52 JOURNAIv OF THE 



There remained to learn the completeness of the pre- 

 cipitation, Another experiment similar to the above 

 was carried out bearing- in mind the sug-gestion of wash- 

 ing with a solution of ammonium hydroxide, at first 

 with potassium sulphate to be sure that there was an 

 excess of that reagent present and then with ammonium 

 hydroxide. The precipitation had not been complete. 

 After several washings, when the original solution 

 might well be presumed to be removed, or the major por- 

 tion at least, the wash water (very dilute ammonia 

 water) g*ave no evidence of the presence of zirconium. 

 Several experiments were carried out, but in only one 

 case was the double salt weighed. It gave about nine- 

 t}" per cent, of the zirconium really present. The fil- 

 trates from several were examined and it was learned 

 that from one to ten per cent, was always lost, the 

 amount depending* on the exact conditions of precipita- 

 tion and the amount of washing* succeeding*. The ob- 

 jection to using ammonium hydroxide as wash water 

 when it was desirable to separate zirconium from iron, 

 aluminium or titanium, is easily seen. 



The conclusion arrived at was, that the precipitation 

 of zirconium as a double sulphate with potassium afford- 

 ed no quantitative means of determination for that metal, 

 nor of separation from aluminium, iron or titanium. 



/v. By SodiicDi Carbonate. 



Sodium carbonate precipitated solutions of zirconium 

 salts completely. A great difficulty arose, however, in 

 the exceeding slowness of filtration and practical im- 

 possibility of washing the precipitate free from the 

 alkaline carbonate. A single result, obtained from sev- 

 eral analyses, was 



