The TropSoils Program
REPORT BY THE EXTERNAL EVALUATION PANEL
The TropSoils Program
REPORT BY THE EXTERNAL EVALUATION PANEL
JOHN K. COULTER, PETER E. HILDEBRAND AND STEPHEN L. RAWLINS
TropSoils is one of the Collaborative Research Support Programs (CRSPs) created to implement Title
XII, "Famine Prevention and Freedom from Hunger" of the Foreign Assistance Act. Primary funding is
provided by the U.S. Agency for International Development under grant no. DAN-1311-G-SS-1083.
This report was printed by the TropSoils Management Entity, Box 7113, North Carolina State Univer-
sity, Raleigh, NC 27695-7113. Charles B. McCants, Director. Neil Caudle, Editor. Jennifer Austin,
THE NATURE OF THE PROGRAM...4
Development Relevance of the Program...5
Problems Being Addressed...5
Human Resources for the Program
CRSP and CPAC...7
THE RESEARCH PROGRAM AND ITS ACHIEVEMENTS...8
Quantitative Description of Water and Chemical Budgets in Acid Savannah Soils...8
Soil Tillage and Root Restriction... 10
FUTURE WORK... 11
The Appropriate Role of Models in TropSoils Research... 12
Deterioration of Soil Physical Conditions Under Long-Term Continuous Cropping...15
Diffusion of Research Results: The Output of the CRSP/CPAC Program... 15
On-Site Management... 17
Summary of Recommendations... 17
SIGNIFICANT RESULTS... 19
THE NATURE OF THE PROGRAM
1. The Panel visited Cornell University September 16-18 and the Centro de Pesquisa
Agropecuaria dos Cerrados (CPAC) at Brasilia, Brazil on September 19-25. At Cornell,
presentations were made by staff and graduate students in the program. At CPAC the panel
had the opportunity to discuss the field and laboratory programs in detail, to visit some of
the farms engaged in intensive production on acid savanna soils, and to discuss the program
with CPAC staff. The Panel wishes to thank the Cornell faculty, the EMBRAPA and CPAC
staff, and the students for the excellent interaction during the visits. This contributed greatly
to the work of the Panel.
2. Several important aspects of the program, including comments and suggestions on
future programs, are discussed in Section 2 and Section 3.
3. The Cerrado of Brazil has about 200 million ha lying south of the Amazon Basin at a
latitude of 60 to 200 S at elevations ranging between 1100 and 1300 meters. Rainfall during
the wet season, October to April, varies from 1000 mm in the south to 2000 mm in the
north, with the CPAC location averaging about 1600 mm. Rainfall is more than adequate
for good production during the rainy season, although there are relatively short dry spells
that can have a severe impact on crop growth.
4. The soils of the Cerrado are dominated by the soil orders Ultisols and Oxisols. In the
Cerrado these soils are extremely low in plant nutrients, particularly calcium and phos-
phorus, and are highly acidic. However, they have excellent physical characteristics, and
when nutrients are supplied are capable of giving good yields. The development of the
Cerrado soils using modern technology of fertilizers and lime, mechanization, improved
varieties and capital has demonstrated that large areas of some of the world's poorest soils
can be reclaimed for a highly productive agriculture.
DEVELOPMENT RELEVANCE OF THE PROGRAM
5. The Ultisols and Oxisols of the Cerrado represent a substantial part of the global area of
these soils, but there are very large areas of similar soils in Africa and Southeast Asia.
Outside Brazil, these are almost entirely farmed by small farmers except in the plantation
areas of Southeast Asia. Mostly such farmers use low or nil inputs of plant nutrients and
depend on long-term fallows to sustain the low yield levels usually obtained. The relevance
of the CRSP work in Brazil to improving the productivity of small farms in Africa and Asia
lies in the importance of the Cerrado research in providing an understanding of the physical
and biological principles involved in reclaiming these soils from poor scrub and grassland
into highly productive agricultural land. The development that has already taken place on
Cerrado soils also provides a unique opportunity to define the second-generation problems
of maintaining soil fertility and increasing productivity, both from the chemical and physi-
cal points of view. It also provides an opportunity to study techniques for reducing the
investments in fertilizers, especially nitrogen, which would be of specific interest to small
6. As knowledge and understanding of the physical and biological problems of farming
these soils are improved, the ideas and techniques generated by the research will be tried in
the very different economic and social setting of small farmers in Brazil and elsewhere
PROBLEMS BEING ADDRESSED
7. An essential part of the work of the TropSoils program is a clear definition of the
problems that are being addressed and the relevance of the solution of those problems to
agricultural development in the various agro-ecological zones. The problem orientation is
implicit in the statements in the Cornell document and the on-site briefing report, but it is
not always explicit. Furthermore, in these documents, descriptions of research topics are
sometimes intermingled with problem descriptions. For example, leaching of nitrates is
stated as a research topic; the problem is that a substantial portion of the nitrogen from
fertilizers or organic matter is not utilized by crops. It is inevitable, of course, that a re-
search program will consist of a number of time-bound, researchable topics, especially
when graduate students are part of the program. All of these topics will have some rele-
vance to the problem in the broadest sense, but some will be of much greater importance in
addressing the problem in hand, and hence the need to select priority topics within the
overall problem area.
8. The Panel also considers that improved presentation would make it easier to understand
the priority being given to the solution of some specific problems. For example, the pro-
gram on deterioration of soil structure after several years of continuous cultivation is based
on a rather vague problem description, the importance of which is somewhat speculative.
As a consequence, the priority of this program in relation to others is not readily apparent.
9. The usefulness of clear problem definition is, however, well illustrated by the program
in Brazil. As indicated earlier, several million hectares of Cerrado soils have been reclaimed
from stunted bush and poor grassland into productive agriculture. This reclamation was
done under an economic climate in which fertilizers and lime were supplied below cost and
an easily saleable crop, soybean, could be grown for a rapidly expanding market. In such a
climate the real costs of reclamation were not a problem.
10. Thus the TropSoils program is not involved in the problem of reclamation (though its
predecessor program was), the technical aspects of which are now reasonably well under-
stood, but in the problems of more efficient management of reclaimed Cerrado soils. In the
Panel's view the more specific problems being addressed include those dealing with the
production and decay of organic matter and the quantitative assessment of anion and cation
movement in relation to water movement in the soils. Improved knowledge of these proc-
esses would lead to more efficient management of fertilizer, green manure and crop residue,
and would provide the technical know-how for developing new cropping or crop-animal
systems that could better use limited amounts of plant nutrients from fertilizers and organic
matter, including that supplied by green manure.
11. A second thrust of the program addresses the potential problems that could arise
through continuous intensive cultivation of Cerrado soils. While such soils appear to have
an extremely stable and favorable structure in spite of their high clay content, there is some
evidence that continued arable farming does lead to some form of structural deterioration in
HUMAN RESOURCES FOR THE PROGRAM
12. Human resources for the program are a combination of Cornell faculty, graduate stu-
dents and staff of CPAC. The management of the program is to be congratulated on an
outstanding job of mobilizing the skill and interests of university faculty. This has been a
major factor in reaching the high scientific quality that the program has achieved. The
program has also captured the interest of CPAC, resulting in strong support for the program
from the Brazilian authorities. The presence of graduate students and Cornell faculty in
Brazil and visits by experienced staff from Corell have provided an intellectual stimulus
much appreciated by Brazilian scientists at the station.
CRSP AND CPAC
13. The CRSP enjoys good relationships with CPAC based on many years of productive
interaction preceding the CRSP. CPAC welcomes the depth with which CRSP studies are
carried out, and feels that these studies are well defined and focused. CPAC is not able to
conduct the more fundamental research work being done by the CRSP because of its broad
mandate and limited resources. CPAC would like to be able to provide more support to the
CRSP, but as their resources have been cut back over time, they have had to reduce support
to the CRSP in order to continue with their own priority projects. A closer agreement on
priorities might alleviate this.
14. CPAC scientists have expressed an interest in being more involved in planning CRSP
research. Some characteristics of the type of research conducted under the CRSP, however,
have held this interaction to a lower than desired level. When graduate students arrive in
Brazil, their individual projects have already been approved by CPAC. Due to graduation
pressures, the students normally have a limited time in which to conduct the research. They
tend to become involved immediately in the implementation of their projects, before strong
personal contacts with CPAC staff have been developed. This problem may be lessened as
CPAC scientists return from studies at Cornell. These returning scientists will be acquainted
with the students, and will have the language capability and knowledge of the Cornell and
CRSP systems to facilitate interaction. Also, assuring that the students work only on proj-
ects of high priority to CPAC will facilitate the provision of CPAC staff time and other
resources for collaboration on CRSP projects.
15. The achievements of the program have been described in detail in the documentation
from Cornell and the research site and the results are summarized in Annex I (documenta-
tion provided by the research group). The comments in this section are confined to two
general areas: quantitative description of water and chemical budgets, and soil tillage and
THE RESEARCH PROGRAM AND ITS ACHIEVEMENTS
QUANTITATIVE DESCRIPTION OF WATER AND CHEMICAL BUDGETS
IN ACID SAVANNAH SOILS
16. The Cornell TropSoils Program Review Document, dated September, 1987 identified
three research topics (page 9) to be pursued by the future CPAC Collaborative Research
Program. These are:
1. nitrogen management of acid savanna soils,
2. quantitative description of water and chemical budgets in acid savanna soils, and
3. soil constraints to management of acid savanna soils.
Specific research projects proposed to implement Research Topic 2 are Project No. 104,
Nitrogen Transformation and Movement in Acid Soils of the Tropics, and Project No. 107,
Crop Water Requirements in Acid Savanna Soils.
The objectives of Project No. 104 are:
a. Develop a comprehensive description in simulation model form of nitrogen
movement and transformation in cropped soils of the acid soil regions of the tropics.
b. Use the model as a guide in interpreting previous studies and in designing further
field experiments to increase understanding of nitrogen fate in these soils crop
c. Develop a simplified model of nitrogen fate in these soil crop systems that will be
useful in guiding nitrogen management programs.
The objectives of Project No. 107 are:
a. Determine the components of the water budget in the cropped Cerrado soils.
b. Quantify the effects of soil tillage practices upon water availability and soil-water
movement and their impact upon crop yield, evapotranspiration and transient soil-
water regimes in a manner useful to irrigation and crop production programs.
c. Develop and test a simplified model relating water use to crop yield, constructed
to be useful in extending experimental results to other locations.
17. Clearly, because of the close linkage between water and nitrogen transport, the objec-
tives of Project 104 cannot be met without first accomplishing the objectives of Project 107.
Likewise, to predict crop yield as a function of water use as called for in objective (c) of
Project 107, a quantitative knowledge of the nitrogen budget is essential.
18. Not only is it impossible to pursue the objectives of the two projects within Topic 2
independently, but there is close linkage between these objectives and the research under
Topic 1 and Topic 3. The quantitative description of nitrogen transformation and movement
in soil called for in Project 104 cannot be made without the information to be developed on
nitrogen management in Topic 1. Objective (b) of Project 107 calling for quantification of
the effects of soil tillage on water movement and crop yield clearly cannot be fully attained
without the results of Topic 3 on soil constraints to management of acid savanna soils.
19. The point is that all of the projects proposed in the Cornell/CPAC collaborative research
program are closely interrelated. The objectives of some individual projects cannot be
attained without first having the results of other projects. It would appear that arranging
projects in a hierarchy would be more productive than pursuing a set of parallel studies
20. It seems to the Panel that rather than standing as one of three parallel topics, Topic 2,
obtaining a quantitative description of water and chemical budgets in acid savanna soils, is
really the overall objective of the Cornell/CPAC program. All of the other projects produce
information to help fill knowledge gaps required to accomplish this.
21. Accomplishing this overall objective will require an integrated approach involving
overall program management from the top. The systems approach outlined in the projects
of Topic 2 provides an excellent mechanism to accomplish this management. First, all
existing information from past experiments and from literature references should be incor-
porated into the quantitative, predictive models. Where gaps occur, new research should be
outlined to fill them. Such overall management will require the long-turn involvement of a
scientist with an holistic view of the program. The pieces can be filled in by graduate-
student programs, but without the overall management these pieces will not add up to a
whole that is greater than their sum.
SOIL TILLAGE AND ROOT RESTRICTION
22. This project, Project 108, is designed to characterize the root-restricting zones in Cer-
23. Circumstantial evidence suggests that the structure of Cerrado Oxisols, which is initially
very stable, deteriorates after a period of continuous cultivation. This deterioration takes the
form of a hard and compact layer below disk-plow depth, resulting in poor root develop-
ment and consequently lower yields. The evidence for this is based on the performance of
soybeans at the National Soya Bean Center, over a period of ten years of mono cropping,
on farmer observations that shallow rooting occurs even in adequately fertilized soils, and
on the fact that the surface soil erodes down to the compacted layer.
24. Experimental evidence supporting this hypothesis has been obtained from pot experi-
ments in which clay soils compacted to differing bulk densities displayed barriers to root
penetration. Yields in the experiment are higher when chisel tillage is used.
25. The Panel recognizes the potential importance of structural deterioration of soils that do
not appear to have any of the recognized mechanisms (freezing, shrinking-swelling) for
improving structure. The apparent absence of soil fauna such as earthworms at depth would
also contribute to poor recovery of structure. Meeting the objectives of this research will
therefore require attention to soil properties other than mechanical compaction.
26. Since the original proposals for the acid savanna soils were developed in October of
1980, there have been major changes in the thrusts and shape of the program. Financial cuts
led to a revised program being presented in April of 1982. This program had thee major
thrusts: to increase the efficiency of fertilizer use in acid savanna soils; to develop improved
methods of soil acidity management, and to develop methods of characterizing soils so that
results may be applied to other acid infertile soils of the tropics.
27. The actual projects undertaken have included only a part of the program as formulated.
As indicated in Section 1, work on nutrient efficiency has concentrated on nitrogen, particu-
larly that from leguminous crops. The work on soil acidity management has been concen-
trated on movement of calcium ions. There is not an active program by the TropSoils group
on characterizing soils, although the project on linking soil color to mineralogy and water-
table measurements could be regarded as part of this.
28. The proposals for the 1987-1992 program follow closely on the same lines. At this stage
the Panel would not comment on details of these proposals, which in any case will be the
role of the peer review panel. Rather our comments are focused on the general direction or
strategy of the program.
29. As indicated above the original and modified proposals were set out as a series of
thrusts each of which contained several research projects,each project having a research
objective or objectives. The letter of July 21, 1987 from the Management Entity to program
coordinators sets out specific requirements for the format for the research proposals. The
Panel strongly endorses these requests and stresses the need for the projects to be set in the
context of development relevance of the subject. It also stresses the need for a well defined
problem statement that differentiates between the problem and the research objectives that
would lead to the solution of that problem.
30. The Panel also considers that more attention should be given to formulating a specific
strategy with a multi-year time horizon. It recognizes that factors beyond the University
control can have a drastic impact on such a plan. Nevertheless the development of this plan
would provide an excellent opportunity for CPAC and Cornell to look at their correspond-
ing long-term needs as related to the development of Cerrado agriculture. Doing so would
provide the umbrella strategy under which the work of graduate students, Brazilian scien-
tists and Cornell faculty could operate, and could lead to stronger synergistic effects from
the participants. To make such a plan worthwhile, there would have to be an obligation on
the part of both the TropSoils and CPAC management to give priority support to the pro-
gram. Clearly the on-site coordinator should play a major role in developing this plan.
THE APPROPRIATE ROLE OF MODELS IN TROPSOILS RESEARCH
31. The Panel has expressed concern in the past that simulation modeling in some TropSoils
programs has not been effectively integrated into the theoretical and experimental research-
programs being conducted. Too frequently simulation models are treated as ends in them-
selves and developed as separated projects that treat one component of the overall system
and end with publication in a scientific paper. The Panel is considerably encouraged by the
presentations made by members of the Cornell team and by the publications made available
in the briefing packet. The capability and desire now appears to exist within this team to do
much more. The modular concept for simulation models developed in the publication by
Buttler and Riha to allow users to incorporate additional procedures and to link existing
models into base models establishes the necessary framework. The comprehension of the
specific biological, chemical and physical processes involved in the soil-plant-atmosphere
continuum demonstrated by the publications of Lathwell and Grove, Lathwell and Bouldin,
Wolf, de Wit, Janssen and Lathwell provide what's needed to put the meat on this frame-
work. The scientific talent is in place, much of the needed basic scientific information has
been developed, and the time is ripe to begin the process of developing an integrated model
for acid savanna soils.
32. This model must include transformation and transport processes that affect crop per-
formance and involve water, nitrogen, phosphorus, calcium and aluminum. Such a model
would provide the predictive capability needed by managers of crop systems on these soils
to assess the potential impact of alternative management choices on crop production under
specific environmental conditions. This would permit a manager to take into account the
cost for each management alternative and choose that with most benefits.
33. The failure of scientists to use models to help solve management problems stems to a
large extent from the fact that they do not perceive this to be their job. Scientists often have
the concept that their job is to discover basic principles and publish them in the scientific
literature. The assumption is that someone else will integrate this information into manage-
ment procedures that will solve problems, and that integration of scientific information into
packages to solve management problems is not science. This orientation fails to recognize
the value of one of the most rapidly emerging and productive scientific disciplines-
34. Systems science is that branch of science concerned with the solution of management
problems by use of the tools of science. it recognizes the fact that real world problems are
not water problems, or pest problems, or plant nutrition problems, or erosion problems;
they are just problems-with water constraints, pest constraints, nutrition constraint and
erosion constraints. Furthermore, solutions to real problems do not come in packages
labeled physics or chemistry, plant physiology or weed science, micrometeorology or
hydrology, soil science or agronomy, economics or marketing. They have interdisciplinary
labels with as many discipline names as are needed to solve the problem.
35. Scientific method requires the formulation of hypotheses and experimentation to test
them. Yet systems within which real problems are embedded are frequently not amenable to
direct experimentation. Direct experimentation on complex systems is often too costly or
too risky, or takes too much time to produce the needed result. When it is not practical to
perform direct experimentation, systems scientists construct a model of the system. The
model stands as surrogate for the real system. A model is a description of the essential
components of a real system and how they interact. In the field of engineering, models
often consist of miniature physical representations of real structures, such as a dam. Today,
with the advent of powerful and affordable computers, models are more frequently con-
structed in numerical or symbolic form.
36. The model allows the scientist to perform experiments by changing either the compo-
nents or the environment of the system and, and to determine the effect on system response.
For a model to be useful it must accurately represent the most important attributes of the
real system. The first step in establishing this is to verify that the output of the model is
consistent with existing observations about the real system under all conditions for which
data are available. Once verified, new conditions can be imposed on the model, or compo-
nents of the model can be altered to predict the system response within a new environment
or with an altered management scheme. Before such predictions can be relied upon, valida-
tion experiments should be conducted on the real system under controlled conditions where
some components of the environment or management are changed and direct system re-
sponse is measured.
37. Ultimately, the goal is to construct models of real systems that can be used by managers
to achieve a specific goal. Clearly, with a system as complex and difficult to work with as
the soil-plant-atmosphere system for acid savanna soils, immediate products that can be
placed in the hands of users cannot be expected. But if such a goal is ever to be achieved we
must begin somewhere. And the process of creating such a model is an effective means of
integrating existing scientific knowledge into a form that can identify gaps. It also can serve
as an effective communication tool to link network members into an effective interdiscipli-
nary team. Effective inter-team communication should be particularly effective if the
modular approach proposed by Buttler and Riha is used. This could permit the fundamental
physical, chemical and biological processes understood best by scientists at the cutting edge
of their discipline to be directly incorporated into the model. Competing modules developed
by other scientists in the same field could then be substituted directly into the model and
evaluated on the basis of improvement in the predictive capability of the overall model.
38. Capitalizing on this systems approach would require organizing the Corell/CPAC team
into a hierarchy. The primary objective of the program would have to be directed toward
the development of a comprehensive model of the soil-plant-atmosphere system for acid
savanna soils. Some person would have to accept responsibility for overall system design
and coordination. Responsibility for design and implementation or specific system modules
would also have to be assigned. Module teams would initially have the monumental task of
incorporating existing scientific knowledge into their sub-model and developing the accom-
panying data base. This integration of existing knowledge into packages frequently is not
looked upon by peers as having the same stature as producing more new parts. But clearly it
is the task of science to build comprehensive predictive theories that explain interactions of
real systems. An immediate return on this investment in integrating existing knowledge
would be a clear identification of knowledge gaps that could be filled by new research. This
could significantly affect the direction of student research programs and provide an imme-
diate use of student research results beyond that of publication in a scientific journal.
39. It would seem most practical to place primary responsibility for overall systems design
and system module development with the Cornell faculty. Equally as important as model
development will be the job of model validation. Since this can only be done on the real
system, primary responsibility for this phase should probably rest with CPAC and the
Cornell permanent staff in Brasilia. Graduate students could effectively work and be super-
vised by staff at both locations. Close working relationships among team members working
on any specific module at both locations is essential. Frequent visits by staff members to
each other's institutions would almost be essential to provide the real-time feedback to keep
such a dynamically evolving project on course.
40. The opportunity to implement a systems approach to solving management problems for
which TropSoils was established appears to exist within the Comell/CPAC team. The Panel
strongly recommends that the team consider this as an alternative approach for team struc-
ture and research management, and that they present an outline in this format at the Decem-
ber TropSoils meeting.
DETERIORATION OF SOIL PHYSICAL CONDITIONS
UNDER LONG-TERM CONTINUOUS CULTIVATION
41. There is considerable speculation but not much hard information about the magnitude
and nature of this problem. Indeed it is possibly of such a nature that a multidisciplinary
team would be needed to tackle it effectively. The problem, as seen by the Panel, is the
need for a better definition of what actually takes place in such soils over time. For ex-
ample, the relatively large amounts of lime may alter the charge characteristics of these
poorly buffered soils, leading to increased dispersion of the clay. This is supported by the
observation that water-dispersible clay in some Oxisols increases with cultivation. Also
changing the pH from that of the natural environment may kill organisms such as earth-
worms that could help alleviate compaction.
42. The panel recognizes that the very limited resources of TropSoils make it possible to
carry on only the limited range of research already under way. But the development of a
multi-year program using the systems approach outlined above would provide the opportu-
nity to involve more support from EMBRAPA. The Panel would endorse work that would
lead to better characterization of the extent and causes of the problem.
43. The Panel would therefore recommend that work continue on assessing the magnitude
of the problem, on attempting to quantify the impact on yields on well documented sites
and the soil characteristics that appear to encourage structural deterioration. The panel
suggests that the research should not focus solely on the mechanical aspects of long-term
cultivation but should also look at the physicochemical and biological aspects of structural
DIFFUSION OF RESEARCH PRODUCTS:
THE OUTPUT OF THE CRSP/CPAC PROGRAM
44. CRSP projects develop three different kinds of research products. One is the develop-
ment of new biophysical knowledge with global applicability. This product is diffused
through theses, journal articles, bulletins, state-of-the-art papers, presentations at profes-
sional meetings, and in general interaction with other scientists. Each individual paper or
presentation can have value in and of itself as a basic knowledge resource. The Cornell/
CPAC TropSoils program reports an impressive array of such products.
45. A second kind of research product is created when a new research method or technique
evolves during the research process. CRSP scientists have proved that a short-cut procedure
can be used for evaluating nitrogen mineralized after incorporation of plant material in
small bare fallow plots as a technique for screening large numbers of potential legumes as
green manures. They were also successful in developing a technique, using polyethylene
bags, to bury soil/green manure samples in the field allowing mineralization to proceed
normally while preventing nitrate escape to the surrounding soil. This is a labor-saving
technique to screen large numbers of legumes as sources of nitrogen. CPAC sees these
products, emanating from the mission-oriented research, as the building blocks from which
further advances in yields on the developed Cerrado soils can be built.
46. However, the collaborative nature of the CRSP demands a third product for the collabo-
rating (and other) countries. This is the development of new technology that helps farmers
solve their most pressing production problems. This requires the integration of biophysical
knowledge into a practice, or set of practices (the new technology), that farmers in the area
are able to use and are willing to adopt. One way to accomplish this integration is through
the development of predictive simulation models as outlined above. To extrapolate the
results the model would require validation at the new locations. But the results from this
approach will take time to develop, and will be more effective in predicting the conse-
quences of biophysical factors than socioeconomic factors.
47. An alternative approach includes a socio-economic component. Essentially this ap-
proach is directed at assuring that a technology which is physically and biologically promis-
ing, will also operate successfully under real farm conditions. The first stage in this is a
preliminary economic evaluation. This does not have to be a sophisticated procedure but
should be sufficient to eliminate those technologies which have no prospect of succeeding
under farm conditions. An important aspect of this is the labor component. For example,
the use of legumes and green manures has a significant cost in terms of farmers' time, both
for growing the crop and incorporating it into the soil. Some of this requirement may come
at a critical time in the farm schedule. For example, incorporation into the soil gives more
efficient use of the N, but mulching the surface needs less labor and hence may be the
48. Clearly this initial screening can be used only to eliminate unsuitable technologies but
the others must be validated under the array of real farm conditions in which they are
expected to be used; farmers must be incorporated in the evaluation process. Though no
validation has been done to date, the Panel feels there is some technology which has been
developed sufficiently on the CPAC station that it is ready to be validated in farmer's
49. One example is that from the research currently being conducted under Project 103. Six
legumes will be selected for further screening for dry-season survival. These six could be
tested against farmers' usual practices on at least ten farms in addition to the CPAC station.
The soils and previous practices would have to be characterized at each location, but no
attempt should be made to standardize or make corrections other than assure that the same
fallow or non-legume crop is used. By selecting a wide range of soil environments, the
selected legumes would be evaluated or validated for a number of biophysical and socio-
economic conditions encountered in the Cerrado soils. Small and large farms, for example,
could both be used for the on-farm research.
50. On-farm research on the Cerrado and related research at CPAC would require an inte-
grator of CPAC to carry out the same kind of function as the integration of modeling on the
Cornell campus. The CRSP team leader would be the logical choice for this assignment,
provided he could be relieved of some of his other duties.
51. The Panel is aware of the fact that on-farm testing would add a new dimension to the
work of the CRSP, a dimension that can make heavy demands on manpower and transport
and is difficult to fit into the research program of a graduate student. CPAC is, however,
strengthening its farming systems unit with a view to doing more work on-farm testing and
it, in cooperation with EMBRAPA should provide the logistical support for the work. -
Involvement of graduate students at the farm level has an important training and orientation
aspect which would be of substantial value in training such students in problem definition.
52. The duties of the on-site coordinator are many and diverse. And the progress reports
indicate that he has to spend a large amount of time on the logistical aspects of the project.
Graduate students need a lot of support in order to get on with their thesis work and there
are many other aspects of the program where a facilitator is essential. Nevertheless it is not
a very productive use of a professional's time to do these essential but non-scientific tasks.
The panel therefore urges the management to explore other means by which these logistical
support tasks could be done, giving the on-site coordinator more time for the research
SUMMARY OF RECOMMENDATIONS
53. The Panel strongly supports the Management Entity in its requests to the Program
Coordinators for a clear statement of the problems being addressed and their relevance to
54. The Panel recommends that attempts to improve the interaction of CPAC staff and
graduate students from Cornell be continued.
55. The Panel recommends that Topic 2: obtaining a quantitative description of water and
chemical budgets in acid-Savanna soils be used as the definition of the overall objective of
the Cornell/CPAC Program. The development of a holistic approach will require the in-
volvement of a scientist with this capability.
56. The Panel suggests the formulation of a longer-range strategy for the joint program. The
development of a multi-year program will provide an opportunity for Cornell and CPAC to
examine their congruent long-term needs. The on-site coordinator should have a major role
in formulating this strategy.
57. The Panel believes that there is a major role for modeling in the program, especially in
the soil-plant atmosphere continuum, and recommends that this work continue to receive
major emphasis; the primary responsibility for overall systems design and system module
development should be with the Cornell faculty.
58. The Panel endorses work that would lead to better characterization of the extent and
causes of the problem of deterioration of soil physical conditions. Work should focus on
both physical and chemical aspects of structural degradation.
59. The Panel recommends that the work at CPAC move to on-farm validation in some of
the projects. This should follow an initial screening of potential technologies with regard to
their likely economic benefits. On-farm validation has an important training and orientation
dimension for graduate students.
60. The Panel recommends that the on-site coordinator be relieved of some of the non-
scientific tasks which, while very important, could be done equally well by someone else.
Significant research results to date from the TropSoils Program and some possible manage-
ment implications include the following items:
1. A green manure, mucuna preta (Mucuna aterrima [Piper & Tracy]), was found
to provide 260 kg/ha as a sole source of nitrogen for production of 6.8 t/ha of maize
2. Incorporated mucuna preta showed a marked residual effect, supplying a limited
amount of nitrogen to a second maize crop while serving to increase the indigenous
population of mycorrhiza in the soil, thus improving phosphorous uptake.
3. The quantity of inorganic nitrogen accumulated in fallow soil was significantly
proportional to nitrogen uptake by maize, proving that a short-cut procedure can be
used for evaluating nitrogen mineralized after incorporation of plant material in
small bare fallow plots as a technique for screening large numbers of potential
legumes as green manures.
4. Burial of soil/green manure samples in the field enclosed in polyethylene bags
allowed mineralization to proceed normally, preventing nitrate escape to the sur-
rounding soil, thus serving as a labor-saving technique to screen large numbers of
candidate legumes as sources of nitrogen.
5. Incorporation of a green manure to Acrustox and Haplustox soil samples re-
duced sulfate adsorption, increased pH levels, and resulted in manure incorporation
other than nitrogen supply.
6. Soil genesis studies verifying the presence of a water table within 2 m of the
surface of a red-yellow Acrustox during a "veranico" dry spell emphasized the
importance of developing management strategies to take advantage of soil water
made available by deep rooting of annual crops grown on these soils.
7. High mechanical impedance in a Haplustox under cultivation for 11 years from
the Sao Gotardo region of Minas Gerais State was partially alleviated by chisel-
plowing, permitting deeper root proliferation, leading to a significant yield increase
in rainfed wheat.
8. Addition of calcium to acid subsoils was unable to overcome the detrimental
effects of even slight soil compaction on wheat root penetration; this serves as an
alert for Cerrado region farmers that physical barriers to root penetration, aggra-
vated by continuous disc tillage practices, may deter root growth in improved
profiles just as severely as chemical barriers constrain root growth in unamended
9. Compaction of a Haplustox with high clay content (70%) in a greenhouse ex-
periment restricted root penetration much more severely than the same soil with
lower clay content. One of the causes of concern with this finding is that even those
soil maps that do exist for the Cerrado region do not specify soil texture classes at a
level of detail useful for delineating these problematic soils.
10. On-farm soil sampling in fields subjected to nine years of heavy-disc primary
tillage confirmed the development of root-restricting high mechanical impedance on
Acrustox sites with high clay content (80%) in the southeast Federal District.
Nearby fields with ten cropping cycles under disc or moldboard plow primary
tillage retained desirable physical characteristics in spite of normal field traffic and
secondary tillage by disc harrowing.