176 cox. 



vary bj' less than 2 jjgv cent on either side, from those given by the 

 calorimeter. 



Other methods have been proposed for the determination of the heating 

 power of coal without the use of the calorimeter. These methods are 

 never more than approximations. 



Tlie oldest of these is tliat of Berthier/" the determination being made by 

 intimately mixing 1 gram of powered coal and 50 grams of litharge together in 

 a clay crucible and covering with a layer of salt. The mixture is heated in a 

 crucible furnace, with a gradually increasing heat until fusion is complete; this 

 requires about fifteen minutes. The crucible is removed, poured, and when cold 

 the buttom is cleaned and weighed. Pure carbon should reduce 34 times its own 

 weight of lead; hydrogen 103 times its own weight. One part of pure carbon 

 can raise the temperature of 8,080-'' parts of water 1°. If the fuel is assumed 



as carbon, its value in heat-units may be estimated by multiplying ^— by the 



34 

 weight of the lead buttom obtained in the assay. As hydrogen is always present 

 in the coal, this method necessarily gives low results.^" Stijlzel ^ observed that 

 between the results obtained by this method and those calculated from the 

 ultimate analysis there was an almost constant diflference; those by Berthier's 

 method being about one-ninth too low. Von John and Fullon*" made a series of 

 comparisons of the results of Berthier's process with those calculated from 

 analysis on European coals. The former were almost uniformily lower than the 

 latter, in some cases 900 calories. In 1895, Noyes, McTaggart, and Graver^ 

 compared the results obtained by Berthier's method, on various coals in the 

 calorimeter and by calculation from the ultimate analysis. They corroborate the 

 statements made in previous work that Berthier's method gives low results. 

 "Theoretically, 1 gram of lead should correspond to a heating effect in the coal 

 of 234 calories. The results calculated with this factor are, however, about 12 

 per cent too low. The average of the results obtained, give an empirical factor 

 of 268.3 calories per gram of lead." The results given were calculated with this 

 empirical formula and agree with those determined by the calorimeter as well 

 as those calculated by Dulong-'s formula. Munroe ^- proposed a "modification of 

 Berthier's process for the valuation of a coal." This has all of the constant 

 error of the conventional process and in general is a more complicated manipula- 



'■'Polytech. Journ., (Dingier) (1835), 58, 391. 



•'' Favre, P. A., and Silbermann, J. T. Ann. clmn. et phys. (1852), III, 34, 

 403. 



^Langbein, H. CJwm. Ztg. (1906), 30, 1116, has pointed out that this 

 method will not even give accurate results for coke. 



=»Stolzel, C. Polyteeh. Journ., (Dingier) (1857), 146, 138; Jahre-sbr. chem. 

 Technol. (1858), 3, 499. 



""Von John, C, and Fullon, H. B. Jahresbr. d. Reichsanst. (1892), 155; 

 Ztschr. f. angew. Chem. (1893), 6, 285. 



=' Noyes, W. A., McTaggart, J. E., and Graver, H. W. Journ. Avi. Chem. Sac. 

 (1895), 17, 847; Analyst (1896), 21, 22. 



'= Munroe, C. E. Am-. Oheni: Journ. (1880-81), 2, 277. 



