216 PERCY SLADEN TRUST EXPEDITION 



a very thin external trabecular layer (o.t.l.), followed immediately by the chamber layer, 

 which occupies by far the greater part of the thickness of the sponge wall. Then, on the 

 inside of the chamber layer, comes a faii-ly thick subgastral or inner trabecular layer 

 (i.t.l.), bounded internally by a thin gastral membrane {g.m.). 



The wide inhalant canals run inwards approximately at right angles from the dermal 

 surface and extend throughout the greater part of the thickness of the sponge wall. They 

 branch moi'e or less and interdigitate with similar exhalant canals, also branched and also 

 extending through the greater part of the thickness of the wall, which open by wide 

 apertures on the gastral surface (whether the gastral membrane ever extends over these 

 apertures as a delicate net I am unable to say definitely, but I think it highly probable 

 that it sometimes does). Owing to the obliquity and irregularity of their direction a 

 transverse section of the sponge wall (fig. 17) shows the inhalant and exhalant canals 

 sometimes cut across and sometimes cut lengthwise. 



The flagellate chambers (f.c.) are very small for a hexactinellid sponge, only just 

 about as large as those of a typical Leucandra, such as L. phillipensis [Dendy 1893], and 

 much smaller than those of L. australiensis [Dendy 1893]. Indeed the whole canal system 

 very closely resembles that of a Leucandra, except that the trabecular layers are represented 

 in the latter by more continuous mesogloeal tissue. The large exhalant and inhalant canals 

 may be regarded as formed by folding of the chamber-bearing layer, and it is probable that 

 in reality there is only a single, much folded layer of chambers. Such appears to be the 

 case at any rate in many places {e.g. in part of the section represented in fig. 17), but 

 usually the arrangement has become greatly confused. 



The chambers come close up to the surfaces of both inhalant and exhalant canals, but 

 the actual surface in both is probably formed by a very thin, net-like trabecular layer. 

 Most of the chambers communicate with the large exhalant canals, but some of them 

 open into the irregular spaces in the subgastral trabecular layer (fig. 18). 



The chambers themselves (figs. 17, 18, f.c.) are oval or thimble-shaped. I have 

 measured them up to 0'2 mm. in length, but usually they appear a good deal shorter than 

 this. The collared cells are very small and indistinct and it is impossible to make out any 

 satisfactory histological details. Such histology as I have been able to observe is repre- 

 sented in fig. 18. 



In addition to the numerous small Anthozoa attached to the outer surface, the sponge 

 wall is penetrated in various directions by the branching stolons of a hydroid colony. The 

 hydranths (fig. 17, hyd.) are elongatedly club-shaped, with few tentacles (two or three?) 

 arranged in a single whorl springing from a short distance beneath the mouth. They are 

 only sparsely scattered at long intervals on the hydrorhiza and appear to be capable of 

 protrusion sometimes from the outer and sometimes from the inner surface of the sponge 

 wall, though all now in a state of complete retraction. There is no distinct horny perisarc, 

 though sometimes a very thin layer can be discerned vs^hich may represent the last vestige 

 of such a structure. The hydranths occupy definite tubular cavities which run inwards 

 from the surface of the sponge and appear to be lined by a continuation of the dermal or 

 gastral membrane as the case may be. This hydroid is probably closely related to AmpJii- 

 hrachium euplectellcB, described by Schulze [1880] as occurring in the soft tissues of 



