CHEMISTRY. 387 



tube, and nitrogen passed over tlie mixture, a cyanide is formed, which, 

 treated in the usual way, gives the Prussian blue reaction. In general 

 he states that when iron b}^ hydrogen and certain non-nitrogenous or- 

 ganic substances arc heated with metallic sodium in an atmosphere of 

 nitrogen, a cyanide is readily formed. [Am. Chcm. J., May, 1881, iii, 

 134.) 



Allary has proposed a simple and effective mode of purifying carbon 

 disulphide, which consists in covering it with a layer of water, adding 

 concentrated solution of potassium permanganate and agitating; re- 

 peating the operation so long as the color is discharged. After wash- 

 ing, it is freed from water and filtered, distillation not being necessary. 

 It should be kept in the dark. {Bull. Soc. Chim., May, 1881, II, xxxv, 

 491.) 



Schiitzenburger and Colsen have described several new comj)ounds of 

 silicon. When crystalline silicon is strongly heated in a current of 

 carbon dioxide the compound (SiCO)^ is produced. When nitrogen is 

 passed over a hot mixture of silicon and carbon (81202^);^ is formed. 

 These bodies the authors regard as the oxide and the nitride respect- 

 ively of the radical carbo-silicon (Si2C2)s. Silicon nitrate (Sig Ns)^ is also 

 described. {Nature, October, 1881, xxiv, 542.) 



Huntington has apj^lied the method of Cooke in the determination of 

 the atomic weight of antimony to the metal cadmium. From the first 

 series of experiments, the atomic weight 112.31 was obtained as a mean, 

 and from the second 112.32. (Awi. J. Sci., August, 1881, III, xxii, 148.) 



Mallet has redetermined with great care the atomic weight of alu- 

 minum. His paper, published in full by the Royal Society, is an admirable 

 example of a thorough scientific research. Three methods were em- 

 ployed: 1st, the'ignition of pnre ammonia alum ; 2d, the precipitation of 

 aluminum bromide by silver; and, 3d, the evolution of hydrogen by the 

 action of aluminum upon sodium hydrate. The greatest care "was taken 

 in obtaining pure materials and \)nYe reagents ; all the operations were 

 conducted with special regard to the elimination of error, and all reduc- 

 tions were made to vacuo. The results were : 1st method, series A, gave 

 27.040±.0073 as a mean of 5 experiments; series B, 27.09Gi.0054, also 

 a mean of 5. 2d method, series A, 27.034±.0049; series B, 27.023± 

 .0052; series C, 27.018±.0069; A and from 3 experiments, and B from 

 5. 3d method, series A, 27.005±.0033, G experiments; series B, 26.990 

 i.004G. As a mean of the whole, the atomic weight is 27.032 ±.0045. 

 But if 1 B be excluded, al=27.019zt.0030. Of the 18 elements whose 

 atomic weights have been carefully determined, ten approach to whole 

 numbers within less than a tenth of a unit. {Phil. Trans., 1880, p. 1003; 

 Am. J. 8ci., April, 1881, III, xxi, 321.) 



Bibart has investigated the conditions under which iron becomes 

 passive. As a resulc he concludes that the passivity of iron is not due 

 to a layer of an insoluble subnitrate, still less to a layer of nitrogen di- 

 oxide, as has been supposed; but it is produced by anj- cause which 



