THE BROADBALK WHEAT SOILS. 
121 
9 inches by 78 pounds, making in the 27 inches just over 500 pounds 
per acre, or nearly one-third of the estimated difference. 
We have yet to consider the two farmyard-manured plats 2b and 2a, 
the former of which lias received continuously throughout the experi- 
ments 11 tons of dung per acre per annum, while the latter has been 
similarly treated since lss4only, having been previously unmanured. 
An examination of the results yields the following data: 
Table Cm.— V> rn< idbalk wheat soils, attmples collected in October, ISO!— Plats :b 
and .'a. 
Plat 8b 
(dunged 
for 50 
years). 
Plat 2a 
(dunged for 
9 years; un- 
manured for 
preceding 
41 years). 
Estimate] oxciwuf pntasli p.-r :ut«> a* ,mi>:ir.<l with plat -\ 
1'ountla. 
10,048 
94M 
I'ouiids. 
1,249 
Excess of potash over that iu plat .1, found soluble in strong hydro 
chloric add in first 9 inches 
Excess of potash over that in plat '.i. found soluble in 1 p»-r cent citric- 
acid solution in 
813 
m 
U6 
746 
289 
<>7 
• Second 9 inches 
Third I inches 
l.-.n; 
1,102 
II will be seen that, in the case of the continuously dunged plat, 
only a very small proport ion «»t' t he est [mated excess of potash is found 
by analysis, either by hydrochloric-acid extraction or i>y citric-acid 
exl raci i<»n. The proport ion of potash in the dung is put by sir Henry 
Gilbei-i at <».7;> percent. The difficulty of forming an accurate esti- 
mate of the average constituents of so variable a substance as dung 
has already been discussed In relation to phosphoric acid, and I can 
not help being inclined to think thai the average richness of the dung 
in potash may have been overestimated. An average error of 0.1 per 
cent in the estimate would amount to 1,570 pounds per acre in fifty 
years, and an error of ().:>") per cent to no less than :),'.»l ) <> pounds per 
acre. 
But even after making a liberal allowance for this, t here is clearly, 
a large proport ion of potash not accounted for in the dl rie-acid -soluble 
contents of the soil, which must either have remained in or reverted 
into an insoluble condition, or have descended still lower into the sub- 
soil. It seems probable thai both suppositions are true, for there is 
certainly, in L893, clearevidence of accumulated potash as far down as 
the third depth. 
In the ease of plat 2a, which has only been dunged in recent years, 
the analytical figures accord much better with the estimate, and it is 
obvious thai whatever error there may be in the potash estimate for 
t lie dung has only been mult iplied ninefold instead of fifty fold; audit is 
