34 COMPARATIVE ANATOMY 



there is a dorsal and a ventral shield (carapace and 2^iastron) con- 

 sisting of numerous pieces and completely enclosing the body 

 (Fig. 21). Both arise independently of the endoskeleton, which 

 is preformed in cartilage : that is to say, they are true exo- 

 skeletal membrane bones (cp. note on p. 71). The exoskeleton, 

 however, comes into the closest relation with the endoskeleton, 

 and may supplant it here and there : thus, in Testudo, for 

 instance, the thoracic and lumbar vertebrae and ribs become quite 

 rudimentary. 



Birds, as already mentioned in the chapter on the integu^ 

 ment, have no dermal skeleton, and this is true of all Mammals 

 except Armadillos (Dasypodidse). In these it consists pf a series 

 of movable transverse bony scutes covering the head and body 

 and of smaller plates on the tail and limbs. Sparse hairs 

 occur between these plates. It is very doubtful whether this 

 exoskeleton has been derived from that of Reptiles : more 

 probably it, like the horny exoskeleton of Manis (p. 26), has 

 arisen secondarily, and in consequence of its development the 

 hairs have become reduced. In Glyptodon, a fossil member 

 of this group, the dermal plates were firmly united together to 

 form a large shield which covered the whole body. 



2. ENDOSKELETON. 



I. VERTEBRAL COLUMN. 



An elastic rod, the notochord or chorda dorsalis, lying 

 in the longitudinal axis of the embryo between the neural and 

 visceral tubes (see p. 9), is the first part of the endoskeleton 

 to be formed, and is the fore-runner of the vertebral column. It 

 is developed as a ridge of the primitive hypoblast, from which it 

 becomes constricted off, and is therefore of epithelial origin. The 

 large parenchyma-like cells of which it is composed consequently 

 do not give rise to any intercellular substance ; vacuoles, however, 

 soon appear within the cells, the protoplasm of which undergoes 

 modification, and thus a retrogressive metamorphosis sets in 

 (Fig. 22). The fact that this occurs at such early stages of develop- 

 ment shows that the notochord must long ago iiave begun to lose 

 its primitive function, whatever that function may have been. 



As these degenerative processes are gradually carried still 

 further, only the walls of the cells persist in the greater part of the 

 notochord ; these become flattened by mutual pressure, so that they 

 appear like a meshwork of pith-cells. At the periphery, however, 

 the cells retain their protoplasm, and become arranged like 

 an epithelium. Around the notochord two sheaths (Fig. 22 A, B) 



