278 M. R. Schneider on Equivalents. 



minatious gave numbers between 213 and 209*9. Although 

 the mean of those experiments in which colom'less chloride of 

 bismuth was employed is almost 210'5, Dumas nevertheless 

 gives the preference to the number 210. He justifies this by 

 admitting that the method employed (the chloride of bismuth 

 was decomposed by carbonate of soda, and the filtered solution, 

 after acidifying with nitric acid, was precipitated by a solution 

 of silver) is subject to a loss of chlorine, and that therefore the 

 lowest number found is the most probable. 



Considering the circumstance that the last traces of chlorine 

 are very difficult to remove from chloride of bismuth by the solu- 

 tion of soda, and that, further, on fractional distillation the oxygen 

 and moisture of the air cannot be fully kept away, scarcely a 

 greater value can be attached to the number 210 than to the one 

 hitherto accepted, 208. This was derived by me from a series 

 of 8 experiments (oxidation of pure bismuth), in which the results 

 varied from 207"5 to 208*5. It is worthy of remark, that the 

 number 208 is in very close accordance with that which Dumas 

 himself derived for the series of the so-called nitrogen series : 

 14 + 17 + 176 = 207. 



The equivalent of antimony also, Dumas has sought to deter- 

 mine by the analysis of chloride of antimony. The substance 

 employed was obtained partly by the treatment of very pure an- 

 timony with chlorine, and the repeated distillation of the pro- 

 duct over powdered antimony, partly by fractional distillation of 

 commercial chloride of antimony. Although the results agree 

 pretty vrell with one another (the limits arel21"63 and 122"32), 

 yet here, for similar reasons to those mentioned with respect to 

 chloride of bismuth, some doubt may be allowed concerning the 

 normal constitution of the chloride employed. It is scarcely likely 

 that different people would obtain exactly the same result by 

 this method. It is worthy of notice, that, by an experiment per- 

 formed by H. Rose in the same manner, the equivalent of anti- 

 mony was found = 120*64. This number agrees pretty exactly 

 with that, ISO'S, obtained by me from the composition of pure 

 sulphide of antimony. 



Even the equivalent of iron, which was determined with such 

 great accuracy and with such concordant results (from the com- 

 position of the oxide of iron) by Erdmann and Mai-chand aud by 

 Maumene, to be 28, Dumas has thought it necessary to check by 

 the analysis of the protochloride and perchloride of iron. These 

 determinations (one with protochloride and two with perchloride) 

 have given the result 28-1 [i. e. 351-25 for = 100). The pro- 

 tochloride of iron employed was coloured liyht yellow, and con- 

 tained therefore perchloride or peroxide of iron. It was only 

 after being dehydrated by hydrochloric acid with which some 



