Notes on the morphology of the Tunicata. 577 



united in this way, and each of these with a torn piece of stolon 

 which must have bound it to another individual. The isolated in- 

 dividuals also show the broken stumps of the stolon. It is certain, 

 therefore, that the colonies consist of at least four individuals, and 

 it is highly probable that all fifteen of my specimens were united 

 into a single colony. It is very possible that the colony was still 

 larger and that only a part of it was caught by the collecting ap- 

 paratus. I have already shown that the stolon consists of an outer 

 tube of tunic with its lining of mantle, perforated by a cord of tissue 

 derived from the wall of the alimentary tube (Fig. 81 C). I am 

 unable to say whether the endodermal cells, or only the mesodermal 

 wall of the alimentary tube, enters into the formation of this cord. 

 This axial cord where it passes across the atrial chamber has a 

 sheath formed by the atrial membrane (Fig. 81 C). This stolon is 

 very similar to the stolon in the Clavelinidae, which unites the buds 

 into a single colony. The union of the individuals in Octacnemus 

 is not so intimate as to suggest comparison with the compound 

 Ascidian colony. In Octacnemus, as in the Clavelinidae, the buds 

 on the stolon are placed belly to back^). 



In my former paper I described a few minor points, such as 

 the thickness of the test etc, to which it is not necessary here to 

 refer, since I desire to describe only such features as have a bearing 

 upon our interpretation of the organization of this very peculiar genus. 



Herdman believes Octacnemus to be a representative of an 

 aberrant family of the Hemimyaria among the Thaliacea. Upon 

 this point he says ^) : "On the whole I regard this form as being 

 allied to Salpa. The condition of the visceral mass is very like the 

 'nucleus' in Salpa, and it occupies much the same position in the body. 

 The musculature might readily be derived from a series of trans- 

 versely running bands by an antero-posterior shortening which would 

 approximate the bands closely, and then by a portion of the muscles 

 being drawn out radially into the eight conical processes. The endo- 

 style and the nervous system are in their proper places, but there 

 seems to be no trace of a dorsal lamina; and the branchial sac is 

 certainly in a remarkable condition. If the obliteration of the side 

 walls of the sac in Salpa has been brought about by the locomotory 

 habits of that form, of course no such change would be necessary 



1) Lefèvre, 1898. 



2) Herdman, 1888, p. 96. 



38* 



