| No. 396.] HYDROMEDUSA, GONIONEMUS VERTENS. 951 
the process until the entire cut edges are drawn together. 
There is no evidence that parts widely separated attract each 
other across the intervening region, as in cases of cytoplasm 
described by Roux. The cytoplasm is confined to those parts 
in contact, or at any rate so near together that they may be 
connected by protoplasmic processes. 
I have said enough to show that the process by which the 
piece closes in is a complex one, in which several factors take 
part. It is towards the discovery of these simpler processes 
that take place during regeneration that we can at present, 
I think, most profitably direct our attention. For if, as ap- 
pears to be the case, many components enter into the process 
that we call regeneration, we can only hope to understand the 
phenomena as a whole when we have resolved it into the imme- 
diate simpler factors of which it is made up. Then no doubt 
we shall find out that we can in turn resolve these factors into 
simpler ones, and our analysis will be carried a step farther. 
As long as we interest ourselves with the facts and factors 
of regeneration the work is not likely to come to a standstill, 
but when we leave the analytical method and attempt to con- 
struct injudicious theories that make the pretense of explain- 
ing a complicated process without attempting to resolve the 
process itself into its factors, then progress stops. Such, I 
believe, to be the case in the attempt to explain the process of 
regeneration by a theory of preformed imaginary germs. A 
theory of this kind is only a pretense; imagination takes the 
place of verifiable hypothesis, and the process that we set out 
to study is explained by saying that there are “ germs ” present 
that have been set aside to bring about the result ! 
