_N. oviferum_ is brown above, with darker patches on the sides of the body and with cross-bars on the limbs. The last two species and _N. fissipes_ of Brazil, near Pernambuco, carry their young in the pouch until the metamorphosis is completed. This long nursing-period necessitates a great amount of food-yolk in the eggs, and this enlargement in turn implies a considerable reduction in their number. The female's load consists of about fifteen eggs only, but these are of a great size, namely one-eighth of the length of the mother's body. _N. pygmaeum_, in Venezuela, is a tiny creature. The female, just one inch in length, carries only from four to seven eggs. It looks then "as if it carried a sac filled with a few gigantic balls." This species is further worthy of note on account of the opening of the brood-pouch, which is a longitudinal slit, whence a kind {203}of thin and slightly elevated ridge or fold of the skin extends on to the neck. The suggestion, that this seam is burst open, in order to set the full-grown young free, instead of their passing through the existing opening, is scarcely credible. These Neotropical tree-frogs seem to be rare, and females with embryos are of course still more uncommon, so that the best account of their structure is still that given by Weinland[88] of _N. oviferum_. How the eggs get into the pouch has not yet been observed, but it is most likely with the help of the male, immediately after fertilisation. The pouch forms two blind sacs which extend forwards over the sides of the back. The eggs are large, 1 cm. in diameter, and the enclosed embryos, or rather tadpoles, had a length of 15 mm., with a large amount of yolk still contained in the spirally wound intestine. The first two gill-arches carried each a double thread, which expanded into a funnel-shaped membrane, not unlike the flower of a _Convolvulus_, and furnished with a capillary network; the stalk contained muscular fibres. These most peculiar structures are of course the much modified external gills. Those of _N. testudineum_ and _N. cornutum_ are likewise bell-shaped. _Hylella_ differs from _Hyla_ chiefly by the absence of vomerine teeth, and consists of about half-a-dozen small species, about one inch in length. The fact that two species live in Queensland and New Guinea, while the others are natives of tropical America, suggests that this genus is not a natural but an artificial assembly, an instance of convergent evolution.