I got a request from Zach to draw Carnotaurus so here is my attempted reconstruction of C. sastrei. Funnily enough I have never drawn or even attempted to draw Carnotaurus previous to this - that is except for the unsatisfactory precursor to this picture...It took me forever to finish but for a first attempt I kinda think it turned out pretty good.
There are some really strange things with Carnotaurus, the first obvious being the skull. Staring at Bonaparte et al.'s (1990) skull reconstruction and photos for a prolonged time and reading their description, I noticed that the postorbital region of the skull in Carnotaurus is really strongly deflected ventrally while the sagittal crest is extremely high. This results in the supratemporal fenestrae having their lateral borders ventrally displaced compared to the medial borders. The sagittal crest in reptiles generally serves as the attachment sites for jaw adductor muscles, namely the M. pseudotemporalis superficialis (MPsTs), M. adductor mandibulae externus medialis (MAMEM), and MAME profundus (MAMEP) (Holliday and Witmer 2007), so an enlargement of the sagittal crest can allow for increased attachment area for these muscles. On the other hand, the posterior face of the transverse projection of the dorsal crest of the parietals serve as the attachment sites for the neck muscles namely the M. transversospinalis capitis (Snively and Russell 2007). The high position of the attachment site allows for a higher leverage for dorsiflexion of the skull by this muscle.
But then, perhaps the most striking feature of Carnotaurus is the frontal horns. For some reason I've always thought the horns were flatter and bladelike but on closer inspection, I was surprised to find how thick they actually are, though the dorsal surface is flat - hmm...I guess my reconstruction kind of looks like they're conical than flat on top...The horns have "a system of shallow grooves" making the surface quite different from that of the rest of the skull (Bonaparte et al. 1990). Apparently, the structure of the horns is not so different from that of horn cores in modern bovid so it is likely that horns in Carnotaurus also had some corneous covering making them longer in life.
Lastly, I was rather intrigued by the strange forelimbs. The ulna and radius are extremely short (1/4 the length of the humerus) and features on the articulation with the humerus indicate that the bones did not rotate significantly. This ristricted movement in the anteroposterior direction. Further, the parallel positions of the ulna and radius forced the palmer side of the manus to face dorsally/anteriorly (and laterally?) as opposed to ventrally/medially as in other theropods (Bonaparte et al. 1990). So apparently, the palms of Carnotaurus face outwards as opposed to inwards. Now this came as a complete surprise for me. The bones of the manus themselves are preserved but not in articulation and the distribution of the bones were interpretted tentatively (Bonaparte et al. 1990). Coria et al. (2002) report of an articulated manus in Aucasaurus providing a better understanding of abelisaurid manual morphology. In Aucasaurus, as in Carnotaurus, there are four metacarpals (mc I ~ IV). It seems unlikely that mc I and IV carried any phalanges, while mc II and III are joined to one and two phalanges respectively, but with no clear indications whether claws were present or not (Coria et al. 2002). Unlike Aucasaurus mc IV in Carnotaurus is more elongate and has a large articulation for the ulna.
So all in all, strange beast...