RE GENERA TION. 2 O I 



extremely complicated, and we might as well fall back at once 

 on the Bonnet-Weismann theory of preformation. 

 • There is another experiment on planarians that has a direct 

 bearing on Bonnet's hypothesis. If a planarian be almost 

 entirely split in two, leaving the halves connected only at 

 the anterior end (Fig. 4), two new heads may develop at the 

 most anterior end of the cut edges (Fig. 4). Van Duyne, 

 who first carried out an experiment of this sort, found two 

 heads developing, and he interpreted their development as due 

 to a process of heteromorphosis. I have repeated this experi- 

 ment a number of times with the same results, but I think 

 there is a simpler and more obvious way to account for the 

 development of the new heads. They 

 appear at the sides of each half, as 

 they would do were a long piece cut 

 from the side of the body ; but in the 

 latter case the result is not due to 

 heteromorphosis. In the former case 

 the two new heads are, after their for- 

 mation, prevented from being carried 

 forward by the presence of the old 

 head. This interpretation is in har- ^^^ ^ 



mony with the results of several other 



experiments. The bearing of this experiment on our present 

 examination is obvious. Two new heads develop, although the 

 old head is present. If the development of the new heads is 

 due to the presence of head-forming substances, as Bonnet 

 supposed, how could they develop as long as the old head is 

 present to use up these substances t The objection might not 

 apply with as much force if the transportation theory did not 

 include the using up of head-forming substances in the old 



head. 



Other examples might be cited, but those given above will 

 suffice, I think, to show the improbability of the stuff-trans- 

 portation theory, or, at least, the results show that it cannot be 

 universal in its application to the phenomena of regeneration. 

 The assumption of head-forming and tail-forming stuffs is too 

 general to explain the results. We have, however, clear evi- 



