366 MOSTLY MAMMALS 
and the third to Phyllobates), to which water is essential 
while in the tadpole stage, adopt the plan of carrying their 
young attached to their backs (either by means of suckers 
or of a viscid secretion), and are thus enabled to transport 
them to another pool when occasion arises. In the case 
of the genus last mentioned, it is the father frog on whom 
the burden of carting about his family falls, but in ‘the 
other instance it is not known to which sex this duty is 
entrusted. A frog (Arthrolepis seychellensis) from the 
Seychelles is likewise in the habit of carrying its young 
on its back, but in this case the purpose of the arrange- 
ment is not to transport them from one pool to another, 
but merely to protect them during development, which 
takes place on land, the tadpoles breathing by means of 
lungs. 
The Coqui frog (Hylodes martinicensis) of the West 
Indies affords, however, the best instance of the manner 
in which these reptiles can develop without resorting to 
the water at all. In this species the eggs are laid on 
the leaves of plants in damp situations, the female parent 
remaining near by on guard until they hatch. This 
takes place in about a fortnight after deposition, but instead 
of tadpoles, perfect little frogs make their appearance in 
the world, all the transformations taking place within the 
egg. A Peruvian species of the same genus (Aylodes 
hineatus) exhibits a precisely similar mode of development ; 
and the same is the case with the curious Solomon Island 
frog (Rana opisthodon). 
In conclusion, mention must be made of the tadpole of 
a South African frog (Dactylethra capensis), not on account 
of any peculiarity in its mode of. development, nor on 
account of its form (although this is strange enough), 
but from- the curious circumstance that it alone, among 
