256 MR. A. K. TOLTON ON 
in front (a.) is incompletely divided into two on the proximal 
side, showing that it represents more than one hypural. There 
is, in the specimen figured, a small piece of cartilage (£.) between 
the “posterior” and “anterior” hypurals. This would appear 
to be exceptional, and may be the vestige of another hypural. 
Hemal arches are now beginning to appear anterior to the 
hypurals, developing from behind forwards. The matrix secreted 
by the skleroblasts diminishes In amount as one goes forward. 
The lepidotrichia are still better developed and extend to the 
margin of the fin. There is figured a row of lateral-line sense- 
organs on each side, which extends on to the future mid-line 
of the caudal fin, dividing the lepidotrichia into a dorsal and 
a ventral series. 
Stage V (1°55 cm.) (Pl. I. fig. 5).—There are now three dis- 
tinct (a rudimentary fourth) epaxial elements (whose appearance 
seemed to give the signal for the secondary flexure of the noto- 
chord). In transverse section they show no signs of forking at 
their proximal ends, and they originate from single median 
masses of cartilage, not from paired pieces as do the arches. 
They are much eloser together than the neural arches, which are 
now appearing anterior to them, much in the same way as the 
hemal arches arose. There is a gap between these epaxial 
elements and the neural arches. These elements, moreover, stand 
quite clear of the nerve-cord, not arching it over as the neural 
arches do. They probably correspond to Huxley’s epiural apo- 
physes in G'asterosteus (Huxley, Q.J.M.S. vol. vii. p. 41). The 
fusion between the neural arch and radial, described as possibly 
existing in Stage III, is now complete: there is no line of division 
between the matrix of one and that of the other. A slight notch 
has appeared in the posterior border, and it is here that the 
caudal artery and vein run out on either side. A radial (7.) at 
the distal end of the hemal arch next anterior to the one just 
described has been formed at this stage, but no fusion of the two 
elements has yet taken place. 
In Stage VI (1'8 cm.) (fig. 6) there is still no sign of the 
cartilages to which the procurrent caudal fin-rays are attached 
in the adult. A typical neural arch, taken further forward than 
those shown in fig. 6, extends at this. stage through a length of 
about 136. It consists of a pair of cartilages lateral to the 
nerve-cord. Their bases do not reach the notochord, and their 
thickness in transverse section 1s about 8p. The hemal arches 
also consist of pairs of cartilages which do not reach up to the 
notochord. They are each about 16 thick, and are separated 
at their distal ends by about 16 » of closely packed skleroblastic 
cells, which have not secreted any matrix as yet. Further 
forward the cartilages are considerably smaller, restricted to the 
sides of the caudal vein, and extend only through a length of 
about 30. Posteriorly the skleroblasts between “the distal ends 
of both dorsal and ventral pairs of cartilages have secreted a 
matrix, and in this way the pairs of cartilages have become 
