566 THE WONDER OF LIFE 



provision. On a point like this it is interesting to get 

 precise facts, and Bordage notes that out of a hundred 

 Phasmids, nine died during moulting, twenty-two tore 

 themselves free with the loss of one or more legs, and only 

 sixty-two survived without any loss. In short, breakage 

 during moulting is a frequently recurrent casualty, pro- 

 vision for which would certainly be favoured by natural 

 selection. ' 



Another difficulty is presented by the regeneration of the 

 abdominal Umbs of hermit-crabs, which are normally pro- 

 tected by the Gastropod shell which the Crustacean borrows. 

 There are five of these — the first very rudimentary in the 

 males, but used for carrying the eggs in the females ; the 

 fourth and fifth used to fix the hermit-crab to the central 

 pillar of the borrowed shell. It is evident that there is 

 little hkeUhood of these Umbs being nipped off — extremely 

 httle in the case of the two hindmost pairs. But Professor 

 T. H. Morgan has shown that all the limbs of hermit-crabs 

 are capable of regeneration, though they do not all grow 

 again equally often, the anterior abdominal appendages 

 being less frequently renewed than the more exposed 

 thoracic limbs, for even these are not always restored after 

 loss. He therefore argues that there is here no relation 

 between frequency of loss and regenerative capacity — a 

 thesis quite counter to the idea of Lessona's law. Weis- 

 mann's general answer is that the regenerative capacity 

 shown by the hermit-crab's abdominal limbs may be a 

 persistent inheritance from ancestral forms which must 

 have had exposed tails. Moreover, one would Hke to 

 inquire particularly into the hfe of hermit-crabs to find 

 out whether there are not now — ^in the combats, in the 

 flitting from one house to another, and particularly in the 



