60 
THE VOYAUE OF H.M.S. CHALLENHEE. 
not explain the origin of the endoderm tissues of the resulting Ascidiozooid. and after 
a careful examination I am unable to find any other elements in the young bud besides 
those mentioned, and yet at an early stage an inner endodermal sac is formed which 
afterwards develops into the branchial sac and the remainder of the alimentar}^ canal. 
This endoderm layer has no connection wdth the wall of the vessel, and is certainly not 
formed from the young ova, but it is, I am convinced, derived from the blood-corpuscles. 
These cells, which are still primitive and undifferentiated, are usually regarded as meso- 
dermal, but E. van Beneden^ has shown that in the Ascidian embryo the mesoblast is 
formed from the primitive endoderm as two laterally placed masses of cells, some of which 
become blood-corpuscles. It is possible that some of these (or their descendants) may 
retain their endodermal characters to such an extent that when they pass into a young 
bud as blood-corpuscles they are still able to act as endoderm cells and form the enteron 
of the future Ascidiozooid. If this view be correct, then the bud is formed by cells 
derived, as we should expect, from all three primary layers of the body of the parent. If, 
on the other hand, the blood-corpuscles cannot be regarded as contributing an endodermal 
element, then it is by no means obvious what the endoderm tissues of the bud are 
derived from. 
The irregular arrangement of the Ascidiozooids in Sarcobotrylloides wyvillii, which is 
rather puzzling as seen from the surface, is not so striking when viewed from the interior 
after the upper layer of the colony has been dissected off. Then the systems can in 
most places be traced, and if the top layer of test be examined from its inner surface, 
the places where the Ascidiozooids were placed, and their arrangement in systems more or 
less like those which are usually found in Botrylloides, can be readily made out (PI. IV, 
fig. 15). In the figure the central area is formed by a thin layer of test, and has no 
common cloacal aperture, while the irregularly rounded projections surrounding it are 
parts of the regions occupied by Ascidiozooids, and have each a somewhat thickened 
margin formed by the test, and in some cases an adhering lining of mantle. 
The mantle is much like that of the genus Botrylloides, but is rather more muscular. 
The branchial sphincter is moderately strong. 
The transverse vessels of the branchial sac are rather wide, and have each a few 
muscle fibres. The stigmata are large and regularly arranged (PL lY. fig. 14, sg.). The 
endostyle is wide and thick. The tentacles are larger than is usual in the genus Botryl- 
loides. 
The infundibulum leading from the aperture of the dorsal tubercle, and its continuation 
the neural duct, are unusually distinct. The funnel is proportionally rather longer than 
in Botrylloides fulgurale (PI, III. fig. 8), but otherwise the relations are the same. The 
neural gland extends for some distance beyond the opaque ovate nerve ganglion posteriorly. 
The stomach is large, and is strongly ribbed on the exterior. In the young 
* Existe-t-il un Coelome chez les Ascidies? Zool. Anzeiger, 1881, No. 88, p. 375. 
