The function of the special spicules with which the gemmules of the Spongillidæ are provided appears to be not only to protect them but more especially to weight them to the extent suitable to the habits of each species. Species that inhabit running water, for example, in some cases have heavier gemmule-spicules than those that live in stagnant water, and their gemmules are the less easily carried away by the currents of the river. The gemmules of sponges growing in lakes are sometimes deficient in spicules. This is the case as regards the form of _Spongilla lacustris_ found in Lake Baa, Isle of Mull, as regards _S. helvetica_ from the Lake of Geneva, _S. moorei_ from Lake Tanganyika, and _S. coggini_ from Tali-Fu in Yunnan; also as regards the species of _Spongilla_ and _Ephydatia_ found in Lake Baikal, many of the sponges of which are said never to produce gemmules. Except in the genus _Corvospongilla_ and the subgenus _Stratospongilla_, in both of which the air-spaces of the gemmules are usually no more than cavities between different chitinous membranes, the pneumatic coat is either "granular" or "cellular." Neither of these terms, however, must be understood in a physiological sense, for what appear to be granules in a granular coat are actually minute bubbles of air contained in little cavities in a foam-like mass of chitin (or rather spongin), while the cells in a cellular one are only larger and more regular air-spaces with thin polygonal walls and flat horizontal partitions. The walls of these spaces are said in some cases to contain a considerable amount of silica. The gemmules with their various coverings are usually spherical in shape, but in some species they are oval or depressed in outline. They lie as a rule free in the substance of the sponge, but in some species adhere at its base to the object to which it is attached. In some species they are joined together in groups, but in most they are quite free one from another. Reproductive buds[O] are produced, so far as is known, by very few Spongillidæ, although they are common enough in some other groups of sponges. In the only freshwater species in which they have been found to form a habitual means of reproduction, namely in _Spongilla proliferens_, they have much the appearance of abortive branches, and it is possible that they have been overlooked for this reason in other species, for they were noticed by Laurent in _Spongilla lacustris_ as long ago as 1840 (CR. Sé. Acad. Sci. Paris, xi, p. 478). The buds noticed by Laurent, however, were only produced by very young sponges, and were of a different nature from those of _S. proliferens_, perhaps representing a form of fission rather than true budding (see 'Voyage de la Bonite: Zoophytologie,' Spongiaires, pl. i (Paris, 1844)).