QUESTION OF THE SOURCES OP THE NITROGEN OF VEGETATION. 51 
take place with heavy manuring for 3 years, and tlie removal of 3 crops of Ijeet and 
one of maize, what would be the result with ordinary manuring and cropping ? 
How are these results to be explained ? The accuracy of the analytical results 
recorded by M. Deherain may be taken for granted. It seems to us, however, that 
in the method of taking the samples of soil for analysis, an exj^tlanation may be found ; 
and we have the less hesitation in suggesting this, since we have found our own early 
results obtained under somewhat similar conditions, quite inapplicable for anything 
like accurate estimates of nitrogen per acre. 
Perhaps it is no undue assumption to suppose that there has been more experience 
of soil sampling at Pothamsted than anywhere else ; and we have, accordingly, learnt 
that very special precautions must be taken, when comparative estimates are to be 
made of the amount of nitrogen in the soil to a given depth, at different periods. 
When this is the object, it is absolutely essential that the samples taken should 
represent very exactly, both the same depth, and the same measure horizontally 
throughout the depth at the two periods. That is to say, it is essential that they 
should contain exactly the same proportions of the corresponding layers at the diflPerent 
dates ; and this is the more important when the layer to be estimated includes both the 
manured and worked surface soil, and some of the unmanured and unworked subsoil. 
Our own plan is to drive down a square iron frame, without top or bottom, having 
an exact measure superficially, and the exact depth for which the result is to be 
calculated. Even when this method is adopted, serious error may arise if at the 
different periods the soil is in a different state of consolidation, the result of manuring, 
the working of the land, the cropping, or the seasons. In other words, it is essential 
that a sample of a given area and dejDth should contain the same weight of dry soil at the 
two periods. W e have given an illustration of the error possible, and of the correction 
necessary, when this is not the case, in a paper we published in 1882.^ 
Now, according to the description of his method given by M. Deherain, he adopted 
the same plan as we did ourselves in our early experiments ; that is, he took his 
samples, not by means of a frame of exact dimensions, but merely with a spade, with 
which it would be quite impossible to take a sample of exactly the same area through¬ 
out the depth adopted. Nor was the depth exactly the same in all cases. It is 
stated that it ranged from 25 to 30 cm. (= 9'8 — 11 *8 inches), whilst the calculations 
per hectare are made for a depth of 35 cm. (= 13 ‘8 inches). It is obvious, that if 
the samples were only taken to a depth of 25 or 30 cm., and upon the results obtained 
the calculations were made for a depth of 35 cm., the amount of nitrogen reckoned 
per acre, or per hectare, must be too high, as the subsoil from 25 or 30 cm. to 35 cm. 
deep would doubtless contain a much lower jiercentage of nitrogen than the layer 
above the depth of 25 or 30 cm. Indeed, M. Deherain’s determinations of nitrogen 
in the subsoils showed less tlian half as high a percentage as in the surface soils. 
* “ Determinations of Nitrogen in the Soils of some of the Experimental Fields, at Rothamstecl, and 
the Bearing of the Results on the Question of the Sources of the Nitrogen of our Crops,” pp. 32 et seq. 
H 2 
