PHILIPPINE COALS AS FUEL. 303 



the efficiency for the second run is slightlj' higher than that for the 

 first, showing the benefit of the first day's experience; however, in no 

 case is the difference much greater than the possible error from other 

 sources. Several preliminary trials were made on the coal regularly 

 used here for firing in order thoroughly to test the working condition of 

 the apparatus. It would have been very desirable to have had duplicate 

 determinations of the steaming quality of each coal, but this was not 

 always possible with the supply on hand; nevertheless it is believed 

 that all the results are complete and sufficiently reliable to show the 

 nature and indicate the real fuel value of the coal ; in fact it has recently 

 been shown ^ that more than one test of a coal is superfluous. Seventy- 

 seven first tests gave an average efficiency of 66.05 and seventy-seven 

 second tests an average of 66.02 and thirty-two third tests one of 65.87. 

 It is evident that promiscuous tests made under different conditions 

 are not at all comparable, for it would be impossible to discover whether 

 the variation was due to the fuel, the apparatus or the manipulation. 

 However, in the work done at this Bureau many factors have been 

 eliminated by using the same plant * and the same personnel ; the others 

 have been carefully controlled by using the same apparatus and main- 

 taining all manipulations and general conditions as nearly uniformly 

 constant as possible, except where a change in the second test was to the 

 advantage of the. coal. With the variable factors eliminated, the coals 

 can be directly compared. 



DESCRIPTION OP APPARATUS AND METHODS EMPLOYED. 



All instruments used were carefully standardized and every precau- 

 tion taken to prevent the possibility of error. As the nature of the coals 

 to be burned was so entirely different, two sets of grates were provided. 



The one was of plain, single bars 1.5 centimeters in width and constructed 

 to give an air space of 1.2 centimeters between each pair, or a ratio between 

 air space and grate surface of 20 to 45. The other, constructed for these tests and 

 used with some of the coals, was a perforated grate with round, tapering holes 

 1.25 centimeters in diameter at the top, the smallest dimension, averaging 25 

 per square decimeter and giving a ratio between air space and grate surface 

 of 18 to 45. 



The two boilers shown in Plate I are exactly alike, the following 

 description apj)lies to both; however, with one exception, the tests were 

 made with the one on the right; they can afford only a clue as to the 

 efficiency of the boilers. This was not sought, for there are no means of 

 comparing the boilers with others fired with Philippine coal, or perhajDS 

 with themselves under different conditions. The boiler was thoroughly 



^ Breckenridge, L. P., U. S. G. 8. Bull. (1907), 325, 32. 



* The losses through radiation and conduction do not vary greatly for any 

 given installation. 



