Photographic records of extending pieces of the wall indicate that the choanocytes were but little affected by the stretching process. By adjusting their contacts mutually and by altering their positions on the wall, they avoided becoming greatly stretched in the direction of the applied tension.
Collar-cells cohere together and appear to be under tension in the expanded oscular tube. They also adhere to the inner surface of the wall. There was no evidence of attachment between spicules and choanocytes, and none for a transformation of the latter into pinacocytes, or conversely.
The pattern of pores changed considerably as the pieces extended, pores becoming obliterated and new ones opening up elsewhere. The pores never became slit-like, despite the considerable extension achieved by the pieces. Where pores persisted for a sufficient period, evidence was obtained to support the hypothesis that the porocytes are interconnected to form an epithelium on which the choanocytes can move about and to which, on its other surface, the founder calcoblasts of developing tri and quadric-radiates are attached.