566 T. A. STEPHENSON 
ancestors of Athenaria and Endocoelactaria. This would 
allow for the retention of the musculature, or of traces of it, 
here and there, with a dominating loss of it along all the lines. 
6. SUMMARY. 
It is difficult to make a concise summary of a paper covering 
a good many inter-related discussions, but the following is 
an attempt to give some of the main points with reasonable 
brevity. 
1. There is difficulty in defining specificity among Actiniaria, 
as in other lowly and plastic animals. Among British forms 
species are well enough marked on the whole, if studied alive 
so that colour and habit can be taken into account. When 
preserved, however, too little is known of possible range of 
specific variation in anatomy for much to be done. Foreign 
forms are so often known in death only that species are some- 
what in chaos and there is little firm ground. Experience 
leads one to the view that among these low and plastic forms 
a species may have its peculiarities of organic constitution 
at an early stage of the development of their expression, such 
expression having affected colour-scheme and general facies 
of the living animal but not necessarily to any extent the 
internal anatomy which can be studied in preserved specimens. 
Much work needs doing by way of studying all forms alive, 
and of killing and preserving numerous individuals which 
belong certainly to the same species, in different ways, and 
studying them so as to reveal effect of reagents, age, state 
of contraction or distension, locality, reproductive maturity, 
and so on, on the anatomy. When a better knowledge of the 
limits of specific variation is gained (and they will be much 
wider in some species than in others) a revision of species 
might be attempted. Especially the value or otherwise of 
measurements of nematocysts as specific characters should 
be looked into. 
2. Although species are in a poor way, genera and families 
are on the whole much easier to understand and make use of, and 
here there are enough data to start a methodical classification 
