ORTHOGENETIC VARIATION IN THE SHELLS OF CHELONIA. 215 



numbers 16 (Willey 11) and 28 (Willey 13), the 5th pair is somewhat smaller than the 

 rest. A reduction in size is likewise present in the 5th costal where this happens to be 

 the last but one, as in No. 12 (left side), No. 43 (left side); and especially in this last 

 specimen it is very obvious how the costals are reduced from six to five in number by 

 the last but one (or 5th pair) disappearing. Moreover a disturbance in this region is 

 indicated by the shape of the last but one neural in so far as this scute has often 

 eight or seven sides instead of its regular hexagonal shape, e.g. numbers 8, 9, 10, 14, 

 21, 22, 23, 26, 27, 43. 



Especially instructive are those asymmetrical specimens with six neurals and with 

 the five normal costal scutes on one side but still with six costals on the other side 1 . 



The marginal bones shew very little variation. The normal number consists of 

 13 pairs, the 5th of which forms the lateral angle of the carapace. Occasionally 

 there are as many as 14 marginals, and in these cases the angle is invariably formed 

 by the 6th; or there are only 12, and there the 4th forms the angle. This shews 

 that the reduction from 14 to 13 or to 12 is caused by the fusion of two neigh- 

 bouring marginals into one. In several new-born specimens such a fusion is still 

 incomplete. 



We know that almost all recent Tortoises and Turtles possess 10 pairs of ribs, 

 but 10 to 12 median dermal bones, eight of which are attached to, or fused with the 

 spinous processes of as many vertebrae, while one dermal bone, in front — the so-called 

 nuchal, — and one to three bones behind — the so-called pygal plates, are no longer in 

 connexion with a vertebra. 



Moreover it is reasonable to assume that originally each complete metamere or 

 segment in the region of the trunk proper had a vertebra, one pair of ribs, one 

 neural and a pair of costal dermal plates. Certain fossil Chelonians shew this arrange- 

 ment, or something approaching it. 



The suggestion is obvious that to each complete metamere belonged also a complete 

 set of epidermal scutes, namely, one median and a pair of lateral or costal scutes, 

 something like the arrangement of the armour in Crocodiles and allied fossil Reptiles. 



It is therefore reasonable to assume that a greater number of successive transverse 

 series of neural and costal scutes represents a phylogenetically older, more primitive, or 

 atavistic stage. 



It is almost certain that the alternating or interlocking position of the costals 

 and neurals is a secondarily acquired feature. Just as the regular hexagonal shape 

 of the neurals is not a primitive but a neat and comparatively late arrangement. 

 This consideration applies equally to the epidermal scutes and to the underlying 

 dermal, neural and costal bony plates. It is well known that the epidermal scutes 

 and the dermal plates do not at all correspond with each other, but scarcely any 

 attention has been drawn to the important fact that the relative position of these 

 various elements changes considerably during the growth of the individual creature; 



1 Of course it is quite conceivable that reduction from 6 to 5 costals cau be brought about in another 

 way. Specimen No. 4i affords such an instance by the vestigial second left costal, but I have at last come 

 to the conclusion that this little scute should be homologised with the 2nd scute of specimen No. 1 as a 

 very old and therefore rare survival. 



VV. III. 31 



