MOLYBDENUM. 141 



sets of ratios, namely : between MoCln and n AgCl ; between 

 MoCIq and MoS,, and between M0S2 and n AgCl. We will 

 use only the first and last of these ; the probable error of 

 the atomic weight deduced from the second being relatively 

 so high as to make the value connected with it compara- 

 tively unimportant. The analyses of the trichloride, being 

 •discordant, are here rejected. 



By reducing the weighings published by Liechti and 

 Kemp * to a common standard we get the following per- 

 centage results. In MoCL the subjoined quantities of the 

 original substance and of MoS, correspond to 100 parts of 

 AgCl: 



MoCl.,. MoS.^. 



58.299 55.762 



58-194 55-591 



58.524 56.065 



Mean, 58.339, rh .066 Mean, 55.806, ± .093 



Hence MoCl, = 166.902, ± .188, and MoS, = 159.652. 



± .2GS. 



With the tetrachloride similarly calculated we get these 

 figures, corresponding to 100 parts AgCl : 



MoCly MoS.^. 



41.492 . 27.957 



41-319 



Mean, 4i-4055» ± -0583 



Hence MoCl, = 236.914, ± .358, and MoS,, if given the 

 weight of a single experiment in the dichloride series, 

 = 159.964, ± .627. 



* These are as follows : 



.2666 grm. MoClj gave .2550 grm. Mo.S.^ and .4573 grm. AgCl. 



.iSri " -1730 



.2530 " .2422 



.4126 grm. MoCl^ gave .2780 



.1923 " 



.5810 grm. M0CI5 gave .3414 



.2466 " -1441 



.3112 

 .4320 



•9944 



.4654 



1.5222 



.6465 



