138 
Fig. 6c. Psammaplysilla purpurea (Carter). Photo- 
micrograph showing branching of fibres and structure 
of the flagellate chambers. One of the ellipsoid masses 
of dense tissue of unknown function is shown at lower 
center. 
open structure and the cortical region is narrow 
(75-100 fx) [e.g., Sta. 60; Dendrilla verongi- 
formis USNM 23104; Druinella purpmea (Bur- 
ton, 1934)]. In those specimens where the 
over-all texture is tough and the branches are 
thinner, the cortex may be up to 350/x thick 
[e.g., Thorectopsamma xana USNM 22994; 
Psammaplysilla kelleri USNM 21241; specimen 
from Sta. 92]. 
endosome: Earlier workers (Lendenfeld, 
1889; de Laubenfels, 1954) have mentioned 
that the flagellate chambers tend to be concen- 
trated around the excurrent canals and are lack- 
ing from large areas of the sponge. After 
examining several specimens, including the 
sponges which de Laubenfels described, it is 
apparent that this is not strictly the case. Cham- 
bers are aggregated around the excurrent canals 
but in all specimens groups of chambers are 
PACIFIC SCIENCE, Vol. XIX, April 1965 
more or less evenly dispersed throughout the 
endosome. In compact specimens (such as the 
type of D. tyroeis ) the relatively great extent 
of the cortex gives the impression that the 
chambers are restricted in occurrence. The 
structure of the flagellate chambers is in no 
way remarkable. They do not have long fine 
prosodal canals, but long aphodal canals do 
occur and vary in length (up to 100/x ) and 
diameter from specimen to specimen. The 
chambers themselves are small and slightly 
ellipsoidal, 23 X 12 to 46 X 30^. A constant 
feature of the species is the occurrence through- 
out the endosome of ellipsoidal masses of rela- 
tively deep-staining cells. These are figured by 
Wilson (1925: pi. 44) but no comment is of- 
fered as to their possible function. They are 
always set off from the surrounding mesen- 
chyme by a concentrated zone of spongin A 
which often extends inward as bands between 
the cells. It is possible that these structures rep- 
resent developmental stages of the fibres. 
DISCUSSION: In order to elucidate the sys- 
tematic affinities of Psammaplysilla purpurea 
(Carter) it was necessary to investigate the 
sponges assigned to the genus Druinella Len- 
denfeld for, superficially at least, the two gen- 
era are closely similar. 
Lendenfeld (1889) described Druinella to 
receive D. rotunda from Australia. This sponge 
was notable for the lobose knotty fibres and for 
the possession of long prosodal and aphodal 
canals. In all features except the possession of 
these canals, D. rotunda is inseparable from 
Aplysina purpurea Carter. Burton (1934) was 
the first worker to definitely associate Aplysina 
purpurea Carter with the genus Druinella. Wil- 
son (1925) had tentatively suggested that 
Psammaplysilla kelleri and P. arabica belong to 
the same genus as A. purpurea, and likened all 
of these species to Druinella and Thymosia 
(Topsent). 
No author since Lendenfeld’s time has seen 
Druinella rotunda nor had similar canal struc- 
ture been described from any other sponge 4 
until de Laubenfels (1948) referred Cacospon- 
gia camera de Laubenfels to Druinella. He 
4 A similar pattern of short canals is well known 
from the work of Schulze (1878) on the Aplysinidae. 
