46 



On the Atomic Weight of Titanium. [Nov. 22, 



instead of merely taking the means of the several observations, the 

 influence of the larger quantities is directly felt. 



Assuming with Lothar Meyer and Seubert that the most probable 

 ratios of Ag, CI, 0, and H, are as follows — 



Ag=6 7456 

 CI =2 -21586 

 O =1 



H =0 '06265 



the various experiments afford the following values for Ti : — 



TiCl 4 



4Ag 



= 34 



•39004 : 



78 -16399 





= 1 



•75989 



4 



TiCl 4 : 







= 11 



•8715 : 



1 



Ti 







= 3 



•0081 ■ 



1 



Ti 



H 



=48 



•014 



1 



TiCl 4 



: 4AgCl 



= 22 



•93796 ■ 



69 -25983 



= 1 



•32475 



4 



TiCl 4 







= 11 



•8716 



1 



Ti 







= 3 



•0082 



1 



Ti 



H 



=48 



•016 



• 1 



TiCl 4 



: Ti0 2 



=49 



•29948 



20 -79032 







= 2 



•37124 



• 1 



Ti 







= 3 



•0053 



1 



Ti 



• H 



=47 



•969 



: 1 



On the assumption that these values have equal weight, the final 

 value becomes — 



1 48 -014 



IT 48 -016 



III 47 -969 



Ti=48 -000 



It would appear, therefore, from, these observations, that titanium 

 must be added to the increasing list of the elements whose atomic weights 

 ■are simple multiples of that of hydrogen. » 



In these observations, which have occupied me many months, I 

 have sought to eliminate such sources of error as were known to me- 

 lt is of course possible, in spite of the agreement between the 

 several values, that the results may be affected by undetected and 

 constant errors. Experience warns us that no determination of 

 atomic weight, however well the individual observations may agree 

 among themselves, can be considered wholly satisfactory if it depends 

 upon a single reaction or is referred to a single relation. It is for this 



