256 BUSH 



in many of the species studied. In S. validus collected the latter part 

 of June, also in S. granulatus, the embryos taken from the opercu- 

 lum had the setae well developed, and each was partially concealed by 

 a conspicuous white patch, which under pressure split uito minute 

 short rods, quickly destroyed by acid. The exact significance of this 

 white patch has not been satisfactorily determined, owing to its being 

 now noted, apparently for the first time, in preserved animals. It is 

 a well-known fact that embryos developed in the tube of the parent, 

 on escaping, immediately on reaching a suitable host, begin to build 

 their tubes, so that it might be possible that these embryos, when the 

 specimens were collected, were about escaping from the pouch, and 

 that their growth or movements were not immediately interrupted and 

 the lime was formed which would have constructed their tubes. On 

 the other hand, the mass may cause the embryo to sink rapidly, en- 

 abling it more readily to find a host, as it must necessarily escape 

 from the brood-pouch into the open sea, at some distance from a 

 suitable object for attachment. Not until a series of animals lias been 

 properly prepared in sections, and a comparative study is made of 

 many species, will it be possible satisfactorily to determine their in- 

 ternal structure. It is found in comparing the published results of 

 the studies of isolated species that there are apparent contradictions, 

 and important points remain unexplained. 



Caullery and Mesnil, like Levinsen, found that the number of tho- 

 racic segments varied, sometimes increasing to 4 on one side ; com- 

 bining this fact with that of the development of the operculum on the 

 right or left side, they divided the genus Spirorbis into new subgenera 

 Paradexiospira, Dexiospira, Paralccospira, Lceospira and Rom- 

 anchella discarding all of Saint- Joseph's generic names, excepting 

 Leodora, which they thought might prove to be the same as the last, 

 stating that they were unable to follow that author in referring the 

 species to his six widely separated genera. 



As it is impossible always accurately to determine the number of 

 thoracic segments, often owing to the very poor condition of the ani- 

 mals in preservation, such a division, although of much interest, does 

 not appear of any more practical value in determining the relationship 

 of species than does the method adopted by Saint- Joseph. The analyt- 

 ical table of species given by these authors includes too many charac- 

 ters to be readily used. 



Saint- Joseph (1894), in his analytical table for the Serpulas, makes 

 the number and size of the teeth on the uncial plates of greatest im- 

 portance. He seems, however, to have erred in describing and figur- 



