158 BARBOUR: ZOOGEOGRAPHY. 



seem that Celebes had been in connection with Java since the Lesser Sunda chain 

 became separated from Java, and that it was probably at the same time in con- 

 nection with Flores. In this way we can account for such a distribution as that 

 of Sphenophryne, a Papuan genus found in Celebes and Lombok, but not occur- 

 ring in Java. If the Java Bridge was Pliocene, then, it would seem probable, as 

 the Sarasins beUeve, that this Flores Bridge was also PUocene; but the Lesser 

 Sunda break with Java was surely much earlier. This interruption was between 

 BaU and Lombok; and hence "Wallace's hne" has the local significance, if 

 nothing more. Though the Sarasins emphasize the essentially Indian character 

 of the fauna of Celebes, still we must bear in mind that the reptiles and amphi- 

 bians show extremely strong evidences of connection to the eastward. 



It is futile to spend time discussing the faunistic relations of Celebes, in 

 view of the completeness with which this has been worked out by the Sarasins. 

 It should be noted, however, that there is good ground for emphasizing the possi- 

 bility of there having been two bridges between Celebes and the Moluccas; for 

 we know that the birds of Buru, and the presence of the Babirusa there, suggest 

 at once a connection with Celebes; while a number of forms known from Hal- 

 mahera suggest a similar connection for that island, even though there is no 

 Babirusa on Halmahera; furthermore the birds of that island are widely different 

 from those of Buru. On the other hand, such forms as Rana moluccana, and 

 Typhlops ater have never been found on Buru. 



It is extremely unfortunate that we do not know more about Obi. There 

 is no evidence that would seem to justify the view that Obi was ever included, 

 as has been suggested, in either one of these two bridges from Celebes to the 

 Moluccas. Obi probably received its fauna by having been connected in the past 

 with the old commissure which stretched from the southern peninsula of Hal- 

 mahera to Papua. The Sula Islands would seem to have much more relation 

 to Buru than to Halmahera, and probably represent the remains of the Celebes- 

 Buru Bridge. There is no geograpliic evidence whatever as to the position of 

 the Celebes-Halmahera Bridge, unless it be that the small islands of Tifori and 

 Majo are the remains of this old land. The water about them is very deep; 

 but, as we have said before, this is not the only place where we must assume that 

 land has existed where now there is deep water. 



The remarkable distinctness between the faunas of Borneo and Celebes is 

 so well known as to need no emphasis here. To return to the Moluccas, we have 

 pointed out in our remarks on the different islands that Ceram had its connection 

 with Papua through the Ke Islands. We were unable to explain, however, and 



