ON THE ANATOMY OF FISHES. 
69 
anterior vertebrae inflexibly united together, but their connection with the skull is of 
an equally rigid character, and this is due to the firm articulation of the transverse 
processes of the fourth vertebra with the skull through the intervention of the post- 
temporal bones, to the articulation of the arch and spine of the third vertebra with 
the exoccipitals and supraoccipital, and to the development of strong, interlocking- 
accessory articular processes on the basioccipital and the centra of the first and 
complex vertebrae. 
We defer for the present any reference to the “auditory ossicles” as constituent 
elements of certain of the anterior vertebrae beyond remai-king that in all probability 
the claustra and stapedes of Weber respectively represent the neural spine and arch 
of the first vertebra, while the ossicles, termed incus and malleus by the same mor- 
phologist, respectively correspond to the arch of the second vertebra and the transverse 
process or rib of the third vertebra. 
The centrum of the first vertebra (figs. 1, 2, 4, and 5, v.') is much smaller than any 
of the normal anterior vertebral centra, being represented by a thiu discoidal and but 
slightly biconcave bone, wedged in between the concave posterior face of the basi- 
occipital and the concave anterior face of the complex centrum. Its ventral surface is 
furnished with a pair of accessory articular processes which fit in between similar but 
much stronger paired processes developed from the ventral margin of the anterior 
extremity of the complex centrum, and also from the ventral surface of the posterior 
end of the basioccipital (figs. 4 and 5, ac., p.). On the dorsal surface of the centrum, 
near its lateral margins, there are two small cup-shaped sockets for the reception of 
the rounded condylar processes of the two “ stapedes.” The centrum is quite distinct 
from the basioccipital and complex centrum, although rigidly articulated to both. 
The large vertebra which succeeds the first was regarded by Weber {loc. cit.) as the 
second. Baudelot (1), Grassi (17), and Nusbaum (29), however, subsequently 
showed that in the Cyprinidm the apparent second vertebra is really formed by the 
complete coalescence of the centrum of tho second vertebra with that of the third; and 
a precisely similar fusion was believed by them to have formed the “second” vertebra 
of Weber in the Siluridee ; but more recently Ramsay Wright (42, 43) has demon- 
strated on developmental grounds that, in the latter family, this vertebra is even 
more complex, inasmuch as it represents, not only the elements already recognised by 
previous writers, but also, in addition, the centrum, arch, and spine of the fourth ver- 
tebra — or, in other words, the “second” vertebra of Weber is formed by the complete 
coalescence of the third and fourth vertebrae with each other and with the centrum of the 
second. From Ramsay Wright’s account of the development of the anterior vertebrae 
in the North American Siluroid Amiurus catus, it would appear that the usual inter- 
vertebral growth of the notochord does not take place between the centra of the third 
and fourth vertebrae, so that the posterior concavity of the former and the anterior 
concavity of the latter are incompletely developed in the embryo and completely 
obliterated in the adult ; and, further, that the second and third vertebral centra 
