No. 1.] ACTINIARIA OF THE BAHAMAS. 29 
tion, the ordinary ectoderm cells surrounding them being on the 
other hand easily separable, a fact which has been observed by 
von Heider (77) as characteristic of the similar cells of the 
verruce of Heliactis. The histological structure of verrucz will 
be discussed more fully later on under Phymanthus cructfer, 
where they are well developed and suitable for investigation. 
The sphincter muscle is strong and of the circumscribed type 
(Pl. III., Fig. 9). The mesogloeal processes supporting the en- 
dodermal muscles become somewhat stronger as they approach 
the region of the circular muscle, and then suddenly become 
much elongated and branched, forming a ridge almost circular 
in cross section, projecting into the body cavity. There is no 
distinct pedicle from which the processes take a common origin, 
as in other Bunodidz which have been examined, but most of 
the processes arise independently from the mesogloea layer, 
only those towards the inner part of the thickening arising from 
a common lamella, or process, which, however, is thin, and not 
at all pedicle-like. ‘Yellow cells” are everywhere abundant in 
the endoderm. 
The margin is somewhat raised, forming a collar, a bare space 
intervening between it and the bases of the outermost tentacles. 
It bears the uppermost of the verruce, which, as already stated, 
are not lobed and do not present any difference of structure 
from those found on the column. The tentacles are cylindrical 
and decidedly entacmzeous, the inner ones measuring I.7cm. in 
length, and the outer 0.8 cm, They are 48 in number, being 
arranged in four cycles, thus: 6, 6, 12, 24. 
The mesenteries are arranged in five cycles, only those of 
the first cycle being complete (Pl. III., Fig. 10), and of these 
the directives are united to the stomodzeum to a greater extent 
than the others, the lips of the gonidial grooves being prolonged 
downwards to form languettes with which the directives are 
united. The mesenteries of the second cycle (II.) are nearly 
if not quite as well developed as those of the first cycle, but are 
all imperfect ; those of the fourth cycle (IV.) have well-developed 
muscle bands, but no mesenterial filaments; while the members 
of the fifth cycle (V.) are very diminutive, not projecting be- 
yond the surface of the endoderm, and in the specimen exam- 
ined, were not present in certain interseptal spaces where they 
should have occurred, the cycle being thus incomplete. The 
