Reilly & Ralph — The System n-Butijl Alcohol— Acetone— Water. 599 



1884, 45, 444). Densities were obtained at 20° C. and then calculated to 



20° 



—jy Only in the case of pure liquids were the densities reduced to vacuum. 



The method adopted was to correct to vacuum each of the weighings (cf. 

 Reilly and Rae, Science Progress, 1918, 47, 434). The densities of the air at 

 the times of the three weighings were found by interpolation from the table 

 given by Wade and Merriman (Trans. Chem. Soc, 1912, 103, 2429). In all 

 the densities recorded, the method of double weighing was adopted. The de- 

 terminations were usually carried out in duplicate, the average variation in 

 the results being only two or three units in the fifth place. 



Several quantities of pure n-butyl alcohol were prepared during the period 

 of this investigation, and the following densities were obtained with separate 



20° 

 freshly-made portions :— 0-81097, 0-81099, 0-81096 at ^. The densities 



agree within the limits of experimental error. 



Preparation of mixtures. 



The samples of both n-butyl alcohol and acetone set aside for the pre- 

 paration of mixtures were kept in a large desiccator over anhydrous calcium 

 chloride. Distilled water, which had stood for several days in a partial 

 vacuum to remove dissolved air, was used in making up the mixtures. All 

 mixtures were made with weighed quantities of the components, the calcu- 

 lated quantities being run into the weighing flask from burettes fitted with 

 calcium chloride tubes to prevent the possible addition of moisture to the 

 acetone or alcohol. 



In preparing tables of densities of acetone, alcohol, and water mixtures, 

 the following procedure was adopted : — To a mixture of pure acetone and 

 water in known proportions, a definite measured quantity of n-butyl alcohol 

 was added. By varying the proportions of any of the constituents a series 

 of mixtures was obtained. For the purpose, however, of obtaining a homo- 

 geneous mixture, the proportions were so arranged that there was no 

 separation into two layers. The densities of these mixtures are given in 

 table I, and a series of curves was plotted (fig. 1) by interpolation from these 

 results. These results were also plotted with triangular co-ordinates (fig. 2) 

 from which for any observed densities direct readings of possible mixtures 

 of n-butyl alcohol, acetone, and water can be obtained. A knowledge of the 

 density alone of a mixture of n-butyl alcohol-acetone-water cannot be used 

 to determine its constitution. When, however, the proportion of one of the 

 constituents is known, the composition of the mixture can be settled by a 

 reference to the densities plotted in fig. 2. 



5 b 2 



