890 STUDIES IN GENERAL PHYSIOLOGY 
ture, it dies. If the experiment is interrupted early—that 
is, when the cell-walls have just begun to become indistinct 
—all, or at least a part, of the cell-membranes again become 
visible upon admission of air. Under these circumstances, 
however, every cell usually divides, not 
into two, but into four cells which cor- 
responds with what has been said before. 
When we wait a little longer before 
admitting air, a circular blastoderm is 
at first formed in which no trace of cleav- 
age is visible. The blastoderm then sud- 
mae denly breaks 
up into a large number of cells at 
once, but curiously enough this 
cleavage is confined, in most cases, 
to the periphery of the blastoderm. 
In this case also the refractive sub- 
stance which has been described 
plays a peculiar réle. Figs. 112- 
17 represent 
the various B1Ge 10 
stages of the renewed cleavage of the 
same blastoderm in which we studied 
the disappearance of the lines of cleav- 
age in hydrogen (Figs. 103-8). Fig. 
108 shows the condition of the blasto- 
derm in hydrogen at 2:10 o’clock. Only 
four large drops of the refractive sub- 
stance, surrounded by droplets of smaller 
size, permit one to recognize the place 
of the blastoderm. At 2:18 pure oxygen was sent through 
the gas-chamber. At first the smaller droplets separated 
from the surface of the large droplets and moved toward 
what had been the periphery of the blastoderm. (Previously, 
FIG. 114 
Digitized by Microsoft® 
