combination give rise to 500° of heat, when the same combine, at say the initial temperature of 400°, the result would be a temperature of 900°; if burnt at this latter degree, then 900° + 500° would be reached, and so on. It would seem as if there were no limit to the temperatures obtainable in this way. But the nature of the materials of which the furnace is constructed imposes a limit, for even the most refractory matters yield at length, and the working would come to an end by the fusing of the brickwork. The diagram is a section through the length of the hearth (for it is usually oblong in plan), and the low arch above H being exposed to the fiercest heat, is formed of the most refractory “silica bricks,” that is, bricks made of coarsely ground silica held together with a little lime; yet this extremely resisting material is acted upon, and the arch has to be renewed every few months or sometimes weeks. The hearth itself is supported by massive iron plates, shown in the diagram by the thick lines, above which is laid a deep bed L, of quartz sand or ganister, or where required a _basic_ lining, beaten hard down, and forming a kind of basin with sides sloping down in all directions to a point immediately below the centre of the fire-brick door D, where is the aperture for tapping, stopped by a mixture of sand and clay until the metal is ready for drawing off, when it runs outside into an iron spout lined with sand and is received into the ingot moulds. B in the figure represents the “bath,” as it is called, of molten metal, which, in the larger furnaces, where 20 tons of metal is operated on at once, may occupy an area of 150 square ft. It need hardly be mentioned that there has to be a certain adjustment between the volumes of air and of gas that pass into the regenerative stoves, in order that the best effect may be obtained. Besides the limit of temperature occasioned by the nature of the materials, there is a chemical reason why the regenerative stoves cannot increase the temperature indefinitely. It is noticed that when the temperature of the furnace has become very high indeed, the flame over the hearth assumes a peculiar appearance, being interrupted by dark spaces. These are attributable to what is called in chemistry “dissociation,”—in this case the dissociation of carbonic acid gas, which by the heat alone separates into carbonic oxide and oxygen gases. In the same way these gases refuse to combine if brought together heated beyond a certain temperature. This phenomenon of dissociation is a general one, for it is found that for any pair of substances there is a characteristic range of temperature above or below which they refuse to combine. The gas used in these stoves is either unpurified coal gas, or that produced by passing steam over red-hot coal or coke.