70 



Records of the Indian Musenin. [Vol. IX 



Types (skeletons). Rept. Ind. Mus, Nos. 1369, 2589, loio-ii, 

 1097 and 830. 



The same specimens form the types of Theobald's Melano- 

 chelys edeniana and of Anderson's Emys trijuga var. hurmana. 

 Apparently the latter name had priority in manuscript, and 

 Anderson ^ refused to resign it. Both in the British Museum 

 "Catalogue" and in the ''Fauna" there is an error as to the 

 dimensions of this race. Mr. Boulenger informs me that the largest 

 specimen in the British Museum has a carapace 29 cm. long by 

 direct measurement. The largest carapace in our coUection is 

 27-5 cm. long. It is that of an adult not yet aged, but two slightly 

 smaller carapaces have the lines of growth on the epidermal scales 

 completely obliterated and evidently belonged^to very old tortoises. 

 There is no specimen on record larger than the one in the British 

 Museum. In a shell 19 cm. long the foramina between the costal 

 and the marginal bones are not completely closed, as they are in 

 one 197 cm. long. They have completely disappeared in much 

 smaller shells of the other races. 



In examining a large series of skulls of the different races of 

 G. trijuga it does not seem possible to find any distinctive racial 

 character, except that skulls of adult individuals of edeniana and 

 thennalis are larger than those of the typical form, as might be 

 expected from the greater size of the whole animal. The thick- 

 ness of the bone that forms the temporal arch varies considerably ; 

 sometimes it is no thicker than that of the lower posterior part 

 of the orbital framework, sometimes it is nearly twice as thick. 

 In some of our older specimens of young skulls the arch appears 

 to be incomplete, owing to the falling away of the quadrato-jugal, 

 which is never firmly anchylosed to the tympanic frame except in 

 old individuals. In such cases, however, it is always possible to 

 detect on the outer margin of the framework a slight projection 

 above the point which had originally corresponded with the antero- 

 superior limit of the missing bone. 



The geographical distribution of the species as a whole is a 

 somewhat discontinuous one, extending all over Peninsular India 

 south of the Indo-Gangetic plain and perhaps penetrating into that 

 plain in the north-west, but not including the valley of the Ganges, 

 although it does include the greater part of Burma. It is a little 

 doubtful whether any race occurs in Chota Nagpur, in which G. 

 indopeninsularis may replace G. trijuga, but Anderson states that 

 the typical form occurs there. 



Specimens in the Indian Museitni. 



I. Geoemyda trijuga madraspatana (Anderson) 

 {= forma typica). 



831 (15b. A.S.B.), 2 Madras. Madras Museum. 



(stuffed). 

 1008-9, 9 S (skeletons) „ ,, 



