476 
SEITARO GOTO 
SO that the coenosarc does not come in direct contact with the 
body of the hermit crab. It goes without saying that the endo- 
derm of the whole colony is continuous throughout. The outer 
surface of the colony is entirely covered over by an ectodermal 
layer, which is comparatively thick in the superficial furrows 
already mentioned and very thin on the ridges, so that the latter 
appear as if directly exposed. In this superficial layer of ectoderm 
are seen numerous endodermal tubes, from which the different 
forms of polyps take their rise. The entire colony may therefore 
be regarded as consisting of two sets of sponge-works, the chiti- 
nous and the coenosarcal. 
The colonies are, so far as I have observed, unisexual ; but her- 
maphrodite gonophores, containing both ova and spermatozoa, 
are occasionally found. A similar case is mentioned by Bunting 
in Hydractinia echinata.^^ 
The individuals of a colony are composed of the following forms: 
(a) gasterozooids,(6) dactylozooids, (c) blast ostyles, (figs. 5 and 6). 
The gasterozooids again occur in three apparently different 
forms. Two of these were noticed by Inaba and regarded by him 
as gasterozooids and blastostylesres, pectively; but a comparison 
of a large number of these polyps has convinced me that they 
are simply different conditions of one and the same form. By 
far the larger number of polyps are long and slender, the tentacles 
very numerous and arranged in several whorls, the hypostome 
very prominent, hemi-ellipsoidal in form and with a small mouth 
in the center (fig. 7). In sections the endoderm of the tentacles 
is seen to be separated from that of the gastric cavity by septa 
of supporting lamella. These gasterozooids are very numerous 
and are found all over the colony, on the general surface as well 
as on the spines (fig. 5). In most of the specimens they are espe- 
cially abundant on the large spines on the lower surface of the colony 
i.e., that side on which the colon}' would slide along when the 
hermit crab moves on. Sometimes they form such thick tufts 
on the spines that the latter are well nigh hidden from view. The 
"Loc. ciL: '94, pi. 9, fig. 7. 
