HYOMANDIBULA OF THE GNATHOSTOME FISHES 581 
ceratal elements of those arches. This sigma-form of arch is said 
by Gegenbaur (’72) to be peculiar to these fishes, but in the Holo- 
cephali (Schauinsland’03), and, as will be later shown, possibly also 
in Ceratodus, the pharyngobranchials project postero-mesially 
as they do in the Plagiostomi. In teleosts also, the pharyngo- 
branchial of the first branchial arch, when it comes into articular 
relations with the neurocranium, may be directed dorso-postero- 
laterally at a sharp angle to the epibranchial (Scomber, Clupei- 
dae), strongly recalling the conditions found in the Plagiostomi, 
but differing radically from the conditions in those fishes in that 
the pharyngobranchial is directed dorso-postero-laterally along 
the lateral surface of the neurocranium instead of postero-mesi- 
ally beneath the vertebral column. 
Gegenbaur considered this sigma-form of arch to be a sec- 
ondarily acquired but archaic feature, for he found it even in 
young embryos of Acanthias; and he thought it caused by the 
dorsal ends of the branchial arches having been pushed posteri- 
. orly by repeated acts of deglutition. But if this be the cause 
there seems no good reason why it should not have been equally 
operative in all fishes, and, furthermore, it certainly could not 
have caused the conditions just above described in Scomber and 
the Clupeidae. In these latter fishes it seems certain that the 
epal and ceratal elements of the first branchial arch have been 
carried forward beyond a point where the pharyngeal element 
of that arch had previously become attached to the neuro- 
cranium, instead of the dorsal ends of the pharyngobranchials 
having been pushed backward; and as the dorsal ends of 
all of the pharyngobranchials are, in Chlamydoselachus, and 
hence probably in other selachians also, attached to the verte- 
bral column and prevented from shifting forward by the interven- 
ing roots of the efferent branchial arteries, it seems to me equally 
certain that here also the epal and ceratal elements have been 
carried forward, doubtless because of the relatively marked an- 
terior growth of the chondrocranium, instead of the dorsal ends 
of the pharyngobranchials having been pushed backward. Where 
the dorsal ends of the branchial arches were not so attached to the 
neurocranium or vertebral column, as in all or most of the arches 
