215 



Precipitate = .09026 .05015 .09156 .05034 



Anhydrous precipitate = 64.28% AlgOg 64.52% Al.Og 



= 35.72% SO3 35.48% Sdg 



Precipitate = 83.84% of total Al.Og 85.05% of total Al.Oj 



It will be seen that the SO^ determinations do not differ appreciably 

 more than the duplicate determinations of SO3 in the same solution ; 

 while the amounts of ALO. obtained, differ by more than a milli- 

 gramme, one part in seventeen. So in another instance, .2893 and 

 .2899 grm. SO3 were obtained in two experiments entirely distinct 

 from the beginning. It would almost seem that the equilibrium were 

 more sharply defined with reference to SO. than to ALOg. Yet it 

 should be borne in mind that the amounts of SO3 actually weighed are 

 always lai'ge, .28 grm. or more, while the amounts of Al^Og weighed 

 were as low as .01 grm., and hence the experimental determination of 

 SO3 is less affected by error than that of AlgOg. 



The final results being dependent on the ratio between differences, 

 small compared with the quantities actually measured, considerable 

 latitude of variation must be allowed. 



The experiment was then repeated, E'", the only variation being 

 that, after the flask was filled to 250 c.c. (at 21°), it was heated to 

 about 50-55° for twenty-five minutes, then turned into a large 

 beaker with glass cover and heated fifteen minutes more, and finally 

 boiled two or three minutes. The liquid on the cover was rinsed off 

 with the solution, and the whole thrown while boiling hot upon a dry 

 filter, which was covered during the ten minutes of filtration. The 

 filtrate was then cooled to 20°, and portions of 50 c.c. taken as usual 

 for analysis. The loss of water by evaporation could not have been 

 great, and would not affect the ratio between AI2O3 and SO3. The 

 results were : 



AUG,. SO3. 



I. = .0148 .30473 



II. = .0148 



Mean = .0148 .30473 



