454 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
Formation of the Endoderm. 
The formation of the endoderm in Turritopsis cannot be adapted 
to any of the schemes of the development of the hydromedusae that 
have been sketched by Metschnikoff. He distinguishes three principal 
methods for the development of the inner germ layer: first, delamina¬ 
tion, a process in which the segmenting blastomeres divide in a plane 
nearly parallel to the surface, and the inner parts or cells become 
primitive endoderm, while the outer parts remain as primitive ecto¬ 
derm. Second, multipolar digression, in which cells migrate into the 
blastoeoele from different regions of the peripheral cell layer, and are 
transformed into endodermal tissue directly. Of this mode he de- 
scribes several subordinate types. Third, unipolar migration, similar 
to the preceding except that the primitive endoderm cells are given 
off at one pole only, at the posterior end of the larva. 
In Turritopsis the endoderm is derived from the syncytial mass of 
tissue left in the center of the embrvo after the ectoderm has been formed 
«/ 
and separated off by the development of the mesogloea. The inner 
germ layer as a rule is formed much later than the ectoderm. Soon 
after the supporting membrane is developed, cell boundaries begin to 
appear in the syncytium in the interior of the larva. The cells thus 
formed are primitive endodermal cells, and are . crowded together 
without any definite arrangement for a number of hours. Stages in 
which the cell walls are reappearing are shown in figures 48, 49 (pi. 34) 
and 50 (pi. 35). When the embryo is about forty-eight to sixty hours 
old, the time at which attachment takes place, a fissure appears in the 
middle of the mass of endodermal tissue. This is the beginning of 
the coelenteric cavity. This separation begins near the anterior part 
and grows toward the posterior end. The coelenteron gradually 
increases in size, and at the same time the endodermal cells begin to be 
rearranged, and finally become situated parallel to each other with 
their bases against the mesogloea and form a definite inner gerfn layer. 
Gerd has observed in Bougainvillia that during the course of cell 
multiplication the boundaries of the cells become indistinct and that 
the peripheral and central nuclei are altogether identical. But this 
species differs from Turritopsis, according to his description, in the 
formation of the compact morula stage, in that it is brought about by a 
multipolar migration of cells into the interior of the coeloblastula; 
