This sponge is allied to _S. bombayensis_, from which it is distinguished not only by the abnormal characters of its gemmule-spicules and the absence of flesh-spicules, but also by the form of its skeleton-spicules and the structure of its skeleton. I have examined several specimens dry and in spirit; but _S. ultima_ is the only Indian freshwater sponge, except _Corvospongilla burmanica_, I have not seen in a fresh condition.

TYPES in the Indian Museum; co-types at Trivandrum.

HABITAT. Discovered by Mr. R. Shunkara Narayana Pillay, of the Trivandrum Museum, in a tank near Cape Comorin, the southernmost point of the Indian Peninsula.

Genus 2. PECTISPONGILLA, _Annandale_.

_Pectispongilla_, Annandale, Rec. Ind. Mus. iii, p. 103 (1909).

TYPE, _Pectispongilla aurea_, Annandale.

The structure of the sponge resembling that of _Euspongilla_ or _Ephydatia_; but the gemmule-spicules bear at either end, at one side only, a double vertical row of spines, so that they appear when viewed in profile like a couple of combs joined together by a smooth bar.

[Ill.u.s.tration: Fig. 20.--Gemmule and spicules of _Pectispongilla aurea_ (type specimen). _a_, Skeleton-spicules; _b_, gemmule-spicules; _b'_, a single gemmule-spicule more highly magnified.]

GEOGRAPHICAL DISTRIBUTION.--The genus is monotypic and is only known from Travancore and Cochin in the south-west of the Indian Peninsula.

15. Pectispongilla aurea*, _Annandale_.

_Pectispongilla aurea_, Annandale, _op. cit._, p. 103, pl.

xii, fig. 2.

_Sponge_ forming minute, soft, cushion-like ma.s.ses of a deep golden colour (dull yellow in spirit); the surface smooth, minutely hispid. One relatively large depressed osculum usually present in each sponge; pores inconspicuous; dermal membrane in close contact with the parenchyma.

_Skeleton_ consisting of slender and feebly coherent radiating fibres as a rule two or three spicules thick, with single spicules or ill-defined transverse fibres running horizontally. Towards the external surface transverse spicules are numerous, but they do not form any very regular structure.

_Spicules._ Skeleton-spicules smooth, sharply pointed, straight or nearly so. Gemmule-spicules minute, with the stem smooth and cylindrical, relatively stout and much longer than the comb at either end; the two combs equal, with a number of minute, irregularly scattered spines between the two outer rows of stouter ones. No free microscleres.

_Gemmules_ minute, spherical, with a single aperture, which is provided with a very short foraminal tubule; the granular coat well developed; the spicules arranged in a slanting position, but more nearly vertically than horizontally, with the combs pointing in all directions; no external chitinous membrane.

Length of skeleton-spicule 0.2859 mm.

Greatest diameter of skeleton-spicule 0.014 "

Length of gemmule-spicule 0.032-0.036 mm.

Length of comb of gemmule-spicule 0.008 mm.

Greatest diameter of shaft of gemmule-spicule 0.004 "

Diameter of gemmule 0.204-0.221 mm.

The gemmule-spicules first appear as minute, smooth, needle-like bodies, which later become roughened on one side at either end and so finally a.s.sume the mature form. There are no bubble-cells in the parenchyma.

15_a._ Var. subspinosa*, nov.

This variety differs from the typical form in having its skeleton spicules covered with minute irregular spines or conical projections.

TYPES of both the typical form and the variety in the Indian Museum; co-types of the typical form in the Trivandrum Museum.

GEOGRAPHICAL DISTRIBUTION.--The same as that of the genus.

_Localities_:--Tenmalai, at the base of the western slopes of the W.

Ghats in Travancore (typical form) (_Annandale_); Ernakulam and Trichur in Cochin (var. _subspinosa_) (_G. Mathai_).

BIOLOGY.--My specimens, which were taken in November, were growing on the roots of trees at the edge of an artificial pool by the roadside.

They were in rather dense shade, but their brilliant golden colour made them conspicuous objects in spite of their small size. Mr. Mathai's specimens from Cochin were attached to water-weeds and to the husk of a cocoanut that had fallen or been thrown into the water.

Genus 3. EPHYDATIA, _Lamouroux_.

_Ephydatia_, Lamouroux, Hist. des Polyp. corall. flex.* p. 6 (_fide_ Weltner) (1816).

_Ephydatia_, J. E. Gray, P. Zool. Soc. London. 1867, p. 550.

_Trachyspongilla_, Dybowsky (_partim_), Zool. Anz. i, p. 53 (1874).

_Meyenia_, Carter (_partim_), Ann. Nat. Hist. (5) vii, p. 90 (1881).

_Carterella_, Potts & Mills (_partim_), P. Ac. Philad. 1881, p. 150.

_Ephydatia_, Vejdovsky, Abh. Bohm. Ges. xii, p. 23 (1883).

_Meyenia_, Potts (_partim_), _ibid._ 1887, p. 210.

_Carterella_, _id._ (_partim_), _ibid._ 1887, p. 260.

_Ephydatia_, Weltner (_partim_), Arch. Naturg. lxi (i), p.

121 (1895).

_Ephydatia_, Annandale, P. U.S. Mus. x.x.xvii, p. 404 (1909).

TYPE, (?) _Spongilla fluviatilis_, auctorum.

This genus is separated from _Spongilla_ by the structure of the gemmule-spicules, which bear at either end a transverse disk with serrated or deeply notched edges, or at any rate with edges that are distinctly undulated. The disks are equal and similar. True flesh-spicules are usually absent, but more or less perfect birotulates exactly similar to those a.s.sociated with the gemmules are often found free in the parenchyma. The skeleton is never very stout and the skeleton-spicules are usually slender.

As has been already stated, some authors consider _Ephydatia_ as the type-genus of a subfamily distinguished from the subfamily of which _Spongilla_ is the type-genus by having rotulate gemmule-spicules. The transition between the two genera, however, is a very easy one. Many species of the subgenus _Euspongilla_, the typical subgenus of _Spongilla_ (including _S. lacustris_, the type-species of the genus), have the spines at the ends of the gemmule-spicules arranged in such a way as to suggest rudimentary rotules, while in the typical form of _S.

crateriformis_ this formation is so distinct that the species has. .h.i.therto been placed in the genus _Ephydatia_ (_Meyenia_), although in some sponges that agree otherwise with the typical form of the species the gemmule-spicules are certainly not rotulate and in none do these spicules bear definite disks.

GEOGRAPHICAL DISTRIBUTION.--_Ephydatia_, except _Spongilla_, is the most generally distributed genus of the Spongillidae, but in most countries it is not prolific in species. In j.a.pan, however, it appears to predominate over _Spongilla_. Only one species is known from India, but another (_E.

blembingia_*, Evans) has been described from the Malay Peninsula, while Weber found both the Indian species and a third (_E. bogorensis_*) in the Malay Archipelago.

16. Ephydatia meyeni* (_Carter_).

_Spongilla meyeni_, Carter, J. Bomb. Asiat. Soc. iii, p. 33, pl. i, fig. 1, & Ann. Nat. Hist. (2) iv, p. 84, pl. iii, fig. 1 (1849).

_Spongilla meyeni_, Bowerbank, P. Zool. Soc. London, 1863, p. 448, pl. x.x.xviii, fig. 4.

_Spongilla meyeni_, Carter, Ann. Nat. Hist. (5) vii, p. 93 (1881).

_Ephydatia fluviatilis_, Weber, Zool. Ergeb. Niederland.