Date Published: February 13, 2019
Publisher: Public Library of Science
Author(s): Kate A. Andrzejewski, Michael J. Polcyn, Dale A. Winkler, Elizabeth Gomani Chindebvu, Louis L. Jacobs, Ulrich Joger.
A braincase of the Cretaceous titanosaurian sauropod Malawisaurus dixeyi, complete except for the olfactory region, was CT scanned and a 3D rendering of the endocast and inner ear was generated. Cranial nerves appear in the same configuration as in other sauropods, including derived features that appear to characterize titanosaurians, specifically, an abducens nerve canal that passes lateral to the pituitary fossa rather than entering it. Furthermore, the hypoglossal nerve exits the skull via a single foramen, consistent with most titanosaurians, while other saurischians, including the basal titanosauriform, Giraffatitan, contain multiple rootlets. The size of the vestibular labyrinth is smaller than in Giraffatitan, but larger than in most derived titanosaurians. Similar to the condition found in Giraffatitan, the anterior semicircular canal is larger than the posterior semicircular canal. This contrasts with more derived titanosaurians that contain similarly sized anterior and posterior semicircular canals, congruent with the interpretation of Malawisaurus as a basal titanosaurian. Measurements of the humerus of Malawisaurus provide a body mass estimate of 4.7 metric tons. Comparison of body mass to radius of the semicircular canals of the vestibular labyrinth reveals that Malawisaurus fits the allometric relationship found in previous studies of extant mammals and Giraffatitan brancai. As in Giraffatitan, the anterior semicircular canal is significantly larger than is predicted by the allometric relationship suggesting greater sensitivity and slower movement of the head in the sagittal plane.
Cretaceous titanosaurid material recovered from Malawi, Africa, was first described by Haughton (1928) and named Gigantosaurus dixeyi based on presumed similarity to specimens collected in Tanzania that were referred to as Gigantosaurus [1, 2]. However, the generic name ‘Gigantosaurus’ was preoccupied and was replaced with the generic name Tornieria by Sternfeld (1911) and G. dixeyi from Malawi became known as Tornieria dixeyi without further justification . The generic name Tornieria was later changed to Janenschia by Wild (1991) . Because the taxon from Malawi is distinct from the titanosaurid genus Janenschia, which was recovered from Jurassic beds in Tanzania, a new generic name, Malawisaurus, was erected by Jacobs et al. (1993) to accommodate the titanosaurid species from Malawi .
The material described, Mal 202–1, is currently on loan to the Shuler Museum of Paleontology at Southern Methodist University (SMU) Dallas, Texas and will be returned to the paleontological collection of the Malawi Department of Antiquities Lilongwe, Malawi. To produce a three-dimensional reconstruction of the endocast and inner ear, the specimen, Mal 202–1, was scanned at the University of Texas High Resolution X-ray CT facility using a voltage of 200kV and a current of 0.12mA producing 1707 slices with a voxel size of 70.6μm. Data from the scan were imported into Amira v 4.2 for analysis and visualization. The model was then imported into MeshLab where a laplacian smoothing algorithm was applied. Final rendering was completed in LightWave.
The bones surrounding the brain of Malawisaurus (Mal-202-1) are well preserved, but with anterior portions missing from the frontals and parietals (Fig 1). Bones present are completely ossified, and sutures are indistinct both optically and in CT data, suggesting the specimen represents a mature or adult individual. The braincase is exceptionally well preserved, but with a small amount of shear, the left side shifted slightly anterior relative to the right, exposing the internal structure of the endocranial cavity.
The braincase of M. dixeyi is well preserved with little distortion allowing for clear resolution of the internal anatomy. The digitally reconstructed endocast lacks the olfactory and cerebral regions (Fig 2). Characteristic traits of sauropods observed in the endocast include the presence of a well-defined and large pituitary fossa and the lack of distinction of gross regions of the brain, presumably obscured by the presence of overlying thick meninges and extensive venous sinuses in life [20, 22, 23, 24, 25].
The cranial nerves have an arrangement similar to other sauropods. The trigeminal nerve (V) is the largest of the cranial nerves and exits caudal to the infundibular region via a single foramen. The endocast shows little evidence for the division of the trigeminal nerve into the ophthalmic (V1), maxillary (V2), and mandibular (V3) branches as observed in Sarmientosaurus ; however, Gomani (2005) noted anterior and posterior grooves that exit ventral to the canal and are visible on the CT scans presented here. These grooves may have held maxillary and mandibular branches. The abducens nerve (VI) originates ventral to the trigeminal nerve and extends lateral to the pituitary fossa rather than entering it, which is a derived character state for titanosaurs [17, 24, 25]. The facial nerve (VII) originates posterior to the abducens and trigeminal nerves and passes ventrolaterally. A large opening posterior to the vestibular labyrinth serves as the passageway for cranial nerves IX-XI. The hypoglossal nerve (XII) exits via one foramen, consistent with most titanosaurs; however, Sarmientosaurus , Jainosaurus , and Pitekunsaurus  contain multiple rootlets.
The vestibular labyrinth of M. dixeyi is intermediate in size compared to the large labyrinth of Giraffatitan and the smaller sizes in advanced titanosaurs such as Jainosaurus (Figs 3 and 4). The rostral (anterior) semicircular canal is larger and is elevated dorsally compared to the caudal (posterior) semicircular canal similar to the condition observed in Giraffatitan (Fig 4). This supports M. dixeyi as a basal titanosaur as more advanced titanosaurs have approximately equal caudal and rostral semicircular canals. The lateral semicircular canal has the smallest diameter of the three, consistent with most sauropods; however, the lateral semicircular canal of M. dixeyi is longer and more slender in comparison to the lateral semicircular canal of other sauropods. The angle between the rostral and caudal semicircular canals is nearly orthogonal and similar to most titanosaurs except Sarmientosaurus.
Using a regression from Campione and Evans (2012), the body mass of M. dixeyi was estimated to be 4.73 metric tons based on the circumference of the humerus (Mal-221) found associated with the basicranium (Table 1). Measurements for the radii of the semicircular canals are shown in Table 2. Comparison of the measured radii of the semicircular canals and the predicted radii based on a regression of body mass from Clarke (2005) reveals a pattern similar to Giraffatitan brancai  (Fig 5). The caudal semicircular canal falls within the 95% confidence interval of predicted size, but the lateral semicircular canal is smaller than the predicted size. The rostral semicircular canal is significantly larger than the predicted size.
The lateral semicircular canal of M. dixeyi is longer and more slender compared to most sauropod taxa although M. dixeyi is a basal titanosaurian as shown by its endocranial structure consistent with other features of the skeleton , (Fig 4). This condition of the lateral semicircular canal is similar to Sarmientosaurus and may indicate increased sensitivity in the mediolateral plane emphasizing lateral scanning movements of the head and eyes . The angle between the semicircular canals of M. dixeyi is nearly orthogonal. A study by Berlin et al. 2013 concluded that deviations from orthogonality of the semicircular canals in mammals was negatively correlated with vestibular sensitivity . Research conducted by Malinzak et al. (2011; 2012) found mammals with the greatest deviations from canal orthogonality experienced slower head rotations during locomotion [29, 30]. Together this suggests Malawisaurus may have experienced higher angular head velocities during locomotion and increased vestibular sensitivity compared to Sarmientosaurus. Furthermore, the rostral semicircular canal of M. dixeyi is larger than the predicted size based on regression from Clarke (2005), which may indicate greater sensitivity. This condition supports behaviors including slower movement of the head in the sagittal plane .
CT scans of the braincase of Malawisaurus dixeyi recovered from the Dinosaur Beds of Malawi reveal insights into the paleoneuroanatomy and physiology of a basal titanosaur. The derived character state of an abducens nerve canal that passes lateral to rather than entering the pituitary fossa places Malawisaurus dixeyi within Titanosauria. The disproportionate size of the semicircular canals of the vestibular labyrinth with a larger rostral semicircular canal than caudal semicircular canal supports M. dixeyi as a basal titanosaur since derived titanosaurs exhibit equally sized semicircular canals. Body mass estimates based on circumference of the humerus are similar to estimates calculated using the radius of the semicircular canals with the caudal (posterior) semicircular canal falling within the predicted mass of 4.73 metric tons. This study has revealed the potential for imaging software in identifying important characters and new insight into physiology and behavior of extinct taxa.