530 On the action of Dung and artificial Manures 
been imputed to accident ; but no one who saw it repeated on 
these particular rows, in six different stripes across the field, 
could doubt that it was caused by the absence of dung, which 
on such sandy land thus appears necessary for carrying a crop 
through to harvest. The land, as I have said, was very light, 
being in part a shifting sand, so that in one part the seed was 
blown out of the ground ; but the dripping summer was favour- 
able, and the crop weis a very good one. 
At the end of October the roots were taken up, and the pro- 
duce ascertained by weighing the yield of half an acre, measured 
across the six sets of rows at one end. The yield was as fol- 
lows : — 
Tons of cleaned 
Dung. 
Artificial Manure. 
Roots, per Acre. 
No. 1. 
26 loads 
281 
2. 
13 loads 
27i 
3. 
13 loads 
7 CNvt. rape-dust . 
27 
4. 
13 loads 
14 bush, bone-dust 
26 
5. 
13 loads 
7 cwt. rags . 
36 
6. 
13 loads 
3 cwt. guano 
36 
7. 
7 cwt. rape . 
20i 
8. 
14 bush, bones 
20 
9. 
3 cwt. guano 
20i 
10. 
15J 
By comparing the two first lots, it appears that in doubling the 
dose of dung we had gained only 1 ton per acre, which is in fact 
gaining nothing. It is proved, therefore, that on some land, 
though poor, if in high condition, there is a point beyond which 
even a large increase of mere dung ceases to act. But the 
clearest mode of examining the result of the trial will be to di\-ide 
the lots into classes, according to the amount of produce, and it 
will be found that they fall easily into classes — an agreement 
which can hardly be accidental, and leads to the belief that some 
uniform causes have been at work. 
First we must of course take the soil in its actual state, not its 
natural state, for it is naturally poor, but in its then state of 
productiveness as it stood without further assistance. 
Soil unmanured. ISj tons of field-beet. ' 
This will of course form the standard by which to judge the 
effect of the other manures. The artificial manures evidently 
follow next, their produce running very even. Rags used singly 
were accidentally omitted from the trial, but having used them in 
the same field and seen their yield, 1 should put them at the 
same amount with the rest. 
The second class then will stand thus : — 
