FORM-REGULATION IN CERIANTHUS. 135 



marginal tentacles I mm.; third piece with tentacle-buds just 

 visible. 



Controls : All well-filled, with expanded discs ; marginal ten- 

 tacles in all cases 3-5 mm.; labial tentacles appearing. 



The experimental pieces were now allowed to close and be- 

 come distended. 



September 2^. - - 12 days after operation. Experimental pieces. 

 All distended, oral ends expanded but less so than in controls ; 

 all with marginal tentacles 23 mm.; one piece with labial ten- 

 tacles just visible. 



Controls : Marginal tentacles 57 mm.; labial tentacles I mm. 



These two series are sufficient to serve as illustrations of the 

 methods. In other series in which the same methods were em- 

 ployed the results were essentially the same, so that it is unnec- 

 essary to give the data for others in detail. . 



In a few series the new tissue at the oral end was punctured 

 with a lancet as soon as it formed and then this end kept open 

 by daily insertion of the lancet between the inrolled margins. 

 Special care was taken in these operations not to injure the 

 tentacle-forming region which lies a short distance from the cut 

 margin (cf. the first paper of this series '030), but to destroy 

 only the delicate membrane and so allow communication between 

 the enteron and exterior. It is unnecessary to give the details 

 of this series since the results obtained are essentially similar to 

 those of other series. In every case the diameter of the oral end 

 remained less than that of the body, the appearance of the ten- 

 tacles was delayed and the regeneration retarded as long as the 

 pieces were kept open. Tentacle-buds when formed on open 

 pieces were blunt and rounded, and after a time became whitish 

 in color and opaque. When allowed to close the oral end began 

 to spread in the typical manner and the tentacles grew as rapidly 

 as in the controls. 



DELAYED CLOSURE IN PIECES OF CERTAIN FORMS. 



In my third paper (Child, '04^) some of the varieties of inroll- 

 ing after section were described and it was shown that in pieces 

 of certain forms, especially those which were split longitudinally, 

 inrolling often occured in such manner as to delay or even pre- 



