The rest of the lines were very much of the same intensity. (For drawing of spectrum of air-tube in capillary part see Plate XV. spectrum 1.) _Violet [negative] Pole, same tube._ [Sidenote: Violet (negative) pole: spectrum described.] I next turned my attention to the violet or negative-pole glow; and here a remarkable change took place in the spectrum, not only in the position of the principal bands or lines, but in their relative intensity (see Plate XV. spectrum 2). The double line α in the capillary part was replaced in the violet glow by a shaded band of second intensity β, the sharp edge of which was extended towards the red, and formed (except for some faint indications) the limit of the spectrum in that direction. The somewhat faint line next α in the capillary tube had its faint representative in the violet pole; but the next two lines (capillary) were represented by the bright band γ in the violet pole lying in a position between them. Next γ in the violet pole came three faint lines, representing β, γ, and δ in the capillary spectrum; and then the bright band α, which was the brightest of the violet-pole group, and represented a medium-intensity band in the capillary spectrum. After this was a faint band near α, representing two rather bright ones in the capillary spectrum, this last being succeeded by other bands in the violet. α, β, and γ in the violet pole were examined carefully for relative brightness, and were, I believe, correctly marked. _Red [positive] Pole._ [Sidenote: Red (positive) pole: spectrum described.] The red [positive] pole was next examined, but presented no peculiar features. It appeared as a fainter representation of the capillary air-spectrum, with some few lines or bands absent, and (as will be seen after) was also a fair representation of a diffused air-spectrum (see Plate XV. spectrum 3). Examined for comparative intensity, at 24 inches from the slit, the whole capillary air-spectrum showed faintly. The marked lines in the centre of the spectrum generally retained their prominence; but after α I judged ε next in brightness. On examining the violet pole at 12 inches from the slit, the whole spectrum was faint and the bands α and β were alone distinctly seen. _Aurora (air)-tube._ (Plate XV. spectrum 4.) [Sidenote: Aurora-tube: discharge described. Spectrum described.]