645 
aborted even in an embryo of stage 47, excepting only the so-called 
lingual and mandibular arteries; which latter arteries are simply branches 
of the external carotid artery of GrEIL’s descriptions. GREIL shows this 
mandibular aortic arch still persisting in an embryo of stage 48. 
In Lepidosteus, a re-examination of my sections shows that there 
is a branch of the hyo-opercularis artery of my earlier descriptions 
of that fish (1908a) that traverses the trigemino-facialis chamber of 
the fish and is strictly similar to the branch of that artery that tra- 
verses the same chamber in Amia, this hyo-opercularis artery thus 
being found in these two holostean ganoids and in the Loricati. This 
artery, in all these fishes, enters the trigemino-facialis chamber by 
its facialis opening, the external carotid entering that chamber by 
the same foramen in Scomber and the Loricati, but entering it, in 
Amia, by a more anterior and separate foramen. The presence of 
these two so similar arteries in these several fishes necessarily raises 
the question as to whether it is the one or the other, or perhaps the 
two combined, that is the homologue of the arteria temporalis of 
Ceratodus, but as this question can not be here decided the hyo-oper- 
cularis is left wholly out of consideration. 
Ceratodus thus, in the arrangement of the arterial vessels here 
under consideration, resembles Lepidosteus more than any other of 
the fishes that have been considered, and comparison of the several 
conditions described would seem to quite certainly establish that the 
arteria orbitalis of Ceratodus is the homologue of the arteria ophthal- 
mica magna of the other fishes. This being so, the topographical 
relations of this artery in Ceratodus to the rudimentary premandibular 
visceral pocket of Grett’s descriptions would seem to confirm the 
opinion that it represents the persisting remnant of the premandibular 
aortic arch. Ceratodus and Lepidosteus then represent, in the relations 
of the mandibular and premandibular aortic vessels to the trabeculae, 
a stage that is intermediate between those found in teleosts and Amia, 
while Amia represents, in this respect, a stage that is intermediate 
between those in Ceratodus and selachians. Whether the teleostean 
and ganoidean arrangements are preselachian or postselachian in origin 
is not evident, but as a marked cranial flexure would not seem to 
be a primitive feature, the teleostean arrangement, at least, would 
seem certainly to be of preselachian origin. 
One other feature of the cranial anatomy of Ceratodus should 
now be mentioned; which is, that in GREIL’s numerous and very 
