One of the principal and most troublesome impurities met with in sand, is iron in the form of oxide. There are two oxides of iron: one, the protoxide, which imparts a green colour to glass; and the other the peroxide, whose staining property is yellow. A very small quantity of the former will give an appreciably green tint, whereas it requires a large quantity of the peroxide to produce even a delicate yellow. In all glass making, it is found necessary to use something which will counteract the colouring properties of these two oxides. The material employed was black oxide of manganese. This is still used in certain glass-works, but from its injurious action on the fire-clay pots, arsenious acid or common white arsenic is employed to effect the same object. The chemical action in the two cases is different: the black oxide of manganese is what is termed an oxidizing agent, and gives up, at a high temperature, a portion of its oxygen to the protoxide of iron, thereby converting it into the peroxide. It thus becomes comparatively harmless, by converting a quantity of that oxide, which gives a green colour, into the other oxide, which has little or no power of colouring, except it be present in large quantities. The difficulty in using black oxide of manganese is, the exact proportioning of it to the quantity of iron present in the sand, a quantity which cannot be easily determined. If the black oxide of manganese be used in excess, some of the oxide of manganese remains unreduced, and, when this is the case, it gives a purple colour to glass. If used in exact proportions, it is reduced to an oxide which does not impart colour to glass. This may be seen in many of the old plate glass windows which were employed for glazing purposes some sixty or seventy years ago, the colour of the panes being generally purple. Since this article was written, I have been consulted by a glass firm of eminence, as to the use of pure black oxide of manganese in the manufacture of flint glass, instead of that ordinarily supplied in commerce. The black oxide of manganese usually sold contains many other constituents besides black oxide of manganese; amongst these are iron, copper, cobalt, and alumina. The iron, as will be seen from what has before been stated, is a decidedly objectionable ingredient to use along with the manganese. Copper and cobalt both stain glass, the former of a bluish-green colour, while the latter makes it blue; and a small quantity of the latter has great staining power. I have thought it advisable to give analyses of the black oxides of manganese, and they are as follows: Binoxide of manganese (Molecule, Mn.O_2), is found native as pyrolusite or polyanite. Appended are two analyses of pyrolusite containing sesquioxide of iron.