The publications in this collection do not reflect current scientific knowledge or recommendations. These texts represent the historic publishing record of the Institute for Food and Agricultural Sciences and should be used only to trace the historic work of the Institute and its staff. Current WFAS research may be found on the Electronic Data Information Source (EDIS)
site maintained by the Florida Cooperative Extension Service.
Copyright 2005, Board of Trustees, University of Florida
The MARIF Maize On-Farm Research Programme 1984
Development of An On-Farm Research Programme
with a Farming Systems Perspective
Farming Systems Support Project
International Programs Office of Agriculture and
Institute of Food and Office of Multisectoral Development
Agricultural Sciences Bureau for Science and Technology
University of Florida Agency for International Development
Gainesville, Florida 32611 Washington, D.C. 20523
NETWORKING PAPER NO. 2
Malang Research Institute for Food Crops MARIF Indonesia
The MARIF Maize On-Farm Research Programme 1984
Development of An On-Farm Research Programme
with a Farming Systems Perspective
C.E. Van Santen Agro-economic Advisor ATA 272 Project c/o MARIF pob. 66 Malang, Indonesia
Malang, October 1984
This second in a series of FSSP Networking Papers contains two parts. The first is a synopsis of the 1984 Maize On-Farm Research Program at the Malang Research Institute for Food Crops (MARIF), in Indonesia. The text was kindly provided by Dr. G.E. van Santen, Agro-economic advisor to the project. In his letter to the FSSP Dr. van Santen stated:
"This OFR Programme is carried out by a multidisciplinary
team consisting of biological and socio-economic researcher
workers and has now run for about one year. At this stage our
team does not claim to have found any solutions or developed a suitable methodology. The motive behind the submission of
this note is that we are at this stage very eager to exchange
experiences with others who may be engaged in similar
activities. The note could possibly assist in stimulating
such a discussion."
The second part of this Networking Paper is a letter by Dr. Dan Galt, Associate Director and Regional Coordinator of the FSSP for Asia and the Near East, in response to Dr. van Santen. These two brief communications are good examples of the kind of dialog that can be initiated through the FSSP Networking Paper Series.
Networking Papers are intended to facilitate the timely distribution of information of interest to the farming systems network of practitioners
throughout the world. The series is also intended to invite response from the farming systems network to help advance the FSR/E knowledge base and state-of-the-art.
Networking Papers do not necessarily present the viewpoints or opinions of the FSSP or affiliated entities, but represent a statement of the author or authors. Comments, suggestions and differing points of view are invited by the author or authors. Names and addresses of the author or authors are given on the title page of each Networking Paper.
Readers wishing to submit materials to be considered for inclusion in the Networking Paper Series are encouraged to do so. Networking Papers are actively solicited by the FSSP core staff. Send typewritten, complete
manuscripts, ready for publication. The FSSP does not perform an editing or production function with Networking Papers other than to reproduce the author's work and send it to a targeted audience. Distribution is determined by geographic and subject matter considerations to help select a sub-group from the FSSP mailing list to receive each Networking Paper on a case-by-case basis.
Execution of on-farm research with a farming systems perspective by interdisciplinary teams of scientists in cooperation with farmers and extension service is being increasingly recognized as the most effective means of attuning agricultural instutional research to farmers' needs, in developing technologies which can be widely used by them.
It involves the successive steps of (1) problem identification by informal surveys and (2) implementation of on-farm research on principal problems, as identified in the informal surveys.
This synopsis described the MARIF* Maize On-Farm Research Programme, being the first out of several on-farm research programmes which are presently being developed within the Institute.
The present Maize OFR programme is being executed in three sub districts of the Malang district by an interdisciplinary team involving agronomists, breeders, and agronomists. The paper summarizes interim results of the informal survey programme December 1983-January 1984, a subsequent round of on-farm trials implemented in the first dry season of 1984 (February-July) and proposals for further research presently being undertaken.
- The Malang Research Institute for Food Crops (MARIF) carries a national
research mandate for increasing food crop production in particular of maize, grain legumes and tuber crops. The Institute's research output
is to contribute to increased crop productivity of farmers with limited
2. Informal Survey
During the informal survey the team arrived through field observations and interviews with farmers at the following:
2.1 Description of a Recommendation Domain
Essential to the methodology followed is the determination of a specific situation for which the research is intended. For this
purpose the concept Recommendation Domain (RD) is used. A Recommendation domain is defined as a group of farmers whose agronomic
and socio-economic circumstances are sufficiently similar that
they will adopt the same recommendations, given equal access to
information (Harrington & Trip 1983).
Thus for the present maize research project a tentative RD was
selected with the following features:
Maize grown: 1. on young volcanic soils
2. on dry land (tegal)
3. as a mono crop
4. during the rainy season and the first dry season 5. at altitudes ranging from 400 to 700 metres above sea level
6. as a semi commercial crop.
The young volcanic soils form a dominant characteristic of this RD.
These soils in the Malang District are probably to a great extend
representative for other areas with young volcanic soils located in
East Java and elsewhere in Indonesia.
2.2 Characteristics of the Farming System of the Recommendation Domain
Average farm size in the area is 0.5 hectare and most farms include both wet k-d (sawahs) and dry land (tegal). The cropping system in
the selected RD includes rice, maize and leguminous crops, while
many farmers grow sugarcane and fruits (commercially grown papaya).
Farmers grow two to three crops per annum, both on dry land and
irri&geble land. Most farmers own some livestock: cattle, goats and
2.3 Summary of field observations derived from the informal survey
i. The general appearance of the maize crop, regardless variety,
was rather spindly with sympthoms of discolouring leaves, in
spite of intensive management including proper soil management,
row planting, proper weeding practices, high N-fertilizer
applications, with up to 280 kg N per ha per crop, and manure
applications once per year.
ii. Plant densities in young crops ranged from 100 000 to 200 000
plants per ha, and consequently signs- of interplant competition were observed. It is to be noted that trials carried out on the MARIF indicate that optiml plant density for this type of plant
material maturing in 90 -o 100 days, is 60 000 70 000 plants/ba.
iii. Farmers apply high quantities of N ranging from 300 to 600 kg/ha
Urea (with N ranging from 135 kg to 280 kg/ha), often resulting in yield losses due to stem lodging. In general, farmers apply
nitrogen fertilizers in equal amounts at first and second weeding/
thinning, approximately at 3 and 6 weeks after planting.
The team hypothized that timing, quantity and method of application of the fertilizer could be improved and that fertilizer also should be given at planting. The optimal fertilizer application rate for
the area concerned, is approximately 300 kg Urea.
The team observed that farmers in general do not apply phosphate
iv. According to the extension service at least one third of the farmers
uses at present improved varieties released some years ago from the
MARIF (Harapan and Arjuna). The team observed that often seed of
these improved varieties was not anymore pure. This was especially
striking when studying the cobs harvested. This impurity could possibly explain for some of the reduced yield potential, which
may be caused by cross pollination with local varieties in adjoining
3. Design of On-Farm Trials
Based on the above mentioned observations a verification and a variety trial were designed.
In the verification trial farmer practices were compared with the following treatments, separately and/or in combination:
3.1 Variety: The farmers' variety, supplied from the farmer's stock (if
possible mue of the varieties released in the past such as Arjuna or Harapan) was compared with pure certified seed from the MARIF of the
same variety (e.g. Farmers supplied Arjuna seed versus ArJuna certified
from the Institute's stock).
3.2 Plant density: Starting with 3 seeds per hill or 90 000 plants per ha
and reduced at 3 weeks after planting to 2 plants per hill equal to
60 000 plants/ha.
3.3 Timing and quantity of fertilizer applications: Reduction of the
fertilizer applications to 300 kgs Urea (138 kg N/ba) from which one third of the quantity applied at planting, one third at 3 weeks and
one third at six weeks after planting.
In the variety trial five improved varieties (l) Barapan certified$
(2) Hybrid Commercial; (3) Muneng Syntheticl (4) Suwan Ii
(5) Arjuma certified; were compared with (6) Arjuna (or Harapan) as supplied by the farmer from his own stock or from his regular source for maize seed.
4. Planting of the trials
During the period February 10 to March 20, six verification and four variety trials were planted in the fields of 10 cooperating farmers. These farmers were selected with the assistance of the resident extension field worker in these sub districts. Planting, other agronomic practices and observations were jointly carried out by farmers, extension workers and MARIF staff.
5. Summary of the observations during the growing period of the trials
5.1 Verification trial
The general appearance of the maize crop, both in the trial fields and in surrounding fields, remained unsatisfactory. The plant stature of
the maize was still spindly with leaves often showing abnormal colouring.
At the plant age of 7 to 10 weeks, in most locations, maize leaves
showed yellow, white striping. In some fields purple coloured leaves
were observed. These observations could possibly point to absolute
or relative nutrient deficiencies.
5.2 Variety trial
Significant (15 %) differences in yield between varieties were found at two locations whereby the farmer's variety Tongkol yielded lowest
and Muneng Synthetic yielded highest.
6. Observations at harvest time
6.1 The average yield of the verification trials amounted to 3900 kg dry
grain per hectare ranging from 1900 to 6160 kg/ha.
6.2 A significant difference in yield was observed between locations.
6.3 The introduction of an improved variety without improved planting and
fertilizer practices did not show any increase in yield,
6.4 The effect of the improved planting method (with lower plant densities)
combined with improved seeds showed, a yield increase of 15 %.
6.5 The introduction of improved fertilizer management, seed and planting
method resulted in approximately 15 % higher yields. The economic
analysis with the partial budget analysis-method indicates that in this
case yields are obtained at 12 % lower costs.
6.6 Comparison between yields obtained in the absence and presence of phosphate fertilizer indicates a considerable yield increase as a result
of phosphate fertilization.
The average trial yield of 3900 kg/dry grain/ha is at least twenty-five percent below the yield potential that should be obtainable with the varieties selected and the level of management applied. Based on the field observations pointing to possible nutrient deficiencies the team formulated a hypothese that these low yields are possibly caused by soil fertility problems requiring further research.
8. Further actions
In view of the above the team plans the following activities starting September 1984 :
8.1 On-Farm trials
- A set of soil fertility trials, studying possible nutrient deficiencies.
- Continuation of the variety trials.
Trials were actually planted in the period September 16-30, 1984 in roughly
the same locations as the first set of trials including four fertilizer
trials and six variety trials.
8.2 Soil Analysis
Soil analysis of samples taken from locations where the first series of
trials were implemented.
8.3 Pot experiment
Identification of possible nu-tient deficiencies in maize grown in pots filled with soil from a farmer's field where leaf discolouring was most
8.4 Socio-eoonomic investigations
To interview participating farmers and their neighbours an issues of manure- and fertilizer application; seed sources; plant densities;
fodder needs supplied from maize crops; and related matters.
- This Synopsis is based on the following working papers prepared by the team
responsible for the MARIF Maize on-farm research programme and consisting of Marsum Dahlan maize breeder; Bambang Sulistyno and J.Ph. van Staveren agronomists and C.E. van Santen agricultural economist.
January 1984 Joint Proposal for the MARIF On-Farm Research Programme for maize
February 1984 Trials descriptions for maize grown on dry lands (verification and variety trials)
March 1984 Proposal for a farmers' field day for the participants
of the MARIF Maize OFR Programme in May 1984 April 1984 Draft Report A Maize On-Farm Research Programme
Development of a methodology
September 1984 Trial results of the MARIE Maize On-Farm Research Programme
- 1984 First dry season.
- In addition the following references were used
- Byerlee D. & Collinson M. Planning technologies appropriate to farmers.
CIMMYT Mexico. 1980.
- Harrlngton L.W. and Trip R. Recommendation Domains s A Framework for
On-Farm Research Draft report, CIMMYT Mexico, November 1983.
- Crops statistics for the Malang district from the Agricultural Services
of East Java.
Telephone: (904) 392-1965
Cable: CENTROP FSSP
Farming Systems Support Project
Institute of Food and Agricultural Sciences
University of Florida
3028 McCarty Hall
Gainesville, Florida 32611
Nov. 16, 1984
Dr. C. E. van Santen
ATA 272 project
P. 0. Box 66
Dear Dr. van Santen:
Your recent letter of Oct. 15 has made its way through my in box after passing through the hands of our project editor. As one of the associate directors of the FSSP in charge of reviewing materials for the newly-established Networking Paper series, and with responsibilities for FSP policy development and implementation in Asia and the Near East, I would like to take a moment to respond to your letter and to your paper, "SYNOPSIS: The MARIF maize on-farm research programme 1984".
In my opinion, this type of short, state-of-the-art FSR project summary is exactly the kind of information the FSSP should be making available to other practitioners via the Networking Paper series. Since other core staff will have to review your paper before its final acceptance as a Networking Paper, I will confine my remarks in the rest of this letter to substantive ones regarding your project. Please be aware that these remarks are no more or no less than my personal opinions and do not represent the collective thought of the FSSP or other FSSP staff members.
I read your project synopsis with great interest, mainly because I conducted my first sets of on-farm trials using improved tropical maize germplasm from CIMMYT versus local (criollo) varieties during the growing seasons of 1975 and 1976 in Veracruz state, Mexico. In regard to your synopsis, I would like to make the following observations:
(1) Your only reference to other components of the farming systems present in the Malang area is on page one. However, your synopsis does not mention whether the system is one of successive monocrops, or if there are other crops grown in association with maize. I assume maize is grown as a monocrop, but it is not explicit in your text that this is the case.
(2) You mention extremely high field-level maize densities
(100,000-200,000 plants/ha). However, such densities seem to have been taken at a young plant age ("Plant densities in young crops...", point
2.3 ii., p. 3). During my three cropping seasons in a tropical environEQUAL EMPLOYMENT OPPORTt)NITY/AFFIRMATIVE ACTION EMPLOYER
ment in Mexico, I found that maize densities taken at theoretical plant establishment were always-significantly less than harvested maize plant densities. Not infrequently these differences were between 50-100%: established plant densities of 50,000-65,000 plants/ha resulted in harvested densities of 25,000-48,000 plants/ha in farmer's fields. My question: did your team measure harvested plant densities and if so, were they significantly lower than establishment plant densities?
(3) If the answer to the second part of (2) above was yes, the next issue I would raise is: what types of routine and special seasonal observations did your team make in regard to pest problems? In a tropical Mexican environment, plant damage by both stem borers (Diatraea saccharalis (F.)) and fall armyworm (Spodoptera frugiperda (J. E. Smith)) was sproadic but occasionally so heavy as to cause dead heart and plant loss even after the theoretical plant establishment stage. For this reason, we dedicated a considerable amount of time to looking into
(a) the quantification of foliar insect damage types, frequencies and severities, and
(b) the observation of the timing of the above damage.
Unfortunately, these are difficult and complex tasks, and my feeling is that we had insufficient manpower and time to do a decent job in these two areas. If your team has addressed these issues, have you had significant success in quantifying economic pest damage levels and timings, or are such problems of little or no consequence in your farmer's fields?
(4) Often, small farm households do not wish to put any inputs (in your case it would be fertilizer) on maize until they are assured of having plants in the field to fertilize which are likely to lead to a harvestable crop. At least this is often the case in the tropical maize production areas of Mexico and Central America. Did either your sondeo or your informal contacts with farm households during the growing season last year ascertain if this might be a reason why your Indonesian farmers do not fertilize at time of planting?
(5) The use-of the word "verification" in a set of trials (pp. 3
and 4) implies more "demonstration" than is usually justified, especially during the initial season of on-farm trials. Our Honduran FSR team discussed alternative terms for a long time before deciding to use the term "exploratory" for the first season's FSR trials. Another similar adjective is "'observation". This is obviously a minor point -- the significance of which is generally lost on both farmers and extension personnel -- but it does relieve the FSR team from some of the responsibilities and related problems of first-season credibility with farmers, regional agricultural research administrators and national agricultural research leaders.
(6) On page 3, section 3.2 under "Verification trials", your paper reads as if your team plans to thin the planted stand down from 3 to 2 plants per hill. Two comments:
(a) unless the Malang farmers practice plant thinning, using this method to obtain a harvested stand of the more optimal 60,000 plants/ha has relevance only for researchers but not for the farming community, and
(b) the same comment as made in (2) above holds here: based
on my personal experience with tropical maize varieties, thinning down a 90,000 plants/ha stand "...at three weeks after planting..." to 60,000 plants/ha is likely to result in a harvested density of less than 45,000 plants/ha. (My experience involving thinning various tropical maize vatiety densities after plant establishment -- with a top planted density of 90,000 plants/ha -- was gained at CIMMYT's tropical maize experiment station in Poza Rica. There, the objective was to attain a pre-selected set of lower densities, the lowest of which was to have been 28,000 plants/ha and the highest of which was to have been 60,000 plants/ha. At harvest, across more than 50 plots, densities ranged from a high of 57,000 plants/ha to a low of 14,000 plants/ha. The average top density was about 48,000 plants/ha. This occurred under the more "'controlled" conditions of an experiment station where plant densities could be extimated with great precision and where thinning was done after there was no longer any doubt that plant establishment should have occurred (at the 10-12 leaf stage, or roughly 0.5-1.0 meters plant height: a stage much later than the three weeks following planting that your team proposes).
I hope I am making too much of this potential problem, and
that your experiences will be much more positive. However, maize agronomists I have worked with subsequently have encountered similar difficulties in trying to obtain "known" stand densities, and so far, to my knowledge, the reasons why this happens are still unclear.)
(7) There are those who would argue (Mosher, for example) that
yield increments of less than 50% (hard-liners would still say 100%) are not worth the effort of the research team, because farmers will be unable to appreciate smaller differences. I maintain that this is simply not true, and that farmers can discern much smaller yield differences. Did your sample of farm households detect the differences of 15% that your indicate in your paper occurred, and were they similarly impressed with the cost decreases of 12%, or do you feel your sample size was too small for your project to make such a statement or that another year's data is necessary to substantiate your experiences of year one?
(8) Adding phosphate to the fertilizer recommendation seems to make sense. Is such a product locally available and reasonably priced? If so, would such an intervention mean that farmers might substitute a N-P fertilizer for urea? Has your team investigated the biological ramifications of this possibility? The economic ramifications?
(9) You paper mentions obtaining an "...average trial yield of 3900 kg/dry grain/ha..." (point 7., p. 5). Two suggestions: first, whenever you present average trial yields, it is always more meaningful to provide the reader with two additional averages:
(a) the average dry maize yield in the area in which your recommendation domain falls, and
(b) the average of all the trial check plots which contain the individual farmer's varieties.
The second suggestion, mainly for the benefit of your non-Asian readers, is that, in the future, it would be helpful to know at what percentage moisture maize is considered to be "dry" in the Asian and Indonesian context. I am assuming it must be somewhere between 12-15%, but between-project and between-country comparisons are difficult without this additional piece of information.
(10) Finally, average trial yields of 3900 kg/ha may be about
average in the Asian context, but are very high relative to both the Latin American and African context. I wonder what importance is accorded to maize in the overall system, and whether rice is usually the predominant cereal crop. It rice is considered to be more important than maize, the farm household may not be willing to invest as much in the management of maize as it does for rice even if the two do not compete directly for the same labor at the same time. There is such a thing as "farm household burnout", characterized by a certain level of management and labor investment in a crop which is not the main subsistence one in the system, but no more than these levels, even though we, as research scientists, would argue that the household is underutilizing both management and labor at the margin in regard to that given crop. This is a difficult issue to grapple with and understand.
I hope you and your FSR team receives these comments in the spirit in which they were set down. That is, as personal reflections which may or may not have relevance in your setting and biological context.
I will enclose with this letter for you a copy of a Cornell University mimeograph which resulted from my work with tropical maize in Mexico ("Yield losses as economic weights in plant breeding decisions on tropical maize"), and a recent publication by Peter E. Hildebrand which addresses the issues of which types of agronomic analyses are appropriate for on-farm trials ("Modified stability analysis of farmer managed, on-farm trials"). Also, for your information, I will be visiting Indonesia (Jakarta and Bogor) for about one week beginning Nov. 27th through Dec. 2nd. During my visit, I will try to give you and/or your project a telephone call. otherwise, I hope to hear from you again.
Thank you again for-sending the FSSP your project's Synopsis. I
assume you will be informed shortly as to whether it will be coming out in the Networking Paper series. I am very pleased that you followed up on the request for material of this nature.
Dan Gait, Associate Director (Asia/Near East), FSSP
cc: Steve Kearl
Peter E. Hildebrand