162 THE DEVELOPMENT OF THE CHICK 
source of much of the mesenchyme, and it is an interesting ques- 
tion whether or not such mesenchyme has a different fate from 
that of different origin. Nothing definite, however, is known in 
regard to this, owing to the impossibility of separating the various 
kinds after they have once merged. 
The Neural Crest in the Region of the Somites. The neural 
crest is very slightly developed in the region of the first five so- 
mites, which is correlated with the fact that these somites are 
devoid of gangla. But the mode of origin is the same through- 
out the somitic region. Shortly after the closure of the neural 
tube in any region the neural crest forms an aggregation of cells 
in the roof, more or less sharply separated from the remainder 
of the tube both by the arrangement of the cells and also by their 
lighter stain (Figs. 107, 109, 112, 113). The early history may be 
followed in a single embryo, by comparing the conditions opposite 
the last somite with that of more anterior somites where develop- 
ment is more advanced. Figs. 107, 108, 109, 110 represent 
transverse sections through the twenty-ninth, twenty-sixth, 
twentieth, and seventeenth somites of a 29s embryo. In Fig. 
107 the cells of the crest are extending towards the upper angle 
of the somite, with which they are connected by protoplasmic 
strands. The aggregation in the roof of the neural tube is thus 
decidedly diminished; the lateral wings of the crest lie in the angle 
between the neural tube and the ectoderm. In the twenty-sixth 
somite (Fig. 108) the lateral wings extend farther from their point 
of origin, and appear to have a more intimate connection with 
the myotome. In the more anterior and older somites, twenty 
and seventeen (Figs. 109 and 110), the process has progressed 
much farther and the neural crest cells are completely expelled 
from the neural tube, which closes after them (Fig. 110). A yet 
later stage is shown in Fig. 111, through the twenty-third somite 
of a 35s embryo. 
The dorsal commissure uniting the right and left sides of the 
crest ruptures, and the cells of the crest aggregate so as to form 
a pair of ganglia in each somite. Thus, although the neural crest 
is primarily a median structure, it becomes divided into two 
lateral halves, and although it is primarily a continuous structure 
it becomes divided into a series of pairs of metamerice ganglia. 
The fate of the interganglionic commissures is conjectural. The 
ganglia are ill-defined from the mesenchyme when they are first 
