92 PROCEEDINGS OF THE NATIONAL MVSEUM. vol. 45. 



examined, otherwise the measurements of the tjjDe represent the 

 average for the species. For individual measurements, see table, 

 page 93. 



Remarks. — Tupaia siaca and T. castanea form a very interesting 

 and at same time puzzling gi'oup of treeshrews. Both are very 

 closely related to one another and geographically they are widely 

 separated. Pulo Bintang is about as far removed from Sumatra as 

 any island of the Rhio Archipelago, and no Tupaias of this group are 

 found on the intervening islands. In fact treeshrews are poorly 

 represented on the islands of the Rhio-Linga Archipelago. On 

 Batam, the next island to Bintang, occurs a treeshrew that is sepa- 

 rable with difficulty from T. glis ferruginea, and on Siiikep, an island 

 close to Sumatra, is another member of the glis group. Other tree- 

 shrews in the Archipelago are without interest in this connection. 

 Tupaia castanea and siaca as judged by color, are certainly related 

 to T. splendidula of Borneo. In point of size they equal T. glis 

 ferruginea, and I have been unable to find any definite constant 

 character in the skulls or teeth, to separate them as a group from T. 

 glis ferruginea. The question immediately arises what is theu" rela- 

 tion to T. g. ferruginea. The two forms, T. castanea and siaca may 

 be geographic representatives of that widely spread species. So far 

 as we know T. g. ferruginea does not occm- at the same localities with 

 them. Onl}^ on Borneo do we find T. splendidula occurring with 

 what is evidentlj^ a representative of T. glis ferruginea, that is T. 

 longipes. Those two forms are very different externally and cranially. 

 On the Anamba Islands are found Tupaias which in point of size and 

 lack of annulations on the tail are certainly members of the splendidula 

 group, but some of them in general body color resemble T. glis ferru- 

 ginea very closely. The available material indicates that typically 

 the glis ferruginea and splendidula groups are very distinct, but forms 

 of each occur strongly suggesting the other group. The relation 

 of T. demissa to the splendidula group is puzzling. It occurs on 

 Sumatra just to the northeast of the range of T. siaca. In the 

 general color of its upper parts it is very similar to T. siaca and T. 

 castanea. It has more grizzling along the sides than has the mfembers 

 of the splendidula group and less on the thigh than usual in T. g. 

 ferruginea. The underparts are more like those of T. g. ferruginea, 

 so is the skull. The hairs of the underside of the tail are certainly 

 lacldng in the annulations even more than the hairs of T. siaca. The 

 species demissa could with almost as much propriety be considered an 

 abnormal color form of T. siaca as it is so considered of T. g. ferruginea 

 Schneider's specimens recorded ^ as Tupaia custanea from the Indraghi 

 region Sumatra are T. siaca. 



Specimen examined. — Eight from the Siak region and two from 

 the Indragiri. 



1 Zool. Jahrb., vol. 23, 1905, p. 87. 



