A STUDY IN POLARITY 73 
the conditions of the experiment the dominance of the head- 
forming material would assert itself, as it appears to do in the 
regeneration of the small triangular pieces, with the result that 
the piece would retain its artificially made form, the new head 
remaining where it was, the point of tissue at the opposite end 
serving for a tail, morphallaxis—which is pronounced in this 
species (Morgan, ’01, p. 48)—taking place in the old tissue to 
make the morphological long axis coincide with the artificially 
established geometrical long axis. No case was found in which 
the piece retained this form. The head swung round so as to 
come into line with the old axis, and at the posterior end appeared 
a tail, which in some cases was so curved that it actually pointed 
anteriorly, but eventually straightened out. If the piece was so 
small that no tail regenerated, enough material appeared along 
the posterior edge to make a perfectly symmetrical worm with 
its original polarity restored. In no case was the polarity per- 
manently changed, but the structural long axis of the regenerating 
worm did not coincide with its geometrical long axis until after 
regeneration was complete (fig. 14d). 
In the regeneration of all triangular pieces there was evidence 
to show that there was a similar tendency on the part of the 
piece to retain its original polarity. The prevention of its full 
expression was probably due chiefly to lack of new material. 
This tendency was indicated in the following ways: (1) by the 
production, in the early stages of regeneration, of a mass of new 
material along the posterior cut distinct from that along the 
anterior cut; (2) by the greater quantity of new material along 
the anterior cut; (3) by the formation of the head nearer the 
anterior end of the piece; (4) by the appearance of the more 
anterior eye first, even in those pieces in which the head was 
clearly at right angles to the old axis; and (5) by the fact that not 
a single specimen whose head did not come into line with the 
old axis ever regenerated a tail. The failure to produce a tail 
is very significant. If the polarity of a triangular piece with a 
head at right angles to the original long axis is such that the 
new head is truly at the anterior end, and the old uncut (right 
or left) margin is truly the posterior end, then either a portion 
