639 
men being the posterior, or facialis opening of the trigemino-facialis 
chamber of the fish, as I have recently shown (AuLis 1914b). Ante- 
rior to the mandibular aortic arch, the short anterior prolongation of 
the lateral dorsal aorta that formed the internal carotid artery of 
earlier stages has become a long artery which first sends an important 
arteria palatina forward along the ventral surface of the chondro- 
cranium and then turns upward along the anterior edge of the para- 
chordal plate. It then runs upward across the internal surface of the 
trabecula and soon gives off a branch, the arteria orbitalis, which 
issues in the orbit through the so-called foramen sphenotrabeculare 
and is said by GrEIL (p. 1364) to be distributed to the eyeball and 
adjacent regions (sich am Bulbus und seiner Nachbarschaft verzweigt). 
Slightly anterior to this artery a second branch of the internal carotid 
is sent to the eyeball, this branch accompanying the nervus opticus 
and being the arteria ophthalmica of GREIL’s descriptions, but more 
properly called, in fishes, either the arteria optica (ALLEN, 1905, ALLIS). 
or arteria centralis retinae (Dourn, 1886). The internal carotid then 
continues onward and upward as the arteria cerebri anterior. The 
arteria communicans cerebri, given off somewhere between the arteriae 
orbitalis and centralis retinae, is apparently the cerebralis posterior 
of the nomenclature that I have employed in my descriptions of 
other fishes. 
If these conditions in this embryo, stage 48, of Ceratodus be 
compared with those in selachians, it is seen that the arteria orbitalis. 
of Ceratodus, which arises from the internal carotid prolongation of 
the lateral dorsal aorta, traverses the cranial wall by a foramen which 
lies in a position that is topographically strikingly similar to that in 
which the efferent pseudobranchial artery of selachians traverses the 
same wall, the dorsal portion of the latter artery being formed by the 
fusion of the proximal portion of the arteria ophthalmica magna of 
embryos with the dorsal portion of the mandibular aortic arch. It is. 
furthermore seen that the arteria orbitalis of the one and the arteria. 
ophthalmica magna of the other have a distribution that is practically 
similar, excepting in that a choroid gland is wanting in Ceratodus. 
This at once suggests the possibility of the homology of these two 
arteries, and in seeking some possible explanation of the conditions. 
I have been led to the following conclusions. 
The lateral dorsal aorta of all gnathostome vertebrates must prim- 
arily have lain along the ventral surface of the neural tube, external 
