The functions of gills are described by the Professor with great minuteness. "The main purpose of the gills of fishes," he says, "being to expose the venous blood in this state of minute subdivision to streams of water, the branchial arteries rapidly divide and sub-divide until they resolve themselves into microscopic capillaries, constituting a network in one plane or layer, supported by an elastic plate, covered by a tesselated and non-ciliated epithelium. This covering and the tunics of the capillaries are so thin as to allow chemical interchange and decomposition to take place between the carbonated blood and the oxygenated water. The requisite extent of the respiratory field of capillaries is gained by various modes of multiplying the surface within a limited space." "Each pair of processes," he adds, "has its flat side turned towards contiguous pairs, and the two processes of each pair stand edgeway to each other, being commonly united for a greater or less extent from their base; hence Cuvier describes each pair as a single bifurcated plate, or 'feuillet.'" The modification which takes place in the respiratory and other organs in Reptilia, is described in a few words. "Many fishes have a bladder of air between the digestive canal and the kidneys, which in some communicate with an air-duct and the gullet; but its office is chiefly hydrostatic. When on the rise of structure this air-bladder begins to assume the vascular and pharyngeal relations with the form and cellular structure of lungs, the limbs acquire the character of feet: at first thread-like and many jointed, as in the _Lepidosiren_; then bifurcate, or two-fingered, with the elbow and wrist joints of land animals, as in _Amphiuma_; next, three-fingered, as in _Proteus_, or four-fingered, but reduced to the pectoral pair, as in _Siren_." In all reptiles the blood is conveyed from the ventricular part of the heart, really or apparently, by a single trunk. In Lepidosiren the veins from the lung-like air-bladders traverse the auricle which opens directly into the ventricle. In some the vein dilates before communicating with the ventricle into a small auricle, which is not outwardly distinct from the much larger auricle receiving the veins of the body. In Proteus the auricular system is incomplete. In Amphiuma the auricle is smaller and less fringed than in the Sirens, the ventricle being connected to the pericardium by the apex as well as the artery. This forms a half spiral turn at its origin, and dilates into a broader and shorter bulb than in the Sirens.