THE ANTHOZOA 55 



3, 3, and we get a form with six mesenteries whose sequence 

 corresponds to the sequence of the first six mesenteries in Actinia 

 equina. But we can go no further. If the mesenteries marked 



4, 4, in 3 were found to be developed before those marked 5, 5, 

 we should get an eight-rayed stage similar in all respects to the 

 Edwardsia stage in Actinia, except for the absence of muscle 

 banners. 



Leiopathes glaberrima has twelve mesenteries in the oral cone. 

 Below the level of the stomodaeum only six are present. Study 

 of serial sections shows that the mesenteries die out in the following 

 order : Firstly, those marked 6, 6, in Fig. XXVI. 5 ; secondly, 

 those marked 5, 5 ; thirdly, those marked 4, 4. It will be observed 

 that the additional pair of mesenteries is the first to disappear, 

 and that the pair which is presumably fourth in order of develop- 

 ment outstays the pair which was presumably developed fifth. 

 In the absence of further evidence it may be conjectured that the 

 first four pairs of mesenteries are formed in the Antipatharia in 

 the same sequence as in the case of Actinia, and that therefore 

 an Edwardsia stage of development may be assumed. It would 

 follow that the normal number of ten, characteristic of the 

 Antipatharia, is arrived at by the development of a single mesentery 

 in each sulco-lateral chamber ; and where twelve mesenteries are 

 present, as in Leiopathes, an additional mesentery is formed on 

 each side between the sulco-lateral and reproductive mesenteries, 

 the arrangement of the last two pairs differing from that in 

 Actinia equina. 



It may be concluded that the existing Zoantharia are derived 

 from a bilateral ancestral form which was provided with no more 

 than eight mesenteries. In this form there was probably no sulcus 

 and no sulculus, and muscle banners were absent. It was, therefore, 

 antecedent to the Edwardsia form, and probably enough was the 

 common ancestor of the Alcyonaria and Zoantharia. From this 

 parent form the Cerianthidea, the Zoanthidea, the Antipathidea, and 

 the Edwardsiidea diverged. From the Edwardsiidea may be derived 

 all the other recent Zoantharia. The Oractidae, Gonactinidae, and 

 Monaulidae appear to have diverged early from the Edwardsian 

 stem, which was continued into the Actinian series which, from its 

 disguised bilateral symmetry, may be called the Cryptoparamera. 

 This gave rise to two main branches: (1) Forms with a stony 

 skeleton, the Scleractineae, equivalent to the Madreporaria of 

 previous authors, and (2) the Malacactineae, equivalent to the 

 Actiniaria of previous authors. From the Malacactineae the 

 aberrant families of the Polyopidae, Sicyonidae, and Tealiidae were 

 derived. These relationships are expressed in the following 

 table : 



