The On-Farm Research Program for palawija based farming systems in Malang uisriL.,,
East Java, Indonesia, 1984-1986
Heriyanto Marsum Dahlan Sri Wahyuni 3unarsedyono C.E. Vag Santen ,
J. Ph. Van Staveren and L.W. Harrington
This paper summarizes work conducted to date in on-farm
research (OFR) by the Malang Research Institute for Food Crops
(MARIF), located near Malang, East Java, Indonesia. MARIF is one
of the six research institutes operating under the Central
Research Institute for Food Crops (CRIFC), a dependency of the
Agency for Agricultural Research and Development (AARD), the
Indonesian organization with responsibility for all agricultural
research within the Ministry of Agriculture. MARIF's research
mandate covers "palawija crops" (non-rice food crops)
particularly maize, soybean, other grain legumes and cassava.
The MARIF OFR program was initiated in January 1984 in
cooperation with the International Maize and Wheat Improvement
Center (CIMMYT) and the ATA 272 Project of the Agricultural
Technical Cooperation Program between Indonesia and The
Netherlands. The interdisciplinary MARIF OFR team consists of
plant breeders, crop protection specialists, agronomists and
The MARIF OFR program has two major objectives : 1. Try out
OFR procedures on a pilot basis in order to ascertain their
appropriate role in MARIF activities, 2. Develop useful
recommendations that farmers in a study area would rapidly adopt.
Because of the pilot nature of the program, it was decided to
initially restrict OFR activities to one relatively homogeneous
study area in Malang district. As the program developed, -OFR
work has begun in other study areas as well. However, this
report only discusses the original OFR pilot effort.
MARIF OFR activities in the study area to date include an
initial exploratory survey, soil analysis, a formal production
survey, additional surveys on specific issues, and five crop
cycles of on-farm experiments. Currently a sixth cycle is being
planned. With two cycles of experiments per year, research has
been underway for about two and a half years. These activities
have been described in detail in a series of nineteen working
papers (see Annex 1).
The MARIF OFR program is based on the procedures for "on-
farm research with a farming systems perspective (OFR/FSP)"
developed by CIMMYT, documented in Byerlee, Collinson et al
(1980) and numerous other publications. The major steps in these
procedures are diagnosis, planning, experimentation, assessment
and formulation of recommendations. The guiding principles of the
OFP/FSP procedures are:
1. A careful focus on important problems and possible solutions
2. On-farm research under representative conditions
3. A focus on defined groups of farmers
4. Farming systems perspective
5. Near ter solutions.
S. Participation of farmers and extension service
- Malang Research Institute for Food Crops (MARIF), Malang, Indonesia
SCIIMYT Economics Program, Bangkok
7. Interdisciplinary cooperation among biological and social-
8. Feed back between on-farm and on-station research.
This paper contains a description of the study area in
Malang, followed by a summary of the diagnosis of farmers'
circumstances, practices and problems. This is followed by a
discussion of the evidence obtained from farm surveys and on-farm
experiments regarding possible solutions to the more important
problems. The paper concludes with an overview of the evolution
of research themes and priorities.
2.0 DESCRIPTION OF THE STUDY AREA
The pilot study area is located in Malang district, E. Java
province. The location of this district is shown in Figure 1.
Farmers in the Malang district operate reasonably complex
farming systems, featuring intensive use of a small land resource
base. Major groups of enterprises include livestock enterprises,
annual crop enterprises, and home garden/perenniel crop
enterprises. MARIF's research mandate is restricted to annual
palawija crop enterprises. Where these interact with livestock
and home garden enterprises, the fact will be noted in subsequent
2.1 Crop-production systems in the Malang district
The crop production systems in Malang district are heavily
influenced by land type sawah vs tegal. Sawah systems are
dominated by wetland rice cultivation, although palawija crops
("secondary", non-rice food crops) are at times grown after rice.
Tegal fields (unbunded, rainfed fields) are responsible for the
bulk of palawija production. A summary of major systems is shown
in Table 1.
For the initial pilot OFR program, the second system was
chosen. This is a relatively simple system (the first, more
complicated system was chosen for study during the expansion of
the MARIF OFR program). Major characteristics of the selected
system are: young volcanic soils; tegal land type; medium
altitude; maize-maize or upland rice-maize cropping pattern.
Maize is by far the most important enterprise in this crop
The chosen system (henceforth referred to as the "study
area" covers around 30,000 ha of physical area in Malang and
includes an estimated 40,000 farms, each operating around 0.8 ha
of tegal farmland (physical area). When extrapolated to other,
similar districts of East Java, the area should cover 100,000 to
150,000 ha and would include up to 200,000 farms.
It should be noted that some farmers (about 28 % of study
area farmers) operate on both tegal and sawah. That is, they
operate system 3 as well as system 2. This leads to certain
H 1, y of toIdmesa m= nat Ja *.
. II.mo Feotlim .t em t op a- "mg sM l wltiw t
system interactions, including increased labor scarcity in the
early stages of the first crop cycle, a lower dependence on maize
as a starch staple, etc. In so far as they affect technology
adoption, these interactions will be pointed out again where
relevant in the sections on diagnosis.
Table 1. Major crop production systems in Malang district
System Land Soils Cropping Altitude Physical Ar--
No. type Pattern (+ aeter) (%)
1. tegal limestone cassava +
-legume) <- 600 43
2. tegal young maize- 400-700 37
3. sawah alluvial transplanted
& young rice-maize
volcanic 400-700 15
4. tegal young maize and
& volcanic tural crops
ash 400-1500 5
Apart from this, the farmers in the study area confront
reasonably similar circumstances, problems and opportunities and
for most purposes (but not all) would likely use similar kinds of
2.2 Physical Conditions
Climate: The average rainfall in the study area amounts to
2130 -m per annum with 5 to 6 wet months (over 200 ma rain/month)
and 2-4 dry months (less than 100 of rain/month). The average
temperature is 24 centigrade with average minimum of 18.6 and
maximum of 25.7 centigrade. The relative humidity ranges from 77
to 85 %. A monthly rainfall distribution is not shown here
because of space limitations.
Soils: The main soil categories in the Malang district
consist of: young volcanic soils (36 %), volcanic ash (18 %),
limestone and lithosoils (37 %) and alluvial (9 %). The study
area focuses on the young volcanic soils which consist of
latosols (60 %), regosols (26 %) and other soil types (14 %).
All of these soils, including the latosols, are known to be
generally deficient in organic material and in P, and are often
deficient in other plant nutrients (K, 3, Mg, etc.) as well.
-2.3 Economic Circumstances
There are a number of socio-economic circumstances that
influence farmer decision-making in the Malang study area. Of
interest are input markets, government price policy for inputs,
marketing, and consumption patterns and taste preferences. Given
the predominant position of maize in the crop production system,
economic circumstances that affect maize are highlighted.
Input markets: Input markets in the Malang area operate
reasonably well. Fertilizer and pesticides are readily available
at highly subsidized prices. For example, a farmer in Malang can
buy one kg of nitrogen in the most readily available form (urea)
in exchange for only 2 kg of maize. In contrast, a farmer in
Thailand can buy one kg of nitrogen in the most readily available
form (21-0-0) in exchange for 9.5 kg of maize. In the past,
improved maize seed was not readily available to farmers in the
study area. This situation has improved considerably since 1983,
when government policy was changed to encourage private firms to
participate in seed production and marketing. Currently, two
hybrids and one variety (Arjuna) are commercially available.
Other input markets (labor, animal traction, etc.) are reasonably
Marketing: The marketing of maize for cash sale is largely
in the hands of private traders, who compete to purchase maize in
order to meet the rapidly expanding demand for maize by feed-
mills in Surabaya. Other maize is marketed through the government
marketing agency (BULOG).
Consumption patterns and taste preferences: Maize
production in Malang has two major end uses: it is a major
starch staple (especially for low income producers) as well as an
important source of cash. About 70 % of the maize that is
produced is used for direct human consumption. Maize is the main
non-rice starch staple in all of East Java. In the study area,
many farm families consume more maize than rice. The importance
of maize as a starch staple raises some concern about cooking and
taste preferences, and how these might effect farmer adoption of
new varieties. In East Java, farmers prefer yellow varieties.
Both traditional and improved varieties, such as Arjuna and the
Cargill C-1 Hybrid, have a yellow grain colour. There are
reports, however, that Arjuna may be somewhat-more difficult to
crack than other varieties for the preparation of local dishes.
These dishes rely on the boiling of maize cracked to resemble
2.4 Crop-livestock interactions
Crop-livestock interactions in the study area take three
1. Cattle provide draft power, for land preparation and weeding
2. Cattle manure is applied as fertilizer to selected fields
3. Maize thinning serve as one source of livestock feed.
Manure: In any given year, about half of the farmers apply
manure to their maize fields. This takes place before the first
plowing for the rainy season crop. Those farmers who do not apply
manure tend to be those without livestock, or those who use the
manure for higher-valued crops. At least once every few years
nearly all fields used for maize cropping receive manure. Manure
use is of interest because of reports of P, K, S and other soil
Fodder: It should be noted that maize thinning do not seem
to be an important source of cattle fodder. Other fodder sources
(including grass and weeds) are reported to be more important,
even in the maize season. Furthermore, all farmers with
livestock report having adequate fodder, even in the absence- of ,
maize thinnings. Many farmers who thin their maize and own
livestock do not even bother to use the thinning as fodder.
2.5 Maize production practices in the study area
Given the dominance of the maize enterprise in the study
area, selected maize production practices deserve some
description. Farmers in the study area follow fairly similar
practices. Some of these are shown in Table 2.
Table 2. Maize production practices in the study area.
------- ------- ------- ---------.-. .-- .- -- ---- --- -- a
Plant local variety of 80-105 days maturity 70 X
Use self supplied seed 75 %
Average seed rate 42 kg/ha
Average plant density at seeding 150,000 seeds/ha
Take out bad or weak plants 83 X
Applied manure last season 57 %
Apply nitrogen 100 X
Average nitrogen dose 162 kg/ha N
Apply phosphate (only in specific villages) 30 %
Average maize yield,local varieties 1.8 t/ha
Source: MARIF (1985).
Land preparation and planting dates: Farmers use fairly
intensive land preparation practices. They plow their maize
fields 2-4 times, using cattle for draft power, then harrow and
level the soil. In case manure is applied, it is spread during
the dry season, before plowing. The rainy season maize crop is
seeded between September and November, depending on the onset of
the rains. This crop is harvested in December-February. The post-
rainy season maize crop is (ideally) seeded immediately
afterwards. However, excess rain at this time frequently forces
some farmers to delay this planting for up to several weeks.
Farmers who grow upland rice as the rainy season crop (which is
usually harvested after 120 days or more) automatically are
forced to plant their post-rainy season maize crop later. Thus
seeding of their maize takes place several weeks after the main
post-rainy season planting period.
Varieties: Relatively few farmers use recently released
improved varieties (23%) although most farmers are familiar with
them (72%). Few farmers purchase new seed from stores, whether
certified or non-certified. Farmers indicate that yield, and
tolerance to insect damage and to lodging are important
considerations in selecting maize varieties. Few farmers
indicate that cooking quality is a major factor in varietal
selection. It should be noted that the "farmer's variety" is a
complex mix of traditional varieties (e.g. Goter and Tongkol),
improved varieties released more than a decade ago (e.g.
Harapan), and a touch of recently released improved varieties
Plant population and thinning: Farmers tend to grow maize at
quite high densities. Seed rates are around 42 kg/ha, and planted
densities often exceed 150,000 plants/ha. Most farmers (83%)
manipulate their plant stands and remove badly developed and
diseased plants. Seeding is done by hand, either in holes made
by- sticks or in the plow furrow. Spatial distribution varies
from 70 x 25 cm to 100 x 45 cm, according to the variety and
Weeding: Maize fields in the study area are rarely weedy,
even in early stages of growth. Weeding is generally done twice.
Cattle are used for weeding between rows, while weeding within
the row is done by hand. The first weeding-takes place about 3
weeks after seeding and is often combined with a fertilizer
application and the commencement of "thinning" (removal of weak
or poorly developed plants). The second weeding usually takes the
form of a "hilling-up".
Fertilizer: All farmers use nitrogen fertilizer with an
average N dose of 162 kg /ha. Nitrogen (in the form of urea) is
usually applied twice, coinciding with the first and second
weeding. These high doses are a reflection on the availability
and low farm-level price of urea. Relatively few farmers use
phosphate. It should be noted that moat phosphate users
identified in the farm survey come from the same few villages.
Pests, diseases and crop protection: Farmers are familiar
with many pests and diseases but report that these seldom
seriously reduce yields. Regarding crop protection, only 13 % of
farmers report using pesticides, mainly Ridomil applied as a
seed treatment for Peronosclerospora maydis (downy mildew).
Researchers noted, however, severe and chronic infestations of
.^- -vr u .t-T -
shootfly, particularly in the post -rainy season.
Table 3. Fertilizer use by variety
Variable Local Improved All
varieties varieties varieties
Farmers using nitrogen 100 % 100 % 100 %
Average N dose in kg/ha 158 168 162
Farmers using phosphate 19 % 50 % 31 X
Average phosphate dose
(users only) 48 69 60
Farmers using potassium 0 % 20 % 8 %
Farmers using manure 60 % 57 % 57 %
Farmers using nitrogen -
at seeding n.a. n.a. 15 %
Farmers using nitrogen
at 3 weeks after seeding
as first application n.a. n.a. 85 %
---------------_------!- -- ---------------------__ __ __ __ _--- --
Source: MARIF (1985). n.a. = not available.
Post harvest operations and disposal of produce: Maize is
harvested by farmers 10 to 20 days after physiological maturity.
Harvesting is done by hand. Ears are carried to the farm house
and sundried for several days. After drying the moisture content
of the grain is often above 18%. Seed for the next planting is
mostly selected from the last harvest. Selected ears or seed for
planting are stored above the cooking place to prevent losses
caused by storage insects and rodents. Only a few farmers buy new
seed after the original purchase of a new variety. Most farmers
in the study area use their maize crop for home consumption, with
occasional sales of surplus quantities.
Differences between rainy- and post-rainy season cropping:
So far the discussion dealt with the rainy season and the post-
rainy season maize crops as being managed in a similar way.
However, differences between the two crops exist and are
Land preparation is usually more intensive for the rainy season
Manure is mainly applied before seeding of the rainy season
Throughout the study area, seeding of the rainy season crop
takes place more or less at the same time, that is immediately
after the onset of the first rains, and cover only a few weeks.
However, seeding of the post-rainy season crop shows a large
variation in planting time, that may cover a period of over one
and a half months.
- Shootfly problems appear to be more severe in the post-rainy
3.0 DIAGNOSIS : PROBLEMS AND THEIR CAUSES
This section emphasizes problems associated with maize
production, because maize is by far the most important crop
enterprise in the study area. Maize is a major starch staple,
and a major source of cash income. The only other crop grown by
study area farmers to any significant extent is upland rice,
which is not included in MARIF's research mandate.
The MARIF team members have continuously advanced in
understanding the problems affecting the productivity of the
maize crop. Initially, the team noted that:
"Despite intensive management, including proper tillage
practices, row planting, adequate weeding practices, high
nitrogen and manure applications, maize crops -in the
domain, -regardless of variety, show symptoms of spindly
stalks, and discoloured leaves. Yields are low, with an
average of -1.8 t/ha grain and only few farmers obtain
yields above 2 t/ha. Yet on-station research indicates that
5 t/ha can easily be obtained from Arjuna (MARIF, 1984).
At first, attention was focused on three factors: variety,
plant population and fertilizer management. However, it became
evident during the fieldwork that other problems required
attention, including- early season insect damage and seed quality.
As research continued, the team gained a better appreciation for
how these problems interact, and the particular causes of each.
The problems affecting maize production, as currently
understood, are as follows:
1. Maize yields are low because most farmers use low quality
seed, of varieties with (possibly) low yield potential.
2. Maize yields are low because of damage during the early
growth stage caused by Atherigona spp. (shootfly) and
Phyllophaga helleri (white grub), resulting in growth
retardation and reduction of plant stands.
3. Maize yields are low because farmers use excessively high
seed rates, resulting in interplant competition during early
4. Maize yields are low (and production costs high) due to low
These problems and some of their interactions are presented
in Figure 2. Note that low seed quality and shootfly damage are
causes of overplanting and that overplanting contributes to a
lower fertilizer efficiency.
Each problem and its respective causes are discussed in more
detail in the following sections.
Figure 2: Components of main problems.
Factors effecting yields
Local varieties witt
3.1) Variety and seed quality
One factor explaining low farm-level yields is seed,
covering the genetic potential of the farmers' variety as well as
the quality of the seed itself at planting time. Regarding
variety and yield potential, many farmers use improved varieties
released many years ago, such as Perta (1956), Harapan (1994) and
Arjuna (1980). These farmers save frmd each harvest some seed
for the next planting. Due to cross pollination with traditional
varieties and land races, genetic erosion may occur and varieties
gradually decrease in varietal purity and yield potential. It
should be noted, however, that in some of the on-farm trials,
these varieties performed very well and had yields not
significantly lower than recently introduced improved varieties,
when grown under conditions of improved management.
Findings from the maize production survey held
indicated that only 23% of the farmers plant Arjuna,
recently released variety, largely due to problems
availability. Arjuna matures in 95-100 days, other
range in maturity from 85 to 115 days.
Regarding seed quality, researchers suspect that low
germination rates and poor seedling vigor enter into farmers'
decisions to overplant. This suspicion has received support from
the formal production survey (42% of farmers mentioned low seed
quality as a reason for overplanting) as well as seed handling
and storage surveys recently implemented by MARIF.
Seed-related problems and causes are shown in Figure 3.
Low seed quality and
low germination j
Insect. problems ---
Plant population Fertilizer I
overplanting low nitrogen
Figure 3: Variety and Seed Quality
Farmers use local varieties with possibly
inadequate yield potential, with seed of
low quality (low germination rates)
*Seed of improved Poor drying & r---
Svarieties not widely, storage practices I Consumer
available I I reduce germination I preferences
L------- ---I rates t L
-- -- -- ---. -L --- - -- - ---- --- -
iPast seed production I
Not supported by research results.
3.2 Insect damage
Maize in the study area is often damaged in the early stages
of growth by shootfly and whitegrubs. Major causes of this
problem include a high natural incidence of these insects, and
late planting of maize. In turn, late maize planting is caused
1for the rainy season) by uneven distribution of early rains and
competition for labor with sawah crops. In the post-rainy
season, late maize plantings may be caused by: (1) a late
planting of the rainy season maize crop; (2) planting maize
after a rainy season crop of upland rice; (3) excessive
rainfall lengthening the turnaround time between the two maize
crops. The severity of insect problems varies greatly from one
year to another. The rainy season maize crop typically suffers
somewhat less from insect damage than the post-rainy season crop.
Interactions between the insect problem and its causes are shown
in Figure 4.
3.3 Plant population management
The farmers' practice of overplanting (at times in excess
of 150,000 plants/ha) then "thinning" (systematically removing
damaged, diseased or poorly developed plants) leads to severe
interplant competition in the early growth stages, as well as a
low harvested density (less than 50,000 plants/ha). There seems
to be a number of possible reasons for this practice:
1) To compensate for expected damage and loss of plant
stand due to shootfly and whitegrub attack
2) To compensate for expected low seed germination rates
and uneven plant vigor, in turn caused by farm-level seed storage
Figure 4: Insect problems
S(shootfly, white grub)
-- Late plantings K- I No insecticides
used on maize
L -- _.1--
SSawah rice competes for labor at maize
planting time (rainy season maize)
------------------------------ -- ----_j
i Late planting of rainy season No experience
-maize affects planting date i with insecticides
Sof post-rainy season maize on maize I
Upland rice in rainy season
delays post-rainy season maize
I- -- ------------------- -' '----- ~-------- '--
I Excessive water delays activities ,
Sfor post-rainy season maize
3) To provide fodder for livestock. (It was originally
thought that farmers might overplant and thin their maize in
order to obtain fodder for their cattle. However, as was seen in I
an earlier section (2.4) this does not seem to be the case.)
Interactions between the plant population problem and its
causes are shown in Figure 5.
In view of these causes, possible solutions to the plant
population problem were seen to involve insect control and I
seed storage practices as well as seed rates.
3.4 Fertilizer management
Farmers apply on average 162 kg/ha N to their maize fields,
but do not obtain high yields. This problem could be explained as
Timing, amount and method of nitrogen application:
Virtually no farmers apply more than a negligible dose of N (3-4
kg/ha) at planting. Most farmers give the first N application 2
Figure 5: Plant population management
Overplanting reduces yields
SCompensation for IPoor seed ,
Expected insect quality a
L damage J
L storage practices
I Farmers 'seed I
Selection practices r-
L- ------- _J
Fodder needs /
Ifor livestock '
Not supported by research results.
to 4 weeks after planting. Farmers, aware of the damage caused
-by shootfly and-white grub and of plant losses due to poor, seed
quality, prefer to postpone fertilizer application until k good
plant stand is obtained. Some farmers also have learned that
urea mixed with seed will burn the seed. They have little
experience with other application methods, e.g. urea placed in an
adjacent hole, 10 cm away from the seed.
Other nutrients and manure: Most farmers-use only nitrogen
fertilizers and do not use phosphate or potassium on their maize
crop, though soil analysis show that most of the young volcanic
soils are deficient in both elements. One source of P and K is
manure, which (during 1984) was applied by about half of the
farmers. It is estimated that most farmers will have applied
manure to their maize crop at least once every five years.
Overplanting: High plant stands during the early growth
season are due to high seed rates. Severe competition between
many plants reduces the grain production per kilogram of N
The problem of fertilizer management and related causes are
shown in Figure 6.
Given the array of causes noted above, possible solutions to
the fertilizer management problem include:
Prior solution of the problems of insect attack and
- Earlier N application
Figure 6: Fer9lizer Management
Figure 6: Fertilizer Management
' applied late
I Lack of information
I on how to apply N
the seeds at seeding
r-- ------ ------.,
SSoils deficient in P,K
and other nutrients
SSome farmers do
not use manure 1
Do not own I Need manure
draft animal high value cl
___ ----- ._
I Adverse growth I
I conditions I
I (-drought, excess
E--J & interplant
SFarmers do not I
apply P and K I
L------ --- ---I
S Never tried
Application of P and K !
-Lower N dose
SApplication of P and possibly other nutrients.
4,0 EVIDENCE FROM ON-FARM EXPERIMENTS
The problems discussed in the previous chapter were examined
ft the five cycles of on-farm experimentation conducted so far.
Sse experiments served to further define problems, while other
*Xperiments looked at possible solutions. As will be seen, the
results of the trials are fairly consistent and tend to support
the hypotheses given in the diagnosis. It appears that farmers'
**ize yields can be doubled through simple improvements in
management practices'at a moderate increase in cost. Altogether,
the team conducted seventy-one on-farm trials in seven villages,
With twenty eight different farmers participating. Some farmers
Cooperated with only one trial in one season, other farmers
participated during all cycles, but not necessarily with the same
~rrr r;cr- -LI-*--C_ ---ljl .~-iLLIL~-I .1_ ____~~_ __ -L-~LII~_