by a method which forms the subject of the second section of this paper.
When a plate of glass is brought to a full red heat, and is then cooled by placing its edge on a bar of cold iron, the same fringes of colour are developed during cooling as by placing cold glass upon hot iron; and in this case the glass retains the property given to it even after it is completely cold.
The author delineates various configurations of colours produced by plates of different forms thus cooled. Among many others, a parallelogram of glass exhibits an inscribed parallelogram, with lines from each angle to the angles of the plate; and when the plate has been divided longitudinally by a diamond, each of the portions again exhibits an inscribed parallelogram, just as if the parts had been separately heated; and in this respect they present a property analogous to that of a divided magnet, each part of which has opposite poles as the entire bar.
A circular plate of glass, cooled with its centre resting on a piece of cold iron, or a cylindrical rod of glass cooled in the open air, when examined by polarized light in the direction of its axis, each present the same appearance of a black cross through their centre, and concentric fringes parallel to their circumference.
Since it is obvious that in these cases of rapid cooling, as well as those of rapidly heating, there must be progressive variations of density of the glass proceeding in a direction from the source of heat or of cold, and since the phenomena exhibited by many crystallized bodies, when examined in the direction of their axis, are precisely similar, Dr. Brewster infers that there exists in these crystallized bodies also a corresponding variation of density, proceeding toward their axes, which will afford an easy explanation of the fringes they exhibit.
In these experiments, the author examines what magnitude of wire and of apertures in the metallic gauze of his lamp is consistent with security against explosion of mixtures externally.
When the gauze is made of wire one fiftieth of an inch in diameter, and at intervals of one tenth, so as to make 100 apertures in the square inch, explosion may take place, either from intense ignition of the top of the lamp, or from lateral currents of air forcing the flame through the interstices.
When the intervals of the same wire were only one fourteenth, though the danger from lateral motion was obviated, still ignition of the wire caused explosion. With intervals of one sixteenth, still there was danger from the same source; but when the distances were reduced to one twenty-fourth on 576 apertures in the square
