[K] Except in "Proterospongia," an organism of doubtful affinities but not a sponge. It consists of a mass of jelly containing ordinary cells, with collar-cells outside.

[L] Cf. Weltner, "Spongillidenstudien, V," Arch. Naturg. Berlin, lxxiii (i), p. 273 (1907).

[M] It is difficult to see any trace of them in thin microtome sections. A fragment of the membrane must be mounted whole.

[N] Rec. Ind. Mus. i, p. 269 (1907).

[O] Proliferation whereby more than one osculum is produced is really a form of budding, but in most sponges this has become no longer a mode of reproduction but the normal method by which size is increased, and must therefore be considered merely as a vegetative process.

[P] P. Ac. Philad. 1887, p. 162.

[Q] Q. J. Microsc. Sci. London, xxii. p. 229 (1882).

[R] Arch. Naturg. Berlin, lix (i), p. 260 (1893).

[S] Journ. As. Soc. Beng. n. s. ii, 1906, p. 189.

[T] See Potts, Proc. Ac. Philad. 1884, p. 28.

[U] Phil. Trans. Roy. Soc. lv, p. 280.

[V] Descriptions of Siberian sponges are not included in these references.

GLOSSARY OF TECHNICAL TERMS USED IN PART I.

SYSTEMATIC LIST OF THE INDIAN SPONGILLIDÆ.

[Types, schizotypes, or cotypes have been examined in the case of all species, &c. , whose names are marked thus, *.]

Genus 1. Spongilla, Lamarck (1816).

Subgenus A. Euspongilla, Vejdovsky (1883).

1.

? S. lacustris, auct. (perhaps in N.W. India).

1a.

S. lacustris subsp. reticulata*, Annandale (1907).

2.

S. proliferens*, Annandale (1907).

3.

S. alba*, Carter (1849).

3a.

S. alba var. cerebellata, Bowerbank (1863).

3b.

S. alba var. bengalensis*, Annandale (1906).

4.

S. cinerea*, Carter (1849).

5.

S. travancorica*, Annandale (1909).

6.

S. hemephydatia*, Annandale (1909).

7.

S. crateriformis* (Potts) (1882).

Subgenus B. Eunapius, J. E. Gray (1867).

8.

S. carteri*, Carter (1849).

8a.

S. carteri var. mollis*, nov.

8b.

S. carteri var. cava*, nov.

8c.

S. carteri var. lobosa*, nov.

9a.

S. fragilis subsp. calcuttana*, nov.

9b.

S. fragilis var. decipiens, Weber

(probably Malaysian, not Indian).

10.

S. gemina*, sp. nov.

11.

S. crassissima*, Annandale (1907).

11a.

S. crassissima var. crassior*, Annandale (1907).

Subgenus C. Stratospongilla, Annandale (1909).

12.

S. indica*, Annandale (1908).

13.

S. bombayensis*, Carter (1882).

14.

S. ultima*, Annandale (1910).

Genus 2. Pectispongilla, Annandale (1909).

15.

P. aurea*, Annandale (1909).

15a.

P. aurea var. subspinosa*, nov.

Genus 3. Ephydatia, Lamouroux (1816).

16.

E. meyeni* (Carter) (1849).

Genus 4. Dosilia, J. E. Gray (1867).

17.

D. plumosa* (Carter) (1849).

Genus 5. Trochospongilla, Vejdovsky (1883).

18.

T. latouchiana*, Annandale (1907).

19.

T. phillottiana*, Annandale (1907).

20.

T. pennsylvanica* (Potts) (1882).

Genus 6. Tubella, Carter (1881).

21.

T. vesparioides*, Annandale (1908).

Genus 7. Corvospongilla, nov.

22.

C. burmanica* (Kirkpatrick) (1908).

23.

C. lapidosa* (Annandale) (1908).

Order HALICHONDRINA.

Siliceous monaxon sponges in which the horny skeleton is much reduced or absent and the spicular skeleton is more or less definitely reticulate. The microscleres are usually rod-like and rarely have more than one main axis.

Family SPONGILLIDÆ.

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

Freshwater Halichondrina which at certain seasons produce gemmules armed with peculiar microscleres. Two distinct kinds of microsclere are often present, that associated with the gemmule sometimes consisting of a vertical shaft at the ends of which transverse disks or rotulæ are borne. There is always at least a trace of a subdermal cavity.

Many authors divide the Spongillidæ into two subfamilies:—Spongillinæ (or Euspongillinæ), in which the gemmule-spicules have no transverse rotulæ, and Meyeninæ (or Ephydatiinæ), in which they have rotules at one or both ends. So gradual, however, is the transition that I find it difficult to decide in one instance to which of two genera, typical respectively of the two "subfamilies," a species should be assigned. Minchin in his account of the Porifera in Lankester's "Treatise on Zoology" (1900) regards the Spongillidæ merely as a subfamily of the Heterorrhaphidæ, and there certainly are few differences of a definite nature between them and the marine family (or subfamily) Remeridæ.

Key to the Indian Genera of Spongillidæ.

The most distinct genus of Spongillidæ not yet found in India is Heteromeyenia, Potts. It is easily distinguished from all others by the fact that the birotulate spicules of the gemmule are of two quite distinct kinds, which occur together on every mature gemmule. Heteromeyenia is represented by several American species, one of which has been found in Europe. Acalle, J. E. Gray, which is represented by a single South American species (Spongilla recurvata, Bowerbank), is related to Heteromeyenia but has one kind of gemmule-spicule tubelliform, the other birotulate. Probably Uraguaya, Carter, should be regarded as a subgenus of Trochospongilla with an unusually solid skeleton; it is peculiar to S. America. Parmula, Carter (=Drulia, Gray) includes South American forms allied to Tubella, but with the shaft of the gemmule-spicule degenerate and consisting of a mere projection in the centre of a shield-like body, which represents the lower rotule. The status of Potamolepis, Marshall, originally described from the Lake of Galilee, is very doubtful; possibly some or all of its species belong to the subgenus of Spongilla here called Stratospongilla (p. 100); but they are stated never to produce gemmules. The same is the case as regards Pachydictyum, Weltner, which consists of a single species from Celebes.

The sponges from Lake Baikal assigned by Weltner (Arch. Naturg. lxi (i) p. 131) to the subfamily Lubomirskinæ are of doubtful position and need not be considered here; while Lessepsia, Keller, from one of the salt lakes on the Suez Canal, certainly does not belong to the family, although it is assigned to it by von Lendenfeld (Mon. Horny Sponges, p. 904 (1889)) and subsequently by Minchin (Porifera, p. 152, in Lankester's Treatise on Zoology, part ii (1900)).

Genus 1. SPONGILLA, Lamarck (Carter emend.).

Type, Spongilla lacustris, auctorum.

Spongillidæ in which the gemmules have (normally) cylindrical or subcylindrical spicules that are sharp or blunt at the ends, without a distinct transverse disk or disks and without comb-like vertical rows of spines.

The skeleton is variable in structure, sometimes being almost amorphous, sometimes having well-defined radiating and transverse fibres firmly compacted with spongin. The skeleton-spicules are either sharp or blunt at the ends. Flesh-spicules are often absent; when present they are needle-like and resemble the gemmule-spicules in general structure; they have not even rudimentary rotules at their ends. The gemmules either lie free in the substance of the sponge or are attached to its support; sometimes they adhere together in free or attached groups.

Spongilla is undoubtedly the most primitive genus of the Spongillidæ, its spicules showing less sign of specialization than those of any other genus included in the family. As a fossil it goes back at any rate to the Upper Jurassic (p. 52).

Geographical Distribution.—Cosmopolitan. In most countries the majority of the freshwater sponges belong to this genus, but in Japan Ephydatia seems to predominate.

Key to the Indian Species of Spongilla.

Subgenus A. EUSPONGILLA, Vejdovsky.

Type, Spongilla lacustris, auctorum.

Spongillæ in which the gemmules are covered with a thick, apparently granular pneumatic coat. A delicate membrane often occurs outside this coat, but it is never thick or horny. The gemmules usually lie free in the sponge but sometimes adhere to its support; rarely they are fastened together in groups (e. g. in S. aspinosa, Potts). The skeleton-spicules are never very stout and the skeleton is always delicate.

The species in this subgenus are closely allied and must be distinguished rather by the sum of their peculiarities than by any one character. They occur in all countries in which Spongillidæ are found. Seven Indian species may be recognized.

1. Spongilla lacustris, auctorum.

[I have not attempted to give a detailed synonymy of this common species. There is no means of telling whether many of the earlier names given to forms or allies of S. lacustris are actual synonyms, and it would serve no useful purpose, so far as the fauna of India is concerned, to complicate matters by referring to obscure descriptions or possible descriptions of a species only represented in India, so far as we know, by a specialized local race, to which separate references are given.]

Sponge soft and easily compressed, very brittle when dry, usually consisting of a flat or rounded basal portion of no great depth and of long free cylindrical branches, which droop when removed from the water; branches occasionally absent. Colour bright green when the sponge is growing in a strong light, dirty flesh-colour when it is growing in the shade. (Even in the latter case traces of the "green corpuscles" can be detected in the cells of the parenchyma.) Oscula star-shaped, of moderate size, as a rule rendered conspicuous by the furrows that radiate from them over the outer surface of the parenchyma below the external membrane; oscular collars well developed.

Skeleton reticulate, loose, with definite radiating and transverse fibres held together by a small quantity of spongin; the fibres slender but not extremely so.

Spicules. Skeleton-spicules smooth, sharply pointed, long, slender. Flesh-spicules slender, covered with small spines, sharply pointed, nearly straight. Gemmule-spicules resembling the flesh-spicules but shorter and as a rule more strongly curved, sometimes bent so as to form semicircular figures, usually pointed somewhat abruptly; their spines relatively longer than those of the flesh-spicules, often curved backwards, especially near the ends of the spicules, at which points they are often longer than elsewhere.

Gemmules usually numerous in autumn, lying free in the sponge, spherical, variable in size but usually rather large, as a rule covered with a thick granular coat in which the spicules are arranged tangentially; a horizontal layer of spicules often present in the external membrane; the granular coat and its spicules occasionally deficient. No foraminal tubule; its place sometimes taken by an open, bowl-shaped chitinous structure the base of which is in continuity with the inner chitinous coat of the gemmule.

S. lacustris is an extremely variable species, varying in the size, proportions and shape of its spicules, in its external form and in the size and structure of the gemmule. A considerable number of varieties have been described from different parts of Europe and N. America, but some of these may represent distinct but closely-allied species; descriptions of most of them will be found in Potts's "Fresh-Water Sponges." The embryology and the earlier stages of the development from the egg have been described in great detail by Evans (Quart. J. Micr. Sci. (n. s.) xlii, p. 363 (1899)), while the anatomy and physiology are discussed by most authors who have written on these features in the Spongillidæ.

Type.—It is impossible to say who was the first authority to use the name Spongilla lacustris in the sense in which it is used by recent authors. No type can therefore be recognized.

Geographical Distribution.—S. lacustris occurs all over Europe and N. America and is probably the commonest species in most parts of both continents. It has also been found in Northern Asia and may occur in the Himalayan lakes and in the north-west of India.

1 a. Subspecies reticulata*, Annandale.

This race differs from the typical S. lacustris in the following particulars:—

(1) The branches are always compressed and anastomose freely when well developed (fig. 5, p. 37);

(2) the skeleton-fibres are finer;

(3) the skeleton-spicules are longer;

(4) the gemmule-spicules are longer and more slender and are never strongly bent.

As regards the form of the skeleton- and gemmule-spicules and also that of the branches the subspecies reticulata resembles S. alba rather than S. lacustris, but owing to the fact that it agrees with S. lacustris in its profuse production of branches, in possessing green corpuscles and in its fragility, I think it should be associated with that species.

A=gemmule-spicules of Spongilla lacustris subsp. reticulata (from type); B=gemmule-spicules of S. alba from Calcutta: both highly magnified.

The branches are sometimes broad (fig. 5, p. 37), sometimes very slender. In the latter condition they resemble blades of grass growing in the water.

Type in the Indian Museum; a co-type in the British Museum.

Geographical Distribution.—All over Eastern India and Burma; also in the Bombay Presidency. Localities:—Bengal, Port Canning, Ganges delta; Rajshahi (Rampur Bhulia) on the Ganges, 150 miles N. of Calcutta (Annandale); Puri district, Orissa (Annandale); R. Jharai, Siripur, Saran district, Tirhut (M. Mackenzie): Madras Presidency, Madras (town) (J. R. Henderson): Bombay Presidency, Igatpuri, W. Ghats (Annandale).

Biology.—This subspecies is usually found in small masses of water, especially in pools of rain-water, but Mr. Mackenzie found it growing luxuriantly in the Jharai at a time of flood in September. It is very abundant in small pools among the sand-dunes that skirt the greater part of the east coast of India. Here it grows with great rapidity during the "rains," and often becomes desiccated even more rapidly as soon as the rain ceases. As early in the autumn as October I have seen masses of the sponge attached, perfectly dry, to grass growing in the sand near the Sur Lake in Orissa. They were, of course, dead but preserved a life-like appearance. Some of them measured about six inches in diameter. At Port Canning the sponge grows during the rains on the brickwork of bridges over ditches of brackish water that dry up at the beginning of winter, while at Rajshahi and at Igatpuri I found it at the edges of small ponds, at the latter place in November, at the former in February. Specimens taken at Madras by Dr. Henderson during the rains in small ponds in the sand contained no gemmules, but these structures are very numerous in sponges examined in autumn or winter.

Numerous larvæ of Sisyra indica (p. 92) were found in this sponge at Rajshahi. Unlike those obtained from S. alba, they had a green colour owing to the green matter sucked from the sponge in their stomachs. The coralloides phase of Plumatella fruticosa (p. 219) was also found in S. lacustris subsp. reticulata at Rajshahi.

So far as my experience goes, this subspecies has always a bright green colour due to the presence of "green corpuscles," even when it is growing in a pond heavily shaded by trees or under the arch of a small bridge. Probably the more intense light of India enables the corpuscles to flourish in situations in which in Europe they would lose their chlorophyll.

2. Spongilla proliferens*, Annandale.

Sponge forming soft, shallow cushions rarely more than 10 cm. in diameter on the leaves of water-plants, or small irregular masses on their roots and stems. Colour bright green. Oscula moderate, flat, surrounded by deep, cone-shaped collars; radiating furrows and canals in the parenchyma surrounding them often deep. External pores contained normally in single cells. The surface frequently covered by small rounded buds; true branches if present more or less flattened or conical, always short, as a rule absent.

Skeleton loose, feebly reticulate at the base of the sponge; transverse fibres slender in the upper part of the sponge, often scarcely recognizable at its base. Very little spongin present.

Spicules. Skeleton-spicules long, smooth, sharply pointed; the length on an average at least 20 times the greatest breadth, often more. Flesh-spicules slender, gradually pointed, nearly straight, covered with minute straight or nearly straight spines. Gemmule-spicules very similar, but usually a little stouter and often blunt at the ends; their spines rather longer than those on the flesh-spicules, usually more numerous near the ends than in the middle of the spicule, slightly retroverted, those at the extreme tips often so arranged as to suggest a rudimentary rotule.

Gemmules usually numerous, lying free near the base of the sponge, very variable in size, spherical, surrounded by a thick granular layer in which the spicules, which are always very numerous, are arranged tangentially, their position being more near the vertical than the horizontal; a few horizontal spicules usually present on the external surface of the gemmule, which frequently has a ragged appearance owing to some of the tangential spicules protruding further than others. Foraminal tubule stout, cylindrical, usually somewhat contorted; its orifice irregular in outline. Sometimes more than one foramen present.

S. proliferens can be distinguished from all forms of S. lacustris and S. alba by the fact that its gemmules possess a foraminal tubule; from S. cinerea it can be distinguished by its colour and its smooth skeleton-spicules, and from S. travancorica by its free gemmules. I have been enabled by the kindness of Prof. Max Weber to examine specimens from Celebes and Java identified by him as S. cinerea, Carter, and have no doubt that they belong to my species.

Type in the Indian Museum; a co-type in the British Museum.

Geographical Distribution.—All over Eastern India and Burma; also in Cochin on the west coast; Ceylon; W. China; Java, Flores and Celebes. Localities:—Bengal, Calcutta and neighbourhood (Annandale); Berhampore, Murshidabad district (R. E. Lloyd): Assam, Mangal-dai near the Bhutan frontier (S. W. Kemp): Madras Presidency, Madras (town) and neighbourhood (J. R. Henderson); Rambha, Ganjam district (Annandale); Bangalore, Mysore (alt. ca. 3000 ft.) (Annandale); Ernakulam and Trichur, Cochin (G. Mathai): Burma, Rangoon (Annandale, J. Coggin Brown); Prome, Upper Burma (J. Coggin Brown); Kawkareik, Amherst district, Tenasserim (Annandale): Ceylon, between Maradankawela and Galapita-Gala, North Central Province (Willey). Mr. J. Coggin Brown has recently brought back specimens from Yunnan.

Biology.—S. proliferens is usually found in ponds which never dry up; Prof. Max Weber found it in small streams in Malaysia. It is common in India on the leaves of Vallisneria and Limnanthemum, on the roots of Pistia stratiotes and on the stems of rushes and grass. So far as I have been able to discover, the life of the individual sponge is short, only lasting a few weeks.

Sexual reproduction occurs seldom or never, but reproduction by means of buds and gemmules continues throughout the year. The former is a rare method of reproduction in most Spongillidæ but in this species occurs normally and constantly, the buds being often very numerous on the external surface. They arise a short distance below the surface as thickenings in the strands of cells that accompany the radiating fibres of the skeleton. As they grow they push their way up the fibres, forcing the external membrane outwards. The membrane contracts gradually round their bases, cuts off communication between them and the parent sponge and finally sets them adrift. No hole remains when this takes place, for the membrane closes up both round the base of the bud and over the aperture whence it has emerged.

The newly liberated bud already possesses numerous minute pores, but as yet no osculum; its shape exhibits considerable variation, but the end that was farthest from the parent-sponge before liberation is always more or less rounded, while the other end is flat. The size also varies considerably. Some of the buds float, others sink. Those that float do so either owing to their shape, which depends on the degree of development they have reached before liberation, or to the fact that a bubble of gas is produced in their interior. The latter phenomenon only occurs when the sun is shining on the sponge at the moment they are set free, and is due to the action of the chlorophyll of the green bodies so abundant in certain of the parenchyma cells of this species. If the liberation of the bud is delayed rather longer than usual, numbers of flesh-spicules are produced towards the ends of the primary skeleton-fibres and spread out in one plane so as to have a fan-like outline; in such buds the form is more flattened and the distal end less rounded than in others, and the superficial area is relatively great, so that they float more readily. Those buds that sink usually fall in such a way that their proximal, flattened end comes in contact with the bottom or some suspended object, to which it adheres. Sometimes, however, owing to irregularity of outline in the distal end, the proximal end is uppermost. In this case it is the distal end that adheres. Whichever end is uppermost, it is in the uppermost end, or as it may now be called, the upper surface, that the osculum is formed. Water is drawn into the young sponge through the pores and, finding no outlet, accumulates under the external membrane, the subdermal cavity being at this stage even larger than it is in the adult sponge. Immediately after adhesion the young sponge flattens itself out. This process apparently presses together the water in the subdermal cavity and causes a large part of it to accumulate at one point, which is usually situated near the centre of the upper surface. A transparent conical projection formed of the external membrane arises at this point, and at the tip of the cone a white spot appears. What is the exact cause of this spot I have not yet been able to ascertain, but it marks the point at which the imprisoned water breaks through the expanded membrane, thus forming the first osculum. Before the aperture is formed, it is already possible to distinguish on the surface of the parenchyma numerous channels radiating from the point at which the osculum will be formed to the periphery of the young sponge. These channels as a rule persist in the adult organism and result from the fact that the inhalent apertures are situated at the periphery, being absent from both the proximal and the distal ends of the bud. In the case of floating buds the course of development is the same, except that the osculum, as in the case of development from the gemmule in other species (see Zykoff, Biol. Centrbl. xii, p. 713, 1892), is usually formed before adhesion takes place.

The sponge of S. proliferens is usually too small to afford shelter to other animals, and I have not found in it any of those commonly associated with S. carteri and S. alba.

Owing to its small size S. proliferens is more easily kept alive in an aquarium than most species, and its production of buds can be studied in captivity. In captivity a curious phenomenon is manifested, viz. the production of extra oscula, often in large numbers. This is due either to a feebleness in the currents of the sponge which makes it difficult to get rid of waste substances or to the fact that the canals get blocked. The effluent water collects in patches under the external membrane instead of making its way out of the existing oscula, and new oscula are formed over these patches in much the same way as the first osculum is formed in the bud.

3. Spongilla alba*, Carter.