SUMMARY. 165 



SUMMARY. 



The following conclusions seem warranted by the evidence obtained. It is 

 to be emphasized that these conclusions are based primarily on herpetologic evi- 

 dence only. 



1. The fauna of the three Greater Sunda Islands, Sumatra, Borneo, and 

 Java, has been entirely derived from the Malay Peninsula region. The supposed 

 Papuasian element in the Javan fauna, which has been emphasized by Werner, is 

 probably entirely non-existent. 



2. The islands off the western coast of Simaatra have probably not formed 

 an ancient highway whereby the migration of animals once took place between 

 Java and the mainland, but have rather formed a part of three great peninsulas 

 which jutted out to the westward from Sumatra, and along which they received 

 the fauna which they support. This explanation is proposed in place of the one 

 previously suggested, that each island received its animal population from that 

 part of Sumatra which was nearest to it in point of distance 



3. The relation of Celebes to the regions lying eastward suggest that there 

 have been two bridges between Celebes and the Moluccas. 



4. The evidence is not conclusive that Obi ever had direct relationship 

 with Celebes or either of the Celebes-Molucca Bridges. The soundings suggest 

 that it may have received its fauna by having been joined to the old commissure 

 which stretched from southeastern Halmahera to Papua. 



5. Mysol supports a most anomalous fauna, if the records for the occur- 

 rences of species on this island can all be believed. The reptiles seem to show a 

 marked relationship with those of Ceram and the southern Moluccas, though 

 hydrographically the islands seem to be intimately related to Papua only, while 

 the mammaUan and avi faunas are both evidently purely Papuan. 



6. The distribution of the genus Casuarius among flightless birds is paral- 

 leled by that of Acanthophis, and by that of certain burrowing skinks; so that 

 Ceram seems to be connected with Papua more intimately and more recently 

 than any other of the Moluccas. This connection was probably by way of the 

 Southwest and Ke Island groups. 



7. The Ke Islands should not be considered equally with the Aru Islands 

 as having a Papuan fauna. The conditions in the two groups are fundamentally 

 different, and the creatures which they support show that the Ke Islands have 

 had direct relations with the Moluccas which the Aru Islands have not had, hence 

 their fauna is much more Malaysian in character. 



