The skeleton of sponges, when it is not composed wholly of spongin, consists of, or at any rate contains, spicules that have a definite chemical composition and definite shapes in accordance with the class, order, family, genus, and species of the sponge. Formerly sponges were separated into calcareous, siliceous, and horny sponges by the nature of their skeleton; and although the system of classification now adopted has developed into a much more complex one and a few sponges are known that have both calcareous and siliceous spicules, the question whether the spicules are formed of salts of lime or of silica (strictly speaking of opal) is very important. All Demospongiæ that have spicules at all have them of the latter substance, and the grade Monaxonida, in which the freshwater sponges constitute the family Spongillidæ, is characterized by the possession of spicules that have typically the form of a needle pointed at both ends. Although spicules of this simple form may be absent in species that belong to the grade, the larger spicules, which are called megascleres, have not normally more than one main axis and are always more or less rod-like in outline. They are usually arranged so as to form a reticulate skeleton. Frequently, however, the megascleres or skeleton-spicules are not the only spicules present, for we find smaller spicules (microscleres) of one or more kinds lying loose in the substance of the sponge and in the external membrane, or, in the Spongillidæ only, forming a special armature for the reproductive bodies known as gemmules. All sponges obtain their food in the same way, namely by means of the currents of water set up by the flagella of the collar-cells. These flagella, although apparently there is little concerted action among them, cause by their rapid movements changes of pressure in the water contained in the cavities of the sponge. The water from outside therefore flows in at the pores and finally makes its way out of the oscula. With the water minute particles of organic matter are brought into the sponge, the collar-cells of which, and probably other cells, have the power of selecting and engulfing suitable particles. Inside the cells these particles undergo certain chemical changes, and are at least partially digested. The resulting substances are then handed on directly to other cells, or, as some assert, are discharged into the common jelly, whence they are taken up by other cells. Sponges reproduce their kind in more ways than one, _viz._, by means of eggs (which are fertilized as in other animals by spermatozoa), by means of buds, and by means of the peculiar bodies called gemmules the structure and origin of which is discussed below (p. 42). They are of great importance in the classification of the Spongillidæ. Sponges can also be propagated artificially by means of fission, and it is probable that this method of reproduction occurs accidentally, if not normally, in natural circumstances.