PKETENDING TO BE A WORM 3oo 



occupies the tenement, but puts on the guise of the original 

 tenant. ' The whole appearance of the animal is singularly 

 like that of a small Serpula, the operculum and branchiae 

 being mimicked by the expanded posterior appendages 

 and telson, and the hairs representing the parapodial set^e, 

 while the small ness of the limbs and the vermiform shape 

 of the body, aid in perfecting the resemblance.' It is a 

 sufficiently curious instance of mimicry that the head of 

 one animal should be imitated by the tail of another. The 

 creature's habit of entering the tube head foremost is 

 quite in accord with its diminutive antennae and abnor- 

 mally small front limbs, and that when not in quest of 

 food it should have the branchial appendages of the tail 

 near the aperture of ts cave is an obviously convenient 

 arrangement. Probably the British Anthura gracilis may 

 use a similar shelter, since it is undoubtedly dredged up 

 in company with shells and stones on which the tubes of 

 Serpul^e are abundant. Many of the Edriophthalma 

 occupy tubes which they themselves construct. This is 

 apparently the case with Typhlotanais brachyurus^ Bed- 

 dard, the tube of which is ' shaped somewhat like the shell 

 of Bentalium.'' The Indian Amphipod, named Goncholestes 

 dentalii by Dr. G. M. Giles, forms its tube within a Dan- 

 talium shell. It seems, however, certainly to belong to the 

 genus Cerajms. other species of which have the tubes free. 



Family 2. — Chiathiidce. 



Though the Gnathiid^e have been long contented with 

 a single genus, the peculiarities of that genus are so great 

 that at one time they were supposed to suffice for the 

 institution of two separate families. There are striking 

 differences that distinguish the sexes from one another and 

 the adults from the young. Characters common to all are 

 as follows. Both pairs of antennae are short, with the 

 flagella clearly marked off from the peduncles; the man- 

 dibles are abnormal ; the first and last segments of the 

 peraBon are very small, and the last per^eopods are want- 

 ing ; an embryonic character permanently retained as in 

 the genera Hyssnra and Cruregens of the preceding family. 



