[Sidenote: Long large tube exhausted and illuminated. Spiral form of discharge.] (4) A large tube, 24 inches long and 2 inches in diameter, with ball and point electrodes respectively, was exhausted, and the current passed through it. The effects were similar in most respects to those produced in the globular and bell receivers, but the streams of light assumed a distinctly spiral form in their passage (see Plate XVIII. fig. 5). This tube when placed between the poles of the magnet showed no effect, except a slight condensation of the streams of light towards the sides of the glass. [Sidenote: Globular receiver again used.] (5) The globular receiver first described was again used (the Tate pump having been cleaned and working easier). [Sidenote: Phosphorescent after-glow succeeding the spark.] (_a_) When exhaustion was as good as it could be got, the spark struck across in a single, slightly expanded, stream of rosy light, having a tendency to curve upwards (see Plate XVIII. fig. 6). The electrodes had but little glow round them, only just enough to distinguish the poles apart. When the flow of the stream was interrupted by breaking contact with one terminal, so that sparks passed in succession, we thought we detected a faint blue phosphorescent after-glow succeeding each spark. [Sidenote: Positive wire only attached.] (_b_) The positive wire only was attached to one electrode, the negative being unconnected. A set of faint whity-blue cobweb-looking streams of light spread from the electrode all over the receiver, having a vibratory motion. The spaces between these were dark, and there was no aura—the effect being similar, but not quite so bright and pronounced, as when both wires were attached (see Plate XVIII. fig. 7). [Sidenote: Negative wire only attached.] (_c_) The negative wire only was attached. The cobweb streams were absent, or only shot out very occasionally. The main effect was a straight nebulous stream of violet light, which commenced at the electrode and spread out in a fan-shape towards the lower brass cap of the receiver; while, at the same time, an aura or glow of similar light, but fainter in quality, spread from the electrode over at least one half of the receiver. This aura would no doubt have filled a small flask (see Plate XVIII. fig. 8). [Sidenote: Effect of gradual exhaustion on the discharge.] (_d_) When exhaustion was first commenced, both electrodes of the receiver (both wires being connected) threw out spider-web-like streams, as in Experiment 1, pale blue from the one pole and somewhat rosy from the other. As the exhaustion progressed the pale-blue streams disappeared, while the rosy flickering ones diminished in quantity and extent until ultimately a single rosy stream of light crossed the receiver as in Experiment 5_a_. Upon admitting the air, these effects took place in an inverse order—the single stream being gradually broken up, and the spider-webs taking its place.