54 PROFESSOR HUXLEY. 
terata. 2. The Scolecimorpha. Under the latter head are 
included the Turbellaria, the Nematoidea, the Trematoda, the 
Eirudinea, the Oligocheta, and probably the Rotifera and 
Gephyrea. In all the other Monostomata the primitive open- 
ing of the gastrula, whatever its fate, does not become the 
mouth, but the latter is produced by a secondary perforation 
of the body wall. In these Deuterostomata there is a peri- 
visceral cavity distinct from the alimentary canal, but this 
perivisceral cavity is produced in different ways. 
1. A perivisceral cavity is formed by diverticula of the 
alimentary canal, which become shut off from the latter 
(Enterocela). The researches of Alexander Agassiz and of 
Metschnikoff have shown that, not only the ambulacral 
vessels, but the perivisceral cavity of the Echinodermata, are 
produced in this manner; a fact which may be interpreted 
as indicating an affinity with the Celenterates (though it 
must not be forgotten that the dendroceele Turbellaria and 
many Zrematoda are truly “ccelenterate”’), but does not in 
the least interfere with the fundamental resemblance of these 
animals to the worms. Kowalewsky has shown that the 
perivisceral cavity of the anomalous Sagitta is formed in the 
same way, and the researches of Metschnikoff appear to 
indicate that something of the same kind takes place in 
Balanoglossus. 
2. A perivisceral cavity is formed by the splitting of the 
mesoblast (Schizocela). This appears to be the case in all 
ordinary Mollusca, in all the polychetous Annelida, of which 
the Mollusca are little more than oligomerous modifications, 
and in all the Arthropoda. 
It remains to be seen whether the Brachiopoda and the 
Polyzoa belong to this or the preceding division. 
3. A perivisceral cavity is formed neither from diverticula 
of the alimentary canal nor by the splitting of the mesoblast, 
but by an outgrowth or invagination of the outer wall of the 
body (Hpicela). The Tunicata are in this case, the atrial 
cavity in them being formed by invagination of the epiblast. 
Amphioxus, which so closely resembles an Ascidian in its 
development, has a perivisceral cavity which essentially cor- 
responds with the atrium of the Ascidian, though it is 
formed in a somewhat different manner. One of the most 
striking peculiarities in the structure of Amphioxus is the 
fact that the body wall (which obviously answers to the 
somatopleure of one of the higher Vertebrata, and incloses 
a *‘pleuro-peritoneal” cavity, in the walls of which the 
generative organs are developed) covers the branchial aper- 
tures, so that the latter open into the ‘‘ pleuro-peritoneal”’ 
