Vascular System of Bdellostoma dombeyt. 39 
number of gill pouches (possibly thirty-five pairs) arise during 
the course of development. Of these, only the posterior 
IO-I4 pairs remain to form the permanent gills of the adult. 
The distribution of branchial vessels, one to each gill cleft 
(gill pouch), is a character in which the Myxinoids differ 
from all other Vertebrates. Even in Petromyzon, each 
afferent and efferent vessel is distributed to the hemibranch 
of two different pouches, the gill arteries corresponding to g7//- 
arches rather than g7z// clefts. It is probable that the Myx- 
enoid condition is not primitive, although the embryological 
development is not known. A probable explanation of the 
manner in which this unique condition has been derived 
from the primitive arrangement is suggested by an examina- 
tion of the efferent branchial vessels and their connections in 
Bdellostoma forsteri (see text figure D.) Here we find, usually, 
two efferent branchial vessels (ef, br) arising from the ring 
vessel (rv) which surrounds the gill passage on the wall of 
the gill pouch. These efferent vessels unite with the lateral 
carotid vessels, and with the dorsal aorta by anastomoses. 
This is an interesting intermediate stage between the arrange- 
ment of the vessels in Bdellostoma dombeyi (see Plate III, Fig. 
XX), and the ordinary vertebrate type. Referring again to 
text figure D, suppose, for example, we take afferent vessels 
from gill 3. If we fuse the two vessels together (which is 
_ almost accomplished in the sixth gill, right side), or obliterate 
one of them, we get ove efferent vessel for each gill, as in gills 
4, 5, and 6, left side. This is the condition in a// the gills of 
Bdellostoma dombeyi. But if we break the connection at k, 
gill 3, and k’, gill 2, and at the same time connect the artery 
between them with the dorsal aorta by a vessel, h (dotted 
outline), this vessel, h, would be an efferent branchial artery 
of the wswal vertebrate type. It would receive the blood from 
the posterior hemi-branch of gill 2, and the anterior hemi- 
branch of gill 3. I have no doubt that this very relation 
would be found as a variation, if a large number of specimens 
of Bdellostoma forstert were examined. 
A consideration of the extreme variability of the blood 
vessels in Bdel/ostoma renders this explanation all the more 
probable. To give a more adequate idea of the nature and 
extent of variation, I have made diagrams to show the rela- 
27 
