20 Records of the Indian Museum. [VOy, (ay 
be bound together by any horny substance. At certain points, 
probably where the aperture fora new papilla is about to be 
excavated, a stout chitinoid covering is secreted over the sponge 
and the macroscleres adopt a convergent arrangement and are 
densely massed together. At such places the nodular microscleres 
are sometimes present in large numbers and form a layer several 
spicules thick over the protecting mass. The papillae are protect- 
ed by a dense ring of vertical macroscleres fortified with chitinoid 
substance and arranged concentrically in several or many circles 
with the heads resting at the base of the very short vertical tubule. 
Within this ring, in the case of exhalent orifices, there is an 
arrangement of convergent macroscleres with their tips meeting 
almost horizontally and their heads set in a broad spiral of about 
I} turns. Presumably the tips can be separated in the living sponge 
by rotation of the heads. The whole arrangement is strikingly 
reminiscent of the diaphragm in the stage of a compound micro- 
scope. The smooth amphiasters are scattered in the flesh of the 
tubules and chambers. Neither they nor the nodular amphiasters 
_play any part in the protection of the external papillae. 
Gemmules.—I have found several gemmules in the specimen 
examined. They are spherical masses of cells of the usual type, 
but have no horny protective membrane. Each is about 0°374 
mm. in diameter. Each gemmule occupies a separate chamber 
which it fills completely. There is a slender strand of cells con- 
necting it with the active part of the sponge. 
Type.—No. Z.E.V. 6430/7, Ind. Mus, in spirit: in a dead Gas- 
tropod shell. 
Locality.—Off Ceylon: 703 fathoms (R.I.M.S. ‘ Investigator’). 
The form of the nodular microscleres is characteristic, in 
particular in the large size of the lateral and terminal bosses; other- 
wise they resemble those of T. socialis, Carter. The species is evi- 
dently related to T. aymata, which, however, has the spicules of 
this type with the bosses perfectly smooth as well as relatively 
smaller. 
A noteworthy feature of T. investigatoris is its power of secret- 
ing a horny covering for its growing-points when they come in con- 
tact with foreign bodies. I hope to show in a subsequent paper 
that it protects itself in this manner against aggression on the part 
of a sponge of the genus Coppatias that is parasitic in its burrows. 
At most of the points at which new galleries are being formed in 
the shell no such covering can be detected, but at some, probably 
where the sponge is in contact with the outer layers of the shell, 
and is about to form a new exhalent or inhalent papilla, there is a 
thick one. It is only where such a covering occurs that the nodu- 
lar amphioxi are found, and if the covering is very thick, a num- 
ber of these spicules can usually be discovered in which the spines 
on the nodules seem to be completely worn away and the nodules 
themselves even to some extent destroyed. Such spicules lie in or 
on the outer or distal part of the covering. ‘These facts would 
