Six or eight typical coals are to be selected, each representing a certain group of nearly the same chemical composition. Each series will consist of several sets of tests, each set being run with all the conditions constant except the one, the effect of which on the size of the combustion space is to be investigated. Thus a set of four or five tests will be made, varying in rate of combustion from 20 to 80 lb. of coal per square foot of grate per hour, keeping the supply of air per pound of combustible and the rate of heating constant. This set will show the effect of the rate of combustion of the coal on the extent of space required to obtain combustion which is practically complete. Other variables, such as composition of coal, supply of air, and rate of heating, remain constant. Another set of four or five tests will be made with the same coal and at the same rate of combustion, but the air supply will be different for each test. This set of tests will be repeated for two or three different rates of combustion. Thus each of these sets will give the effect of the air supply on the extent of combustion space when the coal and rate of combustion remain constant. Still another set of tests should be made in which the time of heating the coal when feeding it into the furnace will vary from 3 to 30 min. In each of the tests of this set, the rate of combustion and the air supply will be kept constant, and the set will be repeated for two or three rates of combustion and two or three supplies of air. Each of these sets of tests will give the effect of the rate of heating of fresh fuel on the extent of combustion space required to burn the distilled volatile combustible. These sets of experiments will require a modification in the stoker mechanism, and, on that account, may be put off until all the other tests on the other selected typical coals are completed. As the investigation proceeds, enough may be learned so that the number of tests in each series may be gradually reduced. After all the desirable tests are made with the furnace as it stands, several kinds of mixing structures will be built successively back of the stoker and tried, one kind at a time, with a set of representative tests. Thus the effectiveness of such mixing structures will be determined.