DOUSOXIA. 453 



understood by an inspection ot ligii. 2ij A & 13. The oiil}- explauatjon 

 required is that the first Pteropine ridge is wanting in Dobsoma, 

 evidently owing to the shortening of the extremitj- of the rostrum. 

 The fifth (last " undivided ") ridge of Dohsonia (f in the figure) thus 

 corresponds to the sixth (first divided) of Pteropus (same letter in 

 figure), and ife is in fact in Dohsonia not infrequently slightly inter- 

 rupted in the middle. 



Ears. — In all species rather abruptly attenuated above, tijj 

 narrowly pointed, both characters slightly less pronounced in 

 D. minor than in the otlier species. 



Wings. — Terminal (third) phalanx of second digit present, about 

 one-fourth or one-fifth of second phalanx, but claw wanting. 

 Wing-membranes perfectly continuous across back, and connected 

 with integument of dorsum only along spinal line (the only parallel 

 to this is seen in the Macroglossine genus Nutopteris) ; posteriorly 

 inserted on distal extremity of first or second metatarsal or on base 

 of proximal phalanx of same toes (the variation in this respect is 

 individual). Katio of metacarpals and phalanges differing from 

 Kousettine type chiefly in two characters: — (1) In typical 2?ouse<iMs, 

 third metacarpal longer than fourth, this again longer than fifth ; 

 in Dohsonia third metacarpal longer than fifth, this again nearly 

 always longer than fourth, the modification due not to a lengthening 

 of the fifth but to a distinct shortening of the fourth metacarpal : 

 (2) terminal (second) phalanx of third digit conspicuously lengthened, 

 longer than metacarpal of same digit, in Sousettus much shorter 

 than metacarpal. — The subjoined wing-indices are calculated from 



situated below the roots of dii-di,. lu Rousettus ii is often slightly smaller 

 than i,^, in Pteropus always conspicuously smaller, in certain species of Pteropus 

 {€. g. the Pt. lomhocensh group) still more reduced than usual, and finally in 

 atyloctcn'nim it has quite disappeared. This evidence in itself renders it 

 almost certain that the missing lower incisor in Dohsonia is ii, not i^, and the 

 correctness of this conclusion is confirmed by the very small size of dij as 

 compared with dij. 



Upper milk canine even more widely separated from di^ than is the 

 permanent canine of a Pteropus from i- (diastema nearly equal to the 

 whole set of four upper inilk incisors), and directed vertically downward, 

 not showing the slightest trace of the conspicuous forwai-d slant of the 

 j)ermanent canine of Dohsonia; crown faintly resui'ved, a little thickened at 

 middle, as if tending to develop a minute secondary cusp from anterior 

 margin. Permanent upper canine situated in jaw above and in front of root 

 of milk canine. Lower milk canine perfectly Pteropine in position, not moved 

 forward to the slightest degree (compare extreme anterior position of ])ermanent 

 canine), separated from di.^ by a diastema of nearly 2 mm., crown styliform, 

 faintly recurved. The large permanent lower canine fills out the entire space 

 in the jaw between the roots of di.^ and dc. — dp', dp^, dp^, and dp^ small, 

 spiculiform. slightly recurved. 



There must be sufficient space, in the maxillary bone in front of dc and in 

 the lower jaw between di. and dc, for the growtii of the powerful npper and 

 lower permanent canines. This may be the simple explanation of the wide 

 diastemas di--dc and di^-dc, in otlier words, of the fact that, though the 

 permanent canines have moved anteriorly, their predecessors have not. 



