User:Abyssal/CMN 8547

Semi-Aquatic Ceratopsid
In 2010 Jordan C. Mallon and Robert Holmes published a description of a partial ceratopsid skeleton. The remains preserved enough useful anatomical information to identify the specimen as a chasmosaurine, but not enough to determine what genus or species it was. Its stratigraphic origins in the second unit of the Horseshoe Canyon Formation suggest that it was either an Anchiceratops or Arrhinoceratops. This specimen has unique anatomical characteristics of the limbs and ribs. It had a unique vertebral count; 10 neck vertebrae, 13 thoracic vertebrae, 12 sacral vertebrae and 39 tail vertebrae. It unique features and "unusually robust" build suggest that it was a semi-aquatic hippo-like animal. Mallon and Holmes find additional support for this hypothesis from its depositional setting, which was a estuary during the Cretaceous.

This specimen was discovered at a Horseshoe Canyon Formation field site near Rumsey, Alberta near the end of summer on a 1925 expedition led by Charles M. Sternberg to the Red Deer River valley. Apart from the mostly missing skull, the specimen was both complete and articulated. Sternberg identified the specimen as belonging to Anchiceratops, although in 2010 Mallon and Holmes noted that he didn't provide a justification for this referral. The specimen was prepared to be used as a panel mount in 1929 and provided a skull from a cast of the Anchiceratops longirostris type specimen. Paleontologists Mallon and Holmes have expressed surprise that this high quality specimen hasn't received detailed examination in the scientific literature, although it received brief treatment in 1933 by Lull. Postcranial remains of ceratopsids that are both complete and articulated are so rare that only one other specimen, described by Brown in 1917, is of "comparable quality".

This specimen is catalogued as CMN 8547. The body is complete, but only fragments of its frill remain of the skull. Sternberg reported the presence of the front portion of the animal's snout in his 1925 field notes, but in their 2010 description of the material, Mallon and Holmes could not find it. The unguals of the left hand's first and third toe are missing, although their former presence with the specimen is documented by photographs from the field and of the original panel mount.

CMN 8547 was discovered on the east side of the Red Deer River 10.3 km southwest of Rumsey at a field locality catalogued as TMP locality L1508. Stratigraphically speaking, the fossils were recovered from near the top of unit 2 of the Horseshoe Canyon Formation. This horizon lay 26 m below Carbon coal zone, also known as coal seam 11. CMN 8547 was also 16 m below a bed of oyster fossils that comprised part of the Drumheller Marine Tongue.

CMN 8547 preserves four fragments of its frill, each of which is "flattened". Sternberg originally described these as originating from the frill's left side, but in 2010 Mallon and Holmes argued that they cam from the right side of the animal's frill based on their curvature. In thickness these fragments range from 28 to 30 mm at the edges of the frill and 7-10mm at the thinnest points. Two of the fragments originated from the edge of the frill where it had a "scalloped" appearance. Mallon and Holmes observed 3mm deep scarring on one of the frill fragments that was left by vascular sulci that ran roughly parallel to one another. These features have been documented in other ceratopsians "but are especially pronounced" in derived chasmosaurines like Anchiceratops longirostris, as seen in the specimen CMN 8535 and Arrhinoceratops brachyops, as seen in ROM 796.

When CMN 8547 was discovered, the specimen was laying on its right side with its left flank exposed. The panel mount was prepared for an exhibit in the Canadian Museum of Nature's Talisman Energy Fossil Gallery. Because the specimen is a panel mount, scientists can only examine its right side. This has hindered multiple attempts at describing the specimen's anatomy in detail, including Mallon and Holmes' 2010 description.

Distortion to the specimen after its burial was "minima[l]", but in order to accurately reconstruct the anatomy of CMN 8547 for thewir 2010 description, Mallon and Holmes had to position the shoulder blade at 45 degrees to the horizontal, space the ribs more widely, and rotate the ischium downward. The vertebral column of CMN 8547 is 4.05m long, with a total of 74 vertebrae. This resemble the vertebrae counts estimated by other workers for different kinds of ceratopsid. Centrosaurus has been estimated as having 77 vertebrae. In 1933 Lull estimated the vertebral count of Chasmosaurus belli at 76 although no full vertebral column is known for this species. The vertebrae of CMN 8547 are spaced roughly 1-2 cm apart except for the far end of the tail where they are spaced closer together.

The distinction between the vertebrae of the neck and body has been a subject of contention for paleontologists. Most researchers define neck vertebrae as those vertebrae with parapophyses on their vertebrae rather than their neural arches. A few have historically defined neck vertebrae based on "the short, straight ribs they support". This minority approach is problematical because "ribs are rarely preserved articulated with their supporting vertebrae". For the sake of consistency with the majority of workers and the aforementioned shortcoming, Mallon and Holmes used the parapophysis-based definition for neck vertebrae rather than the minority rib-based definition.

A ceratopsid neck is usually comprised of 6 typical vertebrae and a syncervical composed of 3 fused vertebrae, although some workers have historically misinterpreted it as being composed of 4 vertebrae. Therefore a typical ceratopsid has 9 total neck vertebrae. However, Mallon and Holmes argue that the syncervical of CMN 8547 is truly comprised of four vertebrae, unlike previous misinterpretations of fosslis that produced the same number. This would make CMN 8547's neck vertebrae count of 10 completely unique among known ceratopsids.

Mallon and Holmes in 2010 defined throacic vertebrae as those with parapophyses on their neural arches. Most ceratopsids have 12 thoracic vertebrae. In 1986 Ostrom and Wellnhofer inaccurately restored Triceratops as having 14 thoracic vertebrae. CMN 8547 has 13 thoracic vertebrae. Plaster hindered Mallon and Holmes' ability to determine whether or not the last thoracic vertebra is co-ossified to the first sacral like Centrosaurus, although it is possible. In 1933 Lull termed this configuration a dorsosacral.

Mallon and Holmes describe the synsacrum as "relatively long" in CMN 8547. Most ceratopsids have 10 sacral vertebrae, but CMN 8547 has twelve. Pentaceratops is known to have 11. The neural spines of CMN 8547's sacral vertebrae are "braced... by a tendon trellis". However, these fossilized tendons were damaged during preparation making it impossible for Mallon and Holmes to compare the tendon trellis of CMN 8547 with those of other ceratopsians. CMN 8547 had 39 tail vertebrae. This number is intermediate between Pentaceratops sternbergii, which had 30 and Centrosaurus apertus, which had 46. P. sternbergii and C. apertus are the only known ceratopsid species with complete tails.

CMN 8547 had unusually thick ribs. They average 39 mm thick from front to back at their midpoints. In Chasmosaurus russeli the average width and midlength is 37 mm and in Styracosaurus it's 27 mm midlength. CMN 8547 is larger than both even though Chasmosaurus russel has been estimated to be about 1.25 times as massive as CMN 8547. Styracosaurus albertensis has been estimated to be 1.5 times as massive as CMN 8547. The ribcage has been described as compact and rigid, although this is partially due to postburial distortion which gives an exagerated appearance of compaction and rigidity by compressing the ribs together.

The shoulder blade was discovered angled at 64 degrees to the horizontal, although Mallon and Holmes interpret it as being more likely that in life this angle was closer to 45 degrees. They compared its outline to those of Triceratops horridus and Pentaceratops sternbergii. The shoulder region of CMN 8547 was general typical compared to other ceratopsids. One of the potential sternal plates discovered near the shoulder region may be allochthonous in origin.

In their 2010 description of CMN 8547, they described its forelimb as being more "stout" than is typical for a ceratopsid. It humerus is almost as long as the typical ceratopsid value but is more robust. The absence of a medial tubercle in the humerus of CMN 8547 distinguishes it from all other ceratopsians. Mallon and Holmes dismissed the idea that this absence is attributable to post mortem damage on the specimen because the medial tubercle is absent on both humeri in field photos and the surface where it should be present seem "complete".

CMN 8547's radius was typical for ceratopsids, but its ulna is unusual. One notable trait is its prominen olecranon process compared to centrosaurines. CMN 8547 preserved two carpal bones, like most ceratopsids. These carpals were probably distal carpals three and four. The bones of its forepaw were unusually short and wide. These elements were generally more "robust" than those found in Centrosaurus apertus, Chasmosaurus belli, or Chasmosaurus irvinensis. Otherwise the forepaw is typical for ceratopsids and has the usual phalangeal formula of 2-3-4-3-2.

The pelvis of CMN 8547 fairly typical for ceratopsids. However, is ischium is unusually robust compared to Chasmosaurus belli, Agujaceratops mariscalensis, and Pentaceratops sternbergii.

CMN 8547 has longer hindlegs relative to the size of its ribcage than Centrosaurus apertus. Its femur does not appreciably differ in length from other ceratopsid, but is thicker. The tibia and fibula of CMN 8547 are proportionally similar to those of other ceratopsids. Its toe phalanges were more "robust" than those of Styracosaurus albertensis due to being shorter and wider.

As of 2010 only four ceratopsid species have been documented in the Horseshoe Canyon Formation. The only known centrosaurine from the formation is unit 1's Pachyrhinosaurus canadensis. Mallon and Holmes characterize the stratigraphic distribution of centrosaurs in the Horseshoe Canyon Formation as "wid[er]" than centrosaurines. Anchiceratops ornatus and Arrinoceratops brachyops are both known from units 1 and 2 of the Horseshoe Canyon Formation. Eotriceratops xerinsularis is known from the middle of unit 5.

From the time Lull, in 1933, endorsed Sternberg's 1925 referral of CMN 8547 to Anchiceratops no researcher has questioned its identity. Of the three known Horseshoe Canyon chasmosaurs, Eotriceratops ' s frill morphology least consistent with the fragments belonging to CMN 8547. Anchiceratops and Arrhinoceratops both have more similar frills and are known from the same unit of the formation as CMN 8547. Nevertheless, CMN 8547 preserves little information that would be useful for classifying at the genus or species level. Further, Mallon and Holmes noted that there were very few reliable characters to even distinguish Anchiceratops and Arrhinoceratops as different kinds of dinosaur.

The extreme rarity of postcranial ceratopsian fossils in the Horseshoe Canyon Formation complicates the identification of CMN 8547. The only specimen besides CMN 8547 with significant postcranial material is ROM 1493. This specimen preserves a skull and part of a forelimb. One of its forelimb elements is a badly weathered incomplete humerus. This humerus has a large deltopectoral crest like that found in CMN 8547. ROM 1493 has been referred to both Arrhinoceratops and Torosaurus, so CMN 8547 might be attributable to them as well, although Mallon and Holmes state that "more conclusive evidence is needed" to be sure.

Mallon and Holmes speculated that Sternberg's original referral of CMN 8547 to Anchiceratops might have been due to its status as the only known Edmontonian ceratopsid at the time rather than any distinguishing anatomical traits. Mallon and Holmes examined some of Sternberg's correspondance from the time period archived at both the Royal Ontario Museum and the Canadian Museum of nature and found no evidence that Sternberg new of the only recently named Arrhinoceratops.

Mallon and Holmes speculated that the increased numbers of vertebrae before the tail and the low count of tail vertebrae compared to Centrosaurus apertus suggests changes in the devlopmental process of the animal. Four of the tail vertebrae could have joined the rear of the sacral vertebrae while two of the front sacral vertebrae could have joined the thoracis vertebrae. This devlopmental hypothesis explains the distribution and number of vertebrae for each segment of the spinal columns. Similar developmental changes could have been at work to form Pentaceratops sternbergii ' s proportional long neck and body but short tail. However, it's likely that in the case of P. sternbergii this process was "less pronounced". Mallon and Holmes cautioned that without additional articulated skeletons to examine it can't be determined whether vertebrae counts varied by individual within this species.

In 1959 Langston concluded that Anchiceratops was common only in areas where other ceratopsids were based their depositional environments and his own personal observations in the field. He hypothesized that Anchiceratops lived only in "marshy" environments in low elevation areas characterized by mainly reductive chemistry. Langston noted that in CMN 8547 its long nose could have been useful for keeping its nostrils above the surface using its frill as a counterbalance. He saw CMN 8547's stocky body and short limbs as evidence it was a "sluggish" protected from predators by the water and isolation of its environment.

The idea that ceratopsids might be semi-aquatic lost traction in the scientific community many years ago. Mallon and Holmes, however, contend that recent arguments in favor of ceratopsids living primarily in wetland habitats call for more attention to be given to the idea. Crocodilians are the closest semi-aquatic modern relatives of ceratopsians, yet their specialized body plans and smaller sizes weaken comparisons between them. Hippos might be a modern analogue for what a semi-aquatic ceratopsid would be like in life. Some researchers have cautioned against ascribing excessive similarities with mammals to dinosaurs, but CMN 8547's has physical features in common with hippos. Other ceratopsians have features in common with hippos as well that may suggest a semi-aquatic lifestyle. Many of the potential adaptations for a semi-aquatic mode of existence in CMN 8547 could be beneficial for a terrestrial animal. However, Mallon and Holmes interpret the large number of candidate adaptations for semi-aquatic living in CMN 8547 as a strong cumulative case that it did so. They argue that the strong arm and chest muscles in CMN 8547 and its "stout proportions" were useful for maneuverability in the mud of its early Maastrichtian lowland swamp habitat. It depositional environment supports this interpretation since the rock's bearing the fossil were deposited in an estuary, the highest of such deposits found in the Drumheller Marine Tongue. Further, there were oysters associated with CMN 8547. In 2008, Tereschenko argued that the heterocoelous vertebrae found near the base of ceratopsids' tails and tall neural spines near their middle for sculling through the water were adaptations for a semi-aquatic mode of life. Mallon and Holmes noted that CMN 8547 did not provide support for this interpretation. However, Mallon and Holmes also cautioned researchers to "limit speculation" about ceratopsid aquatic behvior based on sedimentological evidence since ceratopsids are also known from environments with well-drained soils as well.

The uniqueness of CMN 8547's anatomy stands out among ceratopsians, which typically only varied from eachother in skull anatomy. Mallon and Holmes cautioned other researchers against relying on cladistic analyses that used CMN 8547 as a basis for Anchiceratops ' s postcranial anatomy. Mallon and Holmes describe evidence that some ceratopsids may have been semi-aquatic as "an increasing body of sedimentological, microsite, and paleosol data."

Reference

 * Mallon, J. C., and R. B. Holmes, 2010. Description of a complete and fully articulated chasmosaurine postcranium previously assigned to Anchiceratops (Dinosauria: Ceratopsia). In M. J. Ryan, B. J. Chinnery-Allgeier, and D. A. Eberth (eds.), New Perspectives on Horned Dinosaurs. Indiana University Press, Bloomington, Indiana.