CLASSIFICATION OF ACTINIARIA 539 
Order 1. ZOANTHARTIA. 
Order 2. ACTINIARIA. 
Order 3. MADREPORARIA. 
To this, now, Edwardsiaria would have to be added. In 
this paper Bourne’s division will be used. It is when we come 
to the subdivision of the sub-order Actiniaria that the main 
divergence of opinion begins. 
Carlgren divides into 
Tribe 1. PROTANTHEAE, 
Tribe 2. NYNANTHEAE. 
Another division in use is 
Tribe 1. ACTINIINAE. 
Tribe 2. STICHODACTYLINAE. 
In the following paragraphs I shall indicate the lines of 
grouping which I wish to suggest, giving an outline only. 
Further reasons, filling in this outline, will be found in various 
parts of the paper, especially under the foregoing sections 
dealing with sets of forms individually, and in the later evolu- 
tionary discussions. 
Much has been said about Carlgren’s division into Protantheae 
and Nynantheae, and it has been rejected by some workers, 
at any rate, in the sense in which Carlgren uses it. It is based 
mainly upon the presence or absence of ectodermal muscle 
and a nerve-layer in the ectoderm of body-wall and actino- 
pharynx ; and this, as has been suggested before, is probably 
a universal ancestral character surviving in more or less primi- 
tive forms and, otherwise, in sporadic cases. I cannot accept 
it as a good basis of distinction in itself, although it helps 
to show relationships, in some cases, when taken with other 
things. In this attitude I believe [ am in agreement with 
Haddon (1898, p. 411), Duerden (1900, p. 187, and 1902), 
McMurrich (1904) and Bourne. At the same time I accept 
decidedly Carlgren’s Protantheae, but in a different and much 
more restricted sense. I have tried to show that Carlgren’s 
Protostichodactylines (a sub-tribe of his Protantheae) (and 
also the Discosomidae) are corals (see p. 510), and this restricts 
his Protantheae to Gonactiniidae, Ptychodactidae, and the 
