1905.] BEARING ACTINIANS IN THEIR CLAWS. 509 



reach, passing them towards the middle of the disc, from which, 

 however, they are abstracted by the ambulatory limbs of the crab. 

 Enemies to the crab, too large to be held by the tentacles of the 

 polyps, may nevertheless be warned ofi" by the stinging-cells of 

 the anemone emitted on irritation. A careful consideration of 

 all the circumstances justifies the view that the crab wall secure 

 much of its food through the activity of the anemones, and, 

 further, that the latter will exercise a protective influence upon 

 the crab against larger enemies. The advantages to the actinian 

 appear largely negative. As Mtibius suggests, the movements of 

 the crab will serve to bring the actinian into the neighbourhood 

 of more prey, but its chances of ultimately appropriating to 

 itself much of this seem very small. The feeding experiraents 

 demonstrated very clearly that it is only rarely that the actinians 

 succeed in ingesting their food ere it is withdrawn by the crab. 

 In the case of the actinians Sagartia and Adatnsia, commensal 

 with hermit crabs, it is usually considered that the polyps secure 

 fragments of the food torn up by the masticatory appendages and 

 slipping away, but it is not likely that this occurs with Melia. 

 Independently of the actinians, the crab can only obtain such 

 food as may be lying upon the sea-floor and incidentally come 

 upon the maxillipeds and the ambulatory limbs. 



The acquisition of such a peculiar commensal habit on the part 

 of two wholly distinct types of crabs, Melia and Polydectus, cor- 

 related, in the case of the former at least, with a diminutive size 

 and partial loss of activity on the part of the chelipeds, does not 

 admit of ready explanation. Among the activities of other 

 Crustacea there appear to be no examples which help us to 

 understand how such behaviour and structural peculiarities have 

 become established — no simpler or intermediate stages which 

 suggest the lines along which the evolution has taken place. 

 In the well-known instances of masking-crabs {Stenorhynchus, 

 Dromia) we have the tearing away of suitable objects, such as 

 zoophytes, algse, and sponges, which are then afiixecl to the shell ; 

 but the instinctive processes involved therein are less complex 

 than in the cases under consideration. In the latter the ordinary 

 aggressive and tearing functions of the chelipeds are replaced by 

 those of merely holding a living example of another group of 

 organisms. Even the seizu.re by a crab of an anemone and the 

 affixation of it upon a gastropod shell, as in the well-known 

 hermit crabs Pagurus, and the actinians Sagartia parasitica and 

 Adamsia jxcUiata, involves much less of a departure from the 

 usual activities of Crustacea. 



As in so many morphological and physiological phenomena in 

 nature where intermediate stages are not forthcoming, it is 

 difficult to see how such an instinct cou.ld have been acquired 

 or evolved by slow degrees. For instance, while holding the 

 actinians the crab could not at the same time employ its claws 

 for the usual purpose of seizing and conveying food to its mouth. 

 One is constrained to think of mutation as a possible explanation 



