CLASSIFICATION OF ACTINIARIA 571 
lose their acontia and may lose mobility, and develop stiff 
or thick body-walls, their metabolism slowing down and spare 
energy sometimes being used up in the production of knobs 
and crests and solid horn-like tentacular swellings. This is 
a tendency towards fixity of character and possibly thence 
towards ultimate extinction. It is interesting to note that 
some of the above-mentioned Endocoelactaria have reached 
a similar state, although along an entirely different line. 
The other main line, Endomyaria, went in for endodermal 
sphincters if any, and their special stinging organs are never 
acontia, but they often have acrorhagi and other things. 
Some of them develop vesicular blisters and compound acro- 
rhagi which may reach wonderful complexity of structure ; 
in others the tentacles increase in number and sometimes they, 
not the outgrowths of the body, become complex, at their 
finest with a frill-like effect. These forms, whether it be body 
or tentacles that complexify, are more especially found in 
the warmer seas, and here the tendency to fixity of character 
does not seem much indicated. Along both lines various 
forms halted by the way, of course. 
This idea of the evolution of the group may be helped out 
by the diagram printed below. 
A more detailed outline of the history of Mesomyaria has 
been worked out, and will be found in Part I, p. 495, &e. ; 
a corresponding one for Endemyaria is given in this part, 
p. 563, &e. 
5. Apart from the above considerations, it has been the object 
of the paper to revise and re-define all the families and genera, 
sorting them out in such a way as to make them as homo- 
geneous as possible, and to represent their relationships naturally, 
with the idea of getting the definitions as precise as is feasible 
in order to facilitate identification. It has the advantage of 
collecting all the definitions together, but at the same time 
is not meant to be an exhaustive compilation as regards 
species-lists and so on. Only a minimum of synonymy is 
included, and insufficiently known forms are left alone. The 
classification worked out is, admittedly, complicated rather 
