DYNAMICS OF MORPHOGENESIS 63 



It is of course evident that in these cases the length of the 

 piece is in some way a factor in head-formation, but the ques- 

 tion how it affects head-formation will be answered later. At 

 present we are concerned merely with the fact that the cells at 

 the levels, a, b, etc., when at the anterior ends of short pieces, 

 do not give rise to heads, but when at the anterior ends of long 

 pieces do give rise to heads. In order to find approximately when 

 head-formation is determined, it is only necessary to prepare a 

 large series of the long pieces, such as af or cf in figure 1, and at 

 different intervals after section to remove short pieces such as 

 ab or cd from the anterior ends of a certain number. We know 

 that such short pieces when completely isolated at once do not 

 develop heads. If heads appear at their anterior ends after 

 they have remained a certain length of time as the anterior re- 

 gions of long pieces and have then been isolated, it is evident 

 that the formation of heads must have been determined while 

 they were still a part of the long piece. The records of experi- 

 ments will show the character of results obtained. 



Series ^58. From well fed worms 18 mm. long, about 150 

 pieces, including the region cj in figure 1, were prepared and 

 from these the following five lots were taken. 



Lot I. From the anterior ends of twenty-five of the long pieces, 

 short pieces like cd were removed 15 to 30 minutes after the long 

 pieces were prepared. 



Lot II. From another twenty-five of the long pieces, similar 

 short pieces were cut 3 to 4 hours after the long pieces were pre- 

 pared. 



Lot III. Twenty-five similar short pieces 7 to 8 hours after 

 first operation. 



Lot IV. Twenty-five short pieces 18 hours after first operation. 



Lot V. Twenty-five short pieces 24 hours after first operation. 



These five lots of short pieces were allowed to undergo recon- 

 stitution at a temperature of 20 to 22°C. and when this was com- 

 plete the character of each piece was recorded; the results appear 

 in table 1. 



The five types distinguished have been described in earlier 

 papers (Child 'lib, 'lie). The normal form possesses a head 



