at the highest attainable temperature is only semi-fluid, and is found to be of different densities in the upper and lower portions of the mass, owing to the varying specific gravities of its constituents. A partial admixture slowly going on in consequence of unequal expansion by heat in so bad a conductor as glass, and the motion induced by air bubbles slowly rising to the surface, have the effect of introducing veins, or striæ, consisting of streaks of more or less dense portions carried upwards by the rising air bubbles, running throughout the general mass, and entirely spoiling it for optical purposes. Now Fraunhöfer, knowing no means of preventing the formation of these veins or striæ, proceeded on this simple but laborious mode of counteracting these defects. He made a large potful of glass as perfect as he could by simple fusion; he allowed it to get cold in the pot; he then sawed the mass horizontally into slices, polished their surfaces, and thus examined their internal structure; and wherever there was a line or streak of more or less dense glass, the defective part was applied to a glass-grinder's wheel and cut away, not as a deep narrow notch but by a wide shallow indent; the surface was again polished for re-examination, and this process was repeated until no more veins, or striæ, were visible. The mutilated and indented disc of glass, sometimes cut nearly half-way through, was then put into one of the shallow fireclay dishes already described, gently heated at first, and finally made sufficiently soft to sink down and acquire the form and dimensions of the dish, the impress of whose surface it bore, while its upper surface assumed the polished appearance of ordinary molten glass.

What I desired to achieve was the production, at a small cost, of large and massive discs or lenses, which could not be produced by Fraunhöfer's system. Among the several plans I proposed, I will describe only two, each of which attacked the problem from an entirely different standpoint. First, I may mention that I made a series of laboratory experiments with viscid transparent fluids, contained in glass vessels of various forms and under varied conditions. Venice turpentine was first tried, but very viscid castor oil was the nearest to glass in its indications of movement within itself. Small grains of broken red sealing- wax, by their greater specific gravity, showed well the tendency of the oxide of lead (used in flint glass) to subside; and how, by rotating this vessel with one small fragment of sealing-wax, its movement was restrained within a circle the diameter of which was equal to the subsidence of the particle during a semi-rotation of the vessel containing the oil. The effect of the gentle rotation or rolling of the vessel was also experimented on in various ways. A small portion of the viscid oil was poured out, and a very minute quantity of blue powder ground up in it, just enough to give a faint blue colour. This blue oil was then poured back again into the nearly globular-shaped glass vessel, which must be considered as the glass pot; a little movement of the vessel produced streaks of blue colour like veins in marble, dispersed throughout the general mass of viscid fluid. But by continuing to roll the glass globe slowly for about two or three hours, not the slightest trace of veins or streaks of blue remained visible, while a very slight tint of blue pervaded the whole mass of oil, which was now perfectly homogeneous. It will be observed that the motion so given to the whole mass did not divide it, as the insertion of a stirrer would have done. I also demonstrated the fact that stirring from the surface by a rod was wholly impossible without the introduction of air in large quantities. So extraordinary is this fact that I cannot refrain from putting it on record. Take a glass jar or vessel, say ten inches deep and two inches in diameter, open at the top and closed at the bottom, as shown in Figs. 22 and 23, on page 104. Nearly fill it with clear, but viscid, castor oil, carefully removing all traces of air from the fluid by exhaustion under the glass bell of a common air-pump; place the jar on a table, take a polished metal, or, preferably, a glass, rod about the size of a blacklead pencil, and having a smooth, rounded end, wipe it, and very slowly and steadily lower it some six inches into the oil, as shown in Fig. 22; then as slowly and carefully withdraw it, occupying quite a minute in doing so. There will remain no trace that anything has entered the oil. Now place the jar again under the bell of the air-pump, take a few strokes with it, and there will appear a line of ill- defined mist, standing vertically upwards about six inches in height in the centre of the jar; at each stroke of the pump it becomes more visible, and enlarges in diameter. It soon assumes the appearance of innumerable little globes, like the hard roe of a herring, as shown at Fig. 23. A little more exhaustion, and these still further expand and rush upwards by the thousand, until at last all the air adhering to, and taken down by, the glass rod has been removed. What you may do, and what you may not do, with molten glass was thus beautifully illustrated by some of these preliminary experiments with viscid fluids.