542 
H. V. WILSON 
and one may actually see that these are closed through the creep- 
ing of the niost superficial layer of the dermal membrane (newly 
formed epidermis) across the pore, i.e., over the aperture of the 
pore canal, thus giving rise to a pore membrane. ^ As to the open- 
ing of the pores after closure by the pore membranes, I have no 
actual observations, but it is obvious that the pore must reappear 
as a perforation in the membrane, which then recedes towards 
the margin of the pore canal. 
The question maybe asked, is the pore canal a permanent struc- 
ture, or does it too close up? Since the pore canal perforates 
the dermal membrane its closure obviously could only be brought 
about through an extension of the mesenchyme of that mem- 
brane. In Stylo tella I always find the pore canals distinct even 
when the pores are completely closed. Hence the pore canals 
must be regarded as structures that are permanent in ordinary 
conditions of the sponge. My observations on Reniera and Lisso- 
den doryx (vide infra) nevertheless show that the pore canal itself 
may be partially or completely obliterated in monaxonid sponges. 
Another question may be asked before we leave this matter 
of the pore and its closure. Is there any one nucleus that is 
especially associated with a pore membrane? An examination of 
figs. 1, 2 and 3 shows there is no such nucleus. The pore membrane 
may spread to a considerable distance over the pore canal before 
any of the epidermal nuclei enter it (fig. 3), and when the pore is 
completely closed (fig. 2) the membrane may shov/ a nucleus 
somewhere near its centre or again one or two nuclei near or at 
its margin. The epidermal nuclei are irregularly^ distributed, 
in some spots close together, in others farther apart. This is 
well shown in fig. 2. These nuclei moreover are all alike. The 
facts would seem to indicate that the epidermal sheet of proto- 
plasm spreads of its own initiative over the pore canal, and that 
the nearest nucleus or nuclei are simply drawn into it. In Reni- 
era on the other hand there is always one nucleus at the margin 
2 In the healing up of such a section a considerable rearrangement of the canal 
system certainly' takes place. A new osculum is established, and this apparently 
may develop at any point on the surface of the piece. Pore canals lead through the 
newly formed dermal membrane into what were originally main efferent canals. 
