8 BULLETIN OF THE 
will arise ; and it is necessary to do this ın order to study the origin of 
the gemmiparous cells, and the changes which they undergo preparatory 
to an actual involution. 
The study ot tips of branches shows that the necks of the polypides of 
any branch all lie in one plane, and that this plane also includes the 
youngest poly pide ; also that the youngest polypides always arise distad of 
the next older. Knowing these facts, our observations may be confined to 
a short line running from the neck of the youngest apparent buds to the 
tips of the branches studied. The time at which to search for incipient 
buds and the place in the line where they will be found is illustrated by 
Figure 7 (Plate I.). The youngest developed bud is one the axes of whose 
tentacles are approximately parallel to the axis of the branch, and whose 
brain cavity, gn., is not yet constricted off from that of the oesophagus. 
The place of origin is near the tip, immediately beyond the point at which 
the ectoderm changes rapidly from a columnar to a pavement epithelium. 
Figure 3 is from a section across the branch in the region of an incip- 
ient bud. I have already described the conditions of the cells of this 
region. Those near ex. are larger than the surrounding ones, and show 
signs of cell division both in the ectoderm and mesoderm. In both cases 
shown in the figure, the direction of division is such as will tend to 
increase the superficial area of the layer in which it occurs. The ecto- 
derm seems to be the most important layer of the two in the process of 
invagination which is about to take place. I think one is led to this 
conclusion if one considers a folding of an epithelium to be due to an 
increase ir, the area of the epithelium within a certain circumference 
without a correspondingly great increase in the circumference itself. 
Such a conception implies, first of all, mutual pressure of the cells of 
the invaginating epithelium. The cells of the mesodermal layer do not 
seem to be under mutual pressure; in some cases they are barely in 
contact. The cells of the ectoderm are evidently closely applied, and 
probably, therefore, under mutual pressure. 
The one case of cell division which is occurring in the ectoderm is at 
the inner end of the cell. In fact, the centre of the nuclear plate is much 
nearer the deep end than are the centres of the adjacent nuclei. The 
effect of this division is to increase the area on the inner surface of the 
ectoderm more than that on the outer, as appears from a study of the 
sections shown in Figures 4 and 5. In Figure 4 certain cells lie already 
below the niveau of the surrounding ones, very much as though they had 
moved downward on account of this being the direction of least resist- 
ance. A later stage of this process is shown in Figure 5. Here the 
