The operation of the instrument is as follows: As the cold blast used in the apparatus would be useless for the working of the pyrometer if taken direct from the cold blast main, owing to its irregularity of pressure, the regulator that has been described is employed, and by its means an absolutely steady flow of cold blast air at an unvarying pressure is secured. The diameters of the inlet and outlet nozzles are so nicely adjusted that, so long as both are at the same temperature, the outlet nozzle, which is open to the atmosphere, will pass all the air that the inlet nozzle can deliver without disturbing the pressure in the cistern, O; but if heat be applied to the circulating air through the walls of the spiral heater, the air expands in volume, and is unable to pass through the outlet nozzle in its heated condition as rapidly as it is delivered cold by the inlet nozzle. The consequence is that an increase of pressure takes place in the apparatus between the two nozzles, and it is this pressure that indicates the amount of heat that the air has taken up from the hot blast pipe, in which the spiral heater is fixed. Then, again, as this pressure is directly transmitted to the indicating liquid in the cistern and the vertical tube immersed in it, a rise takes place in the column which is in exact proportion to the expansion of the current of blast passing through the spiral heater. The method of graduating the indicator scales of the Frew pyrometer is worthy of special notice. When the apparatus is fitted up, and before it is permanently fixed in position, the spiral heater is placed in cold water of known temperature, and the point noted at which the colored liquid stands in the indicator tube. The water is then boiled, and the rise in the liquid in the tube is again noted. Suppose, in the first instance, the cold water temperature to be 62 deg. Fahr., and that, from this point up to 212 deg. Fahr., the liquid to have risen 2¼ in. in the tube; this is equal to 1½ in. per 100 deg. Fahr., and from these data a scale is constructed, the correctness of which is easily verified by transferring the spiral heater into an air bath or oil of high boiling point, and then comparing the readings of the pyrometer scale with those of a mercurial thermometer placed alongside of the spiral heater. By this means it can be clearly demonstrated that, up to the highest point to which it is safe to use a mercurial thermometer, the readings of the pyrometer scale and that of the thermometer are identical.