680 RAYMOND PEARL. 



posterior pieces. By varj'ing- the position of the cuts, seg- 

 ments of the body of various lengths may be obtained. All 

 of these, which are about 14- mm. in length, will usually right 

 themselves by as close an approximation to the typical spiral 

 reaction as is possible under the circumstances. The side of 

 the body which is lowest can be seen to elongate in these 

 very short pieces, and just enough of a twist is found to 

 bring the ventral surface of one corner of the anterior end 

 into contact with the bottom. Of course, no complete spiral 

 can be found in such short pieces. Their reaction time is 

 \erj slow. 



Next, experiments were tried with the pieces resulting 

 from splitting longitudinally anterior halves of worms in the 

 middle line. These pieces had the form shown in Fig. 38. 

 Evidently such pieces have only a half of the mechanism 

 necessarj" for the performance of the spiral righting reaction, 



Fig. 38. — Operation diagram (see lext). 



according to the vicAv given above, and therefore should not 

 be able to give the typical response. They have one com- 

 plete side which may elongate, but they have no other side 

 to keep the middle line straight, and so make the elongation 

 effective in forming a spiral. Such pieces, when placed with the 

 dorsal surface down, reacted immediately by bending strongly 

 towards the cut side, i. e., so that the concavity was on the 

 cut side. This was kept up for a time, the animal S(piiruiing 

 about violently, but it was fiually replaced by another reaction. 

 The ventral longitudinal muscles contracted strongly, and 

 raised the anterior end of the piece well up from the bottom 

 (shown in side view in Fig. 39, a). After a strong raising con- 

 traction the piece would extend and settle back again. Then 

 after a time the raising was repeated, and it soon became 

 noticeable that the piece was rising higher each time and 



