TRANSACTIONS OF SECTION D. 523 
There is abundant evidence to prove that the hypural elements in caudal fins 
are the result of fusion of hemal arches with radials; perhaps the best examples 
illustrating this are members of the Gadidz, where various conditions are seen, 
from free radials to fused hemal arches and radials. Hzmal arches alone are 
unsuitable for the support of rays, and there seems no reason to suppose that 
elimination of suitable supports (radials) occurred in favour of arches in a 
position which offered every facility for greater development when we recall that, 
under adverse conditions dorsally, they persisted. : 
Differentiation of the median fins occurred before heterocercy was established, 
and in the specialised homocercal type the original caudal appears to have been 
suppressed and its place taken by the posterior anal fin brought relatively 
further back by the upturning of the caudal extremity. 
Gephyrocercy is understood to refer to perfect external and internal symi- 
metry secondarily acquired; Fierasfer is a good type. The whole of the original 
caudal has been lost, and the dorsal and anal fins have met round the abbreviated 
extremity and share equally the function of a caudal fin. Most probably the 
living Dipnoi are also gephyrocercal., 
3. The Homology of the Gills in the Light of Experimental Investiga- 
tions. By Professor H. Braus. 
The gills develop at a spot where ectoderm and entoderm meet. There have 
been many controversies on the question: which of the two germinal layers pro- 
vides the material for the formation of the gills, or whether they both do so? 
But the fact that they are concerned in the formation of the gills does not tell 
us in which, if in either of them, lies the determining factor for gill-formation. 
Thus, for example, many authors have expressed the opinion that the mesoderm 
(vascular system) contains the primary formative factor. It would follow that 
the covering of the gills with epithelium is a passive process which could be 
performed in one case by the ectoderm, in another by the entoderm, according 
to the topographical circumstances. 
The experimental method now gives us unequivocal mformation as to the 
material necessary for gill-formation and as to the potencies of this material. I 
refer to the results of a number of embryonic transplantations on the larve of 
anura (Bombinator, Rana, Hyla) performed in my laboratory by Dr. Ekman, of 
Helsingfors. 
The gill-ectoderm was detached before the gills had formed, and was then 
transplanted to other parts of the organism. Gill-filaments afterwards developed 
_ from it, but no gill-clefts. Circulation of the blood was also wanting. The fila- 
ments soon perished, and in some species never became large. If the gill-ectoderm 
was lifted, turned round 180°, and replanted in that position at the same spot, 
gill-filaments were formed with circulation and gill-clefts. The latter were all 
twisted 180° from the normal position. This is very clearly seen in the oper- 
culum, which does not begin in front and grow backwards as usual, but in these 
cases begins behind and grows forwards. We may therefore conclude with 
certainty that the ectoderm alone is able to produce gills, and alone determines 
the position and form of these organs. The further development of the gills is 
dependent on the ingrowth of mesoderm (vascular system). A layer of entoderm 
beneath the ectoderm is sometimes present in the species examined, but not neces- 
sarily. Foreign ectoderm—i.e., such as in ordinary circumstances does not develop 
gills—behaves differently, according to the part of the organism from which it 
is taken. If ectoderm is taken from the trunk or from the dorsal part of the 
head and planted in the position of the gill-ectoderm, no gills are formed, but a 
small opercular space appears for the anterior extremity. This is a consequence 
of the formative influence of the rudimentary limbs. If ectoderm is taken from 
the region above the embryonic heart and transplanted to the position of the gill- 
ectoderm, gill-filaments, gill-clefts, and an operculum are formed, as in the 
normal animal. The same effect is obtained from the regenerated ectoderm, 
which results when the ectoderm of the gill-region is extirpated and left to 
itself. It is not yet certain what factors induce this foreign ectoderm to assume 
the functions of ordinary gill-ectoderm, but a similar process has already been 
