(_b_) The anhydride having been washed out, first with plain and afterwards with distilled water, some drops of the latter were allowed to remain in the receiver. On exhaustion a vapour-cloud was formed, and the discharge passed (both terminals being connected with the coil) through this. The rosy stream of light was formed as usual, but was more flickering and unsteady. As the exhaustion was lessened, the rosy stream disappeared, and the cobweb streams began to fill the receiver. These were, however, not so bright and sharp as in a dry receiver, but were faint and broad; while some diffused and nebulous streams of light, running (slightly bent) from pole to pole, and from ¼ to ⅜ of an inch broad, were intermixed with them. When one wire only was connected, the glow and streamers from the electrode were very faint. [Sidenote: Large bell-receiver and plate described. Receiver exhausted and stream of light formed.] (9) A large bell-shaped receiver, 11 × 8 inches, was next used. It was open at the bottom, which was ground as usual; and had a small opening at the top, also carrying a ground edge. A solid brass plate was prepared, ground only round the edge (in order to take the receiver), and in the centre of this brass plate were inserted two disks of soft iron, corresponding in position and size with the poles of the Ladd electro-magnet (see Plate XVIII. fig. 11). When this plate was placed on the magnet, the poles and disks were in contact; and the disks became N. and S. poles within the receiver. A small brass plate carrying a tap and exhaust-tube, and a binding-screw for attaching an electrode within, closed the receiver at top. The receiver and plate being placed on the magnet-poles, the former was exhausted until the discharge became a rosy slightly-diffused stream of light; with a small unilluminated space between it and the negative pole, where the usual violet glow appeared round the wire. This stream of light was used for the experiments after detailed. In some cases the conical armatures were placed within the receiver, in others the disks alone were used as the magnetic poles. [Sidenote: Effect on same when magnet excited.] (_a_) With the apparatus arranged as shown on Plate XVIII. fig. 10, and the magnet excited, a violet glow appeared round the end of the wire which was negative. A small unilluminated space then intervening, the stream ran in a curve between the wire and the armature, which latter was positive. The stream was not steady and had a tendency to rotate; but as this was better observed with the disks only, it is described further on. [Sidenote: Experiments with the conical armature removed. Vibrating stream.]