66 MORGAN. [Vox. IX. 
region of the animal made the same length as the collar, — 
no one would doubt that we were dealing with a metameric 
animal of three segments. And if we look upon Balanoglossus 
as a form profoundly modified to fit it for life in the sand, 
if we regard the proboscis as a pra-oral lobe (and head-seg- 
ment), enlarged as a boring organ, and the posterior metamere 
of the body immensely prolonged, we cannot escape the con- 
clusion that even in the narrowest definition of an animal with 
metameric structure we must include Balanoglossus. 
Collar-pores and Gill-slits. 
Several important questions come up in relation to the 
formation of the collar-pores and the gill-slits. Spengel’s state- 
ment that the collar-pores form from the first gill-slit cannot 
be accepted. It would be as easy to maintain the converse, 
i.e., that the gill-slits come from the collar-pores. We may 
formulate this statement: 7he collar-pores and series of gill- 
openings arise from a pair of invaginations of the ectoderm. 
This puts the problem in another light, and the following 
question is suggested: /s the pair of invaginations to be looked 
upon as a beginning of an atrium ? 
I am not prepared to say that this is true, but it is a view 
that is suggested by the facts of development, and it seems 
worth consideration. 
From the anterior wall of the invagination a forward exten- 
sion opens into the collar body-cavity to form the collar-pores ; 
farther back the first gill-pouch of each side opens into the 
invagination. At a later period the second pouch opens into 
the invagination behind the first opening, and the succeeding 
pairs follow the same law. Now, if the backward extension of 
the ectodermic invagination had pushed back beneath the 
surface ectoderm, and into this pocket on each side the gill- 
slits had opened, there would have been little question as to 
the resemblance between such a structure and the formation of 
the atrium in the Ascidians. The possibilities do certainly 
exist in Balanoglossus, but who can say whether they were 
realized in the higher groups? One observation ought to be 
noticed here in this connection. The lateral walls of the 
