MADREPORARIAN CORALS 39 



order are placed. In other respects it approaches Para- 

 c\'athiis in structure, having well-marked costae and a 

 nodular columella. The special point of interest, however, 

 is that the coral always lives in association with a small 

 worm (Aspidosiphon) belonging to the Order of the Gephyrea, 

 and this is indicated in the dried coral by a small smooth 

 round hole at the side of the base through which in life the 

 worm protrudes.^ 



These corals seem to occur always on sandy bottoms, 

 and have no disc of attachment or other means of fi.xing 

 themselves to rocks, large shells, or other objects of sufficient 

 weight to resist the flow of water. 



The association with the worm is an ingenious arrange- 

 ment to prevent the coral being overturned and smothered 

 in the sand, and to ensure the maintenance of an upright 

 position in which the tentacles of its polyp can catch the 

 floating and drifting organisms on which it feeds, for the 

 Gephyrean worm feeds in the sand and thus acts as a kind 

 of muscular root always ready to drag the coral upright 

 again if it loses its balance. 



The origin of the association can be seen in very young 

 individuals or by making vertical sections of a full-grown 

 specimen. The young worm begins life by sheltering in a 

 small Gasteropod shell {e.g. Cerithium) like a Hermit Crab. 

 The coral larva settles on the outside of the shell by an 

 ordinary base of attachment, and as the coral grows and the 

 base extends it surrounds the shell, leaving only a hole 

 through which the worm can protrude. Growth of the 

 coral does not stop when the shell is completely surrounded, 

 but continues in all directions to form a smooth rounded 

 base perforated on one side by the worm tunnel in the 

 coral substance, which can be traced from the outside to 

 the mouth of the shell. The size and shape of the shell 

 on which the coral larva starts life vary considerably and 

 cause many variations in the subsequent shape of the 

 adult coral. This has led to the splitting of the genus 

 into a large number of quite unnatural species, and subse- 

 quently to the amalgamation of these species into one or 



^ As in Heteropsammia, Fig. 32, p. 79. 



