156 BULLETIN OF THE BUREAU OF FISHERIES. 
afforded by his P. (Psammochela) elegans, has shown that the sponges belong in the 
Desmacidonide, although the chele have apparently been lost by many species. 
The species of this subfamily, except the Beaufort form, are all known to reach 
a large or, at any rate, a good sizer Tubular and flabelliform shapes and a massive 
or irregular form of body with processes are common. At least three species are known 
in an incrusting phase, P. (Sigmatella) corticata papillosa (Lendenfeld, 1889, p. 620), 
P. (Sigmatella) carcinophila on crabs (Lendenfeld, loc cit., p. 615), and P. (Desmacidon) 
psammodes (Hentschel, 1911, p. 322; Dendy, 1916, p. 126). Perhaps the Beaufort 
species will be found in some larger phase. ‘That it breeds in the thin, incrusting con- 
dition is no reason for believing that this is its final state. Mucrociona prolifera, for 
example, breeds in Beaufort Harbor, while a thin incrustation having a skeletal arrange- 
ment much like that of P: osburnensis (Wilson, 1911, Pl. I, fig. 5); but, while the great 
majority of individuals in the harbor do not get beyond an incrusting condition, much 
larger and more complicated phases are reached by some. If P. osburnensis reaches a 
large size, probably its skeleton becomes reticulate. At present it is the only recorded 
form, except P. solida Marshall, which lacks recognizable fibers, in which the skeleton 
is not reticulate, although in P. (Sigmatella) carcinophila Lendenfeld (1889, p. 615) the 
reticulum is confined to the basal portion of the sponge, while from it isolated vertical 
fibers pass to the surface. 
The complete loss of spongin in some forms, if it really occurs, is a remarkable 
fact, especially since the skeletal sand grains continue to be arranged by the sponge 
in bands. Perhaps the spongin is really not absent in these forms but only very scanty 
and transparent. Ridley (1884, p. 439) says that in P. fibrosa the skeletal “fibers are 
wholly composed of foreign bodies united by an almost colorless, not dense, substance,” 
and, as stated above, the spongin in P. osburnensis is easily overlooked. 
The species of the subfamily hitherto recorded are all from Australian waters. ‘Two 
of the forms are, however, thought by Lendenfeld (1889, pp. 613, 620) to occur elsewhere, 
P. (Sigmatella) australis var. tubaria at Nassau (Bahamas) and P. (Sigmatella) corticata 
var. papillosa on the English coast, African coast, in the Indian Ocean, and on the 
Florida coast. The last-named item of distribution rests on the identification of Hyatt’s 
Spongelia kirkii (Hyatt, 1877, p. 539) as a Phoriospongia (Sigmatella), but I can not 
find that Hyatt’s account justifies this step. 
The chele in P. osburnensis are described (see above) as ‘‘tridentate.’”’? Dendy 
describes (1916) in the same way the chele of his P. (Psammochela) elegans, the smaller 
of which (fig. 6c, c’) resembles that of P. osburnensis; and in P. jfibrosa Ridley (Ridley, 
loc cit.) the isochele are said to be ‘‘tridentate.” WLundbeck (1905, p. 4) criticizes the 
use of this term (category) as confusing, since it covers two different forms of spicule, 
chele arcuate and ancore. ‘This is undoubtedly so. On the other hand, it must be 
allowed that when the cheloid is very small, it is difficult to use the Levinsen-Lundbeck 
categories. For, assuming that the “tridentate” spicule is really either a well-defined 
chela arcuata or ancora with three teeth, and not some other form, the decision turns 
on whether the shaft has, in addition to the three teeth, ale or not. And this is not 
easy to determine with certainty when the spicule is very small. Hence it would seem 
allowable, even necessary, to continue to use in practice the term “tridentate” for 
certain small spicules, although it is confessedly somewhat vague. 
