RICHARDS. — U-RXTS OF COMBUSTION OF ORGANIC SUBSTANCES. 583 



The purity of the sugar is indicated by the essential identity of the 

 results obtained from three difterent samples. 



In Determinations Nos. 7 and 8 the bomb was exhausted before run- 

 ning in the oxygen, and in Determinations Nos. 9 and 10 the amount 

 of nitrogen was increased to about 13 per cent by volume in the 

 compressed atmosphere of the bomb, by pumping in air in the first 

 place, and then supplying the oxygen to 35 atmospheres pressure. 

 The essential identity of these results with the others shows that the 

 presence of even this large amount of inert gas does not aflfect the 

 heat of combustion of a substance so easily burned as sugar. 



In these experiments it appears probable that the greatest experi- 

 mental error in a combustion lay in the reading of the thermometer. 

 Probably with a scale divided to jl-g°, even with a good lens, one may 

 make an error of 0.001°. Should this error be made in the same direc- 

 tion in the two extreme readings, the errors would cancel each other. 

 However, should the errors be made in opposite directions, then in a 

 rise of two degrees this error in reading the thermometer would amount 

 to a percentage error of 0. 1 per cent. The average result should be 

 much more accurate than this, however. The agreement of the re- 

 sults of cane-sugar is in accord with this consideration of error. In 

 the future, measurements of the more important substances, when 

 other conditions justify a greater refinement, will be made with a 

 platinum resistance thermometer. 



The Combustion of Benzol. 



The first problem to be solved in the combustion of a volatile liquid 

 was to devise a method by which an accurately weighed quantity of it 

 might be burned. Berthelot ^ determined the heat of combustion of 

 benzol by saturating cellulose with the liquid, which was then ignited 

 in the bomb. This method of procedure is evidently open to the 

 error of a varying loss in weight of benzol by evaporation. Besides, 

 since in this case some of the benzol must burn as vapor, which of 

 course gives a greater heat of combustion than liquid benzol, a cor- 

 rection should be subtracted accounting for the heat of evaporization 

 of that portion of the benzol burned as vapor. These objections ac- 

 count, in part at least, for the irregularity of his results. 



Julius Thomsen * burned benzol both as vapor mixed with air and 

 as a liquid in his Universal Burner. By these methods he obtained at 

 various times four series of results, one of which was not at all in 



3 Ann. Chim. et Phys. (5) 23, 193 (1881). 

 * Thermo. Untersuch., 4, 59. 



