THE RANGE OF DOMINANCE 165 



position of the new head. In a piece achd. Fig. 94, 

 the head develops, as shown in Fig. 96, on the apical cut 

 surface, but in a shorter piece aegi, Fig. 94, the head is 

 likely to appear at an angle to the apical and median cut 

 surfaces, as in Fig. 97. This condition results when the 

 metaboHc rate of the cells on the median cut surface is as 

 high as that of the cells on the apical cut surface, so that 

 both take an equal part in giving rise to the new head. 

 In pieces like afi, Fig. 94, the head often develops nearly 

 or quite in the direction of the transverse axis (Fig. 98). 

 In such pieces there is little difference in metabolic rate 

 between apical and basal cut surfaces, and the cuts are 

 not sufficiently oblique so that the higher level in the 

 major gradient of the lateral as compared with the 

 median region of the cut surface overbalances its lower 

 level in the transverse gradient. Consequently the 

 median regions of both cut surfaces represent the region 

 of highest rate or irritability in such a piece and therefore 

 become the head-forming region. For these and many 

 other experimental modifications of the position of the 

 head in reconstitution no satisfactory general basis of 

 interpretation has heretofore been discovered, but I 

 know of no case which cannot be very simply accounted 

 for in terms of axial metabolic gradients. 



In the bilaterally symmetrical liverwort MarcJiautui 

 (Fig. 23, p. 78), the gradient- relations are apparently very 

 similar to those in Planaria. In these plants practically 

 every cell of the body is capable of giving rise to a new 

 plant, but in pieces without the growing tip new growing 

 tips originate in definite relations to the axes, and their 

 presence inhibits the formation of others. In general, 

 on transverse cut surfaces new individuals arise, like 



