20 THE SALAMANDER 



three asymmetrical cases in which the ilium was attached to the 

 rib of the sixteenth vertebra on one side and of the seventeenth 

 on the other. The vertebra itself conforms to the general pattern 

 of the trunk vertebrae already described, but it is slightly larger 

 and its transverse processes and ribs are very stout and long. The 

 extremities of the ribs are pointed and not flat as are those which 

 support the shoulder. The ilia are closely bound to them by 

 fibrous tissue. 



6. The Caudal Vertebrae. 



The caudal vertebrae (Figs. 1 8 and 1 9) exhibit a gradual reduction 

 in the transverse processes which, however, never entirely disappear 

 even at the tip of the tail. The zygapophyses are also reduced in 

 accordance with the diminishing size of the vertebrae as a whole, but 

 they likewise are never entirely absent. As a rule only the first caudal 

 vertebra bears ribs and they are always very small. The most note- 

 worthy departure from the typical form exhibited by this section of the 

 vertebral column is the possession by all but the three anterior verte- 

 brae of a ventral ov haemal 2irch (h.ar.), through which passes the caudal 

 artery and vein. These arches are doubtless homologous with the 

 chevron bones of Reptiles and some Mammals, but they are com- 

 pletely fused with the centra of the vertebrae. The posterior end 

 of each haemal arch is spread out fan-wise and is slightly lipped, 

 while the anterior end narrows considerably and is splayed off very 

 obliquely on the ventral side so that it does not extend quite to the 

 anterior end of the centrum on the dorsal side. There is a low crest 

 running along the mid-ventral line, while the sides are perforated just 

 below the transverse processes by a foramen (fo.nt.) which permits 

 the exit and entrance of the segmental blood-vessels of the tail. The 

 whole series forms a bony canal as complete as the neural canal, and 

 quite unlike that formed by the slight 'Y'-shaped chevrons of the 

 Lacertilian type found in some Urodeles, e.g. Necturus. 



II. THE SKULL (Pis. I and II). 

 I. The Skull as a Whole. 



One of the earliest investigators to describe the skull in any detail 

 was Duges (i 834). His account is very condensed and is not much 

 more than a catalogue of the several bones, but he was the first to 

 show that the os triangulare is of branchial origin and has nothing to 

 do with the shoulder girdle. Then follows Parker's well-known 

 work (1882) in which the primordial skeleton is distinguished from 



