EMBEYOLOGY OF THE SEA BASS. 
255 
between the alimentary canal and yolk in Fig. 145. It therefore becomes evident 
that the growth of the coelom round the yolk, begun in Fig. 141, is merely a part of 
the geueral growth of the coelom round the ventral surface of the alimentary caual. 
Near its ventral edge the somatopleure shows in the larval stages a band-like 
thickening {m. b., Figs. 141 and 145) which looks in some respects like a muscle band. 
Wolffian duct . — It is generally agreed that the Wolffian duct arises in the Tele- 
ostei as a fold of the coelom. The formation of the dnet begins anteriorly and tr*ivels 
back. Anteriorly the layers of the coelom separate so as to inclose a true cavity, and 
there is formed a well-marked diverticulum (^c. d., Fig. 103, PI. C). But there is 
no stage in which the duct exists along its whole length as a diverticulum. This may 
be due to the possible fact that as quickly as the duct is formed it is constricted off. 
The appearances indicate, however, that in its posterior half the duct is constricted 
off from the coelom as a solid mass, the cells of which are radially disposed round an 
ideal lumen. The position and character of the duct, after separation from theccclom, 
are shown in Fig. 110, PI. ci, n\ d., at the time of hatching, in Figs. 126, PI. cii, 
127, Plate cm. At the time of hatching the ducts extend as far back as the anus, but 
do not form a urinary bladder. The entire course of the duct is straight, and at its 
anterior end it opens into the coelom. 
The urinary bladder begins to form shortly after hatching by the fusion of the 
posterior ends of the ducts. It is very thin walled and opens just behind the anus. 
The only other step in the later development of the duct that I have observed is the 
formation of an anterior loop. On the second or third day after hatching, the anterior 
end of the duct bends round and runs posteriorly for a short distance. This is shown 
in the three successive sections. Figs. 143,, 144, and 145, PI. cv, of which Fig. 143 is 
the posterior. The opening of the duct into the coelom (Fig. 144) at this stage could 
not be made out with certainty, but the part of the duct which I have represented as 
opening is morphologically the anterior end, and in an earlier stage this end very 
plainly opened into the coelom. 
Somites . — The marking off of the somites begins at about the same time as the 
separation of the coelom. It begins anteriorly, but whetlier the first somite formed is 
the true anterior one I do not know. The somites are marked off by dorso lateral con- 
strictions (Fig. 98, PI. cxix, surface view, and Fig. 85, PI. xcvm, vertical longitudinal 
section to one side of median line) from which planes of division run into the sub- 
stance of the mesoderm plates. The somites are, it is needless to say, solid. Hoff- 
mann (17) describes them in the Trout as hollow, but he stands alone in this opinion. 
The formation of somites travels antero-posteriorly, and during the greater part 
of embryonic life there is at the posterior end of the embryo, on each side, a quite long 
tract of undivided mesoderm (tm. mes.. Fig. 85, PL xcviii, and Fig. 98, PI. xcix) which 
at the tip of the tail ends in the caudal mass. New somites are constantly split off from 
the anterior end of the undivided mesoderm, and consequently the formation of somites 
and their gradual alteration may be studied in the posterior part of a single embryo 
quite as well as in corresponding parts of different stages. At the time of hatching 
there remains only a very little of the undivided mesoderm, the somites extending 
nearly to the tip of the tail. 
In Fig. 98, PL xcix, and Fig. 85, PL xcviii, the somites next the undivided meso- 
derm are the younger, and on going forward it is seen that they gradually undergo 
certain changes of shape. When the somite is first formed, it is constricted off from the 
