Marshall—On the Anatomy of the Dragonfly. 761 
this region the dorsal wall of the oesophagus remains nearly 
level, the ventral wall however falls suddenly away from the 
dorsal causing the abrupt enlargement. 
The crop, beginning in the middle of the metathorax, ex¬ 
tends through the first, second and the anterior half of the 
third abdominal segments; at its posterior end it decreases 
slightly in diameter. The thickest part of the crop is in the 
first and second abdominal segments and exceeds in diameter 
any other portion of the alimentary canal. At its posterior 
end the fore-intestine decreases in diameter and, when seen 
externally, passes over into the mid-intestine as the end of 
one tube pushed into another and slightly larger one. The 
boundaries between the different regions of the fore-intestine 
must be more or less assumed as there are no external markings 
or changes which separate one part from another. The same 
is true from a study of the structure of the wall. 
The mid-intestine starts near the middle of the third ab¬ 
dominal segment; its diameter does not change within the 
fourth and fifth segments except^ that, at the posterior end of 
the latter segment, there is a gradual decrease in size. In de¬ 
fining the position of the boundary between the fore- and mid¬ 
intestines one must bear in mind that the presence of the 
oesophageal valve which is pushed into this region of the ali¬ 
mentary canal, makes this boundary a little different in ex¬ 
ternal view to what it is really found to be in longitudinal sec¬ 
tion. We take here, however, the anterior end of the mid- 
intestine as seen in external view and find that this boundary 
lies within the third abdominal segment. In some specimens 
it is well toward the boundary between the third and second 
segment, in others near or at the boundary between the third 
and fourth. In the fourth and fifth segments and also in 
a small part of the third, the mid-intestine is covered with 
transverse folds, each fold appearing as a ridge passing around 
the intestine. The folds have sinuous boundaries and are 
fairly regular. From a surface view many of these folds can 
be traced from side to side across the intestine; when closely 
studied and any one fold is carefully followed it is found that 
