SOME PARTS OF THE SKELETON OF FISHES 279 
The posterior hypural apophysis had greatly enlarged both abso- 
lutely and in relation to the anterior, and traces of a bony deposit 
on its surface were discernible. The number of fin-rays had increased 
to fourteen ; of which two, very short, lay between the end of the 
interhaemal cartilage of the penultimate vertebra and the lower angle 
of the anterior hypural apophysis ; six, gradually increasing in length, 
and becoming jointed superiorly, embraced the posterior edge of 
the inferior hypural apophysis ; and six, of which the inferior were 
long and jointed at their ends, while the superior were simple styles, 
were connected with the posterior edge of the posterior or superior 
hypural apophysis. The terminal osseous ring (2) had in the mean- 
while extended backwards, and now, as a slender tube, tapering 
posteriorly and obliquely truncated behind, embraced more than 
half the length of the previously free part of the notochord. Asa 
consequence, the hypothenuse of the still triangular posterior hypural 
apophysis is now fixed to bone throughout its whole length, for 
the end of the bony sheath in question extends slightly beyond it. 
The remainder of the notochord (c) has its wall still membranous 
and unossified, and ends close to the superior and posterior angle 
of the caudal fin as before. There are no fin-rays above the noto- 
chord, nor is any neural arch developed from the terminal centrum, 
but the rudimentary interneural cartilage of the penultimate centrum 
had greatly elongated, and had taken the same position relatively to 
its superior arch as that occupied by the interhemal cartilage 
relatively to the inferior arch, and had become surrounded by a 
sheath of bone. Behind this two other cartilages (sz, 2) lie parallel 
with one another above the ossified sheath of the chorda, but at 
present they are connected with no fin-rays. I will term these the 
“epiural” apophyses. 
In a half-grown Stickleback (fig. 3) the anterior end of the 
terminal centrum was dilated and cup-like, just as if it were the 
anterior half of one of the ordinary hour-glass-like vertebre, but 
instead of dilating again posteriorly, it is continued into a stout 
style, more than twice as long as the body of the penultimate 
centrum, and curved up so as to make an angle of 45° with the 
rest of the vertebral column. This stout style, with its central cavity, 
looks not very like the previous delicate sheath of the chorda ; but 
such thickened sheath it really is, and with care the remainder of 
the notochord may be traced beyond it between two of the fin-rays 
into the tail-fin itself. The rays between which it lies are the 
uppermost of the superior set in the last-described embryo, and a 
new set, six in number, which have been formed above the noto- 
