VERTEBRA AND RIBS 27 



Other trunk vertebrcC making a dorsal region. In other Tetrapoda 

 the number of cervical vertebrae varies extremely. These are charac- 

 terized by the entire absence of ribs or by extremely short ones, so 

 reduced that they do not reach the sternum. Again, not all post- 

 cervical vertebrae may bear ribs; those which bear ribs form a 

 thoracic region, those without ribs a lumbar region. 



Two of the vertebrae have received special names, the first, to 

 which the skull is attached, is the atlas. In most Tetrapoda this is 

 followed by the epistropheus (axis) which has not only its own 

 centrum, but that of the atlas has joined that of the epistropheus, 

 forming a pivot (dens epistrophei, odontoid process) on which the 

 atlas, and with it the head, turns. In a few reptiles and Erinaceus 

 (mammal) a bone, apparently a neural arch which has lost its cen- 

 trum, occurs between the atlas and skull (fig. 27). This is the 

 proatlas, the morphology of which is uncertain. As it is preformed 

 in cartilage it cannot be regarded as other than vertebral. 



CYCLOSTOMATA have a persistent notochord, but other parts of the 

 vertebral column are poorly developed and no ribs occur. There are some minor 

 differences between the two orders. In the trunk region of Petromyzon (fig. 83) 

 there are two pairs of neurapophyses to a somite, lying in the connective tissue 

 which surrounds the axial structures. They do not meet above the spinal cord, 

 the arches being incomplete. In the caudal region the same parts fuse to a 

 continuous plate of cartilage, with foramina for the exit of the spinal nerves, 

 which bears processes, usually regarded as dorsal spines. Near the end of the 

 tail in the haemapophyses form a similar plate with haemal spines. In Myxinoids 

 (fig. 234) no arches or neurapophyses are developed except in the tail where 

 both neur- and haemapophyses occur, fused into dorsal and ventral plates, 

 Myxine remaining in this respect on a level with the larval {Ammoccetes) stage 

 of Petromyzo)i. 



A small iossil, Paleos pond yl us, from the Scottish Devonian may be mentioned 

 here. Its systematic position is most uncertain. The skull, absence of paired 

 fins, a diphycercal tail and a caudal fin supported by delicate, often branched, 

 fin rays, have led some to regard it as near the Cyclostomes. But against this 

 may be mentioned the perfect vertebral centra (unknown in any Cyclostome) 

 and the dift'iculty of comparing the skull with that of recent Cyclostomes. It 

 is also suggested that it is a larval Dipnoan, but this does not account for the 

 centra. The vertebral column is long and slender, consisting of numerous centra, 

 each with its neutral arch, while both neural and haemal arches occur in the tail. 

 The vertebrtE seem to be more distinct in front. 



PISCES. — The vertebral column differs greatly in different fishes, 

 but some general statements may be made. It is most primitive 



