132 LIBBIE H. HYMAN 



explanation of the formation of biaxial heads has been given by 

 Child ('13 b). The original gradient must be largely eliminated, 

 and this is accomplished by cutting short pieces, — which include 

 such a small portion of the original gradient that there will be 

 little difference in rate of metabolism between the two ends of 

 the piece, — or by depressing the rate of metabolism of the 

 pieces (low temperature, cyanide, anaesthetics, etc.)- If, then, 

 the growing cells at the two cut surfaces of these pieces in which 

 the gradient has been eliminated attain a sufficiently high rate 

 of metabolism, each will develop an independent part, that is to 

 say, a head. If, on the contrary, the rate of metabolism of the 

 old piece remains higher than that of the growing cells, it will 

 dominate over them, and they will be able to give rise to sub- 

 ordinate structures only, namely, tails. Child has been able to 



Fig. 19 Biaxial heads in Derojimosa. 



demonstrate the correctness of this explanation as follows. 

 Planaria dorotocephala seldom gives rise to biaxial heads, but 

 if short pieces are put into depressing agents to eliminate the 

 gradient, and then allowed to regenerate in water, a consider- 

 able percentage of biaxial heads is obtained. As far as I am aware, 

 biaxial heads have not previously been observed in annelids 

 although a reversal of polarity of a similar character has been 

 produced in lumbricids by grafting short pieces in the reverse 

 position upon the anterior end of long pieces (Hazen. '99 and 

 RuttlofT, '08). 



Regeneration in Dero is nearly always complete and 'normal.' 

 In anterior regeneration, the only variation from the normal is 

 the occasional appearance of heads containing fewer or more 

 than the five typical head segments. Such heads may be des- 

 ignated as hypomeric, when the number of regenerated segments 

 is less than normal, or hypermeric, when it is more. Hypo- 



