DR. E. B. WILSON ON THE DEVELOPMENT OF RENILLA. 767 
are clear, with very delicate, rounded or polygonal outlines, and with very conspicuous 
intensely stained nuclei, which are quite similar to those of the entoderm cells (en.). 
Towards the free edge of the septum (e.) the two layers of the lamella disappear and 
the axial cells become confounded with the entoderm cells. 
The iatter point is most clearly shown in transverse sections taken just behind the 
free edge of the septum (fig. 152). We find here that the lamella of the lateral 
portion is simple as in the radial septa, but further inwards the lamella splits into two 
layers between which lie a number of closely packed axial cells. Still farther inwards 
the layers of the lamella entirely disappear and the axial cells graduate insensibly into 
the rounded entoderm cells which form the edge of the septum. In the section 
immediately behind this, the layer of axial cells can be traced quite across from one 
side to the other, but they lie several cells thick in the middle and are scarcely 
distinguishable from the adjoining entoderm cells. In sections further forwards the 
septum entirely disappears and the body of the larva consists of an unbroken layer of 
ectoderm and entoderm enclosing a nearly solid mass of yolk. 
These sections show very clearly that as the septum grows forwards the entoderm 
cells of which it is composed arrange themselves in three layers. The two outer layers 
persist as the entodermic covering of the peduncular septum, and form its main bulk ; 
the cells of the middle layer atrophy, flatten together, and form the axial cells. The 
two layers of lamella which enclose the axial cells are no doubt secreted by the 
adjoining entoderm cells; the appearances indicate that these membranes are simply 
the confluent and much thickened membranes of the cells. 
As in the case of the radial septa, I have studied with care the possibility of 
ectodermic cells passing into the septum at its lateral parts where it joins the body- 
wall, but have been unable to find decisive evidence of such a process. The sections 
show exactly the same appearances as those of the radial septa. The lamella of the 
septum sometimes joins that of the body-wall abruptly, without any infolding of the 
latter ; in other cases the lamella of the body-wall is somewhat infolded, and the angle 
thus formed contains ectoderm cells; in other cases, again, a small triangular space 
appears at the root of the septum, enclosing one or two cells. The latter are quite 
similar to the ectoderm cells which appear in the last-described case, and seem to have 
been introduced from the outside. In rave cases, one of which is carefully represented 
in fig. 156, the lamella has the appearance of folding in so as to leave a narrow 
connexion between the cleft containing the axial cells and the ectodermic layer. 
From these appearances I conclude that the ectoderm cells may in some cases actually 
pass into the septum by an infolding of the lamella, but they can never do so in 
considerable numbers, and take only the most insignificant part in the formation of 
the septum. Far the greater part of the peduncular septum, as of the radial septa, 
is formed from entoderm cells alone. Misled by certain theoretical considerations, I 
was at first strongly inclined to regard the axial cells as ectodermic in origin, having 
been invaginated from the exterior in a fold of the lamella. More careful study 
5 F 2 
