44 BARBOUR: ZOOGEOGRAPHY. 



the wide-ranging Cyclemys amhoinensis (Daud.)- Crocodilus porosus Schn. is 

 also known. 



Of the amphibians, eight in number, we find that four are pecuUar, two 

 representing genera not known elsewhere, Phrj^nixalus and Oreophryne. 

 Broadly speaking, these show affinities much more Papuan than Malayan, as 

 do also the other amphibians, except the two species of Rana. One of these is 

 Rana moluccana Bttgr., closely related to Rana varians Boulenger, which is a spe- 

 cies originally described from the island of Palawan in the southern Philippines, 

 but since discovered on Celebes; so that here again the direct connection with 

 Celebes is emphasized. The other frog, which is recorded as Rana macrodon 

 Dum. & Bibr., more probably represents R. modesta Blgr., which replaces the 

 former species upon Celebes and Talaut, and has been reported by van Kampen 

 from the island of Nusalaut near Ambon. 



To sum up, we see that this group of Moluccas shows conditions decidedly 

 different from those which we have already indicated for Ceram and the islands 

 about .\mbon. In many cases these northern islands show a closer relationship 

 to Papua than do the others; though, strangely enough, such characteristic 

 Papuan genera as Acanthophis and Pseudelaps occur upon Ceram, but not 

 upon Halmahera, to say nothing of the fact that upon Ceram also we find the 

 avian genus Casuarius, which occurs along with Acanthophis in the Ke Islands. 

 These discrepancies can not in this case be explained by probable ignorance 

 of the conditions, for the island of Halmahera we now know well. The only 

 explanation possible is either two independent connections with New Guinea, 

 or else a separation at different times of these two islands from a common land- 

 mass stretching out towards Papua. The hydrography of the region leads one 

 to believe most strongly, one might almost say surely, that there were two, 

 possibly more, separate connections. Halmahera was connected with Salawati 

 along the line of the many small islands now existing, and Obi joined to this 

 same bridge; while Ceram and Ambon may well have received their Papuan 

 types by means of a connection passing through the Ke Islands. 



Ke Islands. 



The Ke (variously spelled Ke, Kay, Kei, Key, etc.), or Ewaf Islands are a 

 small, heavily forested group lying directly south of Cape van den Bosch of the 

 Kumawa district, Dutch New Guinea. They are distant from this point some 

 seventy miles, and from the southeast point of Ceram some 160 miles. The 

 group itself consists of the island of Nuhu Jut, which is fifty miles long and only 



