514 PEOF. GILBERT C. BOURNE ON FIVE 



notice of Boveri (5) and McMurrich (18, 19, 20) ; and these two authors, by 

 ingenious and convincing arguments, showed that the various plans of 

 mesenterial sequence exhibited by the Cerianthidse, Zoanthidpe^ Antipatharia, 

 Madreporaria, and Actiniidea could be derived from tlie Edwardsian type, 

 and claimed that the existing Edwardsidpe are to be regarded as the ancestral 

 type from which all the groups enumerated have been evolved. But in 1895 

 Faurot (10) showed that, in addition to the large macromesenteries with 

 conspicuous muscle-banners, a variable number of microniesenteries, confined 

 to the upper region of the capitulum. and so small as to have escaped previous 

 notice, exist in the two species E. heautempsii and E. adenensls. In the 

 former species the positions of the micromesenteries could be compared with 

 the arrangement seen in Ilalcampa, and it was therefore no longer possible 

 to regard Edxcardsia as an ancestral type. The late Prof. Ed. van Beneden 

 discussed the relationships of the Edwardsida? in his splendid memoir on 

 the Antho/oa of the Plankton Expedition (3)j and argued that they are 

 Hexactinarians simplified by progressive regression of the mesenteries of 

 the second cycle and of the fifth and sixth mesenterial couples of the first 

 cycle. He added : " Au surplus il faut renoncer a voir dans les Edwardsies 

 le tvpe anceslral d'ou seraicnt issus les Hexactiniaires." Van Beneden's 

 view of the relationship of the Edwardsidae has been generally adopted, and 

 for the last eighteen years they have l)een classified among the Actiniaria, 

 and as early as 1898 ("arlgren (8) classified them together with the 

 Halcampoinor[)hidre in his subtribe Athenaria, which corresponds very 

 nearly to the family Ilyanthidfe of (losse. 



It is not possible for me, at the present juncture, to enter into a discussion 

 of the verv debateal)le questions of the classification of the Anthozoa. It is 

 sufficient for me to state that I accejit the principle of Carlgren's (9) division 

 of the Anthozoa into the three subclasses Hexacorallia, Octocorallia, and 

 Dodecacoi-allia, but cannot reconcile myself to the use of the names 

 proposed. (.'arlgren's Hexacorallia includes the ( "erianthidse and the 

 Antipatharia, and is equivalent to van Beneden's group ('eriantipatharia. 

 The classification seems a natural one, but as the Actinians and (Jorals 

 have lono- been known as Hexaetinire and it is now desired to insist on 

 the separation ot' the (VM-ianthidre from these forms, (!arlgi-en's choice 

 of a luinie seems to me unfortunate and liable to cause confusion. The 

 C^eriautipatharia of van BeneJen are a subdivision of the " Scyphactiniaria," 

 a subclass including the Scypbomedusre and Rugosa. In my opinion there 

 are valid reasons for excluding the Scyphomedusre from the Anthozoa, and 

 the position of the so-called "Rugose" Corals is still uncertain. I am 

 inclined to use van Beneden's name of an order, Ceriantipatharia, for the 

 name of a subclass, identical with (Jalgren's Hexacorallia and equivalent 

 in rank to the Octactiniaria and Zoanthactiniaria. The last name is 

 van Beneden's, and is equivalent to Carlgren's Dodecacorallia. The sub- 



