RICHARDS AND BAXTER. — ATOMIC WEIGHT OP IRON. 259 



In order to determine if the reduction had been complete in the last 

 two analyses of this series, the resulting metal was dissolved in cold dilute 

 sulphuric acid. In each case a few tenths of a milligram of black insolu- 

 ble matter remained, which contained platinum. This impurity probably 

 came from the alloving of the boat with the iron during the reduction, 

 and hence introduced no error. A trace of ferric oxide also was found. 

 Even if the platinum had been originally present, the combined weight 

 of the two impurities could not have raised the final result more than 

 0.01 per cent; hence no correction was applied. 



One of us* has recently shown that iron when ignited at a high tem- 

 perature, has practically no tendency to occlude hydrogen, so that in the 

 present case this possible source of error is excluded. On comparing 

 these results with the two determinations of the first series, one is forced 

 to conclude that the oxide made from the nitrate has essentially the same 

 (•(imposition as the oxide made from the hydroxide. The higher value, 

 indicated by Experiment 2, was probably caused by incomplete reduction, 

 and the first experiment agrees with the average of* the second series. 



It is interesting to review the older determinations f in the light of the 

 experience gained in these analyses. Evidently the occlusion (or better, 



* Baxter, Am. Chem. Journ., 22, 363 (1899). 



t A complete list of these determinations is given below : — Fe = 



1814. Wollaston, from Thenard's Results, Phil. Trans., 104, 21 (1814) . . 5-3.2 

 1826. Stromeyer, Pogg. Ann, 6, 471 (1826). Reduction Fe 2 3 : Fe . . . 5-3.5 

 1826. Berzelius, Pogg. Ann., 8, 185 (1826). Oxidation Fe 2 : Fe 2 3 . . . 54.3 

 1830±. Wackenroder, Archive Pharm., 35, 279; 36, 22 (1813). Reduction 



Fe 2 l ';. : Fe 2 55.9 



1844. Svanberg and Norlin, Liebig's Ann. Chem. Pharm., 50, 432 (1844), 

 also Berzelius Jahresbericht, 24, 121 (1845), and 25, 42 (1846). 



Oxidation Fe 2 : Fe., 3 55.87 



Reduction Fe 2 3 : Fe 2 50.09 



1844. Berzelius (the same references). Oxidation Fe 2 : Fe 2 O a .... 50.02 

 1844. Erdmann and Marchand, J. Pr. Ch., 33, 1; Lieb. Ann., 52, 212 



(1845). Reduction of oxide from oxalate Fe 2 3 : Fe._> 56.02 



1846. Mauraene', Ann. Chim. Phys. [3], 30, 380 (1850). Oxidation Fe 2 : 



Fe 2 O s 56.00 



1850. Rivot, Ann. Chim. Phys. [3], 30, 192, (1850). Reduction Fe,, 3 : Fe 54.2 

 18c9. Dumas, Ann. Chim. Phys. [3], 55, 157 (1859). Lieb. Ann.. 113. 26 



(1860). Analysis of Chlorides 56 2 



[The results of Magnus, Dobereiner, Capitaine (1841), and Torrey (1888), alluded 

 to by Clarke, are not worthy of serious consideration. See Smithson. Misc. Coll. ; 

 Constants of Nature, 5, 287 (1897) : also 4, 65 (Becker, 1880)]. 



From some of these values Clarke (1807) computed the value 56.021, while 

 Messrs. Landolt, Ostwald, and Seubert [1898] chose the number 56.0. 



