160 BULLETIN OF THE BUREAU OF FISHERIES. 
together and just covered with spongin; the spicules longitudinal in the fiber or slightly oblique, with 
apex projecting. The fibers enlarge gradually as they approach the surface. At the outer end the 
spicules diverge somewhat, forming a dermal brush, which projects beyond the surface of the sponge. 
Connectives extend between the radial bundles at about right angles to the latter. They include 
each one to three spicules, covered with spongin. 
The dermal skeleton is made up of the projecting dermal brushes and the most superficial connec- 
tives. 
Axinella is often defined as having plumose skeletal fibers. This term applies 
fairly well to the outer parts of the radial fibers in the Beaufort species, but not well 
to the rest of the skeleton, although the axial fibers and the inner parts of the radial 
fibers all show plenty of spicules placed obliquely, with the points slightly projecting; 
and after all, it is to this position of the spicule in the fiber that the ‘‘ plumose”’ character 
is reducible. 
In its external form this sponge closely approaches the European Axinella flustra 
(padina) Topsent (Topsent, 1896, p. 131; 1904, p. 139), a species which has tricho- 
dragmata about 40 wp long. 
If we lay too great a stress on the presence of the spined styles, A. acanthijera might 
be removed to the Ectyonine and put in or close to Raspailia, in some of the species 
of which the acanthostyles are vestigial or absent, as in subgenus Syringella (which, 
to be sure, certain authors separate completely from Raspailia, making it an independent 
genus placed in the Axinellide). This is the familiar form of argumentation based on 
the presence or absence of a particular feature which leads in its application to the 
establishment of ‘parallel genera,” viz, genera assigned to different groups, which 
yet resemble one another except in regard to the feature in question. From one point 
of view such parallel genera are looked on as temporarily defensible because of their 
practical utility, although artificial. From another they are regarded as natural groups 
which owe their general similarity to independent adaptation (or ‘‘convergent evolu- 
tion”). But the character of the skeleton taken as a whole (cf. Vosmaer, 1912, p. 310) 
leaves little doubt that the sponge belongs with the other Axinella species. The 
presence of the spinose styles is accordingly to be interpreted as a vestigial feature 
which has never been quite lost, or, possibly, as an imperfect return (incomplete rever- 
sion) to an ancestral condition which had disappeared in the Axinellids. 
The sponge resembles Raspailia, or, at all events, the species which center around 
R. viminalis O. Schm. (cf. Vosmaer, 1912, p. 313), in yet another respect, viz, in the 
presence of long projecting styles in the dermal brushes. In the R. viminalis type, 
regarded by Vosmaer (loc. cit.) as characteristic of the genus, each dermal brush includes 
a single, strong, far-projecting style, surrounded at its base by a tuft of diverging small 
spicules, generally styles, sometimes oxeas. There are, of course, many species gen- 
erally assigned to Raspailia, in which the radial fibers (or the tufts of spicules which 
represent them) lack the large terminal styles, and a few such as R. irregularis Hent- 
schel from the Antarctic (Hentschel, 1914, p. 121) and R. hornelli Dendy from the Gulf 
of Manaar (1905, p. 172), in which each dermal brush includes not a single such style, 
but a bunch of them. It will be noticed that in the Raspailia species the long-projecting 
styles are stouter than the surrounding spicules of the brush, whereas in the Beaufort 
sponge they are slenderer. Nevertheless, their presence, coupled with the occurrence 
of acanthostyles in the Beaufort sponge, greatly strengthes the already well-supported 
view that Raspailia is an intermediate form, leading up from the Ectyonine to Axin- 
