Kingsley.] 
442 
[March 15, 
“ cells,” the method of their origin, and their purpose or function, 
I have nothing new to offer. I do not believe them to be true 
cells. Semper considers them amoeboid protoplasmic bodies result- 
ing from the yolk under the influence of sea water or from other 
causes. He says that they do not normally occur until the egg is 
nearly mature, though they can be produced in younger eggs by 
the action of salt water. I have tried many times to verify this 
statement but without success. These “test-cells,” as they have 
been called, strongly resemble nuclei with several strongly refrac- 
tive nucleoli, but this appearance has been neglected in preparing 
the drawings for the photographer. In the eggs studied by me they 
were irregularly scattered between the vitellus and the external 
envelope, offering a strong contrast to the description and figures 
given by Lacazc-Duthier of the eggs of a closely allied form. 
The presence of these test-cells prevented my seeing, with any 
degree of certainty, the origin of the polar globules. 
Impregnation and the development of the embryo until the 
tadpole stage occurs within the body of the parent and when the 
embryos are fitted for an active life they may frequently be seen 
passing out of the atrial opening of the adult. The earliest 
stage of the impregnated egg observed had the first segmenta- 
tion completed (fig. 4). Here there are two blastomeres, unequal 
in size, each with one extremity moderately transparent while the 
rest of the cells are so filled with deutoplastic granules as to com- 
pletely hide the nuclei. Ho further notes were made on this egg. 
The next stage observed had four large and four small segmenta- 
tion spheres thus showing a differentation into epiblast and 
hypoblast. The four smaller cells (epiblast) were comparatively 
transparent allowing the nucleus to b£ indistinctly seen while the 
larger or hypohlastic ones were perfectly opaque. In about two 
hours four more small cells were segmented from the four liypo- 
blastic ones, the result being an embryo with four large and eight 
small cells. The observations above described are in close 
accordance with those of Lacaze-Duthier on a species of the 
same genus, but are strongly at variance with the process of seg- 
mentation in other Ascidians as described by Kowalewsky, 
Kuppfer, and others. It is somewhat interesting to note that the 
segmentation as seen in the eggs of Molgula resembles closely 
