cebochoerus and Acotherulum, choeropotamids Choeropotamus and Haplobunodon, tapirulid Tapirulus, anoplotheriids (Dacrytherium, Catodontherium, and Robiatherium, dichobunid Mouillacitherium, robiacinid Robiacina, xiphodonts (Xiphodon, Dichodon, Haplomeryx), amphimerycid Pseudamphimeryx, lophiodont Lophiodon, hyrachyid Chasmotherium, and other palaeotheres (Plagiolophus, Leptolophus, Anchilophus, Metanchilophus, Lophiotherium, Pachynolophus, Eurohippus'').

MP16 also records two species that are restricted to the unit, P. llamaquiquense and P. giganteum, both of which were endemic to the Iberian region. MP17 marks the restricted appearance of another Iberian endemic species P. crusafonti. The endemic species of Palaeotherium were amongst the many taxa of palaeotheres that are recorded from fossils of the Iberian region and are not known from anywhere else. P. giganteum is recorded from the Spanish locality of Mazaterón along with the testudines Hadrianus and Neochelys, alligatoroid Diplocynodon, baurusuchid Iberosuchus, adapoid Mazateronodon, omomyid Pseudoloris, pseudosciurid Sciuroides, theridomyids Pseudoltinomys and Remys, hyaenodont Proviverra, anoplotheriids (Duerotherium and cf. Dacrytherium), xiphodonts (cf. Dichodon), and other palaeotheres (Paranchilophus, Plagiolophus, Leptolophus, Cantabrotherium, Franzenium, and Iberolophus).

After MP16, a faunal turnover occurred, marking the disappearances of the lophiodonts and European hyrachyids as well as the extinctions of all European crocodylomorphs except for the alligatoroid Diplocynodon. The causes of the faunal turnover have been attributed to a shift from humid and highly tropical environments to drier and more temperate forests with open areas and more abrasive vegetation. The surviving herbivorous faunas shifted their dentitions and dietary strategies accordingly to adapt to abrasive and seasonal vegetation. The environments were still subhumid and full of subtropical evergreen forests, however. The Palaeotheriidae was the sole remaining European perissodactyl group, and frugivorous-folivorous or purely folivorous artiodactyls became the dominant group in western Europe.

Late Eocene
The late Eocene marks the first appearances of multiple species of Palaeotherium at the MP17 unit, namely P. magnum, P. medium, P. curtum, P. crassum, P. duvali, and P. muehlbergi. The temporal range of P. siderolithicum, first known in MP16, continued up to MP19, and P. renevieri made its first and only appearance in MP19. Some other species extended in temporal range up to MP19 (P. duvali, P. crassum) while some others lasted up to MP20 (P. magnum, P. curtum, P. muehlbergi). By the late Eocene, the latest species of Palaeotherium were widespread throughout western Europe, including what is now Portugal, Spain, France, Switzerland, Germany, and the United Kingdom. Additionally, the genus is known from as far east as the Thrace Basin of Greece in the eastern European region, beyond the archipelagos of western Europe and Balkanatolia in the middle to late Eocene. Palaeotherium in eastern Europe would have coexisted with vastly different mammalian faunas from western Europe due to the strong endemism in the latter. It is possible that Palaeotherium could have possibly extended as far east as eastern Iran depending on whether the footprints are attributable to it.

Within the late Eocene, the Cainotheriidae and derived members of the Anoplotheriinae both made their first fossil record appearances by MP18. Also, several migrant mammal groups had reached western Europe by MP17a-MP18, namely the Anthracotheriidae, Hyaenodontinae, and Amphicyonidae. In addition to snakes, frogs, and salamandrids, rich assemblage of lizards are known in western Europe as well from MP16-MP20, representing the Iguanidae, Lacertidae, Gekkonidae, Agamidae, Scincidae, Helodermatidae, and Varanoidea, most of which were able to thrive in the warm temperatures of western Europe.

The MP18 locality of La Débruge in France holds fossil records of multiple species of Palaeotherium, namely P. curtum villerealense, P. duvali duvali, P. muehlbergi thaleri, P. medium perrealense, P. crassum robustum, and P. magnum girondicum. The locality indicates that the multiple subspecies of Palaeotherium coexisted with the herpetotheriid Peratherium, theridomyids Blainvillimys and Theridomys, ischyromyid Plesiarctomys, glirid Glamys, hyaenodonts Hyaenodon and Pterodon, amphicyonid Cynodictis, palaeotheres Plagiolophus and Anchilophus, dichobunid Dichobune, choeropotamid Choeropotamus, cebochoerids Cebochoerus and Acotherulum, anoplotheriids (Anoplotherium, Diplobune, and Dacrytherium), tapirulid Tapirulus, xiphodonts Xiphodon and Dichodon, cainothere Oxacron, amphimerycid Amphimeryx, and the anthracothere Elomeryx.

Extinction
The Grande Coupure extinction and faunal turnover event of western Europe, dating back to the earliest Oligocene (MP20-MP21), is one of the largest and most abrupt faunal events in the Cenozoic record, which is coincident with climate forcing events of cooler and more seasonal climates. The result of the event was a 60% extinction rate of western European mammalian lineages while Asian faunal immigrants replaced them. The Grande Coupure is often marked by palaeontologists as part of the Eocene-Oligocene boundary as a result at 33.9 Ma, although some estimate that the event began 33.6-33.4 Ma. The event correlates directly with or after the Eocene-Oligocene transition, an abrupt shift from a greenhouse world characterizing much of the Palaeogene to a coolhouse/icehouse world of the early Oligocene onwards. The massive drop in temperatures stems from the first major expansion of the Antarctic ice sheets that caused drastic pCO2 decreases and an estimated drop of ~70 m in sea level.

The seaway dynamics separating western Europe from other landmasses to strong extents but allowing for some levels of dispersals prior to the Grande Coupure are complicated and contentious, but many palaeontologists agreed that glaciation and the resulting drops in sea level played major roles in the drying of the seaways previously acting as major barriers to eastern migrants from Balkanatolia and western Europe. The Turgai Strait is often proposed as the main European seaway barrier prior to the Grande Coupure, but some researchers challenged this perception recently, arguing that it completely receded already 37 Ma, long before the Eocene-Oligocene transition. Alexis Licht et al. suggested that the Grande Coupure could have possibly been synchronous with the Oi-1 glaciation (33.5 Ma), which records a decline in atmospheric CO2, boosting the Antarctic glaciation that already started by the Eocene-Oligocene transition.

The Grande Coupure event also marked a large faunal turnover marking the arrivals of later anthracotheres, entelodonts, ruminants (Gelocidae, Lophiomerycidae), rhinocerotoids (Rhinocerotidae, Amynodontidae, Eggysodontidae), carnivorans (later Amphicyonidae, Amphicynodontidae, Nimravidae, and Ursidae), eastern Eurasian rodents (Eomyidae, Cricetidae, and Castoridae), and eulipotyphlans (Erinaceidae).

The MP20 unit, the last before the Grande Coupure, marks the last appearances of most species of Palaeotherium, namely P. magnum, P. curtum, and P. muehlbergi. P. medium survived the Grande Coupure event based on its appearance at MP21, making it the last representative of its genus before its extinction by the unit. The extinction and faunal turnover result devastated many of the endemic faunas of western Europe by driving many mammalian genera to extinction, the causes being attributed to negative interactions with immigrant faunas (competition, predations), environmental changes from cooling climates, or some combination of the two.

Researchers have proposed theories as to why both P. medium and Plagiolophus minor survived the Grande Coupure event up to the early Oligocene whereas other species went extinct. Santi proposed that the dentition and cranial musculature of Palaeotherium were generally unsuited for the closed habitat turnovers caused by aridification and expansion of more open habitats, therefore being unable to adapt to the environmental changes. He also suggested that its poorer sight and hearing senses plus slow locomotion could have also made it more vulnerable to immigrant carnivores. The researcher then explained that P. medium could have survived longer than the other species of Palaeotherium because of its cursorial nature, with MacLaren and Nauwelaerts similarly stating that Plagiolophus minor was more well-suited to adapt to open plus drier habitats and immigrant predators than its relatives because of its smaller size and cursorial nature. Sarah C. Joomun et al. determined that certain faunas may have arrived later and therefore may have not played roles in the extinctions. They concluded that climate change, which led to increased seasonality and changes in plant food availability, caused certain palaeotheres and artiodactyls to become unable to adapt to the major changes and go extinct.