244 



SCIENCE. 



[N. S. Vol. V. No. 111. 



being manifestly impossible in the case of 

 iron and steel, we are compelled to judge 

 of the accuracy of the combustion method, as 

 applied to these metals, in some other way. 

 For this purpose, however, we have at hand 

 the results obtained by different chemists, 

 using different methods, but working on the 

 same samples. In the course of the work done 

 by the Committee on International Stand- 

 ards for the Analysis of Iron and Steel, the 

 carbon in four samples of steel was deter- 

 mined : First, bj^ using acid double chloride 

 of copper and potassium as solvent and 

 burning in oxygen gas ; Second, by using 

 the same solvent and burning in chromic 

 acid solution ; and Third, by treating the 

 borings direct with bisulphate of potash and 

 heat, conducting the carbon monoxide and 

 sulphur dioxide formed over hot solid chro- 

 mic acid, which oxidized both gases and 

 retained the sulphur trioxide formed, and 

 finally measuring the volumes of the re- 

 sulting carbon dioxide in an eudiometer 

 tube. Each method was used by a different 

 chemist. The results obtained are as fol- 

 lows: the letters at the side representing 

 the four samples of steel, the figures at the 

 top representing the chemists, and the 

 figures in the columns the percentages of 

 carbon in the steel samples : 



1. 2. 3. 



A 1.455* 1.440* 1.450t 



B 0.815 0.800 0.815 



C 0.450 0.450 0.448 



D...' 0.152 0.185 0.168 



The agreement of the results on the first 

 three samples is quite marked. The dis- 

 crepancy on the fourth sample has not been 

 explained. The matter is discussed in con- 

 siderable detail in reference 1, but we think 

 it safe to conclude that, so far as method 

 goes, the determination of total carbon in 

 pig or cast iron, wrought iron and steel, is 

 reasonably accurate. 



''^Proc Eng of Western Penna. ; 9 [9], 35. 

 ■\Ztschr. anorg. Chem., 4 [3],'und [4], 505. 



The speed of the combustion method as 

 at present worked in good laboratories, is 

 quite remarkable, compared with the pos- 

 sibilities twenty-five years ago. A suflS- 

 cient supply of sample borings being at 

 hand, one operator using two furnaces may 

 readily make from fourteen to sixteen com- 

 bustions in a day of eight hours, it being 

 understood that the bulbs are weighed with 

 oxygen gas in them instead of air, and that 

 the last weight of each combustion, except 

 the last one at night, is taken as the first 

 weight of the succeeding one. It is, of 

 course, assumed that, when turning out the 

 amount of work above described, the fur- 

 naces and apparatus are all in good order, 

 and everything working well. Accidents, 

 an occasional overhauling of the apparatus, 

 blank combustions from time to time for 

 testing purposes, and once in a while an 

 obstinate steel that refuses to dissolve in 

 time or gives trouble in filtration, will all 

 tend to diminish output. The results ob- 

 tained with this rapid work show, when 

 duplicates are made, occasional discrepan- 

 cies as high as three hundredths of a per 

 cent, in a steel containing one per cent, of 

 carbon, but we have seen very large num- 

 bers of duplicates, made as above described, 

 which did not disagree one one-hundredth. 



Again, when work is not so plentiful as 

 to admit of the procedure described above, 

 the method still permits satisfactory speed. 

 Starting with a fresh sample of borings and 

 everything in good order, but cold, it is not 

 difficult to get two closely agreeing deter- 

 minations on the same sample in two hours 

 and a half. Of course, in investigation or 

 referee work more time would, undoubt- 

 edly, be used, especially if the interests in- 

 volved are very great. But we have many 

 times been astonished, in our own labora- 

 tory, at the close agreement between the 

 results obtained in the rapid manner de- 

 scribed above, and the duplicate analysis 

 made on the same sample for confirmatory 



