REPOET ON THE TETEACTINELLIDA. 
cxlv 
In one genus of the Geodiidse, viz., Erylus, which belongs to the subfamily Eiylina, 
the sterrasters, in addition to presenting a different form to those of the rest of the 
fa mil y, differ also in their mode of union : the fibrillated inocytes which bind them 
together are attached to their surface, but do not appear to take a direct course from one 
sclere to the other ; so far as I can make out they wrap round them instead. 
On the inner face of the sterrastral layer is a layer of fibrous tissue in which the 
fusiform cells are tangentially arranged ; this, which differs in thickness in different 
species, will be spoken of as the “ inner fibrous layer ” ; it and the sterrastral layer 
together are probably homologous with the fibrous layer of the Stellettid cortex. 
On the outer face of the sterrastral layer, between it and the outer epithelium, is a 
layer of tissue, which differs in thickness and character with the species, occasionally it is 
collenchymatous, but usually cystenchymatous ; it will be termed the “ ectochrote.” A 
layer of microscleres (somal) invariably occurs in the ectochrote immediately below the 
outer epithelium, and these spicules remain to indicate it in those cases where it is so 
much reduced that but for them it would naturally be set down as absent. In one case, 
Cydonium glariosus, it contains numerous embedded grains of sand. 
The existence of the sterrastral layer gives great definiteness to the drones, which 
traverse it usually as wide, simple, more or less cylindrical canals, closed at the inner 
end by a very evident sphincter ; at the outer end they may either open directly to the 
exterior by a simple large pore (uniporal chones), or break ujr into numerous ramifying 
canals, or extend into wide subdermal sinuses, within the ectochrote, and then open to 
the exterior by many pores with a sieve-like arrangement (cribriporal chones). 
The excurrent canal-system opens to the exterior in very various ways, in some cases 
by chones not differing in character from those of the incurrent system, so that, as well as 
incurrent, we may have excurrent uniporal or cribriporal chones ; in other cases different 
arrangements prevail, thus in Pachymatisma the oscule is the opening of a large chone 
which leads through the sphincter into a subcortical chamber, in which several large 
excurrent canals open by non-sphinctrate apertures, this structure may be regarded as an 
overgrown uniporal chone ; in Erylus and Caminus the oscule is the opening of a large 
cloaca which receives the excurrent canals by non-sphinctrate openings, this may be 
regarded as an overgrown uniporal chone without a sphincter, and in connection with 
this we may notice that the incurrent chones in these genera do not appear to be 
sphinctrate either ; finally, in Geodia the oscule leads into a cloaca, into which numerous 
large excurrent canals open, each provided with a special sphincter, this structure we 
may suppose has been produced by the overgrowth of the centripetal end of a uniporal 
chone and the multiplication of its sphincters. It must not be supposed, however, that 
the cloacal structure of Geodia has been actually produced in this way, in all probability 
it arises in the manner explained on p. xxxi et seq. 
It is on the characters of the excurrent and incurrent openings taken together that 
(ZOOL. CHALL. EXP. PART LXIII. 1888.) ElT t 
