is surrounded by a second wall and air space of 4 inches. The walls of the respiration chamber proper are double, the inner of copper and the outer of zinc, with a 3-inch dead air space between, and through these walls are distributed some 600 iron German-silver couples connected in series with a reflecting galvanometer and serving to indicate any difference in temperature between the inner (copper) and outer (zinc) surface. Any such difference is rectified and the walls of the chamber maintained adiabatic by heating or cooling the air space surrounding the zinc wall-the former by means of an electric current through resistance wires and the latter by circulating cold water through brass pipes (Pl. XXIV). The double wooden wall surrounding the metallic chamber also contains a smaller number of iron German-silver couples, and is in its turn kept nearly adiabatic by regulating the temperature of the second air space. By means of very similar devices the temperature of the ingoing air is maintained the same as that of the outcoming air.

The temperature of the interior of the apparatus is measured by means of a series of copper resistance thermometers connected to a slide-wire Wheatstone bridge, and also by means of two mercurial thermometers. By raising or lowering the shields or varying the flow of water through the absorbers the rate at which heat is removed through the water current may be so regulated as to keep the temperature of the interior constant within very small limits, while the slight variations are made to balance each other in the course of an experiment, so that there is practically no capacity correction. Under these conditions, all the heat evolved by the animal must leave the apparatus either as sensible heat in the water current or as the latent heat of water vapor.

In practice, of course, corrections have to be made for any heat introduced or removed as sensible heat in the feed, excreta, etc. The friction of the water in the coil of copper pipes is also the source of a minute amount of heat, which is computed from the volume of water and the fall in pressure in passing through the pipes. This difference in pressure also affects slightly the readings of the thermometers, tending to make those in the ingoing water relatively higher than those in the outcoming water. The small correction for this difference has been determined experimentally within the range of pressures used.

THE BUILDING.

The building in which the apparatus is housed is of brick, with heavy walls containing an air space to aid in maintaining uniform temperature. The floor is of concrete, and the ceiling high enough to leave about 3 feet clear space above the top of the respiration chamber, which stands upon three brick piers about 18 inches above