Without the magnet, the spectrum consisted of four bright bands, with a continuous spectrum between, resembling that of SO₃ tube No. 1. With the magnet on, the spectrum brightened, especially in the yellow and red, which were dull before; and a set of lines appeared upon it (a line or band in the yellow especially showing) which were not seen before. The lines were distinct, but not very bright. The action on the capillary was noticed to be strongest just between the conical points of the armatures; and, in accordance with this, the central part of the spectrum-band in the red and yellow showed an increased brightness. [Sidenote: Effects when one of the bulbs of the tube was heated. Changes in the spectrum under influence of magnet.] (2) One of the bulbs of the tube was then gradually heated with a small gas-flame. The single stream in the heated bulb became somewhat deflected and broken up into a number of smaller streams; and these, when placed under the magnetic influence, had small spark-like threads of light running among them. The capillary, as the tube was heated, and the sulphur rose in it, changed somewhat in tint, and, under the magnetic influence, became of yellow-rose hue. As the heat was applied to the bulb the bands of sulphur gradually appeared in the field of the spectroscope, until at last the band-spectrum of sulphur entirely took the place of the spectrum seen in the cool tube. The magnet being excited, the spectrum changed at once, a set of bright sharp lines (line-spectrum of S) appearing upon a faint and dull image of the band-spectrum. This effect was constantly repeated upon the magnet being excited. The magnet being taken off, the band-spectrum alone was to be seen. CHAPTER XV. EFFECT OF MAGNET ON A CAPILLARY GLASS TUBE. [Sidenote: Capillary portion of a Geissler tube tested in three ways.] The capillary portion of a Geissler tube was cut away from the bulbs, cleaned, and connected by a small vulcanite tube with the gas-pipe in the room conveying coal-gas at ordinary pressure. The flame was small and oval in shape, 8 millims. high, by 4 millims. wide, and burnt quite steadily. (Plate XVII. fig. 13.) [Sidenote: No effect on flame.] (1) The capillary tube was placed between the poles of the excited magnet, almost, but not quite, touching them; no effect at all was produced on the flame. (2) The tube was placed so that the conical ends of the armatures were allowed to compress the centre of it between them; still no effect was produced on the flame. (3) The tube was placed so that the straight sides of the armatures compressed it between them; still no effect took place on the flame. [Sidenote: Flame between poles of magnet.] (4) The flame itself was placed between the poles of the magnet. It was slightly drawn towards one pole with an inclination to form the magnetic curve. [Sidenote: Quill glass tubing tested. No effect on the flame.]