M. Dumas on the Equivalents. 211 



ammonia was expelled. A source of error in previous determi- 

 nations consisted in using for the reduction vessels of platinum, 

 with w^hich the tungsten forms an alloy. Glazed porcelain also 

 favours the formation of an oxide, which combines with the 

 glaze, and remains unreduced by hydrogen. This w^as avoided 

 by using for the reduction tubes and trays of unglazed porcelain. 

 Another difficulty lies in the readiness with which tungstic acid 

 is carried away in the aqueous vapour formed by the reduction. 

 After all the sources of error had been avoided, results were 

 obtained which gave the number 92 as the equivalent of tung- 

 sten. Hence the equivalents of molybdenum and tungsten do 

 not stand in the relation 1:2; but an instance of this relation is 

 found in the equivalents of oxygen and sulphur, which are as 

 8:16. The equivalent of sulphur has also been redetermined 

 by Dumas, by passing the vapour of perfectly pure sulphur over 

 pure silver heated to redness, and ascertaining the relation be- 

 tween the weight of the original silver and that of the sulphide 

 formed. Previous determinations had given the numbers 16'06 

 and 16" 10; Dumas finds that the equivalent is 16. 



As instances of metals which have the same equivalent, Dumas 

 adduces manganese and chrome, with the equivalents 26. He 

 states that he has redetermined the equivalent of manganese by 

 reducing perfectly pure artificial binoxide of manganese to prot- 

 oxide by means of hydrogen. His experiments leave no doubt 

 that the equivalent is 26. 



With many groups of three elements it is the case that the 

 equivalent of the middle is exactly the arithmetical mean of that 

 of the extremes. For instance, sulphur and tellurium have re- 

 spectively the equivalents 16 and 64; the mean of this is 40, 

 which is very nearly the equivalent of selenium. Similar relations 

 exist between calcium, strontium, and barium, and between 

 lithium, sodium, and potassium. Nevertheless, with certain 

 triads this is not the case. Chlorine, bromine, and iodine have 

 the most marked analogies, not only in the physical and chemical 

 properties of the elements themselves, but also in those of their 

 compounds. Yet their equivalent weights stand in the relation 

 35'5 : 80 : 127 ; and in order to exhibit the numerical relation, it 

 would be necessary to raise the atomic weight of bromine to 81, 

 or to lower that of iodine to 124-5. Dumas has controlled the 

 equivalents of bromine and of iodine by converting the bromide 

 and the iodide of silver into the chloride of silver. The iodide 

 of silver was obtained by converting pure iodine in large crystals 

 into iodide of zinc, and this into iodide of silver. Bromine free 

 from iodine was digested for some time with bromide of silver, to 

 free it from a possible admixture of chlorine, and was then con- 

 verted into bromide of silver. 



P2 



