ARU ISLANDS. 47 



only from New Guinea and Ceram. As a matter of fact there is a record for 

 Ternate, but it is quite unsubstantiated. This distribution is directly compara- 

 ble to that of the Death-adder, and as such is most interesting. 



So far, we know of only two amphibians from the Ke Islands : — Hyla doli- 

 chopsis (Cope) , which we know to be easily carried about, and Cornufer corrugatus 

 (Dum.), which does not seem so likely to have had a similar history. The 

 absence of amphibians in this case, however, must not be taken as offering evi- 

 dence against a land connection. Van Kampen has shown that the land con- 

 nections were probably in many cases so narrow as to lack suitable places where 

 amphibians might breed. Very probably the connections which existed here 

 were of this nature, and so very transitory that amphibians could not pass 

 across them. There is no special need to postulate that these Ke Island con- 

 nections had any special relation to the migration to New Guinea of its amphibian 

 fauna. 



Aru Islands. 



The Aru Islands differ considerably in physical features on the one hand 

 from the Ke Islands, which lie near by, and on the other hand from the opposite 

 coast of New Guinea. The group consists of what was probably recently a 

 single large low-lying island, heavily forested, and with many swamps and 

 estuarine water ways. Lying in general in a north and south direction, it is 

 some hundred miles long by forty-five miles wide. The divisions at the present 

 time are into low-lying islands, simply separated by swampy creeks; and the 

 major divisions are five in number — Kola, Wokam, Kobror, Koba, and Teran- 

 gan. To the westward lies the small island of Wammer, on which is the re- 

 nowned trading settlement of Dobo. Numerous other islands to the eastward 

 and southeastward are separated from the main land-masses by water only a 

 few feet deep. The group lies on the edge of a submerged bank of enormous 

 extent, which reaches out to the northeastward from Melville Island and the 

 Coburg Peninsula of AustraUa away to the western portion of New Guinea. 

 The major portion of the Arafura Sea is from 30 to 70 fathoms in depth; while 

 in the region in which we are interested — that is, between the Aru Islands and 

 the Timoraka district off New Guinea and Prince Frederick Henry's Island — 

 the depth of water runs from 16 to 25 or 28 fathoms. Between New Guinea and 

 Timor-Laut the water is deep — 459, 592 and 650 fathoms being depths which 

 are given on the most recent Dutch Admiralty charts; while the still greater 

 depths previously mentioned occur between the Aru and Ke groups. We see 

 at once, then, from this that these islands may be considered almost an integral 



