Abelisauridae

Skull of Abelisaurus.

Although skull proportions varied, abelisaurid skulls were generally very tall and very short in length. In Carnotaurus, for example, the skull was nearly as tall as it was long. The premaxilla in abelisaurids was very tall, so the front of the snout was blunt, not tapered as seen in many other theropods.[4]

Two skull bones, the lacrimal and postorbital bones, projected into the eye socket from the front and back, nearly dividing it into two compartments. The eye would have been located in the upper compartment, which was tilted slightly outwards in Carnotaurus, perhaps providing some degree of binocular vision. The lacrimal and postorbital also met above the eye socket, to form a ridge or brow above the eye.[4]

Sculpturing is seen on many of the skull bones, in the form of long grooves, pits, and protrusions. Like other ceratosaurs, the frontal bones of the skull roof were fused together. Carnotaurines commonly had bony projections from the skull. Carnotaurus had two pronounced horns, projecting outward above the eyes, while its close relative Aucasaurus had smaller projections in the same area. Majungasaurus and Rajasaurus had a single bony horn or dome, projecting upwards from the skull. These projections, like the horns of many modern animals, might have been displayed for species recognition or intimidation.[6][7][8] In Arcovenator, the dorsal margin of the postorbital (and probably also the lacrimal) is thickened dorsolaterally, forming a strong and rugose bony brow ridge rising above the level of the skull roof.[1] Possibly, this rugose brow ridge supported a keratinous or scaly structure for displays.

Hand bones of Carnotaurus, as interpreted by Ruiz and colleagues (2011)[9]

Data for the abelisaurid fore limbs are known from Eoabelisaurus and the carnotaurines Aucasaurus, Carnotaurus, and Majungasaurus. All had small fore limbs, which seem to have been vestigial.[10] The bones of the forearm (radius and ulna) were extremely short, only 25% of the length of the upper arm (humerus) in Carnotaurus and 33% in Aucasaurus. The entire arm was held straight, and the elbow joint was immobile.[10]

As is typical for ceratosaurs, the abelisaurid hand had four basic digits, but any similarity ends there. No wrist bones existed, with the four palm bones (metacarpals) attaching directly to the forearm. No phalanges (finger bones) were on the first or fourth digits, only one on the second digit and two on the third digit. These two external fingers were extremely short and immobile. Manual claws were very small in Eoabelisaurus, and totally absent in carnotaurines.[10]

More primitive relatives such as Noasaurus and Ceratosaurus had longer, mobile arms with fingers and claws.[11] Paleobiologist Alexander O. Vargas suggested a major reason for the evolution towards vestigial fore limbs in the group was because of a genetic defect; the loss of function in HOXA11 and HOXD11, two genes that regulate the fore limbs' development.[12]

Skeletal diagram of the known material of Carnotaurus

Complete skeletons have been described only for the most advanced abelisaurids (such as Carnotaurus and Aucasaurus), making establishment of defining features of the skeleton for the family as a whole more difficult. However, most are known from at least some skull bones, so known shared features come mainly from the skull.[6] Many abelisaurid skull features are shared with carcharodontosaurids. These shared features, along with the fact that abelisaurids seem to have replaced carcharodontosaurids in South America, have led to suggestions that the two groups were related.[23] However, no cladistic analysis has ever found such a relationship, and aside from the skull, abelisaurids and carcharodontosaurids are very different, more similar to ceratosaurs and allosauroids, respectively.[6]

Known material of Majungasaurus

Below is a cladogram generated by Tortosa et al. (2014) in the description of Arcovenator and creation of a new subfamily Majungasaurinae.[1]

Ilokelesia was originally described as a sister group to the Abelisauroidea.[19] However, Sereno tentatively places it closer to Abelisaurus than to noasaurids, a result which agrees with several other recent analyses.[6][20][26] If a stem-based definition is used, Ilokelesia and Rugops are therefore basal abelisaurids. However, as they are more basal than Abelisaurus, they are outside of the Abelisauridae if the node-based definition is adopted. Ekrixinatosaurus was also published in 2004, so it was not included in Sereno's analysis. However, an independent analysis, performed by Jorge Calvo and colleagues, shows it to be an abelisaurid.[20]

Some scientists include Xenotarsosaurus from Argentina and Compsosuchus from India as basal abelisaurids,[27][28] while others consider them to be outside the Abelisauroidea.[29] The French Genusaurus and Tarascosaurus have also been called abelisaurids but both are fragmentary and may be more basal ceratosaurians.[6]

Skeleton of Eoabelisaurus, a close relative of the Abelisauridae

With the description of Skorpiovenator in 2008, Canale et al. published another phylogenetic analysis focusing on the South American abelisaurids. In their results, they found that all South American forms, including Ilokelesia (except Abelisaurus), grouped together as a subclade of carnotaurines, which they named the Brachyrostra.[30] In the same year Matthew T. Carrano and Scott D. Sampson published new large phylogenetic analysis of ceratosaurian.[31] With the description of Eoabelisaurus, Diego Pol and Oliver W. M. Rauhut (2012) combined these analyses and added 10n new characters. The following cladogram follows their analysis.[32]

Timeline of descriptions of genera

Studies of the abelisaurid Majungasaurus indicate that it was a much slower-growing dinosaur than other theropods, taking nearly 20 years to reach adult size. Similar studies on other abelisaurid genera indicate that this slow maturation may have been a common trait to the whole of the Abelisauridae.[33]