144 
INVESTIGATIONS ON ROTHAMSTED SOILS. 
tion or of purely nitrogenous fertilizers in the other, the more is the 
herbage modified in one direction or another, the proportion of true 
gramineous growth falling or rising. The character of both gramin- 
eous and nongramineous herbage — that is to say, the prevalence or 
scarcity, or even disappearance, of various species of either the gra- 
mineous or the miscellaneous nongramineous herbage — is further 
influenced and modified, according to the nature of the nitrogenous 
manure used or fche constituents of the "mineral" fertilizers apart 
from and in presence of each other. The use of full dressings of 
phosphates and alkaline salts — including especially potassium salts — 
without nitrogen is, on the grass land, found to be particularly favor- 
able to the increase and domination of the LeguminossB, presumably 
because these are, in virtue of the symbiotic functions of the micro- 
organisms in their root nodules, able to gather and accumulate free 
nitrogen. 
It is not singular that what applies to the pasture of a grass park 
should also apply to the natural wild vegetation to which, speaking 
as farmers or gardeners, we apply the generic name of "weeds." 
Now even the careful weeding that is carried out on the Rothamsted 
experimental cereal plats has no magic power to prevent the growth 
of weeds during the actual growth of a grain crop after it is tall enough 
to prohibit the use of the hoe; and Hoos field, like other grain fields, 
grows its inevitable share of weeds. It happens that the selective 
action of manures on weeds is especially marked on plat 4 of section 
O. The most prevalent weeds on the other sections are Atriplex 
angustifolia (narrow-leaved orache) and Convolvulus arvensis (small 
bindweed). On section O, and very especially on plat 4 of this sec- 
tion, however, the prevalent weed is Medicago lupulina (yellow tre- 
foil or "black medick"), a leguminous plant which, in presence of the 
full dressing of superphosphate with potassium, sodium, and magne- 
sium salts, is in some seasons very luxuriant. Thus in 1889 it was 
noticed that half of the weight of the " offal" from the barley crop on 
this plat consisted of the mere seed of this plant. During the past 
season — 1000 — the soil of plat 04 was literally covered with this weed, 
while on plats 4A, 4AA, and 4AAS, with the same mineral manure, 
but with liberal nil rogen supply also, there were only a few individual 
plants of 1h is or any other member of the same natural order, the 
weeds being mainly the two species of nonleguminous vegetation 
just referred to. 
It was also not iced that in this season, which was particularly favor- 
able 1o the growth of the Medicago lupulina, plat 01, quite unma- 
QUred, had only about one-tenth of the quantity of this plant found 
on plal 04, while plat 02, manured with superphosphate only, had 
about one-eighth, and plal 03 (potassium, sodium, and magnesium 
salts without phosphates), about one-fifth as much. This weed, then, 
