778 MR. R. LYDEKKER ON A [Nov. 3, 



readily distinguished by the lower ridges, open valleys, less distinct 

 dentine trefoils, and much snrialler quantity of cement. Compared 

 with the large unworn third right upper true molar of M. latidens 

 from the Punjab, figured in plate xxxix. of the first volume of the 

 work cited ', it will be seen that the Borneo specimen agrees in the 

 number of ridges (although the hind talon is considerably smaller), 

 but is of greatly inferior size, the dimensions of the two specimens 

 being as follows, in inches : — 



Punjab. Borneo. 



Extreme length 8-6 Q-6 



Width of first ridge 4-2 2-9.5 



In plate xxxi. figs. 3, Sa, of Falconer and Cautley's ' Fauna Antiqua 

 Sivalensis ' there is figured on a reduced scale part of the right maxilla 

 of an example of M. latidens from Burma (of which there is a cast 

 in the British Museum) exhibiting two molars, which in the description 

 of the plate " are provisionally regarded as the last milk- and the first 

 true molar. This determination was accepted by the writer (who 

 had not then seen the specimen) in the ' Palaeontologia Indica,' ' 

 although it was remarked that the form of the second tooth (which 

 carries five ridges and a talon) must, on this view, be abnormal. 

 A comparison of that specimen with the Borneo tooth shows that 

 the second tooth in the former must be likewise the last true molar ; 

 and as its dimensions are 5*6 X 2*9 inches, it indicates an individual 

 agreeing very closely in size with the one to which the Borneo tooth 

 belonged. This Burmese specimen is also figured by Clift in the 

 Trans. Geol. Soc. ser. 2, vol. ii. pi. xxxvii. fig. 1, and is one of the 

 types. There is a very similar tooth in the British Museum (no. 

 37253) from Perim Island. 



The descriptive part of this paper may therefore be concluded 

 by observing that the specimen under consideration indicates the 

 occurrence of M. latidens in Borneo ; and also that a small race of 

 the species existed botb there and in Burma. The smaller race was 

 apparently of rare occurrence in the latter countiy ; and it will be 

 extremely interesting if future finds should show that the island 

 form always belonged to this race. 



Coming now to distributional considerations, it may be observed 

 that Mastodon latidens occurs in Perim Island in the Gulf of 

 Cambay *, and thence may be traced ° through Sind, the Punjab, and 

 the Western Himalaya to Burma ; from which point the present 

 specimen extends its range to Borneo. The present writer has 

 already shown ^ that some of the species of Indian Siwalik stegodont 



^ Owing to an inadvertence of the artist the specimen is viewed from the 

 outer instead of from the inner side. 



2 ' Falconer's Palajontological Memoirs,' vol. i. p. 463 (1868). 



3 Op. cit. vol. i. pp. 231-232. 

 ■* Specimens in Brit. Mus. 



= ' Palajontologia Indica,' op. cit. vol.i. pp. 228, 229. 



^ Ibid. vol. i. pp. 256 et scq , 268 ef scq. ; vol. ii. p. 289 (in this passage 

 Elejphas hombifrons is erroneously given for E. insignis) ; and Eec. Geol. Surv. 

 Ind. vol. xvi. pp. 158-161 (1883). 



