208 SHUPELDT. 



of the specimens from the Philippine Islands, at least one-third larger 

 on comparing skulls of equal size. (As 144655 9 Basilan Island and 

 84420 S Trong, Lower Siam.) This character is constant and eminent- 

 ly distinguishing. 



One of the most interesting characters on this superior view of the 

 skull is the variation in the extent of the area of the temporal fossa, 

 and the fact as to whether or not they meet in the median line posteriorly, 

 and to what extent, if they do meet. This character I have carefully 

 examined in the 53 skulls at hand belonging to the National Museum 

 and the Bureau of Science, and in the few skulls of my own collection. 

 On any specimen the surface of one of the temporal fossje, selecting 

 either side, may be more or less rough. In all forms these areas are 

 very distinctly defined, with the limiting boundary line more or less 

 raised. Now in no Malayan form of flying lemur do the temporal 

 fossaj at all approach each other posteriorly : the interval often being 

 nearly one centimeter in adult skulls of either sex. This appears to be 

 the case with the young of Cynocephalus (number 144661 9 ) and still 

 more marked in the young of Galeopterus. ■ 



Among the Philippine forms this character varies, though in none of 

 them do the temporal fossae ever approach each other anywhere as closely 

 as in any of the Malayan forms. 



Eeferring to the numbers in the above tables for the specimens and 

 their localities the following are found to be the intervals between these 

 fossse. In number 144663 it could not be determined as the skull is 

 too much broken up. Of all the skulls from the Philippines the contact 

 is most extensive in number 144655 where the internal margins of the 

 fossffi come in contact for one centimeter, their common borders forming 

 a distinct crest, which extends posteriorly to the occipital crest. In 

 numbers 123433 and 144656, where it is about equal, it is less extensive, 

 the medium crest is less pronounced and does not extend to the occipital 

 crest. The contact is slightly less in number 144658. In numbers 

 144657 and 144659 there is an interval of one millimeter, and in num- 

 ber 144660 about two millimeters. The greatest interval is seen in skull 

 number 106 where it is equal to the interval in the Steere specimen in 

 my own collection, that is five millimeters; it is slightly less in number 

 108, and still less in 105 and 111, the interval still being from one 

 to two millimeters. As a rule, then, the approach of the temporal 

 fossae, posteriorljr, is likely to be far more extensive in Basilan specimens 

 than in those from Bohol ; in the latter they are rarely, if ever, in actual 

 contact. 



Turning to the base of the craniiun we find the palatal region, or roof 

 of the mouth, comparatively broader and much more in front in the 

 Philippine specimens than in any of the Malayan species; the t^Txipanic 

 bullae are much better developed and have thinner walls in the latter than 



