62 HENRY BURY. 
enables us to trace the true relations of this cavity, which 
without the greatest care in tracing its origin might be thought 
to be a diverticulum of the right, instead of the left, body- 
cavity. When it reaches the dorsal side it continues its growth 
past the middle line, till it reaches the water-tube, at the 
level of which it lies. Beyond this I have not been able to 
trace it with certainty; but since at this stage the most anterior 
portion of the right body-cavity is posterior to the water-tube, 
there is scarcely room for doubt that the ** ventral horn” forms 
one side of the mesentery (very short in the pupa) which sup- 
ports this tube—the other side being formed by the anterior 
part of the main mass of the left body-cavity, which gradually, 
during stage A, grows up the left side of the csophagus 
(fig. 12). 
By the time stage B is reached, all trace of separation be- 
tween the right body-cavity and the ventral horn of the left 
cavity has disappeared; so that the mesentery of the water- 
tube (formed, as it seems, between two parts of the left 
cavity) is continuous with the main dorsal mesentery (formed 
between the right and left body-cavities), with which, indeed, 
Semon has confused it. The junction of the two is, however, 
still marked by a change in position—the water-tube and its 
mesentery being adradial (fig. 14), while the true dorsal mesen- 
tery is strictly interradial. It will be seen that the dorsal 
mesentery is even more on the left of the original middle line 
than the water-tube—not, as Semon considers it, in the median 
plane. Its posterior end is, however, nearer to this plane than 
its anterior. 
This dorsal mesentery is first formed early in stage A by 
the meeting of the body-cavities about two-thirds of the way 
down the stomach, and from this point spreads rapidly forwards. 
Nearly at the posterior end of the stomach it passes sharply round 
to the left side, almost on to the ventral surface (fig. 16), thus 
forming a short transverse mesentery, which we shall see again 
in the pupa. 
Very early in stage A the two body-cavities fuse ventrally 
along the ventral surface, at the point at which they so closely 
’ 
