416 WILSON. [Vo. III. 
rows (which here represent the future cesophageal commissures) 
are partly covered in at the sides by the general ectoblast, and 
that their extreme anterior ends are enlarged and partially united 
with the dorsal ectoblast to form the foundation of the cephalic 
ganglia. This section shows also the beginning of the migration 
of the anterior mesoblast cells to form the cephalic mesoblast. 
Figure 76 is a more enlarged view of part of the adjoining sec- 
tion to show the continuity of the neural row with the cephalic 
foundation. Figure 77 is four sections further along, and shows 
the complete disappearance of the neural rows towards the me- 
dian line. Figure 78, four sections further on, passes in front 
through the median line, and shows the mouth-cavity, the group 
of “ Schluckzellen”’ which forms the anterior limit of the mouth 
and were visible already in Fig. 77, and the large ciliated cells of 
the median ventral surface. In sections at a still lower level (z.2., 
on the left of the median line) the neural and mesoblastic bands of 
the left side appear in oblique section. This clearly demonstrates 
the fact that at this period, although the gastrula ts fully estab- 
lished, the germ-bands are completely separated in front. The 
series shows furthermore that there is no apical neural plate, the 
cephalic ganglia arising independently on either side of the me- 
dian line. It shows finally the important fact, that the cephalic 
ganglia arise from the same neural rows that give rise to the 
cesophageal commissures and the ventral chain, though their 
development is slightly in advance of the other parts (see p. 434). 
At a slightly later period the mesoblastic, neural, and ecto- 
blastic elements of the opposite sides of the body grow together 
in the median dorsal line above the mouth (cf Fig. 43.), and the 
unpaired cephalic cavity appears just between the two cephalic 
ganglia (Fig. 56). At this period a bridge of neural tissue, 
closely united with the ectoblast, can be traced across the me- 
dian line from one cephalic ganglion to the other (Figs. 55, 56). 
I have not certainly determined whether this bridge is formed 
by proliferation of the overlying ectoblast or by growth of the 
neural structures already existing. I incline, however, to the 
former view, because there appears to be a complete fusion of 
the neural tissue with the ectoblast in the middle line. 
The question whether the cephalic ganglia, like those of the 
ventral chain, are derived from the neuroblasts or by prolifer- 
ation from the general ectoblast, still remains in doubt. This 
