76 
INVESTIGATIONS ON ROTH AMSTED SOILS. 
receives potassium salts and on which the crops, in virtue of the 
potash supplied, are able to assimilate more nitrogen and are conse- 
quently more luxuriant. This plat shows for the fourth to the eighth 
depths 19 pounds of nitric nitrogen per acre, while plats 12 and 14 
show in the like depths 29 pounds and 30f pounds, respectively. These 
two latter plats received some potash dressings during the first ten 
years, prior to 1852, and have ever since shown greater fertility and 
greater nitrogen-assimilating power than plat 11. This is probably 
not wholly due to the survival of the effects of the early potash dress- 
ings. These two plats have since 1852 been respectively supplied 
with annual dressings of sodium and magnesium salts, the soil-solvent 
act ion of which has probably liberated natural soil potash. But in 
any case they have grown larger crops than plat 11, assimilating, 
therefore, more of their nitrogen supply — a fact which accords most 
interestingly with the smaller quantity of nitrates in their subsoils. 
HOW FAR DO THE NITRATES FOUND ACCOUNT FOR NITROGEN LOST? 
In Table 26 we constructed a balance account for the nitrogen sup- 
plied to various plats, the nitrogen removed in the crops, found in the 
sin face soil, and not accounted for — the unaccounted for nitrogen 
being described as mainly nitrogen lost in drainage and the small 
quantity possibly accumulated in the subsoil. We took as a standard 
plat 3 (unmanured). If we reconstruct the table, taking plat 5 
(minerally manured but receiving no nitrogen) as the standard, we 
get slightly different but equally comparable figures which are more 
convenient for our present purpose. The table will then be nearly 
on the same lines as that of Sir Henry Gilbert, given in his lectures 
to you (p. 156), in which he estimated the annual additions, removals, 
accumulations, and losses of nitrogen for the plats over a period of 
thirty years — 1852-1881. The table was based on the results chron- 
icled and discussed in the paper on the rain and drainage waters at 
Rothamsted, referred to on page 49, and included an estimate of the 
nitric nitrogen contained in the pipe drainage from the various plats 
based upon a large number of analyses, but not purporting (owing to 
the incompleteness of the analyses) to be more than a probably fair 
estimate. The earlier results were, however, supplemented by the 
results obtained by the analysis of the 1881 soil samples. I have 
endeavored, in a corresponding table, to arrive at an estimate for the 
longer period of fifty years, lasting from the beginning of the experi- 
ment i<» L893. The figures for t lie accumulations of total nitrogen are 
based on the analyses of the 1 s<»:j samples, and the crop estimates are 
l);i-c<l on actual analyses of the produce for forty years, andonanesti- 
mate made for I he crop yields of each plat for the earlier years in which 
analyses of the grain and straw were not made. The average annual 
estimates are probably correel within aboiil 1 pound per acre. The 
estimates of loss by drainage, however, are adopted from Sir Henry 
Gilbert's former estimal es. 
