735 
and of the later embryos of the various types, and have observed 
hundreds of them of all stages without once obtaining a gastrula. 
Sooner or later all died, and although it is in the nature of the case 
difficult to reach a decisive result the facts are such as to create a 
strong presumption that the blastomeres of the 8-celled stage have 
lost the power to produce the two germ-layers of a gastrula. This 
is not the place to consider in extenso the significance of the 
partial development, but a brief reference to it may emphasize the 
importance of certain facts that I wish to record. There are two 
possibilities regarding the partial development of the 4 blastomeres. 
Their inability to produce a complete embryo may be due either to 
qualitative or to quantitative limitations. In the first case we may 
conceive the protoplasm to have become so far differentiated as to 
be incapable of regenerating the whole. In the second case the limi- 
tation of development may be caused simply by lack of material, the 
total mass of protoplasm being inadequate to the work it has to per- 
form. Two facts however tell against the second of these possibilities. 
The first is that the ! partial embryos show otherwise no lack of 
developmental vigor or tenacity of life; they swim as actively and 
live as long as the } embryos. The second fact, which seems nearly 
conclusive, is that under certain conditions minute gastrulas are pro- 
duced which may be even less than { the normal size (Fig. 1, D). 
These gastrulas however did not arise from 8-celled stages 
but from 2- and 4-celled, and I can only explain their origin by sup- 
posing either that the + or $ blastomeres underwent preliminary fission 
before beginning their progressive development or that they were me- 
chanically broken into smaller fragments by the operation of shaking, 
as often happens in the case of entire undivided ova. Whatever their 
origin, these minute embryos prove that a mass # the size of the 
normal ovum, or less, is capable of producing a gastrula — a fact 
which contradicts the assumption that the 4 blastomeres fail to de- 
velop through quantitative limitation. 
II. 
Lack of space makes it necssary to pass over at this place the 
somewhat complicated relations of the double and multiple embryos in 
late stages, since they cannot be made clear without numerous figures. 
I pass therefore to the double and multiple cleavage, the study of 
which brings out some interesting results, but it is necessary to give 
first a short sketch of the early normal cleavage, since HaTscHEK’s 
account is inadequate for the purpose. 
48 
