RESPIRATORY ORGANS. oA7 
the adult amniotes the pouches may completely disappear without 
leaving a trace, aside from the Eustachian tube (p. 187) and the thymus 
glands to be mentioned below. The number of clefts or pouches 
varies between considerable limits. The largest number in any true 
vertebrate (there are more in Amphioxus and the enteropneusts) is 
fourteen pairs in some specimens of Bdellosioma. In other cyclo- 
stomes there are seven, eight to seven in the notidanid sharks, six in 
other elasmobranchs, five or six in teleostomes, amphibia and reptiles 
and five in mammals. In this numbering the oral cleft is not included, 
although there is some evidence that the mouth has arisen by the 
coalesence of a pair of gill clefts (p. 206). 
The serial repetition of the visceral clefts does not strictly correspond to the 
other segmentation of the body, their number and position being at variance with 
those of the myotomes. There is a branchiomerism or serial repetition of the 
gill clefts, apparently distinct from the true metamerism of the head. The ap- 
pearance of these clefts or pouches and the relation of aortic and branchial arches 
in the amniotes, where gills are never developed, can best be explained by the 
assumption that these forms have descended from branchiate ancestors. 
Between each two successive gill clefts there is an interbranchial 
septum, covered externally with ectoderm, internally with entoderm, 
and with an axis of mesoderm, the latter in the earlier stages carrying 
with it a diverticulum of the ccelom (fig. 243, c). Later blood-vessels 
(aortic arches) and skeletal elements (visceral arches, p. 63), are devel- 
oped in each septum, the visceral arches appearing on the splanchnic 
side of the ccelom and hence not being comparable to ribs or girdles. 
In the cyclostomes and fishes the gills are developed from the an- 
terior and posterior walls of the typical interbranchial septa. They 
were long regarded as of entodermal origin, but in recent years con- 
siderable doubt has been thrown on this, at least for the fishes, and 
there is some evidence for their ectodermal origin. The matter cannot 
yet be regarded as settled. These gills are either filamentous or la- 
mellate outgrowths of epithelium, each carrying a loop of a blood- 
vessel. Thus each typical cleft is bounded in front and behind by gill 
plates or filaments (fig. 246), those on a side constituting a demibranch, 
the two demibranchs of a septum constituting a gill, while a cleft is 
bounded by demibranchs belonging to two gills. In the young 
elasmobranchs and in the young of a few teleosts (before birth) the 
gill filaments protrude from the clefts as long threads, but later they 
are withdrawn. 
