710 DR. c. I. FORSTTn MAJOR OX THE [June 1, 



As regards the pattern of the Hhizomi/s molars, it results from 

 a comparison with those of BracJu/uromifs and Tachi/ori/ctes, that 

 in Khizomys cusp 2 maintains its independence, whereas in the 

 others, as shown by B. heUileoensis, this cusp is not even in young 

 specimens strongly developed, and becomes soon fused with 4. 

 The homologies of the four outer cusps of lihizonn/s are therefore 

 to be expressed by the following figures: — 1; 4; 2; 54-3 (see 

 PI. XL. tigs. l«-7a). 



Of the lower molars of Rhizomys (PI. XL. tigs. 1 6-8 h) m. 1 is 

 constantly longer and more complicated than either of the two 

 posterior molars, which last agree almost absolutely with each other 

 in size and pattern. Internally some of the species show, when 

 young, four cusps, the two anterior of which are not strongly 

 separated, so that very soon the only remainder of the original 

 separation is a small enamel islet, which too tends to disappear. 

 In lUi. hadiiis, 1 and 4 appear already fused from the beginning, 

 i. e. in the youngest available stages. The homologies of the four 

 cusps, as compared with Brachyuromys, are therefore as follows : 

 1 : 4 (or 1 + 4); 2 + 5; 3. I feel jusliBed in considering the 

 third cusp to be, as in Brnchyuromys, a compound of two (2 + 5), 

 from what is visible in unworn posterior molars as compared with 

 m. 1 (PI. XL. fig. 16, 2 b, 4 b, 5 6, 6 b). M. 1 is besides distin- 

 guished by a surplus on the antero-internal side : not only is cusp 

 1 separated from 4, but, like m. 1 of Eh. betsileoensis, there is an 

 antero-external cusp in addition to what obtains in m. 2 and m. 3. 



The molars of Bhizomys therefore, besides being less hypselo- 

 dontthan those of Tachyoryctes,aTe also more complicated than the 

 molars both of the latter genus and of Brachyuromys, and thus 

 approach more to the brachyodout amongst Malagasy Eodents ; 

 and further on to some members {Trecliomys, Thendomys) of a 

 more primitive group, Winge's Anomalurida. The molars of 

 TacTiyoryctes and Brachyuromys, on their side, show a remarkable 

 likeness in pattern to some other members of the same group, 

 riz. Protechimys ', Arcliaomys. The molars of the former genus, 

 which are much more brachyodont than those of the latter, arri\e 

 at the more simplified pattern of ArcJitfomys, only in a somewhat 

 advanced stage of wear. In comparing the molars of these two 

 genera with each other and with Brarhyin-omys and the Bhizomyes, 

 it becomes evident, beyond doubt, that the simplified pattern of 

 the molars is the outcome of a complicated one. This is further 

 confirmed by the little we know of a Pliocene Rhizomys, viz. : — 



Bhizomys sivalensis, Lydekker '^, less specialized still than the exist- 

 ing Rhizomys. — I have reproduced the enlarged molars (PI. XL. 

 fig. 9 b) from a right mandibular ramus in the British Museum, 

 No. 15925, mentioned in the Catalogue. Lydekker says of the 



• Cf. " Theridomys hlainvilki, Gerv ," in Gerrais, Zool. et Pal. fpan^. 

 pi. 47. fig. 17. — M. Schlosser, " Die Nager des europaischen Tertians," Palseon- 

 tograpliioa. xxxi. 1884, pp. 6.S (4.5)-68 (50), pi. ix. (t). 



2 Records Geol. Surv. India, xi. 1878, pp. 100, 101. xii. 1870, p. 41 ; id. Pal. 

 Ind. X., iii. 1884, p. 107-108 ; id. Catalogue of tlie Fossil Maininnlia in the 

 British Mii.seuni (Natural History), i. London, 1885, p. 233. 



