54 REPOET — 188". 



A.l is evidently completed. The condition at the end of 314 days is 

 practically the same as that at the end of 460 days. The potassium 

 chloride retains only a few milligrams of water. At the end of 534 

 days the potassium chloride has gained 1-4 milligram, while the sodium 

 chloride has lost 2'8 milligi-ams. These numbers represent the change 

 between the 260th and the 534th day. The quantities are so small that 

 they may be due to unavoidable errors of experiment. It is to be noted, 

 however, that the period is a hot one, including as it does the month of 

 July, during which the average daily minimum was 66° F., and the 

 average daily maximum 78° F. Other observations made in the course 

 of these experiments have led us to believe that the distribution of the 

 water is appreciably affected by the temperature. It has been shown by 

 Wiillner (Jahresber., 1860, 47-49), that the effect of salts in solution 

 in decreasing vapour tension is increased by rise of temperature. From 

 this it follows that invapoi-ation is more powerful at high than at low 

 temperatures, and we would expect this effect to be different for different 

 salts. 



A.2 is still in progress, but invaporation from the potassium chloride 

 to the sodium chloride is proceeding so slowly (a little over one centigram 

 during the last 221 days !) that the limit must be nearly reached. 



A.3 is not yet completed. Water is still passing from the potassium 

 chloride to the sodium chloride. 



A.4 is almost, if not quite, in equilibrium, and it is to be observed 

 that the ratio in which the water is divided does not depart far from 

 unity. 



A. 5, in progress for a year, shows that after sufficient dilution, 

 potassium chloride invaporates more rapidly than sodium chloride. 



Series B includes the experiments made to show the effect of in- 

 creasing the relative pi'oportion of sodium chloride. A comparison of 

 B.l with A.2 shows that increasing the relative proportion of sodium 

 chloi-ide causes more rapid invaporation, and, when the quantity of water 

 is small, more comjDlete desiccation of the potassium chloride. The 

 quantity of water remaining with the potassium chloride after the ISlst 

 day is fairlj' constant, and it is to be observed that the maxima come 

 immediately after the hot months (vide supra). 



Series D includes experiments V., VI., and VII., of our previous 

 report (see D.l, D.2, and D.3). A study of D.4 shows that, after a 

 certain degree of diluteness is attained, sodium chloride invaporates more 

 rapidly than lithium chloride, and it seems probable that, given enough 

 water, sodium chloride would invaporate as much as, if not more than, 

 lithium chloride. 



A check experiment was made by using two equal weights of sodium 

 chloride placed in tubes of nearly the same diameter, and allowing them 

 to invaporate water. If the conditions were the same in the two tubes, 

 and did not vary from part to part of the enclosed space, invaporation 

 would go on at the same rate in the tubes. The small differences 

 observed are easily explained. In tube No. 68 a small quantity of the 

 salt remained as fine powder on the sides of the tube. This, deliquescing 

 rapidly, exposed a large invaporating surface, and thus during the first 

 period No. 68 gained water more rapidly than No. 67. This advantage 

 disappeared as soon as the solution adhering to the walls became very 

 dilute. Then, during the second period. No. 67 invaporated more 

 rapidly than No. 68. This was owing to the fact that the diameter of 



