428 NOVITATES ZOOLOGICAF, XXII. IIUS. 



of botb races in the ^^ show ranch more prouonnced bosses or pyramids of the 

 scntes than the long and narrow ones ; ? ? of the smooth-scnted form, however, 

 never show any trace of the scntes being embossed. It would probably be found 

 that the long and narrow tortoises are all hybrids or descendants of hybrids 

 between Testudo daudinii and either gigantea or elephantina as the scntes are 

 smooth or striated. As, however, we can now never trace the exact origin of 

 the tortoises under consideration, I am relnctaiitly forced to inclnde under gigantea 

 all smooth-scnted individuals whether they be short or long, and all the specimens 

 with striated scntes under elephantina. A character noticeable and conspicnons 

 in very old and large tortoises of this group is that the plastral bridge becomes 

 enormously swollen and pushed ont, so that it projects far beyond the lower edge 

 of the marginals and is totally unlike anytliing found in large Galapagos tortoises. 

 It is, however, only in extremely old specimens that it becomes conspicnons, for 

 the very large <S in the British Mnseum (fignred here on Pis. xlix. and l.), 

 though of the largest dimensions, shows hardly any signs of this development. 

 Among the Seychelles and Aldabra-Madagascar tortoises no such definite saddle- 

 shaped carapaces occur as in the other two groups, but the type specimen in Paris 

 and the large adult <S of daudinii in the British Museum have the marginals 

 much everted and the front of the carapace considerably raised, so that we can 

 without impropriety say that daudinii represents the saddle-backed races of 

 the " Seychelles and Aldabra-Madagascar group " abingdonii, becki, ephippiuni and 

 plumtasticus of the " Galapagos group," and the saddle-backed Testudo vosmarri 

 from Rodriguez of the " Mascarene group." 



Dr. Van Denbnrgh was able, thanks to possessing numerous specimens with 

 exact data, to determine four categories of variation : 



1. Variation with age. 



2. Variation with sex. 



3. Variation with distribution. 



4. Individual variation. 



In consequence of there being only a'negligible quantity of Indian Ocean material 

 with any kind of data, and while fully conscious of having before my eyes all 

 these four classes of differences, I can only definitely deal with three — viz. 

 Nos. 1, ~;, and 4. I will take them seriatim : 



1. Variation with Age 



As in the Galapagos group of races, the very young tortoises are all more 

 or less of the same shape ; being more strongly dome-shaped than in most of 

 the adults. One outstanding feature is the enormously deep incised divisions 

 between the scutes in the young ones from J to J grown, so that in many 

 specimens each scute api)ears resting on a high platform. Tliis is not or very 

 rarely found in the Galapagos group. Another difference is that from a very 

 early if not the earliest age the young of T. gigantea have the scutes absolutely 

 smooth. We have no very yonng individuals of either T. micro//// i/es or 7'. giintheii, 

 the two smooth-scuted Galapagos dome-shaped races ; but the youngest we have 

 show stray striatioDS. As before mentioned, extremely old tJ<J of the group 

 we are dealing with have an extraordinary develojiment of the plastral bridge 

 not seen in younger c? cJ or in any ¥ ? . Very old individuals of both sexes show 

 an inclination, often very strong, for the supracaudal to curve round towards and 

 even under the posterior end of the plastron. 



