ON THE VERTEBRATE SKELETON. 277 
the lepidosiren, the amphiuma or the apteryx, with the scarcely more simple 
or less-developed appendage of the thoracic abdominal hzmal arches (ribs) 
of birds and fishes (figs. 15 and 17, a, a) ; and thus we are led to determine 
its general homology, under its manifold forms of fin, fore-limb, wing, or arm, 
as the diverging appendage of the hzmal arch of the occipital vertebra. 
_ Thenatural and typical vertebral segment above-defined cannot bedetached, 
in every fish, hy the mere disjunction of sutures: in the lepidosiren, e. g. the 
ossified part of the centrum has coalesced with that of the next segment in 
advance and would require to be divided by the saw: the same coalescence 
occurs in the human skull, and has led to the definition of the cranial bone, 
called ‘os spheno-occipitale*.’ In osseous fishes, either by connation of 5 
with 9, fig. 5, or by excessive development of bone in the notochordal capsule 
extending forwards from the centrum 5, and producing 9, there results the long 
bone (5,9) continuing the series of vertebral centrums forwards, and corre- 
sponding in position with two segments or arches above. On the hypothesis 
that it represents the central elements of both those arches, it must be divided 
artificially, in order to separate that segment of the cranium which next suc- 
ceeds the occipital one. And, further, either by a similar coalescence of the 
proximal elements of two hemal arches, or by the undue extension of such 
element of one of the arches, interposing itself between the next arch and 
the rest of the vertebra to which that arch belongs, it happens, that unless the 
proximal element or elements in question be artificially divided, as at 28a, 28a, 
fig. 5, two hemal arches (H 1 and H 111) would be brought away, with the 
neural arch detached by the separation of sutures and the division of the 
bone 5,9. If neither that bone, nor 23a were divided, but were, with the 
bones in superior connection with them, separated from the bones anteriorly 
articulated to them by suture, then we should have the group of bones, in- 
cluded by the curved lines marked N u1, N 111, Hut, H 111 in fig. 5. Two 
vertebral segments are plainly indicated in this group by the distinct hemal 
arches and their appendages, H 11 and H 111; but three pairs of bones, 16, 6 
and 10, fig. 5, appear to be in neurapophysial relation with the single and 
symmetrical median bone 5,9. If, however, what has been urged in the 
chapter on ‘ Special Homology’ (pp. 188-196) respecting the petrosal cha- 
racter of 16 be a true interpretation of that bone, then we must eliminate it 
from our present inquiry, inasmuch as being a partial ossification of a sense- 
capsule (and nature herself removes them, as such, in most fishes), it apper- 
tains to.a category of bones (splanchno-skeleton), forming no part of the pro- 
per neuro- or endo-skeleton, in which alone we seek for evidence of asegmental 
disposition of parts corresponding with the segments of the nervous system. 
_ The bony petrosals (is) being removed, let us, then, with the view of ex- 
_ amining the composition of the segment of the skull with which the occipi- 
tal vertebra was articulated, saw across the bones 5, 9 and 28a, and separate 
the bones 6, 7, s from their sutural connections with those in front of them. 
In thus obtaining the segment in question, the opponents to the vertebral 
theory of the skull are entitled to assert that violence is done to nature by 
the sections of the single bones above-cited; the validity of which as an 
objection to that theory will be afterwards inquired into. : 
It is not, however, absolutely necessary to divide the basal bone 5,9: in 
miany osseous fishes a symmetrical bone (fig. 5, 9’) supports the parial bones 
10, and stands in the relation of a centrum tothem ; the neural arch or circle 
of that segment would not, therefore, be broken by the removal with the 
posterior segment of the whole of the bone s, 9. If the corresponding 
* See Table I., Soemmerring. if 
