Ten of these reservoirs are intended for marine animals. The water employed is never changed, but it is kept in continual agitation by circulation, produced by a current of water led from the great pipe which feeds the Bois de Boulogne. This water, being subjected to a strong pressure, compresses a certain portion of air, which, being permitted to act on a portion of the sea water contained in a closed cylinder placed below the level of the aquarium, makes it ascend, and enter with great force into a reservoir, into which it is thrown from a small jet. The sea water thus pressed absorbs a portion of the air, which is drawn with it into the reservoir. A tube placed in a corner of the reservoir receives the overflow, and conducts it into a closed carbon filter, whence it passes into a gravelly underground reservoir, returning again to the closed cylinder. The water is once more subjected to the pressure of air, and again ascends to the aquarium. The cylinder being underground, a temperature equal to about sixteen degrees Cent., which is nearly the uniform temperature of the ocean, is easily maintained. During winter, the aquarium is heated artificially. CHAPTER IV. ZOOPHYTES. "Nature is nowhere more perfect than in her smaller works." "Natura nusquam magis quàm in minimis tota est." PLINY. It will not be out of place here to offer some remarks on animals in general, including the whole kingdom as well as the great divisions which form the subject of this particular volume. But considering the vastness of the subject, and our imperfect knowledge of the whole animal series as a subject of study, nothing is more difficult than to seize upon the real analogies between beings of types so varied,--of organizations so dissimilar. The arrangements which naturalists have established in order to study and describe animals--the divisions, classes, orders, families, genera, and species--are admirable contrivances for facilitating the study of creatures numerous as the sands of the sea shore. Without this precious means of logical distribution, the individual mind would recoil before the task of describing the innumerable phalanges of contemporary animal life. But the reader must never forget that these methodical divisions are pure fictions, due to human invention: they form no part of nature; for has not Linnæus told us that nature makes no leaps, _natura non facit saltus_? Nature passes in a manner almost insensibly from one stage of organization to another, altogether irrespective of human systems.