CLASSIFICATION OF ACTINIARIA 511 
somidae,! both ‘Stichodactyline’. One advantage of placing 
these with the corals is that they are not like the remaining 
true Stichodactylines, which apart from them form a har- 
monious group (see p. 533). 
Two further points arise: (i) are there any corals with the 
Stichodactyline arrangement of tentacles? and (ii) to which 
Madreporarian families do our forms most nearly approach ? 
With regard to the first it does not much matter, for a Sticho- 
dactyline condition of tentacles could arise as a convergence 
anywhere, and has done so among the Ceriantharia. As to 
the second it is for a coral expert to suggest, and pending 
further investigation the families should simply go under 
Madreporaria without closer allocation. 
A vertical section of one of the animals in question is 
shown in Text-fig. 8. It is a cup-shaped form in which the 
tentacles have become reduced to mere knobs. 
What are the points which make these forms like corals ? 
A general statement about them might be made as follows: 
They secrete no horny or lmy skeleton. They may be 
quite solitary, or quite gregarious, sometimes living in sheets 
or carpets. Frequently they reproduce by fission, and often 
compound individuals with several mouths, or individuals 
connected by a basal coenosare are found. The base is adherent. 
The body is without verrucae, variable in form and consistency. 
More than one tentacle connects with at least the older endo- 
coels. The tentacles may be simple, or capitate (cf. Caryo- 
phyllia and others among corals), or branched; or small 
and wart-like, or even reduced to so little as to be invisible 
externally. There are no siphonoglyphes (or rarely ?). The 
mesenterial filaments have no ciliated tracts. Sphincters are 
feeble or absent. Sting-cells of a size characteristic of Madre- 
poraria, but not of Actinians in general, are usually found 
somewhere in the body. ‘There are usually a good many 
1 The Discosomidae as referred to in this connexion means the family 
taken in Carlgren’s sense, 1900, p. 58, and not in the wider sense of some 
authors—including only the genera Discosoma, Paradiscosoma, Actinotryx, 
Rhodactis, Orinia, and Ricordea. 
Mm2 
