PROCESS OF REGENERATION 131 



riorly unless it includes a certain minimum number of body seg- 

 ments. In Dero, this number is three, i.e., the anterior piece 

 must be eight segments long before it will regenerate posteriorly. 

 Anterior pieces shorter than this almost invariably die within a 

 few hours after cutting. In only one case did such a piece sur- 

 vive; it consisted of six segments, and did not regenerate poste- 

 riorly, forming, in fact, a 'tailless' head. Similarly in short 

 pieces, the last piece, containing the posterior end will not re- 

 generate anteriorly unless it contains 12-15 segments; pieces 

 shorter than this invariably die. The reason for the failure of 

 these anterior and posterior ends to regenerate has been dis- 

 cussed in the preceding section. 



As regards the rest of the body, pieces three or four segments 

 long from any part will regenerate complete anterior and poste- 

 rior ends. Is there then in Dero no difference in regenerative 

 capacity along this axis? An axial difference does exist and 

 reveals itself in the following ways; anterior pieces regenerate 

 faster than posterior ones; the regenerated head is larger on an- 

 terior pieces than on posterior pieces; the number of segments 

 regenerated posteriorly is greater in pieces from the anterior 

 regions of the body, and decreases along the axis; smaller pieces 

 from the anterior regions will regenerate completely than from the 

 posterior regions. In regard to this last point, pieces from the 

 anterior half of the body, containing only one or two segments 

 will regenerate into complete worms; while from the posterior 

 half of the body, pieces must be at least three or four segments 

 long before this can occur. These facts indicate an axial dif- 

 ference, which is, however, relatively slight as compared with 

 other forms. 



Biaxial heads arise in short pieces of Dero; two such were 

 observed, and undoubtedly more could be readily produced. 

 One of the observed pieces consisted of about three segments; 

 it regenerated a complete head at each end (fig. 19). The other 

 piece consisted of little more than one segment; in this case, 

 the anterior head was complete, but the posterior one regener- 

 ated onlj^ three setigerous segments, instead of the normal four. 

 This piece was cut in two, but both portions died. An adequate 



