BY JAS. P. HILL. 
15 
thicker in the middle — the continuation of the " keel " — -derived 
also from the throat epithelium as the opening of the notochordal 
lumen gradually moved posteriorly (fig. 16, lps '.). Towards the 
posterior end of the skeleton the "keel" disappears entirely, the 
two " nuclei " separate from each other ventrally, and the 
epithelium of the throat extends up into the cleft between them. 
The continuation of the body lying above the nuclei then thins 
out, and the "nuclei" separate completely from each other, giving 
rise to the " legs " proper which lie under a fold of the throat 
epithelium. The "legs" may pass out at once almost transversely, 
or they may diverge more gradually in different individuals. 
They terminate considerably in front of the middle region of the 
collar. 
The proboscis skeleton is composed of a cuticular substance 
which shows in this species a very distinct stratification, indicat- 
ing the mode of origin of the mass by the deposition of successive 
layers, and the direction of these layers indicates very clearly 
from what source they are derived. As Spengel has shown, the 
"end plate" is derived from the ventral blind sac of the notochord, 
while the "body" is derived from the neck of the same. The 
" keel " on the other hand is formed from the epidermis of the 
proboscis neck, and its posterior continuation from the epithelium 
of the throat. 
In this species the " end plate " and the anterior portion of the 
" body " are separated from the keel by chondroid tissue, and the 
direction of the lines of stratification in these parts indicates 
clearly enough their distinct origin. The relations of the skeleton 
to the limiting membranes are essentially those of the described 
species. The body passes over at its edges into the limiting 
membrane of the notochord, while the wings of the keel pass 
over into that of the epidermis. 
In the proboscis skeleton of this species there occur a few 
small oval cells with non-staining cell bodies and deeply staining 
small nuclei. They are situated between the layers of stratifica- 
tion. Such cells occur in the proboscis skeleton of Pt. clavigera, 
gigas and aurantiaca, and with Spengel I believe they are cells 
