Miracinonyx

Description

Miracinonyx was larger than the modern cheetah. In 1979, Adams claimed that the largest individuals of Miracinonyx would have weighed up to 95 kg (209 lb). Van Valkenburgh and colleagues estimated the size of M. inexpectatus in 1990 based on a relatively complete fossil from West Virginia: skull length of 19.6 cm (7.7 in), backbone length (excluding the tail) of 106 cm (3 ft 6 in), tail length of 85–90 cm (2 ft 9 in – 2 ft 11 in) and body mass of 70 kg (150 lb). The same specimen was estimated to have stood 85 cm (2 ft 9 in) tall at the shoulder by the American paleontologist Alan Turner in 1997. The species was also likely comparable in size to Acinonyx pardinensis, a large extinct relative of the modern cheetah that weighed up to 100 kg (220 lb). The skull length and body mass of M. trumani were estimated around 17.7–18.2 cm (7.0–7.2 in),^{Table S1} and 50–88 kg (110–194 lb), respectively.^MOMv3.3

Skeletal proportion and variation

Anatomical similarities between Miracinonyx and the modern cheetah were enlarged nares and elongated distal (lower) limbs, which may have been adaptations for enhanced air intake and high speed during pursuit predation, respectively. Unlike the cougar, both taxa also had a shortened face and extremely reduced space behind the canine teeth. The postorbital processes, the bony projection on the skull located behind the eye socket, on the frontal bone that forms the forehead of Miracinonyx was less rounded in shape than the corresponding structure in the species of Acinonyx. While the skull morphology and elongated limb proportions were proposed to be likely analogous traits, Miracinonyx differed from the modern cheetah in its limb structure.

Comparison of the radius (outer large bone of the forearm) and muscle attachment sites of the tibia (shin bone) between the two taxa indicated that Miracinonyx had more robust, muscular limbs than the modern cheetah, likely capable of supination (rotating its forelimbs to face the palm forward or upward) with greater power to tackle prey. The fibula (calf bone) and tibia were fused in the modern cheetah but remained separate and parallel in Miracinonyx, which would have allowed more mobility at the ankle for the latter. The muscle attachment sites on the phalanges suggested that the claws of Miracinonyx were fully retractable unlike those of the modern cheetah. A distal phalanx of a juvenile M. trumani, preserved with the keratinous claw sheath and fleshy hood intact in the retracted position, provided further evidence that it had retractable claws. The carpal bone that forms the wrist of M. trumani was similar in shape to that of the lion but different from that of the cougar and cheetah, indicating that it likely used the dewclaw, the vestigial digit connected to the forelimb near the wrist, in a different way from that of the cheetah when taking down prey. The elbow morphology of M. trumani was intermediate between the ranges of variation observed in the cougar and modern cheetah, but more closely resembled that of the cougar. Additionally, the right hip bone of M. inexpectatus was straight in shape, unlike the bent-shaped hip bone observed in the species of Acinonyx.

Both species of Miracinonyx had similar skeletal morphology, but differed in size and robustness of their skeleton. The radius, humerus, tibia and femur (thigh bone) of M. inexpectatus were longer than those of the modern cheetah, cougar and M. trumani, but similar in size to those of A. pardinensis. The ratio between the length of the radius and humerus in M. inexpectatus was approximately 0.95, identical to that of the cougar and smaller than that of M. trumani and species of Acinonyx, 0.97 and 1.0 respectively. The distal ulna (inner large bone of the forearm) of M. inexpectatus was more robust than that of the modern cheetah and M. trumani, but more gracile than that of the cougar. The ratio between the length of the tibia and femur in both M. inexpectatus and M. trumani were about 0.95, larger than that of the cougar and smaller than that of the modern cheetah, 0.90 and 1.0 respectively. The body of femur in the cougar and both species of Miracinonyx were straight in shape, while the corresponding structure in the species of Acinonyx was more curved towards the front. The kneecap of M. inexpectatus was shorter than that of M. trumani, longer than those of the modern cheetah and cougar, and wider than those of all three species.

The ratio between the length of the lumbar vertebrae (spine segment that forms the lower back) and thoracic vertebrae was 0.78 in the cougar, 0.90 in M. inexpectatus, and 0.97-1.0 in M. trumani and the modern cheetah. The relatively long lumbar region of M. trumani was likely an adaptation for flexible forward bending of the spine like that of the cheetah, which would have aided in running at high speeds and producing longer strides. In the 7th lumbar vertebra of M. trumani and the modern cheetah, the transverse processes, small bony projections on each side of the vertebra, were shorter than those of the preceding lumbar vertebrae and directed more towards the front. In M. inexpectatus and the cougar, the transverse processes of the 7th lumbar vertebra were similar in size and shape to those of the preceding lumbar vertebrae. Additionally, similar to that of the cheetah and cougar, the hyoid bone (anchoring bone for the tongue and voice box) structure of M. inexpectatus suggested that this species likely lacked adaptations for roaring, though the corresponding bone structure in M. trumani was not identified.

Paleobiology

Adams interpreted the shared characteristics and proportions of the skull and limbs as probable evidence of close relationship between Miracinonyx and the modern cheetah, and the differences in limb structure as primitive traits of the former. Van Valkenburgh and colleagues argued that the similarities in cranial and limb morphology cannot be used to support Adams' conclusion, as they could instead be independently evolved adaptations for cursoriality (specialization in running) in distantly related and geographically separated carnivores due to similar ecological pressures rather than shared ancestry.

Brain

In 2022, Spanish paleontologist Borja Figueirido and colleagues compared the brain endocast of M. trumani to that of other felids, and found that the general brain shape of the modern cheetah was more highly domed (curved more upward) than that of M. trumani. They estimated the total endocranial volume (the internal capacity of the braincase) of M. trumani to be higher than that of the cheetah, the latter of which has the smallest brain size relative to total body size among living felids. The authors claimed that M. trumani may not have been as cursorial as the modern cheetah, under the assumption that the relatively small brain size of the latter was an adaptation to save energy and conserve body weight. The somatosensory cortex and visual cortex, the region of the brain that processes sensory information from the body and visual information respectively, were more developed in M. trumani than in the cougar, which would indicate that M. trumani may have relied more on eyesight than the cougar to detect prey due to enhanced vision. The auditory cortex, the region of the brain that processes sound, was more expanded in M. trumani than in the cougar, so the former possibly had better perception of sound as well.

Figueirido and colleagues also estimated the paranasal sinuses (air-filled spaces within the bones of the skull and face) volume of M. trumani, and compared it to that of the cougar and cheetah. The size of paranasal sinuses relative to the total cranium volume was 25% larger in the cheetah, but only 15-20% larger in the cougar and M. trumani. The authors argued that the larger, more vascularized (containing more blood vessels) paranasal sinuses of the cheetah were likely an adaptation to prevent the brain from overheating during high-speed chases, possibly indicating that M. trumani with relatively reduced volume in such spaces likely did not run as fast as the modern cheetah, or that it did not need to evolve such specializations as North America during the Pleistocene was colder than the current habitat of the modern cheetah, Africa and Iran. They further proposed that skeletal adaptations for high-speed locomotion in M. trumani may have developed before modifications in its brain structure, given the similarities in the limbs and differences in the brain and sinuses between M. trumani and the modern cheetah.

Diet and competition

M. inexpectatus likely occupied a similar ecological niche to contemporary felids and canids in the open grasslands of North America. Isotopic analysis of carnivorans from the Middle Pleistocene of Fairmead Landfill in Fairmead, California suggested that M. inexpectatus and other predators from this region primarily consumed herbivores that grazed on open prairies. Contemporary herbivorous mammals from the Fairmead Landfill were the antilocaprids Capromeryx and Tetrameryx, horse Equus, large camelids Camelops and Hemiauchenia, peccary Platygonus, ground sloth Paramylodon, Columbian mammoth and white-tailed deer. The carbon isotope values of M. inexpectatus were similar to those of the saber-toothed cat Smilodon and dire wolves, indicating dietary overlap and possible competition for the same prey. In contrast, its carbon isotope values were higher than those of the scimitar-toothed cat Homotherium, suggesting differences in their diets and potentially reduced competition. Fossils of jaguars have also been found in association with M. inexpectatus, so the two species coexisted and may have competed for prey.

M. trumani may have been able to hunt larger prey than the modern cheetah. A 47 kg (104 lb) modern cheetah and 88 kg (194 lb) M. trumani were estimated to be capable of hunting prey weighing up to 213.5 kg (471 lb) and 372 kg (820 lb), respectively. M. trumani was probably a generalist predator, with regional differences in diet indicated by variations in carbon and nitrogen isotope compositions between specimens from Yukon and Wyoming. The isotopic values of the Wyoming specimens overlapped with that of large, contemporary carnivores like wolves, bears and American lions, so the population from this region likely occupied similar trophic positions and may have exploited similar terrestrial herbivore prey that lived on temperate grasslands. Isotopic analysis also suggested that M. trumani from Wyoming preyed not only on pronghorns, but also on larger herbivores such as the bighorn sheep, bison and horses Equus and Haringtonhippus. In contrast, the Yukon specimens exhibited approximately 6% higher nitrogen isotope values than contemporary terrestrial carnivores and herbivores. These values fell within the range observed in piscivores (fish-eaters), so the Yukon population was likely specialized in hunting aquatic prey, potentially including anadromous fish like salmons which migrate from the sea to the freshwater to spawn. In Yukon, M. trumani experienced harsh winters and coexisted with numerous large carnivores, including the giant short-faced bear Arctodus, scimitar-toothed cat Homotherium, cave lion Panthera spelaea, brown bears and wolves, so the population likely evolved such strategy to reduce competition with other predators. M. trumani also coexisted with large herbivorous mammals of the mammoth steppe in Yukon, including the woolly mammoth, Yukon horse, steppe bison, muskox, moose, wapiti, caribou, dall sheep and saiga. In the Grand Canyon of Arizona, M. trumani inhabited high-altitude mountainous environments and may have preyed on bighorn sheep and Harrington's mountain goat, whose fossils were the most abundant among ungulates in this region. Puncture marks found on the skull, lower jaw and backbone of the subadult specimen from the Grand Canyon matched the shape and spacing of the canine teeth in M. trumani, suggesting that the species also engaged in conspecific aggression, fighting between members of the same species.

Disputed theory of evolutionary arms race

In 1988, American paleontologist John Chorn and colleagues reported fossils of modern pronghorns from the Natural Trap Cave of Wyoming and argued that the pronghorns developed adaptations for high speeds to escape from the contemporary M. trumani, whose fossils were the second most abundant among predators in the cave after those of wolves. The authors claimed that M. trumani and the modern cheetah filled the same ecological role and niche, preferably preying on relatively small, fast-running ungulates, and proposed that M. trumani and the pronghorns evolved cursoriality through an evolutionary arms race, a competition between two groups of organisms developing adaptations and counter-adaptations against each other over generations. In 1997, American biologist John A. Byers argued that pursuit predators, particularly the species of Miracinonyx, were the primary reason that caused the selection for cursorial adaptations in the species of Antilocapra, the genus that includes the modern pronghorn. This theory was depicted in the 2002 six-part BBC nature documentary Wild New World, which focused on the Pleistocene megafauna of North America.

However, American biologist Gina M. Semprebon and Spanish paleontologist Florent Rivals disagreed with Byers' theory in 2007. The authors contended that the Antilocapridae, the family that includes the modern pronghorn, developed cursorial adaptations before the existence of pursuit predators in North America, and cautioned against attributing such adaptations to a particular selective pressure without sufficient evidence. In 2017, American paleontologist David Levering and colleagues claimed that North American ungulates diversified in the late Miocene epoch, probably by evolving longer limbs for fast, efficient movement to adapt to the expansion of open grasslands. The authors argued that longer limbs would have allowed ungulates to take longer strides in open habitats, reducing the energetic cost of long-distance travel while increasing running speed, and suggested that predation was unlikely to have been a major selective pressure driving the evolution of these cursorial adaptations. However, based on morphospace analyses (graphical representations of variation in organism's morphology), Levering and colleagues noted that the Miocene antilocaprids and contemporary North American camelids and equids already possessed cursorial limb adaptations before or during the Miocene vegetation changes, possibly indicating that these vegetation changes were not the selective force driving the adaptations, or that open habitats existed in North America before the expansion of grasslands.

To identify the factors that shaped the evolution of antilocaprids, Brazilian paleontologists João C. S. Nascimento and Mathias M. Pires compared the estimated diversification and extinction rates of the antilocaprids with those of other contemporary North American fauna and flora in 2025, and concluded that environmental changes, competition with other herbivores and predation were all significant factors. The authors supported that the diversity of Antilocaprinae, the subfamily of Antilocapridae, likely increased during the late Miocene epoch, approximately from 10 to 7 million years ago, due to the competition among herbivores driven by rapid increase of open grasslands which coincided with the estimated time of diversification in antilocaprines. They also suggested that the feline predators like Miracinonyx possibly influenced not only the cursorial adaptations of antilocaprines, but also the decline in their diversity based on the correlation between the estimated diversification rates of felines and extinction rates of antilocaprines in North America. To observe whether cursorial adaptations developed gradually within antilocaprids or were present early in their evolution, American paleontologists Fabian C. Hardy and Anne E. Kort compared the ankle bone proportions of the modern pronghorns with those of the Miocene antilocaprids in 2026, since they considered the relatively short ankle bones as adaptations for high speeds in artiodactyls (even-toed ungulates) like pronghorns that live in open environments. The authors concluded that no significant difference in ankle bone proportions can be found between the modern pronghorns and their early relatives of the Miocene, suggesting that the antilocaprids developed cursorial adaptations prior to the first appearance of Miracinonyx and that their limbs were already adjusted to open habitats. Based on isotopic evidence and fossils of M. trumani from the Arctic environments of Yukon, Cassatt-Johnstone and colleagues proposed that the species was not exceptionally specialized like the modern cheetah for hunting cursorial prey such as pronghorns, but had a more flexible diet comparable to those of modern lions and cougars, both of which also have broad geographic ranges.

Distribution

Fossil remains of Miracinonyx have been found as far south as Mexico, across the United States and as far north as Yukon, Canada. Fossils of M. trumani have been found in Arizona, California, Mexico, Nevada, New Mexico, Wyoming, and Yukon. The specimens from Arizona, California, Mexico and Yukon were initially misidentified as those of the cougar, but have since been reclassified as M. trumani, indicating a broader geographic range for the species in the Late Pleistocene than previously thought. Specimens comparable to and likely representing M. trumani were known from phalanx assigned to as cf. M. trumani in Colorado and partial tibia assigned to as M.? trumani in South Carolina. Some Late Pleistocene fossils originally identified as M. cf. M. inexpectatus from Florida and Texas probably represent M. trumani instead.

Fossils of M. inexpectatus have been found in Arkansas, California, Florida, Maryland, Pennsylvania, South Carolina, Texas, and West Virginia. An isolated 3rd premolar from the Pliocene-aged Hagerman Fossil Beds National Monument of Idaho was assigned to M. inexpectatus, dated between 4.2 and 3.1 million years ago, but its specific depositional age (the exact time when the rock layers were accumulated) within this range cannot be determined due to the lack of locality data. Specimens comparable to M. inexpectatus were known from teeth assigned to as M. cf. M. inexpectatus in Colorado, teeth and phalanges assigned to as F. cf. F. inexpectata in Georgia, and partial skull and teeth assigned to as F. cf. F. studeri in Mexico. A possible skeleton of this species has also been informally reported from southwestern Virginia.

M. inexpectatus lived between the Late Pliocene and Middle Pleistocene (Blancan and Irvingtonian faunal stages), with the earliest definitive record in the Cita Canyon of Texas, dated between 3.04 and 2.58 million years ago, and the youngest fossil record in the Conard Fissure of Arkansas, dated to approximately 600,000 years ago. M. trumani lived during the Late Pleistocene (Rancholabrean faunal stage), with the oldest possible fossil assigned to as cf. M. trumani and radiocarbon dated to 52,200 years Before Present from the Mineral Hill Cave in Elko County, Nevada.^MEGA14C The youngest possible phalanx bone assigned to as cf. M. trumani was discovered from the Haystack Cave of Gunnison, Colorado, specifically within the lower rock layers of the cave numbered 4-8. Radiocarbon dating on the bone fragments of an unidentified mammal from the rock layers numbered 3-4 of the Haystack Cave suggested that the depositional age of these rock layers can be possibly dated to 19,659-16,604 calibrated years Before Present (14,935 ± 610 uncalibrated years Before Present), which has been considered as the latest occurrence of M. trumani.^{Table S3}