No. 2.] THE ORIGIN OF THE SEX-CELLS. 227, 
tion beyond this point. There were no quarter planulae 
obtained. 
In order to find out whether when the two blastomeres of 
one ovum were separated they divided synchronously, a very 
fine needle was carefully sharpened until a small knife edge 
had been formed, and then the two blastomeres cut apart with 
this. At first they were flattened each on one side where they 
had been in contact, as is shown in Pl. II, Fig. 37; in five 
minutes they became globular (Fig. 38), and in ten minutes 
for both blastomeres the first cleavage (or properly the second 
cleavage) started in one, forming the two-celled stage, while in 
the other it did not pass through to the opposite pole, but 
started to form buds. Again, in ten minutes more, the second 
cleavage appeared, and while in one blastomere this resulted 
in the formation of a four-celled stage perfectly normal in 
appearance ; in the other one a constriction formed that cut off 
buds, making three small cells at the upper pole, and one large 
cell at the lower pole ; both of these conditions were completed 
at the same period of time. As this latter form is one often 
seen under normal conditions, we cannot conclude that this 
was due to any injury caused by the separation. 
Another fact which was very noticeable in these blastomeres 
which had been shaken, or cut apart, was, that the blastomeres 
derived from them, seemed to have lost, at times, their power 
of holding together. Very often they would segment for a 
certain distance, and then, the cell mass would break up into 
numerous minute blastomeres. 
I regret that I was not able to separate more of these eggs, 
and to follow this work further, but it was not until the last 
week of my stay at Wood’s Holl that it was found possible 
to observe them other than at night, and as the segmentation 
was carried on so rapidly, there was not as much opportunity 
for repeating the experiment as often as desired. However, 
since in five cases planulae were reached, it seems to prove 
that in Hydractinia, as Driesch has proved for Echinus, and 
EK. B. Wilson has shown for Amphioxus, each blastomere of 
the two-celled stage may develop into a half embryo. 
