374 RECORDS. 



four cells, and each develops, not into an abortive monster, but 

 into a perfectly formed though dwarf larva (Fig. i, B). Thus 

 it is possible to produce from a single Qgg from one to four per- 

 fect animals ; and in case of certain species (hydromedusae) it is 

 theoretically possible by a similar method to produce from a 

 single Qgg as many as eight or even sixteen perfect dwarfs. 

 Again, in some of these cases, for instance in the nemertine, the 

 undivided Qgg may be cut to pieces in any planes taken at 

 random ; yet every piece, if of sufficient size, may upon fertili- 

 zation develop as if it were a whole tgg and produce a perfect 

 dwarf Here is an astounding contrast to the results of our first 

 experiment. What becomes of our theories of prelocalization 

 here, and what becomes of our mechanical theory of develop- 

 ment, if we hold such a theory ? Neither the cells nor the 

 regions of the tgg seem to have any predestination such as is 

 shown in the molluscan Qgg. It is the essence of a machine or 

 automaton that its operation is due to its structural configura- 

 tion. Impair or destroy that configuration and the action 

 ceases. But from these eggs we may take away any of the 

 parts, or the whole may be cut to pieces, yet there is no impair- 

 ment of action, but only a readjustment to form smaller sys- 

 tems like the original whole. The ^gg, therefore, says the 

 vitalist, can not be an automaton and its development is inex- 

 plicable upon a mechanical theory. 



Such is the paradoxical result to which a superficial compar- 

 ison of these two cases leads us — a kind of embryological anti- 

 nomy, as it were, which at first sight may seem to take away 

 all hope of finding law or order in these phenomena. I will 

 undertake to show you speedily that the apparent contradiction 

 is easily explicable. I have placed the two cases side by side 

 because each seems to demonstrate the truth of one side of an 

 ancient embryological controversy ; and we shall presently find 

 reason for the conclusion that each of the opponents, like the two 

 knights and the shield, have recognized but a part of the truth. 



The probable explanation of the difference of the behavior 

 between the eggs of Doitaliinn and of Ampliioxus is a very 

 simple one. When we closely study eggs of this type we find 



