534 SMITPISOXIAX MISCELLANEOUS COLLECTIONS VOL. 54 



In this combination only the series by Jones and Holmberg are im- 

 portant. The other figures count for little or nothing. Welsbach's ' 

 determinations, also by the sulphate method, cannot be safely utilized, 

 for lack of details. He fo;md ISTd^ 144.55, 144.53 and 144.57; in mean, 

 144.547, ±.0103. 



Peit and Przibylla ' give the following data for ncodymium, obtained 

 bv their volumetric method : 



Mean, 144.516, ± .0111 



Calculating with 8 = 33.0667, ±.00075, we now have two distinct 

 values for neodymium, as follows : 



From the sulphate Nd = 143.752, ± .0057 



From the oxide 144.516, ± .0111 



General mean, Nd = 143.910, ± .0051 



If we assume that Welsbach's latest determinations were based upon 

 essentially the same value for sulphur as that given above, his mean, 

 144.547, ±.0103, may be combined with the other values. In that case 

 the general mean becomes Nd = 144.037. The round number, Kd=:144, 

 is as near the truth as the present evidence will permit us to approach. 

 It is possibly some tenths of a unit too low. 



1 Sitzungsb. Wien. Akad., 112, 1037. 1903. 



- Zeitsch. anorg. Chem., 50, 259. 11X16. For the process, see under lantlianum. 



