(5) A piece of quill glass tubing was selected, 5 millims. in diameter and 1 millim. thick, and drawn out to a point, the end of which was snapped off and the tubing connected as before. The flame was 20 millims. high, and 5 millims. across, and somewhat lambent. On being placed (1) between the conical ends and (2) between the flat ends of the armatures, no effect could be seen on the flame. [Sidenote: Effect on taper and spirit-lamp flames.] (6) A small taper-flame was placed between the poles of the magnet: no effect was produced, except that the flame gave a slight “jump” each time the magnet was excited. A spirit-lamp flame was tried with a similar result. _Action of Magnet on a bar of heavy glass._ [Sidenote: Heavy glass bar and mounting described.] A piece of heavy yellow-tinted glass was selected, being a bar 10 centimetres in length, and 8 millimetres square. This was mounted in a frame with a Nicol prism at one end, and a double-image prism (next the eye) at the other. [Sidenote: Placed along poles of magnet. Effect of magnet on candle-images.] (1) The glass bar and mounting were placed upon and along the poles of the magnet (in the direction of the magnetic curves), and the double-image prism and Nicol were so adjusted that two images of a candle were seen—the one below bright and normal, the one above, by rotation of the prism, as nearly as possible extinguished (Plate XVII. fig. 4). On exciting the magnet the faint image at once conspicuously brightened, at the same time assuming a slightly green tinge. To get full effect of brightening, it seemed necessary to have good pressure-contact between the battery-wires and the binding-screws. [Sidenote: Effect on using a tourmaline as analyzer.] (2) Using a tourmaline as analyzer in lieu of the double-image prism, the candle-flame was seen alternately brightened and darkened, as the tourmaline was rotated; and when the image was obscured by rotation, excitation of the magnet caused it to brighten strongly. This effect was accompanied by the apparent removal of a dusky red patch or spot, which occupied the centre of the field when the flame was obscured. [Sidenote: Bar placed at right angles to the poles: no effect produced.] (3) The bar of glass and double-image prism being placed between the conical ends of the armatures, but at right angles to, instead of along, the poles, upon excitation of the magnet no effect at all was produced. (4) The bar and prism being placed in the same position between the flat ends of the armatures, no effect at all was produced. [Sidenote: Slight effect on second experiment.] (4_a_) Experiment No. 4 was repeated. It was thought that on excitation of the magnet the secondary image slightly brightened; but there was a doubt about it, and the effect (if any) was slight. [Sidenote: Effects produced when a biquartz was introduced.]