CLASSIFICATION OF AOTINIARIA 517 
to have more than one family, it seems wise to retain the old 
plan to the extent of having a group to cover them, the principle 
of which is good. This group must be labelled by Carlgren’s 
name Athenaria, with the Edwardsiids of course excluded. 
The rank of this group will be discussed in a later section, 
but here we may consider the general characters justifying it. 
The Athenaria appear to be the representatives of those 
forms which, being the outcome of a muscular Haleampa- 
stage, have retained more similarity to their ancestor than 
the majority of other forms, and have kept to a more or less 
burrowing life. There is variation in size; the predominating 
shape is vermiform, the relation of length to diameter varying 
in different cases and different states of expansion, diameter 
sometimes considerable. Text-fig. 7 shows the contrast between 
some of these and one of the ordinary adherent anemones with 
short wide form. In these Athenaria the aboral end is not 
a definite base, but a rounded physa, which is sometimes able, 
however, to adhere to small objects. There is little or no 
sphincter. Often there are cinclides. The number of tentacles 
is usually small, and at most does not pass about forty. The 
number of mesenteries is similarly limited, and either these 
all have the grade of macrocnemes, or else there is a division 
into macro- and microcnemes—and in Peachia the state of 
affairs is intermediate. The mesenterial filaments have ciliated 
tracts. 
The above may be taken as a sort of definition of the 
burrowers or Ilyanthids. The subdivision of the group remains 
to be discussed. 
Of course, some of the forms formerly included here have 
long since been removed, others more recently—the Cerian- 
tharia and Edwardsiaria. Forms with no base but with 
acontia are little known, but seem to fit in quite well with the 
Phelliidae (see Part I, p. 524), though possibly a new family 
may later on be needed for them. Carlgren has suggested 
a Haleampactidae, but it is here treated as coming under 
Phelliidae. Andvackiidae is not yet established. For Hal- 
campactis see Part I, pp. 499, 509, 525. 
