EEPORT ON THE TETRACTINELLIDA. 
XXVll 
Epallax in general terms to a Placina detached from its seat, turned sideways np, and 
furnished with additional oscnles opening through the hypophare, one for each excurrent 
sinus. In Placina the roots of the excurrent evaginations are continued into the 
hypophare by trabeculae, these are present in Epallax, simulating the pillars of sub- 
dermal cavities. The addition of au ectosome and of secondary folds or invaginations 
in the walls of the main ones has already been mentioned. 
In other plate-like sponges {Pcecillastra, Astropeplus, among the Choristida, 
Azorica, Corallistes, among the Lithistida) the plan of the evaginations is not so clearly 
displayed as in Epallax, but that it is very much of the same nature appears from the 
fact that the oscules and pores are similarly distributed, and that the excurrent and 
incurrent canals run more or less transversely across the plate. In some massive 
sponges, which have originated in plate-like forms, such, e.g., as Pachastrella abyssi, it 
would appear as though the evagination proceeded on a similar plan. 
Canal-System in Spherical Sponges. — In small spherical sponges, such as Myihastra 
clavosa, and in young examples of Stelletta, the axes of the evaginations radiate towards 
the periphery ; as a consequence, from those excurrent canals which open nearest the 
margin of the oscule and are consequently the most superficial, secondary evaginations 
proceed radially towards the surface, while those which open nearer the centre of the 
oscule and run almost axially through the sponge, frequently expand at their distal ends 
into more or less concentric canals, from which again radial canals proceed towards the 
surface. In Stelletta phrissens the evaginations, at first radial, appear at a very early 
stage to curve round spirally, so as to acquire a more or less concentric arrangement 
(PI. XVI. fig. 19); subsequently, no doubt, radial canals proceed from them, but in 
the fully grovm sponge it is difficult to discover any definite arrangement. In most 
spherical sponges a general tendency towards a concentric and radiate arrangement 
of the canals is, however, observable, and the concentric arrangement is shown in 
an illustration given of part of a radial segment of Anihastra communis (PL XIII. 
fig. 8), where concentric excurrent and incurrent canals are shown alternating with 
each other. 
Excurrent imthout corresponding Incurrent Canals. — In many sponges in which the 
oscules are collected in a special area [Synops vosmaeri, PI. XXIIL), or in which numerous 
excurrent canals open into a common cloaca [Caminns spheroconia, PI. XXVII. ), the 
primary excurrent canals are without corresponding incurrent canals ; if uow we return 
to the Khagon we shall find that every incurrent canal derived from it must by the 
nature of the case involve the existence of a corresponding excurrent canal, but there is 
one excurrent canal that does not involve the existence of a corresponding incurrent 
canal, and this is the remains of the paragaster itself ; it would thus appear probable 
that each of the large excurrent canals in the case under consideration represents the 
remains of a paragastral cavity, produced by a process of budding from the margin of 
