VERTEBRJi: OF THE TORTOISE. 127 



three explanations may be given: one in conformity 

 with the structure of the thoracic-abdominal cage in the 

 crocodile ; the other based upon the analogy of that part 

 in the bird ; and the third agreeably with the phenomena 

 of development. According to the first, the median 

 piece of the plastron, called " entosternal," S, answers to 

 the sternum of the crocodile, or " sternum proper," and 

 the four pairs of plastron-pieces, es, hs^ ps^ xs^ answer to 

 the " hsemapophyses" forming the so-called sternal and 

 abdominal ribs of the crocodile. Most comparative ana- 

 tomists have, however, adopted the views of Geoffroy St. 

 Hilaire, who was guided in his determination of the 

 pieces of the plastron by the analogy of the skeleton of 

 the bird ; according to which all the parts of the plas- 

 tron are referred to a complex and greatly developed 

 sternum, and the marginal plates are viewed as sternal 

 ribs (hsemapophyses). The third ground of determina- 

 tion refers the parts of the plastron, like those of the 

 carapace, to a combination of parts of the endoskeleton 

 with those of the exoskeleton. 



In Fig. 21, the marginal plates m 1 to m 12, are 

 twenty -four in number, or twenty-six if the first (nuchal, 

 ch) and last (pygal, ]jy) vertebral plates be included. 

 Omitting these in the enumeration, three marginal pieces 

 intervene on each side at the angles between the first 

 median plate and the point of the first costal plate formed 

 by the end of the second dorsal rib, which point enters a 

 depression in the fourth marginal piece, m 4 ; the fifth, 

 sixth, seventh, eighth, ninth, and tenth marginal plates 

 are similarly articulated by gomphosis to the six suc- 

 ceeding ribs ; the eleventh marginal plate has no corre- 

 sponding rib ; the twelfth is articulated with the point of 

 the ninth dorsal rib supporting the eighth costal plate. 



