202 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
and it may also be inferred that the median portion of the connect- 
ing stalk is morphologically the undifferentiated anterior portion of 
the chorda, while the more lateral portions of the connecting stalk 
may be regarded, as they have been by Killian (91, p. 102), as 
representing the sclerotome of the somite. Furthermore, the inference 
drawn by Froriep ('92*), on the ground of evidence presented by 
Kastschenko (’88) and Kupffer (88, *90, 94), that the lumen of the 
connecting stalk must be ventral and morphologically a part of the 
proccelom, receives no support. If Kupffer’s statement that the pre- 
mandibular cavities of Ammocostes are formed as diverticula from the 
alimentary canal is correct, their development in Ammocostes must 
differ essentially from that in Squalus. Goette (’90), however, flatly 
contradicts Kupffer’s statements. My own observations on Ammocoetes 
lead me unhesitatingly to accept the evidence presented by Goette.! 
Besides, the criteria furnished by the study of the early stages of devel- 
opment of the premandibular cavity in Squalus seem to me more satis- 
factory, because more decisive, than the evidence used by Kupffer (93°, 
p. 522) to demonstrate the ventral nature of the connecting stalk of 
the premandibular cavities in Ammocostes, viz. the relation to a blood- 
vessel which is only hypothetically the complete homologue of the dorsal 
aorta, I find this blood-vessel in embryos of Ammocates of somewhat 
advanced stages of development (4 mm.) extending above the connecting 
stalk of the premandibular cavities, as the apparent anterior continuation 
of the dorsal aorta, as stated by Kupffer. But there is also ventral to 
the connecting stalk a similar blood-vessel, which unites with the dorsal 
vessel both anterior and posterior to the connecting stalk. It is conse- 
quently difficult for me to comprehend why the more dorsal vessel rather 
than the more ventral one is to be regarded as the anterior continuation of 
the dorsal aorta. Kupffer gives no reasons, simply stating that the ven- 
tral vessel can be homologized, if at all, with the carotis ventralis of Mam- 
malia. Now, if we are to apply rigidly such a criterion as Kupffer’s to 
am unable to accept Hoffmann’s conclusion on the basis of the evidence he pre- 
sents, I believe there are good grounds for holding that a visceral arch, which 
once existed between the mandibular and the hyoid (first and second visceral) 
arches, has disappeared in phylogeny. The evidence in favor of this view will be 
summarized later. 
1'That Kupffer has not in his studies come to a right understanding of the 
development of the anterior head mesoderm seems to me certain from a comparison 
of my sections with those figured by him (’90, Figg. 81 und 82, Taf. 28). The 
cells which he calls ganglionic are in my opinion the anterior mesoderm. This 
appears to me to be Kupffer’s fundamental error. 
