310 AUGUST WE ISM ANN [april 



and gave rise to the new mucous membrane. These " primitive cells 

 of the epithelium of the mid-gut in the imago," as Eengel 1 calls 

 them, thus form in this case the distinctly demonstrable regeneration- 

 apparatus of the epithelial wall, and we may assume the existence of 

 a similar regeneration-apparatus in the eye of the newt. 



The results of an investigation bearing specially on the question of 

 the adaptive nature of regeneration have been recently published by 

 T. H. Morgan, 2 and these results are undoubtedly of value, although I 

 find myself unable to agree with his interpretation of them. He 

 selected the hermit-crab {Pagurus) as the subject of his investigations, 

 and this was certainly a very suitable choice, since the experimenter has 

 here at his disposal a series of appendages which, on the one hand, are 

 very different in their biological value, and on the other are exposed to 

 the danger of loss in widely varying degrees. Examination of freshly- 

 caught animals showed that, out of a hundred hermit-crabs, nine lacked 

 one of the three anterior legs, while none were without the small fourth 

 and fifth, which probably serve to fix and support the animal within 

 its gasteropod-shell. This result is only what might be expected when 

 we bear in mind that the hermit-crab withdraws into its dwelling- 

 place with great rapidity when danger threatens, and that the fourth and 

 fifth legs must be the first to reach the shelter of the shell. In any 

 case, these limbs are unlikely ever to be the aim of an attack, which is 

 probably always directed against the three larger and more exposed 

 limbs in front of them. As the posterior part of the body remains 

 entirely within the shell, the abdominal appendages are perfectly 

 protected, and could be injured only during a change of dwelling, or 

 if an attempt is made to draw the animal forcibly out of its shelter. 

 Accordingly Morgan found that only one out of a hundred specimens 

 lacked the second or third abdominal appendage, and there is still 

 the possibility that this defect was congenital, and not due to injury. 

 The problem these experiments were intended to solve was this : Is the 

 power of regeneration graduated according to the probability of loss ? 

 Are those limbs most easily and most frequently regenerated which 

 are most frequently injured ? and are those which are never injured, 

 perhaps never regenerated at all ? 



Experiments in cutting off the various limbs showed that they 

 were all capable of regeneration, though they did not all grow again 

 equally often, the anterior abdominal appendages renewing themselves 

 less frequently than the thoracic limbs, though even these did not 

 become renewed in every instance. I need not go into further details ; 

 it is sufficient, as far as the main question is concerned, to know that 

 all the limbs possess the power of regeneration, those most liable to 

 injury and those naturally well protected alike. The relative biological 

 importance of the different limbs in Pagurus, too, does not seem to be 



1 Zeitschr. wiss. Zool. 1896, Bd. lxxii. 

 2 "Regeneration and liability to injury," Zool. Ball. Boston, 1898, vol. i. No. 6. 



