SIPHON-WORMS AS MESSMATES 183 



mensalism as exists between the Aspidosiphon and the coral is 

 an interesting one. Commensalism is usually looked upon as 

 conferring some mutual advantage on the contracting parties, 

 and one or the other of these usually seeks the other out. But 

 in the case in question the mutual advantage is far to seek. It 

 can hardly help the coral to have a large proportion of its base 

 burrowed by a spacious canal, but the fact that the Gephyrean 

 pulls the otherwise immovable coral about may be, and probably 

 is, an advantage to the Ccelenterate. On the other hand, the 

 Gephyrean gains protection and a home more spacious than the 

 Gastropod shell affords. The Aspidosiphon can hardly find or 

 attract the larval coral to come to rest on its borrowed shell, and 

 it is unlikely that the larva is especially on the outlook for such 

 shells as are inhabited by Gephyrea. It seems more probable 

 that the Aspidosiphon may select for its home a Mollusc shell 

 which already bears a young coral, but the whole matter seems 

 to demand more careful study. It is certainly remarkable that 

 three distinct genera of coral, each with but one species, should 

 be inhabited by three distinct species of Aspidosiphon, and that 

 neither commensal has hitherto been found apart from the other." 

 In some cases, at least, there would appear to be a third partner 

 in the concern, for numbers of a kind of minute bivalve mollusc 

 were found closely attached to the outside of the siphon-worms. 

 Regarding them Shipley remarks: " These were so closely 

 adpressed to the skin of the Aspidosiphon as to indent it, appear- 

 ing as little pearls set in a matrix. The advantage they obtained 

 by taking up such a position is not very evident, but there they 

 were, and as far as one could judge they were, until Professor 

 Herdman dropped them into his collecting-jar, flourishingo 89 



