144 INDIVIDUALITY IN ORGANISMS 



or nearly spherical in form. In most cases, however, 

 there is no difficulty as regards this point, because the 

 longest diameter of the pieces coincides in direction with 

 the original axis. A comparison of the direction of the 

 new axis wliich arises after return to water with that of 

 the original axis shows that the former is at right angles 

 with the latter. The new hydranth develops without 

 relation to either of the cut ends from the uppermost 

 region of the piece a,s it lies in the aquarium, and this 

 region was originally its lateral surface. In these cases 

 the alcohol not only inhibits the increase in metabolic 

 rate in relation to the terminal cut surfaces, which 

 determines the development of hydranths at the two 

 ends, but decreases the rate throughout the piece. In 

 this way it obliterates the original gradient and domi- 

 nance to such a degree that when the metabolic rate 

 rises again on return to water the original axial relations 

 do not reappear, but a new gradient and a new dominance 

 arise in relation to the external conditions to which the 

 piece is subjected, and the axis of the new individual 

 coincides in direction with the new gradient. In all 

 cases, so far as my experiments go, the new hydranth 

 arises from the uppermost part of the piece, no matter 

 what region of the piece in its original condition this 

 part represents. 



When short pieces, which have already produced 

 biaxial hydranths (Fig. 75), are used for this experi- 

 ment, the changes are very similar to those described 

 for longer pieces. In alcohol the tentacles and the 

 apical regions of the two hydranths die and disintegrate, 

 but the more basal portions gradually lose their hydranth 

 structure and the pieces become small rounded masses 



