No. 405. ] DEVELOPMENT AMONG ANURA. 699 
The most remarkable structures were the gills. Three vis- 
ceral arches were present, and according to one statement three 
gill-slits (three slits appear in the figures, but in another part 
of the description it is said that the first and third arches 
have each only one free edge — a condition that would admit of 
only two slits). The gills consisted of a pair of very thin 
bell-shaped membranes, each gill connected with the gill arches 
by two filaments, one ending on the first and one on the second 
arch. This peculiar arrangement led Weinland to the conclu- 
sion that the bells represent two gills fused together. He 
regarded them, in function at least, as external gills. On the 
free edge of the third arch was found the rudiment of what 
was supposed to be an internal gill, which suggested that the 
bell-gills are temporary larval organs. The bell-gills were 
veined by a capillary network, and the vessels were filled with 
blood, so that the gills were probably already functioning, 
although pressed close to the embryo in the egg; it was only 
after floating them out in water that their windflower shape 
and large size were discovered. The gill measured across, 
when expanded, not less than three-quarters of the diameter of 
the egg. The gill filaments were provided with striped muscle 
fibres; these muscles, Weinland argued, could not be of use in 
the egg, but the gills may be retained a short time after hatch- 
ing, when the muscles might function in water, whether the 
_ tadpole remained within the pouch or escaped. 
Weinland knew nothing of the subsequent history of the 
tadpoles, as to when they hatch and whether they ever live in 
water, either free or within the pouch; Boulenger makes some 
unqualified statements in regard to these questions. 
He has noted the occurrence of eggs in the dorsal pouch of 
the female of Nototrema fissipes ; he observed a single specimen, 
? which there were sixteen ova, measuring each 10 mm. in 
diameter. «From the size and small number of the ova,” he 
writes, « it may be safely predicted that the young undergo the 
whole of the metamorphosis, within the pouch, as in Nototrema 
oviferum, which is the nearest ally of Mototrema fissipes” Of 
ye species of Nototrema found in Ecuador and Peru, also pro- 
vided with pouches, Boulenger says in an earlier work (82), 
