474 
SEITARO GOTO 
is lined by a subcontinuous thin layer of chitin which stains 
deeply with haematoxylin. In favorable places the trabeculse 
are seen to consist of superposed lamellae around an axial substance. 
The spines, which form a characteristic feature of this species, 
are very numerous and of various sizes and shapes from simple, 
short, rounded conical projections to large, stout branching, antler, 
like spines nearly 2 cm. in height and 1 cm. or more between the 
tips of the outermost branches. The inner structure is essentially 
the same as in the body of the chitinous shell, with the difference 
that in the spines the meshes are largely elongated and are arranged 
with their long axes parallel to the length of the spines, so that in 
a longitudinal section, such as is reproduced in fig. 4, it is common 
to see long tubular cavities, especially in the axial parts. The 
formation of the chitinous trabeculse from concentric layers around 
an axial substance is especially clear in sections of the spines, 
the axial substance being sometimes very conspicuous by its 
opacity. Stechow mentions in cross-sections of the spines '^zwei 
deutlich geschlossene Rohre mit etwas dickerer Wandung," 
which he considers as possibly the first formed part of the spine, 
but according to my observations they are constant neither in 
number nor occurrence. 
As to the mutual relation of the chitinous skeleton and the soft 
parts, it may be said that it is in general similar to what obtains 
in Solanderia;^2 that is to say, each mesh is filled with ectoderm 
and endoderm, the latter forming, in sections, a well stained core 
with a small central lumen, surrounded by tall, somewhat vacuo- 
lated cells of the ectoderm which is in direct contact with the sur- 
rounding chitin. Where two meshes communicate with each 
other the ectoderm is seen to be directly continuous. Colcutt^^ 
says that in Hydractinia echinata the meshes of the deeper part 
of a colony are filled with degenerating coenosarcal masses; but 
in the present species the deepest meshes are filled with coenp- 
sarc which is in no way different from that of the upper part. 
It has already been mentioned that the meshes of the deepest 
. layer are closed by a thin, deeply staining chitinous membrane, 
" Ibid: '09, p. 22. 
12 Goto: '97 
13 Colcutt: '98, p. 85. 
