Fig. 97. Longitudinal section taken rather to the right of the middle line of the skull of a young Indian Elephant (E. Indicus) × 1/8. (Camb. Mus.)
| 8. anterior nares. | 12. pterygoid. |
| 10. periotic. | 17. nasal. |
| 11. palatine. | Other numbers as in Fig. 96. |
The mandible has a high ascending portion, is rounded off below and has no angle. The symphysial portion is long, narrow, and spout-like, and the coronoid process is small. The thyro-hyals are ankylosed with the basi-hyal, which is connected with the large forked stylo-hyals by ligament only.
Rodentia. The cranial cavity is depressed, elongated, and rather small, and the cerebral fossa lies entirely in front of the cerebellar fossa. The occipital plane is vertical or directed somewhat backwards, and the supra-occipital does not form much of the roof of the cranium. The paroccipital processes of the exoccipitals are generally of moderate size; in the Capybara (Hydrochaerus), however, they are very long, and are laterally compressed and directed forwards. The parietals are small, and often become completely fused together; there is sometimes a small interparietal. The frontals in most genera have no trace of a postorbital process; in Squirrels, Marmots and Hares, however, one occurs, but in no case does it meet a corresponding process from the zygomatic arch, so the orbit and temporal fossa are completely confluent. In Hares the postorbital process of the frontal is much flattened, and has an irregular margin. The temporal fossa is always small, and in Lophiomys is arched over by plates arising respectively from the parietal and jugal; a secondary roof is thus partially developed in a manner unique among mammals, but carried to a great extent in many Chelonia. The nasal bones and cavities are large, attaining their maximum development in the Porcupines (fig. 98, 1). The premaxillae is always very large, and sends back a long process which meets the frontal. The vomer is occasionally found persisting in two separate halves, a feature recalling the arrangement in Sauropsids. In many Rodents there is an enormous vacuity at the base of the maxillary portion of the zygomatic arch. It is sometimes as large as the orbit, and attains its maximum development in the Capybara and other Hystricomorpha; in the Marmots, Beavers, and Squirrels (Sciuromorpha), and in the Hares it is undeveloped. In Lagostomus the maxillae bears an upwardly directed plate of bone, shutting off from this vacuity a space which is the true infra-orbital foramen.
Fig. 98. Half front view of the skull of a Porcupine
(Hystrix cristata) × ½. (Camb. Mus.)
| 1. nasal. | 5. premaxillae. |
| 2. maxillo-turbinals. | 6. jugal. |
| 3. infra-orbital vacuity. | i 1. upper incisor. |
| 4. maxillae. |
The zygomatic arch is always complete, and in many cases the jugal extends back to form part at least of the glenoid surface for articulation with the mandible. In Coelogenys the jugal and maxillary portion of the zygomatic arch is greatly expanded and roughened, and the maxillary portion encloses a large cavity. The palate in Rodents is narrow, and the space between the incisor and molar teeth passes imperceptibly into the sides of the face. The anterior palatine foramina form long, rather narrow slits in this region. The bony palate between the grinding teeth is sometimes as in the Hares very short, sometimes as in the Capybara very long. The maxillae extends back beneath the orbit to unite with the squamosal. The pterygoid is always small, but sometimes has a well-marked hamular process which in Hystrix, Lagostomus, and some other genera unites with the tympanic bulla. The periotic is large, and fused with the tympanic, which forms a prominent bulla, and is generally drawn out into a tubular meatus. The bulla attains its maximum development in Chinchilla and Dipus.
The mandible is narrow and rounded in front, the two halves meeting in a long symphysis. The angle is generally drawn out into a long backwardly-projecting process, which is often pointed and directed upwards. In the Hares the angle is rounded. The coronoid process is never large.