In order that this arrangement may give perfect satisfaction the two vessels (C and V) must each be of at least half the capacity of the aquarium itself, and the total quantity of salt water sufficient to fill the aquarium together with one of them. It should also be remembered that since the pressure of water depends not on its quantity, but on its height measured perpendicularly, it follows that the height to which the fountain will play is determined by the height of the vessel C above the level of the jet. [Illustration: FIG. 38.--METHOD OF AËRATING THE WATER OF AN AQUARIUM A, aquarium with fountain; C, cistern to supply the fountain; O, pipe for overflow; V, vessel for overflow] Let us now suppose that the aquarium and the upper vessel have both been filled with sea water. The fine jet from the pipe plays into the air and returns with a supply of oxygen to the aquarium, while the excess above the level of O passes into the concealed vessel below the table. If the two vessels are as large as we recommend, and the jet a very fine one, the fountain may continue to play for hours before C is empty, the animals of the tank being favoured all this time with a continuous supply of air. And when the supply from above is exhausted, the contents of the bottom vessel are transferred to the top one, and at the same time so effectually strained by the layer of muslin that no sedimentary matter passes down to choke the fine jet of the fountain. One great advantage this method possesses is that the living creatures derive the benefit of a much larger quantity of water than the aquarium alone could contain; and thus, apart from the aërating effects of the fountain, the result is the same as if a much larger tank were employed. In our next illustration (fig. 39) we give a modified arrangement based on the same principle which may commend itself by preference to some of our readers. Here the supply pipe to the fountain passes through a hole in the bottom of the aquarium instead of into the top, and the outlet pipe is bent downward within so as to form a syphon. Those who are acquainted with the principle of the syphon will understand at once the working of such an arrangement as this. Let us suppose the vessel _c_ to be full of water, and the fountain started, while the water in the aquarium stands no higher than the level _l_. The water slowly rises until the level _h_ of the bend of the outlet tube has been reached, and during the whole of this time no water escapes through the exit. As soon, however, as the latter level has been attained, the water flows away into the lower vessel, into which it continues to run until the lower level is reached, and then the outflow ceases, not to commence again until the fountain causes the water to rise to the upper level.