1 20 MCMURRICH, [VOL. I. 



With these results I turned to the related genus Rhodactis 

 and examined three specimens of R. Sancti-Thomae. In my 

 original description ('89) of this form I described one pair of 

 directives as being present, and somewhat hastily assumed the 

 presence of the second pair, though I could not be sure of its 

 existence. A further examination showed that in one specimen 

 but a single pair of directives was present and that pair was situ- 

 ated some distance from the end of the long axis of the stoma- 

 todasum ; in a second specimen one pair was again present, 

 also to one side of the axis of the stomatodaeum, but at the 

 opposite end of the stomatodaeum there was some irregularity 

 in the arrangement of the mesenteries, and I could not be cer- 

 tain that a second pair of directives was not present there ; 

 and finally in a third specimen I found no directives, though 

 here again, owing to certain mesenteries not being cut exactly 

 transversely, it is impossible to affirm their absolute non- 

 existence. 



These results, especially those obtained from Ricordea, seemed 

 to me to have considerable bearing on the validity of the pro- 

 posed order Holactiniae. Recently two other observations of 

 forms lacking directives have been recorded. Haddon and Duer- 

 den ('96) failed to find directive mesenteries in Cystiactis tuber- 

 culosa, and Kwietniewski ('97), finding none in Thalassianthus 

 senckenbergianus and T. aster, has established for these forms 

 an order Thalassianthae. We have, then, four different genera 

 belonging to as many distinct families, according to the usual 

 classification, and yet presenting the essential peculiarity upon 

 which the new order is based. 'If the order is a valid one, 

 then we must separate Gyractis from the Bunodidae, which it 

 seems to resemble most nearly. Ricordea must be removed 

 from its association with Rhodactis, to which it is closely 

 related even in its histological peculiarities ; Cystiactis must 

 be regarded as belonging to an entirely different group from 

 Alicia ; and Thalassianthus must be placed in a new order 

 even though it should prove necessary to separate it from 

 Actineria. But this is not all. We must place one specimen 

 of Ricordea florida in the new order and leave another among 

 the Hexactiniae, and we must similarly place one specimen of 



