148 On the Composition of Wdtcrs of Land-Drainage. 



samples is, no doubt, due to the sulphate of ammonia with 

 which some of the plots were manured every year, for it is well 

 known that when a solution of sulphate of ammonia is filtered 

 through a soil containing; carbonate of lime, the sulphate is 

 decomposed ; and whilst the ammonia is retained in the soil the 

 sulphuric acid contained in the sulphate of ammonia passes away 

 in combination with lime. 



As regards the distribution of rain between December 6th, 

 1866, and May 21st, 18G7, it may be stated that a good deal of 

 rain fell in December. In January we had heavy falls of snow, 

 alternating with rapid thaws, warm weather, heavy gales, and a 

 good deal of rain. A large amount of rain fell in the beginning 

 of February, and a moderate quantity during the remainder of 

 the month. A good deal of snow fell in March. April and the 

 beginning of May were unsettled, and there was rather more 

 than the average fall of rain. 



Third Series of Drainage-Water Experiments. 



During the summer months little or no water passes from the 

 land into the drains, nor could sufficient be collected from all 

 the Experimental Plots on Mr. Lawes's Broadbalk wheat field 

 to enable me to make a full and strictly compai'ative series 

 of analyses. 



In January, 1868, all the drains at Rothamsted began to flow, 

 and on the 13th of that month Dr. Gilbert kindly collected 

 for me samples froin 14 different plots, which, like the previous 

 samples, were submitted to careful and full analyses, when the 

 results shown in Table III. (pp. 150, 151) were obtained. 



The preceding analyses suggest the following observations : — 



1. The total amount of solid constituents in the drainage- 

 waters fi'om the different Plots of the Experimental Wheat-field 

 varied from 18'11 grains to 57*55 grains in the impei'ial gallon. 

 The drainage from the highly manured plots in some instances 

 contained o times as much saline matters as that from the 

 continuously unmanured Plots 3 and 4. On Plot 16, highlv 

 manured with mineral and ammoniacal manures previous to 

 1865, and since that year left unmanured, the drainage con- 

 tained somewhat more solid constituents than on the continuously 

 manured plots, but less than on the remainder of the experi- 

 mental plots. 



2. The drainage-water from all the plots of the same .field 

 contained merely traces of ammonia, although ammonia-salts 

 had been applied to several plots in large quantities. 



3. Appreciable quantities of nitric acid occurred in all the 

 samples, even in those from the unmanured Plots 3 and 4. 



