360 G. H. PARKER 
half of this structure. When the cannula of the injection tube 
was tied into the superior canal of the distal quarter of the 
peduncle, the same freedom, of communication between the two 
_ canals was demonstrated as when it was tied into the proximal 
end of the peduncle. It is, therefore, evident that the pores of 
the septum occur not only in the distal half of this structure, but 
also in its distal third and fourth quarters. They probably form 
a more or less continuous series in this region. 
If the converse of the preceding experiments is tried, in that 
the injections are made into the inferior canal and the superior 
canal is cut open, fluid flows freely from the inferior into the 
superior canal, and it is fair to conclude that the pores in the 
septum between these canals are unprovided with valves that 
limit the direction in which fluid may pass. In the peduncle, 
then, the septum between the superior and inferior canal is im- 
pervious to injection fluids in the proximal half of its extent and 
freely pervious in its distal half. In this half the pores must be 
relatively large, for such coarse material as india ink is freely 
transmitted. 
As Agassiz (’60, p. 208) long ago pointed out, the canals of 
the peduncle communicate freely with the spaces in the rachis. 
In all my work on the canals in the rachis of Renilla I have failed 
' to find more than extensions and ramifications of the two canals 
of the peduncle of that sea-pen, and I am, therefore, unable to 
confirm Hisen’s declaration (’76, p. 12) that where the peduncle 
joins the rachis in Renilla there are four canals, a number agree- 
ing with that reported in the peduncles of many sea-pens. As I 
have worked upon the same species of Renilla that Eisen did, I 
suspect that in this particular his account is inaccurate. 
If the peduncle of a Renilla is cut across about midway on its 
length and the injection cannula tied into the superior canal of 
the part which is still connected with the rachis, an india-ink 
injection can be made to flow through the superior canal and out 
at the pore of the axial siphonozoéid. The injection, as a rule, 
will not spread further. If methylen blue is used, a much more 
general injection is obtained. The methylen-blue solution flows 
first into the intermesenteric chambers of certain autozodids, 
