|Fossil forelimbs of specimen IGM 100/15 at Nagoya City Science Museum|
Therizinosaurus (//; meaning "scythe lizard") is a genus of very large therizinosaurid dinosaurs that lived in Asia during the Late Cretaceous period in what is now the Nemegt Formation around 70 million years ago. The first remains of Therizinosaurus were found in 1948 by a Mongolian field expedition at the Gobi Desert. Fossils of this species were first described by Evgeny Maleev in 1954 and originally thought to belong to a turtle-like reptile. It is known only from a few bones, including gigantic hand claws, from which it gets its name. Additional findings comprising forelimb and hindlimb elements have been discovered during the 1960s-1980s.
Therizinosaurus comprises the single species T. cheloniformis, which could grow up from 9 to 10 m (30 to 33 ft) long and weigh possibly over 5 t (5,000 kg). They had the longest known claws of any land animal, reaching up to 1 m (3.3 ft) in length. Unlike other therizinosaurs, the claws were very stiff and elongated. But like other members, they would have been slow, long-necked/high browser herbivores equipped with a keratinous beak and wide torsos. The feet were equipped with four weight-bearing toes, resembling a little bit the unrelated sauropodomorphs. Therizinosaurus were very tall animals, likely outmatching predators like Tarbosaurus.
They were (along with Nanshiungosaurus) some of the last and the largest representatives of its unique group, the Therizinosauria. Therizinosaurus are classified as therizinosaurids, meaning they were more advanced than the primitive therizinosauroids. At first, therizinosaurs had very complex relationships due to the lack of genera and findings at the time. Originally, they were thought to be some kind of Cretaceous sauropodomorphs or transitional ornithischians, but after years of taxonomic debate they are now placed within the Theropoda, specifically as maniraptorans.
The first fossil remains of Therizinosaurus were discovered in 1948 by the Mongolian Paleontological Expedition of the USSR Academy of Sciences at the Nemegt Formation of the Gobi Desert, southwestern Mongolia. The expedition unearthed enormous manual unguals (claws) from the stratotype locality, Nemegt. This unusual specimen was labelled under the number of PIN 551-483 and consisted of three partial manual unguals, a metacarpal fragment and several rib fragments. Later on, the fossils were named and described by the Russian paleontologist Evgeny Maleev in 1954, who thought they belonged to a large, 4.5 m (15 ft) long marine turtle that used the giant claws to harvest seaweed. This specimen became the holotype for the new genus and species Therizinosaurus cheloniformis. The generic name, Therizinosaurus, is derived from the Greek θερίζω (therízo, meaning scythe, reap or cut) and σαῦρος (sauros, meaning lizard), in reference to the enormous manual claws and the specific name, cheloniformis, is taken from the Greek χελώνη (chelóni, meaning turtle) and the Latin formis, since the remains were thought to belong to a turtle-like reptile. Maleev only depicted a fragmented rib and the most complete claw and noted some inconsistencies regarding the affinities of this new taxon. Though it was not fully understood what general kind of creature these claws belonged, in 1970, the Russian paleontologist Anatoly K. Rozhdestvensky was one of the first authors in suggest that Therizinosaurus was a theropod dinosaur and not a turtle. Contrary to Maleev, he illustrated the three holotypic manual unguals and reidentified the metacarpal fragment as a metatarsal. The unusual shape of both metatarsal and ribs fragments led Rozhdestvensky to classify them as sauropod remains.
Further expeditions in the Nemegt Formation unearthed more fossils of this particular genus. In 1968, prior to Rozhdestvensky statements, the upper portion of a manual ungual was found in the Altan Uul locality and labelled as IGM 100/17. Another fragmented ungual was discovered at the Hermiin Tsav locality in 1972; this specimen, IGM 100/16, was found preserving its lower portion. During the year 1973, the fairly more complete specimen IGM 100/15 was also recovered from Nemegt, and consisted of the left and right arms compromising both scapulocoracoids (left is fragmented though), both humeri, right ulna with radius and left ulna, two right carpals, the right metacarpus including a complete digit Il, and some ribs with gastralia. All of these specimens were first described by the Mongolian paleontologist Rinchen Barsbold in 1976. In this new monograph, he observed that the rib fragments in IGM 100/15 are more slender than the ones of the holotype, and identified IGM 100/16 and 100/17 as pertaining to the first and third digit, respectively. During the same year and also from Hermiin Tsav, the specimen IGM 100/45 was discovered by the Joint Soviet-Mongolian Paleontological Expedition. Unlike the previous findings, this specimen is represented by a hindlimb composed of a very fragmented femur with tibia, astragalus, calcaneum, a lower tarsal, a tetradacyl pes compromising four partial metatarsals, nearly complete digits I, II and IV (although II and IV are missing the unguals) and the presumed second phalanx from the digit III. These newer remains were described by the also Mongolian paleontologist Altangerel Perle in 1982. He referred the specimen to Therizinosaurus based on the striking resemblance to Segnosaurus, another therizinosaurid genus also known from limb elements.
For maniraptoran standards, Therizinosaurus obtained enormous sizes, estimated to have reach 9 to 10 m (30 to 33 ft) in length with ponderous weights from 3.6 to 5 t (3,600 to 5,000 kg), or possibly over 5 tonnes. These dimensions make Therizinosaurus the largest therizinosaurs known and the largest known maniraptorans. Along with the contemporaneous Deinocheirus, these were the largest maniraptoriforms. Though the body remains of Therizinosaurus are relatively incomplete, inferences can be made about their physical characteristics based on more complete and related therizinosaurids. Like other members of their family, Therizinosaurus had proportionally small skulls bearing a keratinous beak atop long necks, with bipedal gaits and heavy, deep, broad bellies for foliage processing (as evidenced by the wide pelvis of other therizinosaurids) with the addition of sparse feathering. Their hindlimbs ended in four functionally, weight-bearing toes unlike other theropod groups in which the first toe was reduced to a dewclaw. In 2010, Senter and James used hindlimb length equations to predict the total length of the hindlimbs in Therizinosaurus and Deinocheirus. They concluded that an average individual of Therizinosaurus may have had approximately 3 m (9.8 ft) long legs. More recently, Mike Taylor and Matt Wedel suggested that the whole neck would be 2.9 times the size of the humerus, which was 76 cm (760 mm), resulting in a 2.2 m (7.2 ft) long neck based on comparisons with the cervical vertebrae series of Nanshiungosaurus.
Therizinosaurus can be distinguished from other therizinosaurids in having very straight manual unguals that are side to side compressed (flattened) and hypertrophied (elongated) with poor curves, an extended deltopectoral crest on the humerus, the length of metacarpal I is larger than 2/3 the length of metacarpal III, and the metacarpal I has an enlarged inner crest that connects the inner lower condyle and upper inner lobe. The most distinctive feature of Therizinosaurus was the presence of gigantic claws on each of the three digits of their front limbs. These were common among therizinosaurs but particularly large in Therizinosaurus, and they are considered as the longest known from any terrestrial animal.
The most complete arm of Therizinosaurus is represented by the specimen IGM 100/15 which preserves an almost complete right arm except for the left one which is more fragmented. The right arm covers 2.4 m (7.9 ft) in total length. The scapula measures 67 cm (670 mm) long with a stocky and flattened dorsal blade, wide acromial process and a very widened ventral surface; the blade seems to be broken at the end though. Near the anterior edge of the scapular widening and near the scapulocoracoid suture, a foramen is visible; it likely functioned as a channel for blood vessels and nerves when alive. The posterior edge of the scapula is robust and the acromion is lightly built, likely fused into a cartilaginous system with its periphery in life. The coracoid measures 36 cm (360 mm) in length, it has a broad and convex lateral surface that forms a slightly inclined concavity near of the scapulocoracoid suture. This concavity bends down towards the scapular widening. Near the scapulocoracoid suture, this edge turns very thin and possibly into cartilage along with the periphery of the coracoid in life, as the case of the scapular edge. A large foramen is also present on the coracoid. The glenoid is broad and deep, slightly pointing to the outer lateral side. It has robust, convex crest-like borders. The supraglenoid thickness is developed in a convex crest-shaped form, it is divided across the top of the scapulocoracoid suture. The attachment for the biceps muscle is prominently developed by a large tubercle with a stocky top, indicating powerful muscles in life.
The humerus is robustly built, measuring 76 cm (760 mm) long. It has a broad proximal end. The deltoid crest is tall and thick, its top is located approximately 1/3 from the upper end, the length of the crest is no less than 2/3 the length of the whole bone element. The lower end of the humerus is very expanded and flattened in a posterior direction. The condyles are developed onto the anterior side of the lower expansion while the epicondyles are very broad and project over the limits of the articular areas. The ulna is measured at 62.02 cm (620.2 mm) and most of its length is occupied by the straight shaft. The ulnar process is very wide. The upper articular area can be divided into the inner and outer lateral sides. The lateral side has a triangular-shaped border and is slightly concave; it is limited in a top view by the fossa for the upper articulation of the radius. The inner side forms a semilunar-shaped fossa that caps the lunar-shaped condyle of the humerus. The radius is 55.04 cm (550.4 mm) long and slightly S-curved. The upper end is flattened in a lateral direction, very wide and the distal end is highly robust.
The preserved right manus is composed of two lower carpal bones, complete three metacarpal bones and a complete second digit. The first lower carpal measures 8.23 cm (82.3 mm) tall and 8.53 cm (85.3 mm) wide having two articulations on the lowermost end. The upper surface of this carpal is divided by a broad fossa that forms the articulation of the carpus. On its inner side it has a triangular-shaped outline and attaches to the upper surface of metacarpal I occupying a little bit less than the lateral side, which articulates to metacarpal II. When they cross, these areas are separated by an oblique projection. The second lower carpal is smaller than the first one, it measures 5.6 cm (56 mm) tall and 5.93 cm (59.3 mm) tall. Its lower surface is flattened and the articular surface of the carpus extends from the first carpal to the second carpal over the articulation of the two bones.
The metacarpal I is 14.55 cm (145.5 mm) long and compared to the others is more robust and stocky. Overall, the lateral side is broad, especially on the uppermost area; the inner border is thin and narrow. The upper articulation is configured into three parts. The lower articular top is somewhat asymmetric, and bends to the inner side from the left, along with a wide, deep opening. The metacarpal II measures 28.68 cm (286.8 mm) in length and is the most elongated and robust. It has an inclined, square and flattened upper articulation. The articulation on the lower head has very symmetrical condyles, being divided by a broad, deep fossa. The lateral connecting openings are poorly developed. The metacarpal III covers 19.16 cm (191.6 mm) in length, it has a very thin shaft compared to the other metacarpals, and features an upper articulation that splits into three parts. The lower articular head is asymmetrical with deep and broad openings. As in metacarpal II, the lateral connecting openings are poorly developed. Only the second digit of the right manus was preserved, consisting of two phalanges and a large ungual. The first and second phalanges are somewhat equal in shape and length (14.17 cm (141.7 mm) and 14.38 cm (143.8 mm), respectively), they also share the robust and stocky structure. The upper articular facets are very symmetrical and preserve a crest that is particularly taller in the first phalanx. The top border of this crest is very pointed and thick; it likely served as the site for attachment of the extensor tendons in life. The lower heads are nearly symmetrical, but the central fossa is considerably wider and deeper in the first phalanx. The preserved unguals are specially enormous, estimated to be approximately 52 cm (520 mm) in length. Unlike other therizinosaurs, they are very straight, laterally flattened and not particularly curved with very sharp point tips. The upper articulation is symmetrical and has a cross-ridge that divides the surface into two asymmetrical parts. The size of the unguals IGM 100/16 and 100/17 may indicate that they belong to manual digits I and III, respectively.
The affinities and relationships of Therizinosaurus and kin were problematic in the early years of research. In 1954, Therizinosaurus were considered to have been giant marine turtles and the genus was assigned by Maleev to a separate Therizinosauridae given how enigmatic the type specimen was. The fossils remained with uncertainty among the scientific community, however, in 1970 Rozhdestvensky was one of the first paleontologists suggesting the idea that Therizinosaurus were actually theropod dinosaurs instead of some turtles. He considered that the large claws were used for a herbivorous life-style and open termites mounds. Also, the supposed ribs of the holotype were likely from a different dinosaur, possibly some sauropodomorph. With the discovery and description of Segnosaurus, in 1979 Perle named a new family of dinosaurs, the Segnosauridae. He tentatively placed the family as theropod dinosaurs given the similarities of the mandible and dentition to other members. A year later, the new genus Erlikosaurus was named by Barsbold and Perle in 1980. They named a new infraorder called the Segnosauria, composed by Erlikosaurus and Segnosaurus. They also noted that while aberrant and having ornithischian-like pelves, segnosaurs featured similar traits to other theropods With the discovery of the referred hindlimb to Therizinosaurus in 1982 by Perle, he concluded that Segnosaurus were very similar to the later based on the pes morphology and they possibly belonged to a single, if not the same group. In 1983, Barsbold named a new genus of segnosaur Enigmosaurus. With this discovery, he analysed the pelvis of this new genus and pointed out that segnosaurids were so different from other theropods that they could be outside the group or represent a different lineage of theropod dinosaurs. Later on the same year, he intensified the exclusion of segnosaurs from being theropods by noting that their pelves resembled those of sauropod dinosaurs.
Consequently, the assignment of segnosaurs started to shift towards sauropodomorphs. In 1984, Paul claimed that segnosaurs rather than being aberrant theropods were indeed, sauropodomorphs which successfully managed to remain into the Cretaceous period. He based this idea on anatomical traits such as the skull and similar pes configuration. He maintained his position in 1988 by placing the Segnosauria into the now obsolete Phytodinosauria, in addition, he was one of the first in suggest a segnosaur assignment for the long forgotten Therizinosaurus. Other prominent paleontologists like Gauthier or Sereno supported this vision. However, with the unexpected discovery and description of Alxasaurus in 1993, the widely accepted, sauropodomorph affinities of segnosaurs were highly questioned by Rusell and Dong. This new genus was far more complete than any other segnosaur and multiple anatomical features indicated that is was a member of the group and related to Therizinosaurus. With this, they identified the Therizinosauridae along with the Segnosauridae to be same group, the former name having historic priority. Due to some primitive characters present in Alxasaurus they coined a new taxonomic rank, the Therizinosauroidea, containing this new taxon and the Therizinosauridae. All of this new information provided new data on the affinities of the new-named "therizinosaurs"; they concluded that these animals were actual theropods, although uncertainties were still present. In 1994, Clark and colleagues redescribed the very complete skull of Erlikosaurus and even more theropod traits were found this time. They also validated the synonymy of the Segnosauridae with Therizinosaurida and classified therizinosaurs as maniraptoran dinosaurs. Nevertheless, against every opposition, a small and feathered therizinosauroid was described in 1999 by Xu and colleagues, the new genus Beipiaosaurus. It definitely confirmed the placement of therizinosaurs among theropods and also their taxonomic place on the Coelurosauria. This discovery also indicated that feathers were highly distributed among theropod dinosaurs.
In 2010, Lindsay Zanno revised the taxonomy of therizinosaurs with extensive detail. She found that many parts on therizinosaur holotypes were lost or damaged, and sparse specimens with no overlapping elements were disadvantages when concluding the relationships of the members. As an example of this, she accepted the referral of the specimen IGM 100/45 to Therizinosaurus since it matches multiple therizinosaurid traits but decided not include the specimen on its taxonomic analysis due to the lack of forelimb remains. She also excluded the supposed ribs that were present on the holotype since they likely came from a different animal and not Therizinosaurus. In 2019, Hartman and colleagues also performed a large phylogenetic analysis of Therizinosauria based on the characters provided by Zanno in her revision. They found similar results to Zanno regarding the family Therizinosauridae but this time with the inclusion of more taxa and specimens.
In 2012, Stephan Lautenschlager and colleagues analyzed the complete endocranial anatomy of Erlikosaurus and other therizinosaurs that preserve braincases. They found that therizinosaurids had well-developed senses of smell, hearing, and balance. The former two senses may have played an important role during foraging, predator evasion, and complex social behavior. These senses were also well-developed in earlier coelurosaurs and other theropods, indicating that therizinosaurs may have inherited many of these traits from their carnivorous ancestors and used them for their different and specialized dietary purposes. Lautenschlager in 2014 tested the function of various therizinosaur hand claws. He noted that while the claws of other therizinosaurs remained standardised and not highly specialized, Therizinosaurus used their claws when foraging vegetation and probably as threat display or gripping mates during mating as the claws are more elongated, straight and built to withstand higher levels of stress than other members. These would not have been used for digging, which would have been done with the foot claws because, since as in other maniraptorans, feathers on the forelimbs would have interfered with this function. However, he could neither confirm nor disregard that the hand claws could have been used for sexual display as in some species of turtles due to the lack of more complete specimens.
The discovery of fossilized embryos at the Nanchao Formation and multiple egg nests from other formations indicates that therizinosaurs had a colonial nesting-style similar to other herbivores like some sauropodomorphs, titanosaurs and hadrosaurs. Apparently, nestling therizinosaurs were capable of movements after birth and did not necessarily depend of their parents. It is also indicated that therizinosaurs did not have a philopatric behaviour when nesting.
The remains of Therizinosaurus were found in the well-known Nemegt Formation in the Gobi Desert, dating back to the Maastrichtian stage about 70 million and 68 million years ago. Over the time therizinosaurs have been suggested to have been slow, high browser theropods supported by the morphology of their dentition and broad torsos. The environments that Therizinosaurus inhabited consisted of large meandering and braided rivers with highly wooden terrains that supported diverse herbivorous dinosaurs like Therizinosaurus. Large, enclosed, canopy-like forests composed by Araucarias have been determined by the δ13C level preserved on the tooth enamel of many herbivorous dinosaurs and the numerous petrified wood across the formation. In addition, the presence of oasis-like formations are also reported. These structures may have served as areas of interest for the smaller oviraptorosaurs.
The dinosaur fauna of the Nemegt Formation was very diverse. Therizinosaurus shared its habitat with other theropods such as the deinonychosaurs Adasaurus and Zanabazar; ornithomimosaurs Anserimimus, Deinocheirus and Gallimimus; oviraptorosaurs Nemegtomaia, Elmisaurus, Gobiraptor, and the apex predators of the formation, Tarbosaurus. Other herbivorous dinosaurs include Barsboldia, Tarchia, Prenocephale, Nemegtosaurus, Opisthocoelicaudia and the most common herbivores, Saurolophus. Senter and James have suggested that Therizinosaurus were among the tallest dinosaurs in the Nemegt Formation. Due to their prominent heights and high browsing life-styles, these animals probably had no competition with other herbivores over the foliage. However, they also suggested a potential niche partitioning with the sauropods of the formation. If engaged in combat, a large individual of the predatory Tarbosaurus may have been not able to bite any higher than the thighs or belly of an adult standing Therizinosaurus. The gigantic claws likely functioned as optimal defense against this situation.