306 



ELLINOR HELENE BEHRE. 



ends are as nearly as possible at the same body level as the 

 anterior ends of the long pieces give practically 100 per cent, 

 headless forms. Moreover, we find th'at when a piece such as 



a-b has remained for a few hours as the 

 anterior portion of the long piece a-x 

 and is then isolated, it usually gives rise 

 to a head. It is evident that at the 

 anterior end of the piece conditions de- 

 termining head formation have been so 

 fixedly established during these few hours 

 that they are not essentially altered by 

 the later isolation of the short piece, 

 although when the short piece is iso- 

 lated at once head formation is inhib- 

 ited. 



The experimental procedure is as fol- 

 lows: A lot (A) of 25 or 50 long pieces 

 (a-x, Fig. 6) are cut and allowed to reg- 

 ulate, and head-frequency noted. A 

 second lot (B), short pieces (a-b, Fig. 

 6) , are likewise cut and allowed to regu- 

 late. These serve as controls. Several 

 hundred long pieces (a-x, Fig. 6) are 

 now cut, and from the anterior regions 

 of these, lots (C, D, E, etc.) of short 

 pieces (ab, Fig. 6) are cut at different 

 intervals following the first section. In 

 this way the head forming region is left 

 for a certain length of time as the ante- 

 rior end of the long piece before it be- 

 comes the anterior end of the short 



piece. Table IX. gives the results of such a series of experi- 

 ments. 



Evidently the process of head determination then begins almost 



immediately after section and the conditions existing in the piece 



during the first few hours after section must constitute the 



most important factor determining the character of the result. 



The controls A give 100 per cent, heads, whereas the controls 



