Charles R. Stockard 509 
as in the case of the csophago-cutaneous duct. If such did occur the 
pouches could searcely form. But since the pouch is well formed before 
the cartilage appears, and as the cartilage formations are not so extensive 
as to extend around these huge pouches so as to become intra-branchial 
structures they must remain about the external gill tube as extra-bran- 
chial. It may also be remarked, not that I think it a very strong point 
in the argument, that cartilage supports are entirely unessential to the 
efficient action of these pouched gills, while such supports are very 
useful in maintaining a fully spread condition of the typical fish gill. 
Further, when we remember that only the Marsipobranchii have pouched 
gills and likewise extra-branchial skeletons, while all fishes have intra- 
branchial arches and unpouched gills, one can scarcely doubt that some 
correlation exists between the processes that have given the pouches and. 
the causes that have made the skeleton extra-branchial. Of course the 
various stages of the struggle, if we may so call it, between the pouches 
and the intra-branchial arches may well have been more complex than 
the present explanation supposes. 
I am inclined to consider the condition of the cesophago-cutaneous 
duct as one of arrested development, 7. ¢., it remains in the tubular stage 
through which all of the other gills pass before forming their pouches. 
This might be called a modification but it is one which has not greatly 
affected the primitive condition of this structure, and argues only to a 
slight degree, if at all, against the above position. Further if it be 
claimed that the extensive cartilage support of this duct was a new 
or special acquisition, one might then reply that this indicates the fact 
that in Bdellostoma we have an animal on the road to possessing an intra- 
branchial skeleton, which is even now partially formed. The extra- 
branchial condition from this point of view would be, therefore, primary 
and not secondary. But since these animals possess a mandibular arch 
they must have also possessed intra-branchial gill arches because the one 
is to be considered only as a modification of the other. At any rate 
according to our present knowledge there is clearly no good reason for 
accepting Dohrn’s view that the arches were lost on account of a change 
in the animal’s life-habits. | 
From such considerations as the above one must at least admit that 
these animals are contrasted with fishes as having extra-branchial skele- 
tons on rather flimsy grounds. To any one that will study these structures 
it will, I believe, become clear that marsipobranchs have in part, and 
probably once had entirely, an intra-branchial skeleton. 
The Shifting of the Gills in the embryo of Bdellostoma was correctly 
described by Dean, 99, as mentioned at the beginning of the discussion 
