Depleted fisheries can be restored by fixing to tiles cut pieces of sponge, each of which grows into a new sponge. Some sponges detach portions of themselves as buds. Leucosolenia (Vasseur, 188o) breaks off the end of a branch, which falls blind-end down and grows into a complete sponge. In Oscarella lobularis the tip of a lobe will become thin-walled and then be nipped off; having no skeleton it does not sink, but may exist for a fortnight as a free-swimming sponge before it fixes on the bottom and grows into a flat lobed sponge like its parent (F. E. Schulze, 1879 [0]).
The form of reproduction especially characteristic of sponges is gemmulation, found in Glass Sponges, Four-rays, Needle sponges, Horny Sponges and Lace-chalks, but never recorded in Banana-chalks, among which sex is highly developed. A sponge gemmule arises by a number of undifferentiated cells coming to gether and enveloping themselves in some sort of membrane or epithelium. In the Sea-orange the growth of a radial bundle of spicules in the parent forces the gemmule to the exterior, where it drops off; in Grapnel-sponges the gemmules appear to be thrown out in the cloacal current; in Glass Sponges and in Claw-sponges the exterior cells of the gemmule become flagellate and it swims away like a "planula" larva; in the Sea fig (fig. 2) the gem mules remain during healthy life sealed in envelopes of spongin just above the sponginous base of attachment, ready to repro duce the sponge if it be destroyed. Gemmules of Fresh-water Sponges are similarly protected by chitinous shells, reinforced by various rods or spindles (Spongilla) or amphidiscs (Ephydatia). After years of desiccation they grow new Spongilla when water returns. Fresh-water Sponges of temperate climate die in winter, and the gemmules show like seed-pearls among the remnants of the skeletons, which they reclothe in spring. Young sponges of the same species, growing near each other, generally fuse together if their growth brings them into contact (not in the case of bath sponges) ; when this happens to two grown from gemmules of different parents, the resulting sponge will contain cells of two heredities and in its gemmule both strains may be present.
There is also, at least in some groups of sponges, true sexual reproduction. The development of ova into larvae may be ob served readily in Chalky or Horny Sponges, in Halisarca (a slime covered purple crust, without skeleton, probably closely allied to Bush Horny Sponges) in Oscarella, Spindle-sponges and Claw Sponges. It has only recently (Burton, 1928) been recorded in a Sea-orange, though this species has been observed constantly for 150 years. It has never been recorded in Grapnel-sponges or Fan sponges, it is rare among Pin-cushions and Ijima was unable to find it among Glass Sponges.
Many descriptions of spermatogenesis in sponges appear to the descriptions of parasitic flagellates (Syncrypta spongiarum, see Linnean Soc. Journ. Z. xxxiv.) and when spermatozoa are figured
there is rarely evidence that they are not alien. In Grantia com pressa Gatenby (192o, ibid.) has shown both the spermatogenesis and the fertilisation of the ovum by the spermatozoon through the intermediary of a collar-cell, in Sycon raphanus; Bidder (ibid.) has shown the direct fertilisation of the ovum under the cloacal surface by the spermatozoon. The spermatozoon was stiff-tailed when seen alive; it is possible that the tail only comes into play when the ovum is near.
Segmentation of the ovum in the Calcaronea produces a ring of eight similar cells (recalling the colonial Flagellates Stephano sphaera and Cyclonexis), each of which then divides horizontally into a smaller and a larger cell ; from the eight smaller cells is produced a hemisphere of flagellate cells and from the eight larger a hemisphere of rounded cells. This "amphiblastula" larva swims for a day or two, then the flagellate hemisphere is sucked inside the other and becomes concave, the two-layered cup so formed fixing by its lip to a rock. The rounded cells secrete the spicules and form the dermal layer, a vent bursts through both layers, pores are formed, and the young sponge is a two layered cylinder whose lower end is closed by the rock, the perforate walls and flagellate lining giving a resemblance to the Olynthus stage of Clathrina (fig. 8). Spicules sometimes appear while the larva is still free-swimming, and (Bidder, Brit. Assoc., 1928) the early division into two classes of cells occurs from the necessity of separating the spicule-forming cells from the locomotor cells, so that the activity of the flagellate cells shall not be hindered by the deposit of calcite or opal in their substance. There is no such division in Oscarella, which has no skeleton, but in most sponges the larva is egg-shaped with a surface of flagellate cells covering a solid interior of spicule-forming cells.
The fixed "Olynthus" larva of Sycon is subsequently meta morphosed into a Sycon by the flagellate cells retreating into radial tubes ; this may be considered to repeat a change which took place in racial history.
B1BLIOGRAPHY.—References given above by author's name and date will be found in full in (I) G. C. J. Vosmaer, Bibliography of Sponges 1551-1913, 1928, Cambridge Univ. Press. GENERAL: (2) I. B. J. Sollas, Porifera, 1906, Cambridge Natural History, vol. i. (easy and good) ; (3) E. A. Minchin, Porifera, 1900, Lankester's Treatise of Zoology, vol. ii., London (the standard text-book) ; (4) E. Hentschel, Porifera, 1924, Kiikenthal's Handbuch der Zoologie, Band i., Berlin (good modern German text-book) ; (5) Y. Delage, Spongiaires, 1899, Delage and Herouard, Traite de zoologie, tome ii. (French ; full, good and well illustrated). All of these text-books can be bought as a single volume.