Solubility of Salts in Water. 375 



that the water present should not be lost by evaporation, 

 otherwise the composition of the solvent solution would be 

 altered. The thoroughly efficient method of mixture, by 

 blowing air through the solution, previously employed by me, 

 could not therefore be applied. Recourse was therefore had at 

 first to what may be termed saturation by gravitation. Tubes of 

 the form shown in fig. 1, PL VIII., were made, and were used 

 in the following manner. The bend a of the capillary-tube was 

 filled with mercury. The solvent solution was poured in till 

 it just covered the surface of the platinum gauze b. The salt, 

 previously fused and cast into thin sticks, was then placed on 

 the gauze and the open end of the tube closed with a cork. 

 The whole was then placed in a wide test-tube, which was also 

 corked and placed up to the mouth in the water of a bath, kept 

 at the constant temperature of 20° C. by the method previously 

 described by me. After the lapse of twenty-four hours the 

 saturation-tube was quickly withdrawn. The mercury was 

 expelled by blowing into the open end ; and in the same way 

 a portion of the saturated solution was transferred to a weighed 

 nickel crucible, and the amount of salt dissolved was deter- 

 mined in the usual way. 



The above method yielded excellent results, but was some- 

 what troublesome, and was after a time discarded for one 

 which did not require fusion of the salts employed. In a 

 constant-temperature bath similar to the previous one was 

 placed a copper water-wheel, to the axle of which were 

 attached clips to receive the saturation-tubes. The wheel was 

 caused to revolve by means of a current of air entering at the 

 bottom of the bath below the inverted buckets of the wheel. 

 The revolution of the wheel caused such a circulation of the 

 water in the bath, that the temperature at all parts was the 

 same, and never varied more than + o, 02 C. from 20° C. 

 The wheel revolved, on an average, 18 times a minute, or 

 1000 revolutions per hour ; and as saturation extended over 

 a period of 20 to 24 hours, the salt and solvent solution con- 

 tained in the saturation-tubes attached to the axle of the wheel 

 were shaken together 20,000 to 24,000 times. The solvent 

 solution and the salt in excess were placed in narrow test- 

 tubes closed with rubber stoppers, and enclosed in wider tubes, 

 also closed, and placed in the clips on the axle. After satu- 

 ration had gone on for the required time, the tubes were 

 removed from the wheel and placed in an upright position in 

 the bath, and, when the excess of salt had settled down, 

 leaving the solutiou perfectly bright and clear, a portion was 

 quickly decanted into nickel crucibles and estimated in the 

 usual way. In every case two or more determinations were made . 



