Centre for Soil Research
Field Research Brief No. 29
TITLE: Continuous Function (many-mini) Trial: Soybean responsee
to Lime, P, K, Mg and Green Manuring
EXPERIMENT NO: 1703 (Kuamang Kuning)
RESEARCHERS: Mike Wade, I P. Gedjer, and Putu Wigena
A series of experimental trials are being initiated by the
Center for Soils Research in a transmigration area in Jarbi known
as Kuamang Kuming, as part of the Trans II program. Because this
is a recently opened area, there has been very little agronomic
research done in the past. From soil analysis we can estimate
the soil fertillity, and probable management practices required
to achieve good crop yields. However site-specific field data is
the most reliable method of determining what nutrients are defic-
ient and the rates of inputs required to minimize or overcome
The purpose of the trials reported here was to rapidly and
economically (in both time and money) establish an estimation of
the fertilizer and lime inputs required to produce healthy, high-
yielding crops. Such information can provide an indication of
the more pertinent fertility research needed for the area, as
well as "fine tune" the range and/or basal rates of inputs for
subsequent research trials.
The design of these "many-mini plots" is non-conventional
and violates certain assumptions and criteria normally applied to
agronomic field research. It is not in the scope of this report
to argue the pro's and con's of the methodology, but we call it
to the attention of the reader so as to be aware of the unusual
design and any implications that may be imparted to the results.
Two experiments were established; one a lime by phosphorus
and the other a potassium by magnesium trial. All fertilizers
and lime were applied at six rates each. For experiment 1 lime
rates were: 0, 373, 750, 1500, 3000, and 6000 kg/ha, and P rates
were: 0, 10, 20, 40, 80 and 160 kg P(as TSP)/ha. For experiment
2, K rates were: 0, 10, 20, 40, 80 and 160 kg K(as KCl)/ha, and
Mg rates were: 0, 4, 8, 16, 32 and 64 kg Mg(as MgSO4.H2C)/ha.
Treatments were made in a rate-continuous, non-randomized,
strip-plot arrangement. Each strip was 1 m wide, and a 1 m
border surrounded each set of treatments. This resulted in a
given replication layout as shown in Figure 1. There were 3
replications for each experiment. However replication 3 in each
trial was given a blanket application of green manure, i.e. 10 t
The soil chemical analysis prior to initiation of the
experiment is given in Table 1. The soil classification is
tentatively an Inceptisol (dystropept?). The site of the trials
was cleared manually in 1983 but logs were windrowed with a
bulldozer. Since clearing no cropping has taken place, as the
land was fallow with mainly grass and some brush regrowth.
Soybeans (cv. Wilis) were planted in November, 1985 after
treatment applications were made. They were sown in 25X25 cm
hills (16 hills/m2) with 3 seeds/hill. Base fertilizer for the
lime-by-P trial was 80 kg K and 32 kg Mg/ha. Base treatments for
the K-by-Mg trial were 3 t lime and 80 kg P/ha. The central four
hills of each m2 were used as the observation plants. As of this
reporting, grain harvest and sample processing were incomplete
and therefore only plant height and pod set data are presented.
Figure 2 shows the soybean response, by plant height and pod
set, to lime and P fertilization. Clearly P deficiency was more
serious than soil acidity in limiting soybean growth. Pod set
more than quadrupled with rates of 80 and 160 kg P/ha. Such a
response is to be expected on a soil with only 5 ppm P (Mehlich I
extract). Liming increased the growth parameters by approximately
40%, and relatively low rates (750 kg/ha) produced near optimum
(plateau) yields. This relatively modest response to line is
also not unexpected, as the soil had only 1.7 meq Al+H and this
is the first crop after fallow.
As has been seen in trials in the Sitiung area, a green
manure application alleviated the lime response, and produced
overall higher yields than the un-manured plots. However the
green manure had little effect on the nature of the P response,
except again to produce overall better growth.
Figure 3 indicates that there was no significant effect of
lime on the P response, i.e. no interaction. At all lime levels
the response was nearly linear to the 40 kg P/ha rate, and rates
of 40 and above were sufficient for near maximum (plateau) pod
Figure 4 shows the effects of the K and Mg fertilizer
applications on soybean growth. There was basically little con-
sistent response to either element. There seemed to be a
tendency for a negative effect of K on the un-manured replica-
tions, while there was a positive effect on pod set when green
manure was applied. The fairly high rates of fertilizer on this
low ECEC soil (3.1 meq/100g) might have produced salt injury.
Whether it was this or some other reason, the green manure appli-
cation apparently alleviated it. As with experiment 1, there was
a very consistent enhancement of both the vegetative and repro-
ductive parameters due to the green manure across both treatment
These trials appear to have been quite successful in
achieving the stated objectives.
A. A number of soil fertility points were elucidated:
1. P deficiency is very serious, and the most critical
(limiting) soil factor for soybean production.
2. Rates of +/- 40 kg P (200 TSP)/ha may be required for
maximum soybean growth.
3. Additional in-depth research on the establishment and
maintenance of adequate soil P is warranted.
4. Liming is necessary for optimizing growth of soybeans,
but rates of only +/- 1 t/ha may be adequate.
5. There was no apparent K or Mg deficiency on this initial
6. Green manuring increased soybean growth for all
treatments (responsive and non-responsive) compared to the
un-manured plots. The practice appeared to alleviate the
soybean response to lime.
7. Further investigations to substantiate the response to
green manure, and why, are warranted. Although only on one
replication, none of the 72 inorganic treatment combinations
equalled the green manuring plateau mean, indicating
something other than soil macro-fertility was enhanced.
B. The trials were very inexpensive, as less than $200 cf
materials and labor (not including technical supervision) were
required for the installation, care and harvest of this first
Table 1. Selected Chemical Properties of the Unamended Soil
Prior to Experimentation.
Rep. H Ca+Ma A+H ECEC Acid Sat.
meq/100ml % ppm
I 4.3 1.0 1.8 2.8 60 3
II 4.3 1.6 1.6 3.2 47 7
III 4.2 1.2 1.9 3.1 58 4
mean 4.3 1.3 1.8 3.0 55 5
t<-------*- 0 -------- Ut
s- _ih4 -
o s- n as 61 Ps
C= -fa.. cv I &>x (.0 Sicp r
1o C4-- 2for .- 1. p/o $
Prlre I. Plot Ug of
Lc o slc desi u'A ect I o .
c L .m Ky C-~ials.
o4 o(4 L'o4 1
,q ,,i .,)
o0 oo o0 go
P RCfes CKJA/L)
I l I I I t
a .?5 1.S 3.0 .6
0 *7s I~r 30 .
Lime a~Q~s ~s
&fegc- of 7' fJ ;/inz 2 o-cy -t, plc'L Pc, Se- 5)
Spf/r e) oye .Ys/Apc s CC.)
-q-- (-g'p 1)S G C!k> S (/ G3,Pa^ w/^u^
- ?- 1
0 210 qo 80
P ca4e s
o4Lrey i 3
( Lp -ePCe =
EfQC4J P e-fs ),- .,
J +i-Aee /eve/f e'I
fCG) IE -
I I I
o 0 o40
I I G b U
Fp 0a w- 4.
fec+ of N/2/,/) g po^dfSo )
a~f p/a.fd helAetf- i) -^ /p k9//> .
ped sefC ) c^ 7/c. hi P (s) s ebas.
I r G
K Rct+es (k-/A)
q c )D g
'6 G- G -- -- -- Ci
-------- it- L 1L4-)