VI CVSTIGNATHIDAE 209 



Thoro2)((. — The fingers ami toes are free, the tips sinqdy 

 swollen and not dilated into discs. Closely allied to Chorojihilus. 

 111. miliaris, of Brazil, the only species, has very long toes. The 

 head is broad and flat. The upper, nearly smooth surface of the 

 body is flesh-coloured, with brown marblings ; the limbs are 

 cross-barred ; the under parts whitish, granular on the belly. The 

 male is devoid of vocal sacs. The total length may be 2 inches. 

 Hensel has published the following notes of this species, under the 

 name of Hylodes ahhreviatus. The tadpoles are quite flat, their 

 bellies forming a kind of sucking disc, so that these creatures, 

 even before the appearance of the hiud-liinbs, can quickly wriggle 

 up vertical walls of stones, provided these are covered with a 

 little water. In correlation with this habit, the root of the tail 

 is not compressed laterally, but is as broad as it is high, and the 

 usual vertical fln is restricted to its distal third. On the prox- 

 imal portion of the tail the ventral fin is flattened and broadened 

 out so as to form almost the continuation of the peculiar disc- 

 like belly. The anal opening is not a projecting tube, but is a 

 flattened transverse slit. 



Fam. 5. Cystignathidae. — This is one of the largest families, 

 and also one of the least satisfactory. Its numerous members, 

 more than 150, exhibit such a versatility in adaptation to 

 circumstances (there are aquatic, terrestrial, arboreal, and burrow- 

 ing species), with a corresponding development or loss of 

 anatomical characters which we should like to rely upon as 

 taxonomic marks, that the numerous genera not only run into eacli 

 other, but also get entangled with those of other families. In 

 fact the whole family is ill defined. It can be characterised as 

 follows : — The shoulder-girdle is arciferous ; the sacral diapophyses 

 are cylindrical or Imt slightly dilated; the metasternum has 

 either a bony style or it forms a cartilaginous plate ; the 

 terminal phalanges, although they sometimes carry adhesive 

 discs, are never claw-shaped. 



The last statement is, of course, intended to separate tiie 

 Cystignathidae from the Hylidae, of which, however, the three 

 genera Thoropa, Chorophilus, and Acris stand on debatal)le 

 ground (cf. p. 186, Hylidae), while, on the other hand, most of 

 the Australian genera, notably Chiroleptes, have unmistakably 

 dilated sacral diapophyses. The difference from the Pelobatidae 

 can in this case be one of decree onh'. 



VOL. vrir ^ 



