i9°s-] SMITH— COLUMBIUM AND TANTALUM. 155 



TaF5.CsF 2TaF5.3RbF 



TaF5.2CsF TaF5.2NaF 



TaF5.2NH4F TaF5. 3 NaF 



TaF5.3NH 4 F TaFs.2KF 



The existence of several such double fluorides with each of the 

 alkali metals naturally raises the question whether these salts ought 

 to be used for the determination of the atomic weight of tantalum, 

 inasmuch as each salt is likely to be contaminated with smaller or 

 larger quantities of the other, depending upon the condition or the 

 care with which they are prepared. Marignac used potassium 

 tantalum fluoride and ammonium tantalum fluoride in his re-deter- 

 mination of the atomic weight of tantalum. It would seem, from 

 the study of the salts just mentioned, that even this skilled and 

 careful analyst could not have been sure that he had a definite, 

 homogeneous body in the determinations which he made. Of 

 course, if there was even a slight amount of a second salt in the salt 

 used for the atomic weight work, it would naturally vitiate the final 

 result. Of all the double- fluorides of the alkali metals and bases 

 with tantalum which have thus far been studied by Mr. Balke, that of 

 sodium and tantalum, of the ratio 3 to 2, seems to be the one having 

 some definite and most stable characteristics. We hope to re-deter- 

 mine the atomic weight of tantalum, but it is not probable that we 

 shall use any one of the double fluorides, of which mention has been 

 made, although they appeal strongly because of the ease with 

 which they can be crystallized. The uncertainty, however, as to 

 whether they are really absolutely of one definite ratio every time 

 that they are crystallized is uncertain. Hence they had better be 

 abandoned in atomic weight determinations. 



The question may also be asked, may not the double fluorides of 

 colurnbium with the alkali metals, which have been used for atomic 

 weight purposes, been contaminated with salts of varying ratios ? 

 This point will receive attention. 



Turning again to colurnbium, it seems proper to record that 

 having eliminated the tantalum completely from a mixture of oxides 

 obtained from Haddam columbite, the remaining potassium colurn- 

 bium oxy-fluoride was crystallized a number of times from water 

 and also from solutions containing much hydrofluoric acid. This 

 procedure finally gave a mother liquor that was decidedly acid. A 

 metallic acid remained in this mother liquor. According to Her- 



PROC AMER. PHILOS. SOC XLIV. l8o. K. PRINTED JULY 31, I905. 



