Foreign inorganic bodies, such as grains of sand, spicules of other sponges, etc, commonly occur in the medulla of the fibres, either of the principal alone, or of all without distinction, in the Spongidae and Spon-gelidae. They are probably taken up by the caps of polygonal cells from the mesoglaea. Their amount may be trifling or considerable, and the connecting spongin much reduced and of soft consistence, e. g. in Dysidea, Psammoclerna: or it may even disappear altogether as in Psammopemma. Foreign bodies frequently occur scattered through the mesoglaea, taken up from the surface of the sponge, and it has been supposed that a given sponge exercises a selective action on the material imbedded2. The peculiar 'skipping rope' fibres, i. e. filaments with a knob at each end, which occur dispersed in variable quantities in the mesoglaea of certain Ceratine sponges, are supposed to be either parasitic organisms or pathological formations caused by the presence of parasitic algae. They attain some length, e. g. 8 - 10 mm. (2/5 in. circa) in Hircinia variabilis, and were at one time thought to be characteristic of a family Hircinidae 3.

1Cells, probably imbedded spongoblasts, are found between the lamellae of the cortex in Ianthella (Darwinellidae) and in the medulla of Dendrilla. According to von Lendenfeld, who has studied the fibres of Dendrilla, these cells, which form layers between the caps of spongin, are destructive in nature, - hence spongoclasts. They destroy the caps of spongin and the inner layers of the cortex, and grow centrifugally, retaining their cap-like disposition. They also bore through the cortex at a spot where a branch is forming, and give origin to its medulla. Their powers become exhausted at last, and then they give origin to caps of spongin. Hence though the structure of the fibre resembles that which it has when first formed, it is not absolutely identical with it. Aplysilla violacea and Aplysina resemble Dendrilla in the structure of their fibres. The exact mode of formation of the medulla, how it becomes continuous throughout the fibres, how it increases, if it does increase in other sponges, are points unknown.

See on the subject, von Lendenfeld, Z. W. Z. xxxviii. pp. 267, 285, 304; Id. Z. A. viii. p. 469; Polejaeff, Keratosa, op. cit. pp. 4-12.

2On the inclusion of foreign bodies, see Schulze's remarks, Z. W. Z. xxxiii. p. 14. In Halme nidus Vesparum (Auleninae) the cortex of the sponge is filled with sand grains, large on the exposed portions, small in the sheltered or internal, cemented into a crust by spongin. Dysidea, Psammoclerna, Psammopemma are described by Marshall, Z. W. Z. xxxv, together with Phoriospongia. The last-named affords an instance of a skeleton of foreign bodies, plus proper siliceous spicules. Some true homy sponges possess spicules; see von Lendenfeld, Proc. Lin. Soc. New South Wales, ix. p. 493; x. p. 490. Hence it is possible that Phoriospongia is a Geratine in which the spongin is lost, as in Psammopemma. The relationship of the horny to other sponges is a vexed question. See Polejaeff, Keratosa, op. cit. p. 76 et seqq.; Vosmaer, Biol. Centralblatt, vi. p. 187; von Lendenfeld, op. cit. supra, x. p. 483; Id. Z. A. vii. p. 201; ibid. viii. p. 484.

3See Schulze, Z. W. Z. xxxiii. p. 19; Polejaeff, Keratosa, op. cit. pp. 12-16; and for the second view mentioned, von Lendenfeld, Z. A. viii. p. 483. Carter maintained, in 1878, the view

Many sponges have been found unisexual, others hermaphrodite, e. g. Sycandra (Sycon) raphanus, Aplysilla violacea, Dendrilla; others sometimes in one, sometimes in the other condition, e.g. Oscarella lobularis, Halisarca Dujardini1. The male element in hermaphrodite species usually ripens first. A coloration indicative of sex has been observed in Chalinula fertilis by Keller though not by Vosmaer, and the male and female Spon-gilla lacustris are said to differ structurally by Marshall. The sexual cells are wandering mesoglaeal cells which ripen in the mesoglaea. The spermato-spore of the Calcarea and Verongia (Aplysinidae)2 becomes binucleate: its protoplasm so divides that one portion becomes a covering or enveloping cell, whilst the nucleus of the other multiplies and the nucleated mass gives origin to the spermatozoa. The whole spermatospore in other sponges divides into spermatoblasts. The ovum of the Calcarea remains during growth an amoeboid wandering cell, but in other sponges it becomes spherical and acquires a fixed outline. Both spermatospore and ovum in the Non-Calcarea are invested by a capsule of flat epithelium derived from mesoglaeal cells and generally arranged in a single layer, but in Aplysilla violacea and Dendrilla in several layers.

And in the two exceptions named both sexual products collect in small masses, each mass invested by a common epithelial capsule with processes separating the individual spermatospores or ova one from another3. The ripe ovum is naked4, more or less granular, with a clear exoplasm, sometimes pigmented. It undergoes impregnation, segmentation, and development into a ciliated embryo in situ, except in the burrowing sponge Clione, where it is expelled as soon as it is ripe enough to undergo development (Nassonow). Segmentation -is total and as a rule regular; variably but slightly irregular in Oscarella lobularis; somewhat irregular in Halisarca that the filaments in question, to which he gave the name of Spongiophaga communis, were vegetable parasites; see his paper on 'Parasites of the Spongida,' A. N. H. (5), ii. p. 165.

1The male or female element predominates in a given specimen of Sycandra raphanus. Unisexual individuals of Oscarella, etc, may have shed one or the other element. It is possible that all sponges may produce at one time sperm, at another ova. The sexual products ripen in the walls of the cones in Syconidae, among the ampullae of other sponges, in the base of incrusting sponges, or in the endogastric septa, e. g. in Oscarella.

8Polejaeff, Keratosa, op. cit. p. 72. In the Calcarean Asconidae the ripe sperm-balls may frequently be found projecting into the gastric cavity. Hence Haeckel thought they were derived from endoderm cells. See Polejaeff, Calcarea, op. cit. p. 33, and Vosmaer, Porifera, p. 413.

3In A. violacea the number of ova in each mass appears to increase by fission (?) to about forty, and subsequently to dwindle to four. The ovum of the same sponge is separated by a space from the walls of its capsule to which, however, it is suspended by a single peduncle cell. The contents of each mass of capsules ripen simultaneously in Aplysilla but not in Dendrilla. The ova of Euspongia are aggregated 10-30 in number within a small area near an exhalent canal, and the mesoglaea in which they lie is surrounded by a richly anastomosing network of canals. Polejaeff states that the epithelial cells of the capsule of an ovum grow in volume during the formation of the embryo (Keratosa, op. cit. pp. 52-3).

4There is an egg membrane in A. violacea (von Lendenfeld), and a calcareous shell in Sycaltis (Amphoriscus) testipara and S. (A.) ovipara (Haeckel, Kalkschwamme, i. p. 157)Dujardini, extremely so in Chalinida fertilis. The diameter of the embryo increases during segmentation, and when fully grown it passes into the gastric cavity and escapes by the osculum or by the inhalent canals. The parent sponge may then perish (Chalinida fertills).