CALOEIFIC VALUE OF PHILIPPINE COALS. 183 



an exceedingly close agreement with the generally accepted number 

 of Berthelot/=' of Petit " or of Favre and Silberman *^ of 8,140, 8,103, 

 and 8,080 calories, respectivel}', for pure carbon. The Philippine coals 

 are known to be very low in heavy hydro-carbon compounds '"' and 

 naturally the iised carbon would likewise be low, since it would be 

 the more easily degasified. The above results indicate that they are 

 entirely devoid of hydrocarbons that have high heats of combustion, 

 or else that the pure fixed carbon is pure carbon. Therefore, at present 

 it is impossible to establish a generally applicable formula for calculating 

 the heat of combustion of all t3rpes of coals from the results of proximate 

 analysis and instead of using eighty-two times the per cent of fixed 

 carbon for its calorific value, for our coals, we must use eighty-one. 



Bement says "moisture is the most variable" " factor in Illinois coal and "it 

 is a fact that the coal from a general locality has been analyzed over and over 

 again, with always a more or less different result when presented in onty the 

 moist coal composition, so that after all of the multiplicity of work, a final 

 conclusion or full xmderstanding is still unattained." '"' 



It is probable that the great variability in the moisture is accountable 

 for the different results. I have shown '"' that when coal is analyzed 

 according to the ofQcial method, a variation of one and a half per cent in 

 the volatile combustible matter, respectively fixed carbon, is produced if 

 the percentage of loosely held water is varied five per cent; moreover, 

 in the analysis of Philippine coal the official method gives large mechanical 

 losses. The smoking ofE method which was devised and substituted for 

 the official method gives no mechanical loss and also by its use no 

 variation in the fuel ratio is produced when the content of loosely held 

 water is varied. In the light of this research it seems proper to attribute 

 the great discrepancies which were noted in Bement's paper, to error in 

 the method of analysis rather than to inherent properties of the coal. 



There are other errors in the proximate analysis of coal which, though 

 recognized, are not often capable of calculation as they represent the 

 sum of a number of variable chemical changes. It was |)ointed out 

 above that the presence of carbonates in coal introduces the error of 

 high carbon in the ultimate analysis. This is also true of the proximate 

 analysis. Carbonates which give 'off carbon dioxide, g3'psum and silicates 

 which give up their combined water, and pyrites which is converted to 

 iron oxide, when the coal is burned to ash, are all sources of error. 



" Berthelot, loc. cit. 



"Petit, loc. cit. 



"Favre, P. A., and Silbermann, J. T. Ann. cMm. et pliys. (1852), III, 34, 403. 



""Cox, A. J. This Journal (1906), 1, 877. 



"Bement, A. III. Geol. Surv. (1906), Bull. 3, 23. 



"'Bement, A. Journ. Am. Chem. Soc. (1906), 28, 637. 



"»Cox, A. J. This Journal, Sec. A (1907), 2, 60. 



