As the size of the plant and the fuel consumption increase, such a check of losses and recording of results becomes a necessity. In the larger plants, the saving of but a fraction of one per cent in the fuel bill represents an amount running into thousands of dollars annually, while the expense of the proper supervision to secure such saving is small. The methods of supervision followed in the large plants are necessarily elaborate and complete. In the smaller plants the same methods may be followed on a more moderate scale with a corresponding saving in fuel and an inappreciable increase in either plant organization or expense. There has been within the last few years a great increase in the practicability and reliability of the various types of apparatus by which the records of plant operation may be secured. Much of this apparatus is ingenious and, considering the work to be done, is remarkably accurate. From the delicate nature of some of the apparatus, the liability to error necessitates frequent calibration but even where the accuracy is known to be only within limits of, say, 5 per cent either way, the records obtained are of the greatest service in considering relative results. Some of the records desirable and the apparatus for securing them are given below. [Illustration: 2400 Horse-power Installation of Cross Drum Babcock & Wilcox Boilers and Superheaters at the Westinghouse Electric and Manufacturing Co., East Pittsburgh, Pa.] Inasmuch as the ultimate measure of the efficiency of the boiler plant is the cost of steam generation, the important records are those of steam generated and fuel consumed Records of temperature, analyses, draft and the like, serve as a check on this consumption, indicating the distribution of the losses and affording a means of remedying conditions where improvement is possible. Coal Records--There are many devices on the market for conveniently weighing the coal used. These are ordinarily accurate within close limits, and where the size or nature of the plant warrants the investment in such a device, its use is to be recommended. The coal consumption should be recorded by some other method than from the weights of coal purchased. The total weight gives no way of dividing the consumption into periods and it will unquestionably be found to be profitable to put into operation some scheme by which the coal is weighed as it is used. In this way, the coal consumption, during any specific period of the plant's operation, can be readily seen. The simplest of such methods which may be used in small plants is the actual weighing on scales of the fuel as it is brought into the fire room and the recording of such weights. Aside from the actual weight of the fuel used, it is often advisable to keep other coal records, coal and ash analyses and the like, for the evaporation to be expected will be dependent upon the grade of fuel used and its calorific value, fusibility of its ash, and like factors.