230 THE ATOMIC WEIGHTS. 



Since this value is a little under and Cleve's a little over 

 139, the latter figure may fairly be used in all calculations 

 involving a knowledge of the atomic weight of lanthanum. 



DIDYMIUM. 



The atomic weight of didymium has been determined by 

 Marignac, Hermann, Zschiesche, Erk, and Cleve. Mosan- 

 der's early experiments v,^e may leave out of account. 



Marignac * mixed a solution of the sulphate with a slight 

 excess of barium chloride, filtered, weighed the precipitate, 

 and estimated the excess of barium in the filtrate by the 

 ordinary method. The first precipitate always contained 

 didymium, and therefore weighed too much. By deducting 

 the weight of the second precipitate, representing the excess 

 of the barium chloride, from the weight of barium sulphate 

 theoretically formable, the weight of the latter proportional 

 to the quantity of didymium salt taken was found : 



2d BaSO^. 

 .084 grm. 



.075 " 

 .088 " 

 .014 " 



These figures give us a ratio between the sulphates of 

 didymium and barium which we may express as follows. 

 Column A gives the 013(804)3 proportional to 100 parts of 

 BaSO^, as calculated from the first precipitate of the latter. 

 Column B gives a similar ratio calculated with the second 

 BaSO^ precipitate, this being deduced from the total BaSO^ 

 which the chloride used could form : 



A. B. 



82,344 84.685 



82.539 82.626 



82.692 85.545 



82.451 84.425 



82.247— Erk. 



Mean, S4.320, rfc .414 



Mean, 82.455, i .052 



* .A.rch. des Sci. Phys. et Naturelles, (i,) 11, 29. 1840. 



