178 THE BIOLOGY OF STENTOR 



ment after deletion of parts, these tests show that the oral primor- 

 dium need not be complete or remain always continuous for the 

 full differentiation of its separate parts. 



It is equally clear that in these interrupted primordia there is a 

 strong tendency for the parts to rejoin. This association is perhaps 

 best shown by tandem grafts in which the ends of two separate 

 primordia were made approximate (unpublished). Two stentors 

 could be grafted together in homopolar telobiosis without any 

 disturbance of the lateral striping or injury to the primordia, which 

 were brought into alignment. In almost every case in which the 

 anlagen were in mid-stage development, they fused together as a 

 continuous membranellar band (c). A complete set of mouthparts 

 was formed only at the posterior end, although almost always at 

 the point of joining there appeared an accessory oral pouch (see 

 also Fig. 4 if). 



We have therefore learned much of how the oral primordium 

 develops under both normal and abnormal conditions, but how 

 the precise and elaborate feeding organelles are guided to their 

 perfection remains a mystery. 



because one section (x) which may have been pushed into 

 heteropolar orientation was not incorporated. 



B. After splitting, a: Stage-5 reorganizer with anlage sliced 

 through three times, b: Reorganization proceeds, and a well- 

 formed membranellar band is produced but no mouthparts. c: 

 Specimen regenerating because differentiation was incomplete. 



C. Joining of tandem primordia. a: Tandem graft of two 

 stage-3 regenerators cut just beyond the ends of the primordia. 

 b: Anlage join though not originally touching, development 

 continues, and a space clear of lateral striping (x) develops 

 alongside as in Fig. 4 if. c: Complete mouthparts formed only 

 at posterior end, only an extra oral pouch (y) being produced at 



mid-level. (After Tartar, 1957c, in part.) 



