THE ANCESTRY OF THE CHORDATA. 553 
gill-slits are formed as regular fusions and perforations of the 
body wall and gut from before backwards. Hence the three 
features which alone distinguish Chordata from other animals 
are present, and associated from an early period in develop- 
ment. Added to this the minor features of Chordate anatomy 
are also represented by (1) the origin of the mesoblast; (2) 
the remarkable asymmetry of the anterior parts; (3) the 
opercular fold; (4) the excretory funnels opening into the 
atrial cavity thus formed. From all these facts we may form 
a preliminary conclusion that the Enteropneusta bear some 
relation to the Chordata. We will now discuss what relation 
this is, and before doing so we must determine what relative 
importance is to be attributed to the two modes of develop- 
ment known to occur, the one largely embryonic the other 
pelagic. 
In our present state of ignorance as to the mode of develop- 
ment of Tornaria and of the details of its later stages, it is 
difficult to compare these two modes, but the question as to 
which is to be regarded as primitive is probably a part of the 
larger question as to the comparative likelihood of the pre- 
servation of ancestral features in the free or in the pretected 
developments. This question cannot be fully gone into here. 
No general answer has as yet been given to it, and since the 
balance of probability is very nearly divided between these two 
possibilities we may be right in assuming either of them to be 
correct. For the purposes of the following argument it will be 
assumed that, on the whole, development within an egg-shell, 
as involving a less complicated struggle with environmental 
forces, is less subject to variation than that in the open sea, and 
consequently is more likely to preserve ancestral features. 
Besides this, in the special case before us, the adult structure 
is practically conclusive against Echinoderm affinities, to 
which the pelagic development would point if regarded as 
primitive. 
Assuming, then, that the development of B. Kowalevskii 
is more primitive than that involving a Tornaria stage, the 
following features are of great importance : 
