SIWALIK AND NARBADA CHELONIA. 
9—163 
genus, and accordingly, at least as a provisional measure, tlie name Colossochelys is 
retained. Whether the non-union of the pygals indicates that Colossochelys participated 
in the subaquatic habits of Manouria cannot be determined. In the form of the 
epiplastron G. atlas differs very widely from the Aldabra and other recent gigantic 
land-tortoises, and apparently comes nearest to Manouria emys and perhaps the 
Madagascar Testudo radiata ; although in both cases the resemblance is a remote one. 
The xiphiplastron most resembles the corresponding part of Manouria emys and 
Testudo horsfieldi^ and differs from all the gigantic recent tortoises. The cranium 
provisionally referred to this species seems to indicate a form intermediate between 
Manouria emys and the gigantic tortoises of Aldabra. In size the largest individuals 
of C. atlas were about double the size of the largest males of T. elephantina ; while 
the smaller (? female) individuals were about one-half larger than the males of the 
latter, or about double the size of the females. 
Some features in the structure of the present species suggest a few considerations 
of a somewhat wider import. In the first place the non-union of the pygal plates 
in this pliocene form, as well as the occurrence of the same feature in all the emydine 
tortoises, which are very probably the older forms, may suggest that the union of 
these plates in all recent land-tortoises, except Manouria emys, is a specialized 
character of late origin ; in which case the last-named species will be the solitary 
survivor of a very ancient type — this being borne out by its subaquatic habits and 
wide distribution. 
With regard to the epiplastron, there can be little or no doubt but that the form 
which the gular plates present in Testudo horsfieldi, T. grceca, and numerous other 
species, is the most generalized, since it obtains in the majority of the emydine 
tortoises. In the prolongation of the gulars in front of the postgulars, with the 
retention of the V-shaped sutures between the two, the epiplastron of Colossochelys 
atlas and the other Siwalik species described below is evidently an advance on this 
generalized form. The next step is the epiplastron of Manouria emys, where the 
suture between the gulars and postgulars forms a very obtuse angle, and the former 
are almost entirely in advance of the latter ; and from this to the epiplastra of the 
Aldabra Testudo elephantina and T. ponderosa is but one step more, where the gulars 
are very small, and the suture dividing them from the postgulars forms almost a 
straight line. 
Bearing in mind this relation and taking into consideration the intermediate 
character of the cranium described above and provisionally referred to Colossochelys 
atlas between the crania of Testudo ponderosa and Manouria- emys, it appears not 
improbable that the Aldabra tortoises are a branch which having taken origin from the 
old Indian stock of gigantic land-tortoises,^ has developed very highly specialized 
characters in the cranium and plastron, and in common with other recent tortoises, 
has lost the divided pygals characteristic of at least one of the older Indian forms. 
1 Vide infra. The extreme specialization of the epiplastron of Colossochelys atlas probably indicates that this species 
forms a separate divergent branch. 
C 
