It will be seen from the above description that the sponge is a sort of living filter. As the water passes in through the pores, it deposits in the substance of the sponge all the little organisms that it contains; on these the sponge feeds. It will naturally be asked, how does this living filter work? Water will not pass through small holes to flow out again at large ones in an upward direction, unless helped by some mechanism. How is this supplied? By the industry of the cells of the sponge. Its canal-system includes a set of wide chambers, lined with cells which have long cilia, called flagella. These flagella, constantly moving in one direction (like the fan of a ventilator), create a current, which passes the water on with such force that it reaches the central cavity, whence it is expelled through the oscula. These chambers do not communicate directly with the exterior. They are closed, except at certain small holes, the "prosopyles," where they take in the water that enters from spaces connected with the pores. At the main end of the chamber is an aperture called the "apopyle," capable of being partly closed, and leading into an excurrent passage. This last communicates with the central cavity of the sponge. It will be seen that the topography of the sponge is a very complicated business. All its details have been studied by means of thin sections specially prepared and placed under the microscope (see p. 183); in these the labyrinth of canals and chambers is seen cut through at various points; the cells lining them and dividing them may be individually studied. The passage of water through the sponge was first observed by Robert Grant; many of the most recent discoveries regarding the structure of sponges we owe to Professor Sollas. We have not yet explained what our living filter does with its food when it gets it. The ciliated cells of the internal lining take in solid particles just as Amoeba does; and from these they may be passed on to the cells of the middle layer, amoeboid cells, which can move about. These cells are considered to be derived from the primary layers of the body, especially the inner one, and to have wandered into a cellless middle layer, comparable in nature with that of some Coelenterates. The sponge is full of firm or gritty particles, which form its skeleton, and remain when the sponge is dead, and the softer parts decayed. These, when magnified, often present beautiful and curious shapes. The use of them is not only to support the body, but also to prevent the sponge from being eaten by other animals.