No.34] TUBULARIAN HYDROIDS. 421 
immediately after cutting, by placing them on one end, under 
a coverslip, in a cell-slide, and providing them with water enough 
to cover fully. The interior could thus be easily studied. 
Each fragment appeared at first as a simple ring (see Fig. 1), 
in some cases with a system of endodermal cells across the 
middle — open freely at both ends; no signs of any somatic 
fluid were visible at first ; in about five minutes the peripheral 
cells of the cut ends were seen to elongate gradually (Fig. 2) 
and to extend in an amoeboid manner towards the center ; 
this continued until they met in the center, thus closing in 
the cut ends entirely. The cells at this point presented a 
distinctly elongated radiate arrangement (Fig. 3). The mem- 
brane thus formed over the cut ends was very thin and allowed 
the vigorous circulation of somatic fluid (which appeared very 
soon after the closing) to be readily seen. In an hour’s time 
this circulation had become very rapid, and many masses which 
looked like cells, and pigment granules were seen whirling 
about in the central cavity. 
In order to observe these first stages of healing more fully, 
fresh stems were taken, placed under the microscope and longi- 
tudinal fissures made in them with needles. At first a slight 
gaping of the wound appeared; very soon, however, the cells 
along the edges elongated towards each other until in a very 
short time the wound was completely closed, and circulation of 
the fluid, which had apparently been inhibited by the wounding, 
was again established. In all cases the healing was very rapid. 
Evidently both ectoderm and endoderm had taken part in this 
process as the wounded ends of the coenosarc (seen in Fig. 
4), when viewed from the side, showed the pigmented layer 
beneath the colorless ectodermal layer. 
After the healing of the cut ends no changes were visible 
for some time; the membrane at the ends of the coenosarc 
remained thinner, and the somatic cavity larger. Whether 
this was due to the distension of cells in that part, or an actual 
thinning of the walls, owing to the wandering away of some of 
the endodermal cells into the central cavity, was not deter- 
mined. The rapid circulation and next visible change in 
the regenerating portions indicated that some changes in the 
