The experimental tube being again cleansed and exhausted, the mixed air and nitrite-of-butyl vapour was permitted to enter it until the associated mercury column was depressed 1/10 of an inch. In other words, the air and vapour, united, exercised a pressure not exceeding 1/300th of an atmosphere. Air, passed through a solution of hydrochloric acid, was then added, till the mercury column was depressed three inches. The condensed beam of the electric light was passed for some time through this mixture without revealing anything within the tube competent to scatter the light. Soon, however, a superbly blue cloud was formed along, the track of the beam, and it continued blue sufficiently long to permit of its thorough examination. The light discharged from the cloud, at right angles to its own length, was at first perfectly polarised. It could be totally quenched by the Nicol. By degrees the cloud became of whitish blue, and for a time the selenite colours, obtained by looking at it normally, were exceedingly brilliant. The direction of maximum polarisation was distinctly at right angles to the illuminating beam. This continued to be the case as long as the cloud maintained a decided blue colour, and even for some time after the blue had changed to whitish blue. But, as the light continued to act, the cloud became coarser and whiter, particularly at its centre, where it at length ceased to discharge polarised light in the direction of the perpendicular, while it continued to do so at both ends. But the cloud which had thus ceased to polarise the light emitted normally, showed vivid selenite colours when looked at obliquely, proving that the direction of maximum polarisation changed with the texture of the cloud. This point shall receive further illustration subsequently. A blue, equally rich and more durable, was obtained by employing the nitrite-of-butyl vapour in a still more attenuated condition. The instance here cited is representative. In all cases, and with all substances, the cloud formed at the commencement, when the precipitated particles are sufficiently fine, is _blue_, and it can be made to display a colour rivalling that of the purest Italian sky. In all cases, moreover, this fine blue cloud polarises _perfectly_ the beam which illuminates it, the direction of polarisation enclosing an angle of 90° with the axis of the illuminating beam.