IIARPYIOXYCTERIS. 



803 



the bitid outer cusp, the posterior (i)e) the postero-exteriial basal 

 cusp ; the anterior of the inner set {a) is probablj^ the anterior 

 basal cusp, which in all lower jnolars has moved from the anterior 

 to the antero-internal corner of the teeth *, the two posterior (i\ i'^) 

 represent the bilid inner cusp, m^ is similar in structure and size 

 to m,. m^ smaller than m^, but not differing in structure. 



Palate-ridges. — Unknown. 



K.vternal characters. — Nares prominent, subtubular, as in Doh- 

 snnia and many olher Fruit-bats. Ears moderate, somewhat 

 broadly rounded off above. Tibiae shorter than in any other Fruit- 

 bat, between one-third and one-fourth of forearm, only a little 

 longer than foot with claws, Interfemoral nearly obsolete in 

 centre ; calcar present. Membranes arising from sides of dorsum, 

 and inserted posteriorly on first phalanx of second toe (or junction 

 of first and second toes, as in Dohsonia). Second digit with claw ; 

 third metacarpal longer than fifth, Avhich is a little longer than 

 fourth, as in Dohsonia, but total length of these digits considerably 

 increased (as compared with Dohsonia) chiefly by a lengthening of 

 the metacarpals and proximal phalanges (see table below, giving in 

 upper row the wing-indices of Harpy ionycteris, in lower row those 

 of Dohsonia for comparison). Fur of upperside (in single species 

 known) extending on proximal three-fourths of forearm and the 

 whole of the tibia and foot ; colour brownish above, dull drab 

 beneath ; size as Dohsonia minor. 



Hahits. — Unknown. 



Affinities. — Harpyionycteris has no closer relative among living 

 Fruit-bats than Dohsonia. So evident is the intimate phylogenetic 

 connection between these two genera that IJarpyionycteris may be 

 said, almost with certainty, to be the peculiarly modified Philippine 

 representative of the Austro-Malayan Dohsonia. It follows that 

 the present genus ought to be classed in the subfamily Pteropodinae, 

 immediately after Dohsonia, and it would have been so here, if not 

 for the fact that the plan of this Catalogue (subdivision into sub- 



* It looks as if the three inner cusps in m,, nij, and m, of Harpyionycteris 

 were simply the tricuspidate inner ridge of the ordinary Megaciiiropteraa 

 molar; a comparison witli m^ of Dohsonia renders it probable, however, that 

 only the two posterior cusps (/', i-) represent the inner ridge, while the anterior 

 cusp (rt) is in reality a cusp developed from the antero-internal basal ledgo. 



t In Harpyionycteris with claw. 



3f2 



