On the Rain and Drainage - Watei-s at Rothamsted. 
33 
runnings in the autumn of the same year ; see Table XLVI. 
The increase of chlorine observed in these instances is much 
the greatest in the case of Plots 3&4, 5, 9, 17, receiving no 
chlorides. Supposing soluble salts have been washed below 
the level of the drains, it is natural to suppose that they would 
to some extent rise again by diffusion when the downward 
passage of water ceases, or that the water containing them 
in solution would be brought again to the surface by capillary 
attraction during dry weather. Simple diffusion would seem, 
in the examples just quoted, to have been the most active 
agent at work, the increase of chlorides being by far the most 
considerable in the case of plots very poor in this constituent ; 
these plots have clearlv, during the period of rest, obtained 
chlorine from the subsoil water. 
The average fisrures at the foot of Tables XLVI, and XLVII. 
are the means of 49 analyses in the case of Plots 5, 6, 7, 9, 11, 
12, 13 ; of 48 analyses in the case of Plots 3 & 4, 10, 17, 18 ; 
of 47 in the case of Plots 8, 15, 16 ; and of 43, 32, and 25 
analyses in the case of Plots 14, 19, and 2 respectively. 
The average amounts of chlorine are, with one exception, 
considerably below the earlier results of V'oelcker and Frank- 
land (Table XLV.). This is chiefly due to the very wet 
character of the last three seasons, and the consequent dilution 
of the drainage-waters. 
Looking first at the plots receiving no chlorides in the 
manure, we see that the unmanured Plot 3 & 4 has given on 
an average 4-9, and the unmanured Plot 16, 5*1 of chlorine per 
million of drainage-water. Plot 5, receiving mineral manure, 
gives 5 2 ; and Plot 9, with nitrate of sodium, and a dressing 
of mineral manure on half its surface, gives 5*7 of chlorine per 
million. 
The absence of any considerable amount of chlorine in the 
waters of these plots is the practical test that a mixture of 
drainage-waters has not taken place, but that each water fairly 
represents the plot to which it belongs. Plot 9 is most capable 
of acting as a test plot in this respect, lying as it does between 
two plots receiving chlorides, and one of them (Plot 8) receiving 
the largest amount of chlorides of any in the field. In earlier 
years the drain-water from Plot 9 occasionally contained much 
chlorine, probably through mixture with surface-water. In a 
running on December 8, 1868, Frankland found 60 0 per million 
of chlorine ; in a running on April 11, 1878, we found 30-3 per 
million. Since, however, the improved drainage arrangements 
have been in action, the highest amount of chlorine found in the 
drainage for this plot has been 14- 7 per million, which occurred 
VOL. XVITI. — S. S. D 
