PHALANSEElDiB. 129 



tliis family, I have found it convenient to represent each tooth as 

 usual by its own serial number, but to put this number in 

 italicized figures when the tooth is minute, as compared to the other 

 larger and more obviously functional teeth, and to add to it one 

 asterisk when it is occasionally, and two when it is generally absent. 

 Thus the ordinary normal premolar formula of PJudanger orientalis 

 would be P- j ' ' g» " 4 as compared, to P. j ' " " g*I ' 4 ^^ P^- wuicidatus, 

 such exceptional variations as that of the Ph. orientalis above 

 referred to (P. j " g ' g ' 4 ) being looked upon as too abnormal to have a 

 place in the specific or generic description. 



Finally, it should be specially noted that the numbers and 

 positions of the minute teeth, while affording to the species and 

 genera additional characters in a broad and general way, are yet too 

 variable to be of any systematic importance in individual instances. 

 All the chief specific characters should therefore be taken primarily 

 from the large and obviously functional teeth. 



In the homologies of these, the diagnostic teeth, no diflSculty can 

 occur except in the case of the lower p.^ and m.' in species with 

 only four large and functional cheek-teeth. They may, however, 

 always be distinguished by the fact that m.' is tricuspid, with one 

 high anterior and a pair of low posterior cusps, while m.'' is quadri- 

 cuspid, and p.* never has more than one cusp in the same transverse 

 line. 



As in other climbing forms (see below, under Bidelphyidce, 

 p. 317), the hind feet are here so shortened for the purpose of 

 grasping that their length is of little use in determining the general 

 size of the animal. The length of the lower leg, measured from the 

 top of the knee (when bent) to the sole, is therefore used as the 

 standard measurement for the members of this family. 



Some of the members of the present family present a very note- 

 worthy example of the comparative readiness with which lateral 

 parachute-like extensions of the skin may be developed for the 

 purpose of aiding in flight. Thus three difierent genera possess 

 such lateral - flying- membranes, these three having apparently 

 developed their membranes independently, as is proved by the fact 

 that each of them is exceedingly closely allied to some non-flying 

 form, while their relations to each other are of the most distant 

 kind. Thus Petauroides and Petaurus present absolutely no generic 

 cranial or dental differences from Pseudochirus and Oymnobelideus 

 respectively, while Acrohates is but little further removed from 

 Bistcechurus. 



