THE WONDER OF LIFE 567 



moultings — some very good present reasons for the retention 

 of the regenerative capacity in the abdominal appendages. 



Imperfect Regenerations. — Another objection to the 

 theory which interprets the distribution or occurrence of 

 the regenerative capacity as adaptive is found in those 

 strange and highly interesting cases where the re-growth 

 takes place, but not according to pattern. As Spallanzani 

 showed in 1768 and T. H. Morgan in 1899, a decapitated 

 earthworm may grow a second tail (as shown by the dis- 

 position of the excretory tubules or nephridia) instead of 

 replacing its lost head. But this only serves to show that 

 the regenerative process is hable to go wrong at times just 

 as the embryonic development does. The fact that a 

 headless creature is sometimes born does not affect the 

 general conclusion that development is a regulated and 

 harmonious process. 



Werner points out that when a hzard re-grows its tail, it 

 does not always adhere to the pattern. When the scales 

 are comparatively simple, the regeneration is almost perfect, 

 but when the scales are complex and there is much orna- 

 mentation, the regenerated tail is simpler than the one 

 that was lost ; it tends to he an ancestor's type of tail. Hence 

 the wit has suggested that to find out a hzard's pedigree, 

 you have only to pull off its tail. Perhaps a truer way of 

 stating the case, however, is that the regenerated tail is 

 nearer the embryonic type, which is not surprising if 

 regeneration be due to a local persistence and re-awakening 

 of embryonic growing powers. 



There seems to be a widespread tendency towards the 

 reproduction of a simpler or ancestral form, or, in some cases, 

 of a simpler and more embryonic form. Thus in cock- 

 roaches and walking-stick insects (Phasmids), which have 



