FIELDS AND GRADIENTS 279 



duced is normally controlled in relation to the polarity of the 

 fragment. 



It is well known that the bodies of Coelenterates, Planarians and 

 Annelids are polarised. A differential of some sort exists between 

 different levels, so that in cut pieces the end nearest the apical 

 region usually regenerates a new apical region (see p. 271), while 

 the other cut surface usually regenerates a posterior end. (The ex- 

 ceptions to this statement are treated later (p. 296), and it will be 

 found that they can all be formulated in terms of another general 

 rule.) 



In these cases, the cut piece contains a portion of the general 

 gradient-system of the entire individual, which piece by the fact 

 of cutting becomes isolated as a separate field-system in which the 

 factors determining polarity are still graded from apical to basal end. 



This rule, however, needs some amplification. In Planaria, if a 

 transverse fragment is divided in the middle line into two halves, 

 both will form a head at the anterior end. But if it is divided into 

 three pieces, the central of which includes the main longitudinal 

 nerve-trunks, the two outer pieces will, if below a certain length, 

 form heads either obliquely at the medio-anterior corner, or at 

 right angles to the original main polarity, on the median cut surface : 

 the percentage of medianly directed heads increases with decrease 

 in the length of the piece.^ It would appear probable that in this 

 case the new heads are determined in relation to the cut ends of the 

 lateral nerves, which come off transversely from the main longi- 

 tudinal nerve-trunks, and are of the same essential structure, con- 

 taining cells as well as fibres. There is of course also a secondary 

 medio-lateral susceptibility gradient in the intact animal, and this is 

 presumably correlated both with the course of the lateral nerves and 

 with the determination of medianly directed heads. 



(ii) The second general rule is that the origin of polarity is to be 

 sought in external factors. Either the polarity of the regenerating 

 fragment is taken over from that of the whole organism, which is 

 derived from that of the embryo, which in turn is due to factors 

 external to the ^gg (pp. 36, 60); or the regenerating fragment ac- 

 quires a new polarity under the influence of the external agencies 

 acting upon it after its isolation. In some cases, although the frag- 



^ Beyer and Child, 1930. 



