518 MISS JOAlSr B. PROCTER ON THE 



the two forms of bone, seem to indicate that these are inde- 

 pendent structures. It will also be remembered that in strip- 

 ping the carapace from the young specimen of 42 mm., some 

 of the neural plates adhered to the carapace rather than to 

 the vertebra. Another point against endo-skeletal origin is the 

 alternating level at which the costal plates arise on consecutive 

 ribs, according to the arrangement of the epidermal shields 

 above them. 



This brings up the question, also unsettled, regarding the corre- 

 lation of epidermal shields to the underlying plates. If the plates 

 are of dermal or fascial (sub-dermal) origin, one presupposes such 

 a correlation ; if of endo-skeletal origin, there should be none. 

 Various authors argue the case accordingly, H. H. JS^ewman 

 opposing any idea of correlation very strongly, whilst Gadow 

 treats it as an accepted fact. G. H. Parker (1901), in his paper 

 on the Sculptured Tortoise, attempts to explain the alternating 

 arrangement of these two structures by applying Harrison's 

 observations on the regenerating tail of a tadpole (3899). 

 He says (p. 23) : — " The tail of a developing tadpole is com- 

 posed of an outer covering of ectoderm — which ultimately 

 gives rise to the outer layers of the skin — and of a core of 

 mesoderm. These two masses of tissue grow in very different 

 ways, so that as the tail lengthens, the ectodermic covering, 

 which is most actively produced anteriorly, slips posteriorly 

 over the underlying mesoderm, whose region of growth is chiefly 

 at the posterior end. Although this posterior migration of the 

 ectoderm has been actually demonstrated only in the tadpole, 

 there is reason to believe that it occurs in other vei-tebrates." 

 The present study points to a very marked and obvious correla- 

 tion between shields and plates in Testudo loveridgii, which, 

 having its carapace in what one might term an incipient stage of 

 development, thr-ows much light on the subject. A careful study 

 of man}'- species of Testudo and Cinixys during the period when the 

 plates are developing, brings forth a wealth of evidence that this 

 correlation is an indisputable fact in these genera. In newly- 

 hatched individuals of T. loveridgii (and also those of T. ihera, 

 grceca, horsjieldii, leithii, etc., gtc.) the bony plates are lacking 

 beneath the areolars, developing apparently as the shields them- 

 selves commence to form their growth-rings. In specimens of 

 different species the amount of plate which has accrued at the end 

 of the season's growth varies, but is at first in all cases greater 

 beneath shield-sutures (that is around the shield-edges) than else- 

 where. The neural plates beneath a shield-suture are developed 

 laterally, whilst the alternating ones are not. The costal plates 

 beneath the costal shield-sutures are at least twice as broadly 

 developed as those between ; their proximal portions, however, 

 alternate in breadth and narrowness inversely in strict accordance 

 with the ovei'lying shield-sutures. Moreover, in the male type- 

 specimen the X-ray plate {vide PI. III.) shows that the bony 

 plates are in perfect harmony with the abnormal third vertebral 



