ZOOGEOGRAPHY. 157 



must have been I'onnecl by subsidence. Islands presenting such deep areas all 

 about them are simply, according to Suess (4, p. 638), great horsts, and "the 

 theory of the permanence of oceanic basins represented by Wallace, is for this 

 reason alone untenable." There is hydrographic evidence in the shape of 

 extensive shoals for most of the greatly extended Papuan land which we assume 

 existed between New Guinea and New Zealand, reaching out to and including 

 the Fiji Islands. There are deep clefts between the Solomons and New Cale- 

 donia, but they are comparatively limited in extent. Torres Strait is so shallow 

 as to need no comment. The deep water between Ke and Aru is remarkable, 

 as there can be no doubt that in comparatively recent geologic times there was 

 dry land either from Ke to Aru, or from Ke to Papua. 



Mention must be made of the extremely fragmentary data which we have 

 regarding the geology of many of the islands. This applies especially to many of 

 the Moluccas. Suess (2, p. 171) says that the observations made on Halmahera 

 are not sufficient to hazard even a conjecture as to the structure of the island, 

 though regarding other islands we have a certain amount of geologic or palaeonto- 

 logic information. We have spoken of the Tertiary deposit in Java, decidedly like 

 that of the Siwalik fauna of India, which is also of late Tertiary age. No such 

 extensive deposits ha\'e as yet been founel on Borneo, though Sumatran beds are 

 known. Mastodon, however, is known to occur on Borneo, Banka, and Sumatra, 

 as well as on Java. If we assume that at the time of the laying down of the 

 Javanese Tertiary beds Java was still in connection with Sumatra and the main- 

 land, and if we also agree that Java was the first of the Greater Sunda Islands 

 to break away from the mainland, then we see that this break must have taken 

 place in times more recent than Tertiary. This is not a convincing argument, 

 but it affords a working hypothesis. From the impoverished fauna which we find 

 upon the Lesser Sunda Islands we must assume that they wei'e separated from 

 Java earlier still. All are agreed to this. Celebes probably existed in about 

 the same shape that it now has, and it also became separated from the Greater 

 Sunda Island region at this time. We know from the researches of the Sara- 

 sins that in Eocene times Celebes was covered by the sea. Its connection with 

 Java was Pliocene (fide Sarasins), and along this connection came most of the 

 animals which we find in Celebes now. The existence of both a Java Bridge 

 and a Flores Bridge from Celebes seems now to be beyond contention. We have 

 hydrographic evidence for both bridges in the shape of chains of islands and 

 shoals; besides there are many species of animals from Java and Celebes which 

 do not, so far as we know, occur in the Lesser Sunda chain. Thus it would 



