CLASSIFICATION OF THE ANIMAL KINGDOM, 55 
cavity. This occurs in no other vertebrated animal. Kowa- 
lewsky has proved that this very exceptional structure results 
from the development of the somatopleure as a lamina which 
grows out from the sides of the body and eventually becomes 
united with its fellow in the middle ventral line, leaving only 
the so-called “‘ respiratory pore” open. Stieda has mentioned 
the existence of the raphé in the position of the line of 
union in the adult animal. Rathke described two “ abdo- 
minal canals” in Amphioxus; and Johannes Miller, and 
more recently Stieda, have described and figured these canals. 
However, Rathke’s canals have no existence, and what have 
been taken for them are simply passages or semi-canals 
between the proper ventral wall of the abdomen and the in- 
curved edges of two ridges developed at the junction of the 
ventral with the lateral faces of the body, which extend from 
behind the abdominal pore where they nearly meet, to the 
sides of the mouth. Doubtless, the ova which Kowalewsky 
saw pass out of the mouth had entered into these semi- 
canals when they left the body by the abdominal pore, and 
were conveyed by them to the oral region. ‘The ventral 
integument, between the ventrolateral lamine, is folded, as 
Stieda has indicated, into numerous close-set, longitudinal 
plaits which have been mistaken for muscular fibres, and the 
grooves between these plaits are occupied by epidermic 
cells, so that, in transverse sections the interspaces between 
the plaits have the appearance of glandular ceca. This 
plaited organ appears to represent the Wolffian duct of 
the higher Vertebrata, which, in accordance with the 
generally embryonic character of Amphioxus, retains its 
primitive form of an open groove. ‘The somatopleure of 
Amphioxus, therefore, resembles that of ordinary Vertebrata 
in giving rise to a Wolffian duct by invagination of its inner 
surface. But the Wolffian duct does not become converted 
into a tube, and its dorsal or axial wall unites with its fellow 
in the raphé of the ventral boundary of the perivisceral 
cavity. 
In all the higher Vertebrata of which the development has 
yet been traced, the “‘ pleuro-peritoneal”’ or pervisceral cavity 
arises by an apparent splitting of the mesoblast, which 
splitting, however, does not extend beyond the hinder por- 
tion of the branchial region. But, in many Vertebrata (e. g., 
Holocephali, Ganoidei, Teleostei, Amphibia) a process of the 
integument grows out from the region of the hyoidean arch, 
and forms an operculum covering the gill-cleft. In the frog, 
as is well known, this opercular membrane is very large, 
and unites with the body wall posteriorly, leaving only a 
