THE COMMON FRESH-WATER SPONGE. 
5 
which they first appear ; but in what way successive additions 
are made to them after their dependence on their parent cell 
ceases has yet to be ascertained. As these spicules are formed 
they are removed by the sponge-cells to the place where they 
are ultimately needed, as Mr. Carter observes, ‘‘ with as much 
instinct as that which characterises the arrangement of the bits 
of stick in an ant-hill.” 
When the growth of the sponge-mass has made some progress, 
the formation of a distinct investing membrane (fig. 17a), out 
of what was the flat transparent border, becomes obvious. This 
membrane is gradually detached from the central ovi-bearing 
cells, either by the shrinking of the latter or by the protrusion 
of bundles of large smooth spicula (fig. 176), which force it 
outwards, leaving, here and there, open spaces between the 
membrane and the central cell- mass. The membrane contains 
numerous small spiniferous spicules (fig. 18a) scattered through 
its substance. There are also spread over the membrane nume- 
rous leaf-like nucleated cells (fig. 186), resembling a compressed 
layer of multifid leaves, but which are polymorphic, constantly 
changing their forms, as well as their positions, and contain- 
ing one or more contractile vesicles. Amongst these cells are 
numerous apertures (fig. I7c), which, according to Carter, have 
the power of opening and closing (fig. 18c, d, e ) ; but which at 
their maximum rarely exceed inch in diameter. It appears 
to be through these apertures that the food passes into the cavity 
between the investing membrane and the parenchyma. 
The parenchyma (fig. 17 d) consists of a cellular mass per- 
meated by canals (fig. 17c,/), respecting the ultimate structure of 
which there is much reason for doubt. The canals are regarded 
by Carter as of two classes; the one a series of afferent canals 
fig. 17/) opening into the parenchyma from the cavity of the 
investing membrane ; whilst the other he considers to be a 
distinct efferent set (fig. 17c), not communicating with the 
afferent ones, but terminating in a single tube (fig. 17,^^), which 
projects from the periphery of the organism and constitutes an 
excretory outlet. The spaces between these canals are occupied 
by a gelatinous substance full of peculiar cells or sarcoids and 
large smooth spicules. Some of the cells are ciliated (fig. 6c-d), 
each protoplasmic atom having a single cilum, whilst others 
are unciliated (16a-6), but all are polymorphic. The exact 
arrangement of these cells in the parenchymatous mass is uncer- 
tain. At one time Carter thought he had satisfactory evidence 
that they were grouped in the interior of a parent cell, in the 
centre of which their undulating cilia were continually vibrating ; 
but he subsequently ascertained the incorrectness of this ex- 
planation, and inclined to the belief that they were arranged 
round the exterior of primary cells, the cilia being external to 
