.. .33.
$3.;

 

if 3?...
13:...) 31;;11!

«. . .n...

.5? int... a:

 

 

MICHIGAN 8

ll l l "‘ “Will?

\ll llllllllll

3 1293 01020 7599

“L

 

 

 

 

 

 

‘l

 

This is to certify that the

thesis entitled

A Review of a Silvopastoral Research Program
For Seasonally Dry and Drought Prone Areas of Jamaica

presented by

Abigail Susan Eaton

has been accepted towards fulfillment
of the requirements for

 

Masters degreein Forestry

 

   

Major professor

[hue September 30, 1994

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

Ll RARY
Michigan State
Unlverslty

 

 

 

PLACEII RETURN BOXtomnwombcnockunnomywncord.
TO AVOID FINES return onorblforoddoduo.

DATE DUE DATE DUE DATE DUE

i " 7'“th 0 901993‘

E-

l I l

MSU IoAn Alanna». Mon/Equal Opportunity lnotltwon
mm:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A REVIEW OF A SILVOPASTORAL RESEARCH PROGRAM FOR SEASONALLY
DRY AND DROUGHT PRONE AREAS OF JAMAICA

By

Abigail Susan Eaton

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Forestry

l 994

ABSTRACT

A REVIEW OF SILVOPASTORAL RESEARCH PROGRAM
FOR SEASONALLY DRY AND DROUGHT PRONE AREAS OF JAMAICA

By

Abigail Susan Eaton

A review of a silvopastoral research program was conducted. The program's
overall goal was the integration of tree fodder growth and management knowledge with
current small-scale silvopastoral management systems in Jamaica. This review of the
program had three objectives: 1) examining whether the five phases of the program
facilitated the on-farm research process; 2) determining whether the five phases of the
program were usefiil in assisting the program to meet its overall goal; 3) consolidating
findings gathered during these phases and providing fitrther direction should similar
agroforestry projects be taken on elsewhere. Key factors for conducting on-farm -
agroforestry research are presented. Among other things, findings indicate that a
continuous exchange of information between on-station and on-fann research, good
communication between researchers from different disciplines, timing of the phases of
research, and feedback to and from the local community are important to the outcome and

continuity of agroforestry research.

ACKNOWLEDGEMENTS

The research process. Over the last three years I've learned that this process can
mean very different things to different people. All of their perspectives are valuable, and
I‘ve tried to incorporate what I've valued most about each individual's perspective into my
own philosophy of what the process should be. It has been a learning experience.

There are many people to whom I am grateful. My committee members Dr.
Maureen McDonough, Dr. Douglas Lantagne, and Dr. George Axinn all enlightened me
with their personal perspectives and guidance on the process. I thank them for that. Dr.
Stanley Ries instilled in me a constant reminder of a need to maintain ethics and personal
integrity throughout the process. It is something I shall never forget. Dr. Michael Gold
added moral support and guidance and provided a "push" at times when it was sorely
needed. The support staff in the Department of Forestry can not be recognized enough
for their help in moving me and fellow graduate students through the maze of procedures
required to receive a degreeua special thanks to Barb Anderson for her patience.

It is not possible to adequately thank all of my fellow graduate students and other
fiiends for their support over the last few years. I only hope they know that I appreciate
all that they've done for me and that in some way I have been able to return the favor.

Finally I want to thank my family, especially my parents Wayne and Gwen, for
their unquestioning material and moral support at the times when it was most nwded. It
has made my life much easier. I regret that my father is no longer here to see the
culmination of the process. But I believe that in some way he will know how much I
appreciate what he has done for me and that he knows that I love and miss him.

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

1. INTRODUCTION AND OVERVIEW OF PROJECT

1.1
1.2
1.3

Introduction to the Project
Site Overview
Study Objectives

2. LITERATURE REVIEW

2.1
2.2
2.3
2.4

2.5
2.6
2.7
2.8
2.9
2.10
2.11

Introduction

Agroforestry - Why and How?

On-Station and On-Fann Research Linkages

Diagnosis: Assessing Knowledge, Constraints, and Needs
for Research

Experimental Concerns in On-farm Research

On-farm Research and Farmer Participation

Selection on On-farm Research Collaborators

Research Teams and Collaboration of Researchers
Evaluating On-farm Research and Agroforestry Interventions

Summary
Research Questions

3. MATERIALS AND METHODS

3.1
3.2
3.3

Materials
Methods
Organization of Findings & Discussion

4. FINDINGS
4.1 Phase One and Two: On-station Research and Indigenous Knowledge
4.2 Phase Three: Socioeconomic Factors and Farming Systems
4.3 Phase Three Continued: Analysis of Socioeconomic Data

Gender Issues
Farmer Cooperators

4.4 Phase Four". On-farm Research
4.6 Phase Five: Further Ventures in On-farm Research

4.8 Summary

iv

iii

iv

H

14
20
23
28
29
32
36
38

40
42
47

48
54
56
62
63
65
71
77

5. DISCUSSION
5.] A Review of the Research Questions
5.2 Climate variation: On-station and On-farm Sites
5.3 Farmer Collaboration
5.4 Tractor Use and Other Inputs
5.5 Disciplines, Training and Communication
5.6 Feedback to and fiom the Community
5.7 Conclusions

6. RECOMMENDATIONS
6.1 Farmers' Input
6.2 Further Study
APPENDIX

BIBLIOGRAPHY

79
83
85
88
90
91
92

94
94

96

98

Table
Table
Table
Table

Table
Table
Table

Table

PENN?"

39‘.“

LIST OF TABLES

Agroforestry Practices

Types of Farmer Participation

Sources of Information for Research Questions

Biophysical Characteristics of Moneague, St. Ann Parish and

Green Park, Trelawny Parish, Jamaica

Potential Fodder Tree Species Tested During Phase One

Summary of Selected Characteristics at the Green Park Community
Sources of New Technology Information Cited by Green Park Cattle
Farmers

Fodder Tree Species Tested On-farrn in Green Park, Trelawny Parish

25
42
49

51
57
60

68

LIST OF FIGURES

Figure l. JADF / MSU Agroforestry Program - Project Tirneline

vii

Chapter 1

INTRODUCTION
1.1 Inflation to the Project

Between March 1990 and July 1993, extensive research was conducted in Green
Park, Trelawny Parish, Jamaica and Moneague, St. Ann through a cooperative effort of
the Jamaican Agricultural Development Foundation‘ and the Department of Forestry,
Michigan State University (the J ADF /MSU Agroforestry Program). The project started
with the premise that fodder trees are extremely valuable livestock feed resources that are
underutilized in Jamaica (Roshetko 1991; Torres, 1983; Le Houe'rou, 1980; Pandy, 1982;
Malla, 1988; JLA 1983). Under this premise the systematic incorporation of tree fodder
into livestock production systems, particularly to improve dry season feed availability,
might be warranted. Thus the ultimate goal of JADF/MSU Agroforestry Program was
integrating tree fodder growth and management knowledge with current small-scale
farming systems to develop a sustainable small-scale silvopastoral management system
(Roshetko, 1991).

Five phases of research have taken place under the JADF/MSU Agroforestry
Program. Two initial phases to study the indigenous knowledge of small cattle holders in
Jamaica and conduct trials of potential fodder tree species appropriate to the area were
developed. These two phases were designed to establish baseline information (biological
and social) upon which fixture efforts would build (Michigan State University proposal to
JADF, 1989). While a desire to lead into on-farm research, based on the results of the
trials existed at this time, the additional three phases that were added to the program
evolved as new students were recruited into the project (Dr. Michael Gold, personal
communication). An on-station trial of potential fodder species in Moneague, and a study

of indigenous knowledge of fodder trees and indigenous silvopastoral systems in Green

 

1 The Jamaican Agricultural Development Foundation is a Jamaican non-governmental organization that
focuses on assisting small-farmers.

l

2

Park, were conducted concurrently (Roshetko, 1991; Monison, 1991). These phases
were followed by a year long study of socio-economic and farming systems with the intent
of incorporating additional knowledge gained about local conditions into the on-farm
research component (Andreatta, 1993). On-farm research began in 1991, taking the most
promising species from the on-station trials, along with some favored indigenous species
noted by farmers in the area, and testing them under farmers' conditions. Phase four
tested fodder tree establishment and species' response to coppicing and other silvicultural
practices (Krecik, 1993). Phase five continued the studies fi'om phase four and
investigated two species more in depth. In addition, the feed trials were conducted to
observe palatability and cattle response to the fodder tree species (Morikawa, 1993).
Figure 1 illustrates the sequence of the phases of research conducted from March
1990 to July 1993. This is the period spent conducting actual field research in Jamaica.

1.2 Site Overview

Both Moneague and Green Park were selected by J ADF at the outset as the sites
for on-station and on-farm research. The willingness of Alcan (Jamaica) Company's Rio
Hoe Farms to provide Michigan State University (MSU) full access to some of its land
and resources, facilitated the execution of controlled on-station experiments and provided
the primary impetus for JADF's selection of Moneague (Dr. Michael Gold, personal
communication). Green Park, on the other hand, was selected as the site for on-farm trials
because it is a drought-prone area, comprised of small-scale farms (Morrison, 1991), with
a relatively large number of resource users raising livestock. In this instance, small scale
cattle farmers are defined by Morrison (1991) as those having fewer than 40 head of

cattle.

ES» :2. a 38 28 3 «.63 20c he 830.980 05 has 3:88
Bea 80.: 9.3:! g .63 3 38—950 803 5.2—2 6323 3:222: «53033 was amen—28 8.5 .BEEun E 383:8 5.3% «60%. 085 3 28 5&2 03—05: 3E. a

 

 

3.5. woo..— d .8355 EBB—33.3m—
AI 5.88. 5.5.8 3580 iv

8882.. a ismfiam 8:. 33.2
r {a 580 a. as: 55.8 aaom .v

etc.— 505
AI 8.31m weigh. as via—808.com iv
fem 50.5
$322.53
.853...

03382 . 53:0
A: gig 3:. .636..— 1v

 

 

goons-moonszg m <33: a: flinchananzzfio
Maw Na— na. ga—

. 8:95.... .825 . acuEm cease? am: a .35. ._ 2am

 

4

Green Park is a former sugarcane plantation. Bought in 1958 by the Kaiser
Bauxite Mining Company, it was originally designed as a resettlement area for displaced
residents fi'om St. Ann Parish, living in an area that Kaiser wanted to mine. Potential
resettlers were given the choice of land in Green Park or monetary compensation for land.
In some cases the resettlers took the land in Green Park and later sold off parcels to
inhabitants nearby, or to Jamaicans working abroad who returned and were looking for a
place to settle down (what Andreatta (1993) terms ”circular migrants"). It is unlike most
areas in Jamaica, where farmers typically acquire their holdings through ancestral
inheritance (Morrison, 1991; Andreatta and McDonough, 1992). Most of the community
is composed of those who either relocated fi'om nearby parishes, grew up in the area, lived
outside the valley and purchased parcels fiom resettlers or other residents, or circular

migrants. Only about 20 percent of the Green Park community are resettlers.

1.3 Study Objective
This paper reviews the JADF/MSU Agroforestry Program. There are three study

objectives:

1. To examine whether the various phases of the program facilitated the on—farm
research process.

2. To determine whether the various phases of the program were usefiil in
assisting the program to meet its overall goal: the integration of tree fodder growth and
management knowledge with current small-scale farming systems to develop a sustainable
small-scale silvopastoral management system appropriate to dry and drought prone areas
of Jamaica.

3. To consolidate the findings gathered during these phases in order to better
understand reasons for either succeeding of failing to meet the goal of the project as well
as provide further direction should similar agroforestry projects be taken on elsewhere, if
not retuming to Green Park itself. Were project phases effective at building a base of
knowledge and awareness for facilitating the implementation of subsequent project
phases?

5

Thus, by examining the research conducted from 1990 to 1993 under the J ADF/
MSU Agroforestry Program, information on indigenous knowledge, socioeconomic
factors and farming systems, and biological research and tree fodder production systems
are consolidated to further define strengths and weaknesses that may be built upon (or

avoided) in the future.

Chapter 2
LITERATURE REVIEW

2.1 Int e 'on

The J ADF/MSU Agroforestry Program was an on-farm research program for the
eventual development of sustainable small-scale silvopastoral management systems
appropriate to dry and drought areas of Jamaica. The literature review focuses primarily
on lessons learned about on-farrn research to date. Initially the literature review addresses
the field of agroforestry. Discussed are the importance of agroforestry, how agroforestry
should be conducted, and why on-farm research is important in agroforestry. Literature
pertaining to on-farm research for both agriculture in general and agroforestry are tapped
because there is a great deal to be learned from the longer history and experience of on-
fann research for agriculture. Factors pertaining to on-station and on-farm research
linkages, farmer participation, assessment of characteristics, needs and constraints of
farming systems and communities, experimental considerations in on-farm research,
selection of farmer cooperators for on-fann research, and research teams and
collaboration for on-farm research also are addressed. Finally, considerations for
evaluating on-farm research and agroforestry programs are discussed. Based on findings

fiom the literature review, research questions are then formulated to guide the study.

2.2 A rofo -Wh and How?

Agroforestry is a collective name for land use systems and
technologies in which woody perennials (trees, shrubs,

bamboo, etc.) are deliberately combined with herbaceous

crops and/or animals, either in some form of spatial arrangement
or temporal sequence. In agroforestry systems there are both
ecological and economic interactions among the different
components (Raintree, 1987).

7

As a focus of organized scientific activity, agroforestry is relatively new in
comparison to that of agricultural crops (Raintree, 1991). Yet farmers worldwide have
utilized agroforestry techniques as an integrated approach to land use for hundreds of
years. Indeed, as Nair (1993) states, agroforestry would seem to be a new name for a set
of old practices in many nations.

Agroforestry represents an interface between agriculture and forestry and
encompasses mixed land use practices. These practices have been developed primarily in
response to the special needs and conditions of resource poor farmers in tropical
developing countries that have not been satisfactorily addressed by advances in
conventional agriculture and forestry (Nair, 1993). During the Green Revolution2 many
high yielding cereals and related technologies were developed that placed a high demand
on increased fertilizers use and other high-cost inputs. While these crops are very
productive, they are also beyond the reach of many resource-poor farmers in developing
countries (Nair, 1993).

Realizing that something must be done to address the needs of the resource-poor,
there was a resurgence of interest in the concepts of intercropping and integrated farming
systems. These were recognized as systems that are both familiar and relevant to the
resource-poor farmer. This resurgence, combined with increasing concerns about the
environment, especially tropical deforestation, eventually led to many studies that showed
the advantages of combined production systems of crops, trees and animals (Nair, 1993).

There is a consensus that agroforestry is practiced for a variety of objectives (Nair,
1993). Nair (1992) identified at least 18 common yet distinct agroforestry practices
(Table 1). Of interest to this study are the variety of practices that come under the label of

silvopastoral or agrosilvopastoral systems. In silvopastoral systems, trees are

 

2 The Green Revolution represented a period during the 60's and early 70's when several International
Agricultural Research Centers (IARCs) and the Consultative Group on International Agricultural
Research (CGIAR) undertook research with the objective of enhancing the productivity of major
agriculturalcropsandanimalsofthetropics.

Table 1. Common Agroforestry Practices in the Tropics (Nair, 1992)

 

 

 

 

 

Agroforestry Practice Arrangement of Components
Agrisilvicultural Systems
Improved tree fallow Woody species planted and left to grow during the
"fallow phase.”
Modified taungya Combined stand of woody and agricultural species

duringbcarly stages of plantation establishment.

 

Alley cropping/hedgerow intercropping

Woody species in hedges; agricultural species in
alleys between hedges; microzonal or strip
arrangement.

 

Home gardens

Intimate, multistory combination of various trees
and crops around homesteads

 

Plantation-crop combinations

Shade trees for crops; intercropping with crops;
integrated, multistory mixtures of plantation
crops; mixtures of plantations in alternate
arranggnents.

 

Multistory tree gardens

Multispecies, multilayer dense plant associations
with no yanked planting arrangements.

 

Windbreaks and shelter-belts, livehedges

Trees around farmlands / plots.

 

Trees in soil conservation and reclamation

Trees on bunds, terraces. raisers, etc., with or
withoutjrass strips.

 

 

Multipurpose trees on croplands Trees scattered haphazardly or systenuuically on
bonds terraces. or plot / field boundaries.
Fuelwood lots Interplanting fuelwood species on or around

 

agricultural land

 

Silvopastoral Systems

 

Trees on rangeland ourastures

I Trees scattered irregularly or systematicalL

 

Silvicultural Systems

 

Fodder banks

Production of protein-rich tree fodder on farm /
rangelands for wt-mdmfofier production.

 

 

Agrosilvopastoral Systems

Plantation crops with fodder and livestock

Cattle under coconut crops (example).

 

Home gardens with animals

Multistory combination of various trees and crops,
as well as animals around homesteads.

 

Multipurpose woody hedgerows around
homesteads

 

Woody hedges for browse, mulch, green manure,
soil conservation, etc.

 

 

 

 

 

Other (Special) Practices
Apiculture with trees Trees for honey
Aquaforestry Trees lininLfishponds
Multipurpose woodlots Various purpose (wood, fodder, soil reclamation,

 

etc.)

 

 

9

incorporated into pasture and/or animal production systems. Practices range from
relatively simple, i.e. planting trees on rangeland or pastures, to combining plantation
crops with fodder and livestock.

A great deal has been written about the potentials for agroforestry over the last 15
years. However there are several characteristics of agroforestry that complicate the
research process including poor understanding of farmers' socioeconomic backgrounds
and technical agroforestry strategies, lack of basic technical information about
agroforestry systems and components, system complexity and variability, lack of locally
validated agroforestry technologies, and lack of data for agroforestry research and
development policy (Scherr, 1991). The extremely site-specific nature of agroforestry,
conditioned by biophysical and sociocultural characteristics, poses serious difficulties in
defining precise recommendations of wider applicability (Nair, 1992; Scherr, 1991).
Despite this, development agencies and donors with a global mandate to addressing the
needs of both the resource poor and the environment, need information to formulate
policy on agroforestry with a wider applicability to more regions in the world (Nair,
1992). Nair (1992) asserts that in order to channel development assistance to agroforestry
programs, researchers must develop and provide specific recommendations on the most
appropriate agroforestry systems and practices for different parts of the world.

Raintree (1983) states that the design of a program for research on agroforestry
interventions must be based first upon the diagnosis of the farming system. Unlike much
agricultural commodity research that is both established and has research validated
technologies for production of individual crops (Roling 1982, p98), agroforestry has only
a few such technologies. In the former, farming systems research (F SR) was developed to
identify constraints to adopting existing technologies and as an attempt to adapt
technologies to meet farmers' circumstances. Farming systems research departed from
previous agricultural research approaches by focusing analysis on households rather than
solely on crops (Buck, 1990). In agroforestry, however, in view of what Raintree (1991)

10

terms the ”relatively undomesticated nature of the vast majority of trees used in
agroforestry," technologies and designs need more research before they can be
recommended for adoption. This factor, along with the complexity and scope of
agroforestry systems, led the International Center for Research in Agroforestry (ICRAF)
to a variation of the farming systems perspective as a method that could deal
comprehensively with the field of agroforestry.

In 1981, ICRAF began to develop a FSR based approach appropriate to
agroforestry research known as "Diagnosis and Design" (D&D) (Raintree, 1987). Beyond
the diagnosis of problems or constraints in standard FSR methodology, D&D must also
design and evaluate potential technologies fiom which research needs and requirements
might be derived. Raintree (1990) adds that at a micro-scale (e. g. household and farm
level) D&D is quite consistent with the general pattern of FSR and that D&D methods
might be considered an agroforestry-specific variant. Nevertheless, when ICRAF initiated
systematic work on the development of a diagnostic methodology for agroforestry, there
was the thought that this work would eventually lead, through trial and feedback, toward
a single optimal set of procedures for agroforestry research (Raintree, 1990). However,
as researchers gained experience with D&D in different settings with different
collaborators, rather than deve10ping one optimal set of methods, users had different needs

and resources and preferred to use different adapted forms of the same underlying logic

 

(Raintree, 1990). In comparison to FSR methodologies (Shaner et a1, 1982), Raintree
(1990, p. 44) notes that the most distinctive features of the D&D methodology that

clarify its underlying logic include th following:

1) a unique focus on the role of trees within the whole
system (and hence a larger diagnostic scope relative to agricultural
commodity-based technology).

2) a basic needs approach to identifying and evaluating
household production subsystems in terms of the farmer's
objectives (i.e. specifying components of the farming system in terms

1 1
of their role in satisfying basic household needs).

3) a trouble-shooting procedure for identifying critical constraints,
leverage points, and functional specifications for agroforestry
interventions (through a great emphasis on the iterative nature

of the basic D&D process).

4) a separate assessment of the sustainability of the land-use system
under the most likely firture scenario.

In a review of agroforestry practices in sub-Saharan Afiica, Cook and Grut (1989)
found that, to be successful, projects introducing new agroforestry technologies should
follow at least the basic established patterns of farming systems research. This pattern
starts with diagnostic research at the farm level to identify problems, moves to on-station
research to find possible solutions, then to adaptive on-farm research linked to extension,
demonstration and adoption. Findings indicated that a failure to follow this pattern leads
to a poor ”fit” between projects and farmers needs (Cook and Grut, 1989).

Whatever methods ultimately are used to identify relevant agroforestry
technologies and initiate trials with the intended users, it is almost certain that the initial
technology that is identified can be improved over time (Raintree, 1990). Because of this
a special emphasis is placed on the need for an iterative process, or "re-diagnosis” and "re-
design," in agroforestry research. Introducing a new technology may, by itself, modify the
diagnosed situation and warrant a repeat diagnosis. In addition, no technology that is
selected solely through a discussion with the intended users can be regarded as properly
evaluated by the intended users until they have had time to live with it on their farms for
an adequate trial period (Raintree, 1990).

Scherr (1991) stresses that the several characteristics of agroforestry that
complicate the research process, as mentioned earlier, tend to make on-farrn research

unusually important in agroforestry research. Indwd, Agroforestry Systems ( 1991)

12

devoted an entire issue to discussing this topic due to its relative importance to the
development and improvement of agroforestry technologies and systems.

In her discussion of on-farrn research for agroforestry, Scherr (1991) highlights
five important components of the agroforestry research agenda, all emphasizing studies

undertaken on-farrn, or in close interaction with farming communities:

1) diagnostic studies to determine the potential relevance and use of agroforestry
systems for specific farming systems;

2) studies to document basic information about trees and shrubs used in
agroforestry;

3) strategic studies on basic processes of agroforestry response and interactions3;

4) design and testing of agroforestry technologies with farmers, to assess the
technologies' development potential for specific farming systems (participatory
research);

5) studies of changing patterns of agroforestry in land use and policies to
effectively support farmer adoption of agroforestry.

Scherr (1991) adds that, as in D&D, these complementary components of research
should ideally run in parallel, with close linkages and integration to allow for an "iterative

re-diagnosis and redesign" based on new findings.

2.3 Qn-stgtion and On-farm Research Linkages

"On-farm research (OFR) is a problem-oriented approach to agricultural research that
begins by diagnosing the conditions, practices. and problems of particular groups of
farmers. Once the problems are identified, a research program is designed to address
them. A key part of any such program is conducting experiments on farmers' fields
under farmers' conditions and management. Those experiments are then evaluated
usingcriteriathatareimportanttothefarmer.andtheresultsareusedtomake
recommendations.” (Tripp and Woolley 1989).

 

3 Due to the system complexity and variability found in agroforestry, such systems present greater
challenges for research design and implementation. "Strategic” in this sense means that scientists must
move cautiously into committing resourees for long-term agroforestry experiments. Each experiment
needstobedesignedtomeetpriority reswrchobjectivesforawell-definedresearchdomain (Scherr,
1991).

13

Despite a recognition for the need to conduct experiments under ”farmers'
conditions and management," on-farm research most commonly seeks to validate and
demonstrate technologies developed under controlled experimental conditions (Sumberg
and Okali, 1988; Francis and Attah-Krah, 1989). Sumberg and Okali (1988) argue that
on-farm research has a role in all stages of technology development, but objectives and
methodologies transplanted from the research station are inappropriate, particularly given
the restrictions they impose on farmers' capacity to experiment with new technologies.

Before agroforestry technologies can be taken to the farmer, even on an
experimental basis, questions pertaining to appropriate tree species, and reliable,
economical establishment methods must be addressed (Sumberg and Okali, 1988).
Inevitably, new research questions will arise from the experiences of farmers that may
most appropriately be explored in highly controlled research plots (Collinson, 1987). It is
this dynamic and iterative interplay between a well balanced mix of on-station and on-
farm research that can help realize the potential of on-farm research for agroforestry
(Sumberg and Okali, 1989; Kerkhoff, 1990; Scherr 1991). Scherr (1991) emphasizes that
this dynamic interaction must be continuous and not sequential. Regular farmer visits to
station trials will enhance early feedback on technologies under testing before trials are
attempted in farmers' fields (Pinners and Balasubrarnanian, 1991).

According to Kerkhof (1990), the question of the relative merits of "on-station"
and "on-farm" research is still hotly debated in agroforestry circles. It is very difficult to
design trials that provide statistically valid results, while at the same time reflect the
complexity and variability of local farming conditions. The only way around this problem
seems to be a combination of both on-farm and on-station trials (Kerkhof, 1990).

Flora (1992) notes that before a technology is developed and tested the challenge
for farming systems research is to determine what is crucial about the environment before
a technology can be, 1) effective (ofien based on agroecological variables) and, 2)

adopted (often based on socioeconomic variables). The constant challenge is to determine

l4

gm'ckly and parsimoniously the key difi‘erentiating characteristics in each situation that will
allow the development of the best technology most likely to be adopted to meet the needs
of the farm family (Flora, 1992).

The following two sections discuss some methods and considerations for acquiring
needed information (both socioeconomic and biophysical) for agroforestry technology

development and on-farm research.

2.4 Diagnosis: Assessing Knowledgg, Constraints and Needs for On-farm Reseagh

Tripp (1991) describes the following problems as seemingly endemic to on-farm
research. First, although on-farm research was developed to address the needs of
resource-poor farmers, defining client groups has been a major problem. This in itself is
not surprising, as Rolings (1988) notes that targeting research and extension towards
resource-poor farmers has always been difiicult. However, even when researchers may be
able to describe farming practices in the area, they often locate trials on unrepresentative
fields, ofien producing results of little relevance to the majority of farmers (Tripp, 1991b).
Second, poor diagnosis is often a problem. Extensive formal surveys, covering every
facet of a farming system, have been a feature of many on-farrn research projects. Often,
however, there is little connection between the data generated and the research priorities
that eventually emerge ( Flora 1992; Tripp, 1991b).

Tripp and Byerlee (1988) make several suggestions, based on past experience, that
would allow the social scientist to play a more effective role in on-farm research. To be
efi‘ective, social science research methods need to be flexible, relatively simple, well
focused and rapid. Efi’ective participation (of farmers) in diagnostic activities for applied
research requires timely results for planning experiments, and informal but well-focused
surveys are best suited to meet that need (Tripp and Byerlee, 1988).

In very general terms, surveys can be separated into two types: informal and
formal (or structured). Simply put, an informal survey begins with conversations with

15

farmers (or other informants), based on an initial set of guidelines or hypotheses, and
centers on questions or topics which are of interest to a researcher or program. As
information is gathered, other topics or concerns that arise out of conversation are
pursued. Results are then written up and should be used for project planning (Tripp,
1991a). As Tripp (1991a) notes, ”informal" is perhaps a misnomer, especially if it gives an
impression that surveying is done in a casual or haphazard manner. Considerable
experience is usually required to plan and carry out an informal survey (Tripp, 1991a).
Informal surveys are often followed up by formal surveys, although this need not be the
case if informal surveys are implemented carefirlly. (Tripp, l991a&b).

A formal survey, which uses a structured questionnaire, is nevertheless usefirl for a
variety of reasons: 1) it provides good baseline data against which progress of adaptive
research can be measured; 2) it presents the opportunity to quantify the most common
practices in an area and to form a more precise idea of what constitutes "typical" or
"average" farmer practices; 3) the results can provide a better understanding of variations
in farming practices; 4) a more precise estimate of farmers' perceptions of production
problems can be developed; and 5) if well conducted, formal surveys can contribute to
further exploring the causality of production problems, which helps identify problems
(Tripp, 1991a).

Kerkhof (1990), after surveying 21 agroforestry projects in Africa, found that well-
designed surveys followed by proper monitoring and evaluation are crucial in putting
projects on the right track. However, Kerkhof(1990) also found that projects had a
tendency to implement elaborate, detailed surveys that risked taking huge quantities of
time and effort while yielding few practical results. Such situations infer a need to strike a
proper balance between the information needed and the amount of time needed to collect
the data. Where resources are limited, the best approach is to keep things simple and
focused (Kerkhof, 1990). This means carefirlly designing a survey that focuses as closely

as possible on the issues (biophysical or socioeconomic) most relevant to the project. ”It

16

is important that surveys (regardless of what type) are treated as an integral part of the
project if they are to provide usable results. Involving extension staff in carrying out
surveys, and in analyzing their results, is usually the best way of doing this; it also has a
usefirl stafi‘ training fimction" (Kerkhof, 1990 p.7).

Structured surveys may be important for verifying selected findings of an informal
survey, quantifying some important variables, providing a cross-check for the informal
survey, and lending greater credibility to a diagnostic exercise (Hildebrand, 1986).
Hildebrand (1986), suggests that a formal, structured survey may be replaced by a slightly
longer and more carefully managed informal survey than would be otherwise conducted,
or by two or more informal surveys. Implementing two or more informal surveys would
be most relevant in areas with more than one cropping season, or with distinct seasonal
weather patterns if addressing issues farmers face at those different points in time.

Another alternative for diagnosis is the Rapid Rural Appraisal (RRA), which now
refers to a wide range of techniques and methods. Several principles illustrate the key

features of RA (Scoones and MacCracken, 1989):

1. Rapid Rural Appraisal comprises a series of choices for any given situation.

 

2. The process of RA is iterative. Questions are added or revised by the team as
information progressively is revealed

3. A multidisciplin_ag team is utilized, allowing for a broad spectrum of insights.

4. RA is semi-structured - it is systematic while at the same time maintaining
flexibility and adaptability.

5. Triangulation is used to achieve accuracy. This involves using diverse
methods and information sources rather than statistical replicability.

6. Unnecessary detail is avoided through optimal ignorance.

7. Trade-offs between precisicn-breadth--depth-timeliness are made explicit
through appropriate imprecision.

l7
8. Recognizing that researchers interfere, RRAs attempt to make biases explicit.

9. The process actively seeks out dialogue with local people and attempts to
involve local people in the research process and planning. Thus, it is participatog,
and as such it is best carried out by and with local people.

RA is not a substitute for other information gathering techniques. RRAs seek to
inform, usually early in a project's development, and can complement more conventional
surveys or anthropological techniques. Weaknesses in RRAs may arise if team members
fail to appreciate each others' disciplines, if team members insist on rigid control of
variables, or the process is allowed to become too open-ended (Butler and Butler, 1987).
Nevertheless, when time and person power are limited, RRAs may be a superior
alternative to either conventional surveys or anthropological techniques (Gibbs, 1985).

The sondeo (Horton, 1986; Hildebrand, 1981) is one method ofRRA To identify
the specific goals, local conditions, and resources available, sondeos usually involve
open-ended questions and dialogue with farmers, as opposed to long structured
questionnaires. Usually conducted in interdisciplinary teams, a focus on specific topics or
hypotheses intensifies as an informal survey progresses and new information is gathered .
RRA is not a standardized method. There are a wide range of possible activities that may
be utilized, but it attempts to be systematic in order to be replicable (Gibbs, 1985).

Participatory mral assessment (PRA) was developed in response to some of the
weaknesses and other criticisms of RA. PRA is RRA with firll participation of the
community (Etling and Smith, 1994). PRA is designed to focus on rural communities,
systematize rural participation, and help communities establish resource management
plans. PRA focuses on natural resource management and involves specialists from various
disciplines and organizations (Ford, 1989). Local villagers cooperate actively in each step
of PRA. Theoretically these steps include the following: 1) site selection, 2) preliminary
visits by the PRA team, 3) collecting data, 4) data synthesis and analysis, 5) making
problems, 6) making opportunities, 7) adopting a village resource management plan, and
8) implementing the plan. Teams are usually composed of four to six specialists (Etling

18

and Smith, 1994). Spatial, temporal, social, and technical data are emphasized in
collecting data. Spatial data come fi'om village sketch maps compiled with village leaders,
a village transect with land uses, and individual farm sketches (usually six to eight farms).
Temporal data include a seasonal calendar depicting land use, food surpluses and
shortages, cash availability and shortages, etc., and a time line of events depicting changes
in resources (i.e. deforestation, soil loss, rainfall, population). Social data is collected
through farm interviews carried out at the households where individual farm sketches are
compiled (Efling and Smith, 1994).

Once data is collected, it is organized and lists of problems and opportunities for
action are compiled with village representatives. Villagers are then gathered to discuss the
listed problems and assign priorities to them, as well as to define and rank opportunities
that may address the most severe problems. A written plan is then developed that
describes the opportunities to address problems, specific actions to be taken, individuals or
groups responsible for various actions, resources needed, and a timeline for completion of
activities. A village leader usually guides the implementation of the plan (Etling and
Smith, 1994).

The advantages of PRA are numerous: 1) the creation of visual materials that
villagers can easily understand; 2) involvement of villagers, village groups, and interested
agencies through systematic participation; 3) interactive problem solving and
interdisciplinary problem solving; 4) identification of village based priorities; and 5)
application in the field quickly and inexpensively (Etling and Smith, 1994). The primary
disadvantage of PRA is that it omits evaluation, as it tends to end with implementation. In
addition, PRA fails to address the problem of a lack of leadership and program
management skills that exist among the farmers. These skills are needed in implementing
plans, and the lack of which usually cause greater problems than the lack of resources or
technical expertise (Etling and Smith, 1994).

l9

Rocheleau (1991) suggests that researchers need to expand their repertoire of
appraisal techniques, to both improve capabilities for participatory research and so that
individuals and institutions have a wide range of choices to use within a multiplicity of
local and national conditions. Several appraisal methods are listed that can help to
describe a particular place and situation and to direct research and development plans to fit
rural peeple's realities, including some of the methods mentioned previously (i.e. Rapid
Rural Appraisal, D&D, Agroecosystem Analysis, Participatory Rural Appraisal,
Ethnoecology, and syntheses of some or all of these). The danger with m appraisal
method is that a single appraisal or survey may not be adequate for long-term planning or
may raise expectations and leave residents with what a fi'ee "diagnosis" and a prescription
for "medicine" that is not locally available (Rocheleau, 1991 ). There is also the risk that
information collected will be used in an "extractive" (as opposed to interactive) manner by
researchers‘. Whatever appraisal method is chosen, to be effective it must lead to action
with continuing participation of rural people (Rocheleau, 1991).

There has been too much emphasis on single rapid appraisal techniques to
understand priority roles for agroforestry in farming systems (Scherr, 1991). Commonly
used appraisal methods were originally intended only as a starting point for researchers to
construct a common understanding of a system's problems and identify priority topics
requiring further in-depth diagnostic research (Raintree 1987; Scherr, 1991). The iterative
process of "re-diagnosis and re-design" emphasized in D&D has often been only
emphasized in principle, leaving the initial rapid appraisal as the basis of an entire program
of experimental research (Scherr, 1991).

While Scherr (1991) contends that current levels of understanding of the role of
agroforestry in land use is still too underdeveloped to be trusted to rapid appraisal alone,

she agrees that farmer surveys also are not especially reliable for understanding farmers'

 

4 Extractive in this case pertains to the use of information for a reseocher's own information or agenda
Local people may contribute work, knowledge, and other resources, but the information acquired is not
necessarily used for the benefit ofthe people from whom it is collected.

20

strategiesjin establishing and managing existing agroforestry systems. Instead, researchers
and farmers will have to work together on-farm, in exploratory or "diagnostic" trials with
new components, sites, configurations or management systems, to determine jointly which
are suitable for more research (Scherr, 1991). Flora (1992) notes that a major
recognition of farming systems research in recent years is the iterative nature of diagnosis.
Conventional wisdom among practitioners (of farming systems research) now holds that
the best diagnosis occurs when working with the farmer or farm family in actual on-farm
trials. As Raintree (1990, p.50) asserts " on-farm researchers are, almost be definition, in
the community for the purpose of acquiring information. " In addition to initial appraisals,
several years of on-farm diagnostic research may be needed before investment in long-
temr formal experiments can be justified, or the specifics of experimental design be
appmpriately selected (Scherr, 1991).

Nevertheless, whether sooner or later, experimental concerns for both long-term
and participatory experiments in agroforestry must be considered. The following section

discusses some of these experimental concerns.

2.5 Exmrimental Concerns in On-farm Research

Rocheleau (1991) notes that much of the literature on on-farm research discusses
methods of reconciling statistically valid experimental designs with field conditions. The
most fi'equently used methods are those that allow for control plots, some variation in
treatment, and statistical analysis of variable performance by different treatments within or
between farms.

In some concluding observations from Farmer First, Chambers et al (1989)
comment that almost all the on-farm experiments described within their book were
designed to produce statistical data in some form. They indicate that there is a need for
on-farm numerical (statistical) results that can be used by other agricultural institutions,

but that many people would agree that the most important indicator in evaluating new

21

technologies is farmers‘ adoption. Where other statistical results are needed, the analytical
rigor required differs according to whether the analysis is to help farmers or to help the
experiment station. Where the primary purpose of analysis is to help the farmer,
"techniques in experimental design and analysis which were thought not to be very
relevant or helpful (in on-farm research) included standard randomized block designs,
Latin squares, and factorial and multiple treatment structure with analysis of variance. In
contrast, some techniques found useful to farmers' understanding include scatter diagrams
for presenting results from a number of farms, and linear regression, for investigation into
the stability of biological systems" (Chambers, et al, 1989, p 159).

In a survey of on-farm research, it was found that common weaknesses in the
research process included unanticipated high variability between farmers' management
practices, insufficient participation of the farmers, and loss of farms fi'om the study.
Successful trials were simple, with few treatments and close interaction with farmers
coupled with a flexible design and assessment (Lightfoot and Barker, 1988).

Rocheleau (1991) notes that farmer and researcher preferences for different
designs may differ substantially. Just as outside researchers can gain substantial
information by varying treatment between farms, farmers may gain and share more insights
by having controls and a range of experiments to compare close-up. Farmers may not
appreciate the placement of various treatments within a randomized block design,
especially agroforestry treatments that cannot be divided into small portions of a cropped
plot (Rocheleau, 1991). Additionally, standard statistical approaches do not overcome the
problems of high levels of within and between farm variation often encountered in on-farm
trials.

Nothing can substitute for knowledge about levels of variability in trial sites
(Pinney 1991). Pinney (1991) suggests that since neither a researcher controlled, transfer
of technology approach, nor farmer controlled experimentation seem satisfactory for

22

agroforestry; a clear need exists for joint researcher-farmer trials to answer certain types
of questions.

To link farmers' research interests in their own land-use system with more formal
on-farm testing procedures, Ashby (1986) used "the decision-making approach." Under
this approach the essential features of a technology are communicated to farmers who are
considered both innovative and representative of a wide range of socio-economic levels.
Farmers then teach researchers about existing local practices relevant to the proposed
technology, enabling researchers to lead discussions with farmers about possible
treatments or innovations. From common elements of various farmers' ideas, the
researcher designs treatments to be tested. Rocheleau (1991) contends that group-
focused farm trials that combine different real-scale treatments on various members' farms
with regular group meetings to observe and compare all treatments in the multi-farm
experiment provide another alternative to more formal on-farrn testing procedures.

The approach of using farmer groups to define on-farm trial treatments is now
being used by many projects (Pinney, 1991). However, some researchers are reluctant to
involve farmers in trial design and management that stems from an inaccurate view (by
researchers) of the rigidity of statistical analysis. As one possibility, Pinney suggests the
use of farmer augmented trials to overcome the potential problem of alienating individual
farmer innovation in trials. In augmented trials, farmers gain familiarity and experience
with a technology and then suggest amendments to on-farm experimentation. Augmented
trials contain both treatments that are replicated in every block and those that are not. For
the researcher who must still meet the requirements of a more rigorous scientific analysis,
the existence of statistical computer packages ensures that most trials have complete
blocks and equal replication of all treatments by removing conservative computational
design restrictions of orthoginality and balance. ”Designs that are not balanced. but
constructed to make pair-wise treatment occurrences as possible, achieve little variation in

the precision of treatment differences fi'om balanced designs” (Pinney, 1991 p.264).

23

Thus, still retaining some structure for analysis, on-farm research under this design
enables more flexibility and participation fiom farmers while enabling researchers to

conduct scientific comparisons.

2.6 On-fgrm march and fgrmgr participgtion
On-farm and farming systems research literature has traditionally placed a strong

emphasis on farmer participation and collaboration and on talking to farmers about their
needs, problems, and reactions to technologies (Biggs, 1989; Cernea et al. 1985). As
Rocheleau (1991) notes, however, participation is subject to a broad range of
interpretation.

The International Service for National Agricultural Research (ISNAR) uses the
term on-farm, client-oriented research (OFCOR) to define an on—farrn research approach
to help meet the needs of specific clients, most often resource-poor farmers. OFCOR
complements and is dependent upon experiment station research. It involves a specific
research approach and methods involving a client-oriented philosophy. A series of
activities are carried out at the farm level that range from diagnosing and ranking problems
through the design, development, adaptation, and evaluation of appropriate technological
solutions. Farmers can be directly involved at various stages in the process, i.e. planning,
designing, managing, monitoring and evaluation (Merrill-Sands and McAllister, 1988). In
a study of nine national agricultural research systems (N ARS) worldwide, however, it was
found that no one strategy was used to incorporate farmer participation in the research
process (Merrill-Sands and McAllister, 1988; Biggs, 1989). This suggests that there is no
one ”recipe" for eliciting the participation of farmers. Rocheleau egg]. (1989) note that
the possible types of participation and collaboration on agroforestry trials range fi'om
researcher-designed trials on-station to rural peoples' own experiments that are

”discovered” and documented by research institutions.

24

Biggs (1989) defines four types of farmer participation in on-farm research which
he utilizes as a fiamework for understanding farmer participation in an additional study of
the same nine national agricultural systems cited above. The modes are distinguished by
difi‘erences in objectives and the organizational and managerial arrangements required for
implementation (Table 1).

In the contractual mode the involvement of farmers is minimal. They provide the
resources such as land and labor, that scientists need for on-farm research. This was
found, to a varying degree, in all of Bigg's case studies. The consultative mode is
characterized by a sequence of research stages: diagnosis, design, technology
development, testing, verification of trials, and diffirsion of technologies. Farmers are
interviewed about their problems at the start, after which scientists accord priorities to the
problems, make most of the decisions about what trials will be conducted, and design
trials and surveys. Farmers involvement increases again towards the end of the research
process when they are asked to evaluate new technologies. This is the dominant mode
found in most studies (Biggs, 1989; Ashby, 1991).

The collaborative mode involves continuous interaction between researchers and
farmers, who are seen as partners in the research process. Emphasis is on treating farmer
participation as a monitoring firnction to help plan on-farm and on-station research each
year. Diagnosis and evaluation are conducted continuously with farmers, not only at the
beginning or end of a project. Maurya, gal ( 1988) used this method as farmers in one
on-farm research project were brought to the research station to select, with scientists,
rice varieties which scientists then experimented with on farmers' farms. In this mode,
scientists recognize that indigenous knowledge is always changing, partly as a result of
farmers' informal research and development (R&D) systems, and as such it requires that
scientists and farmers constantly exchange new information and insights (Biggs 1989).

Finally, the collegig mode emphasizes activities that are designed to increase the
ability of the informal research systems (i.e. farmers) to do research and of informal

25

systems to request information and services fiom formal ones. The informal and formal
research systems not as "complementary colleagues." Research-minded farmers have
control over the research site, while scientists serve as periodic guests. The collegial
mode is incorporated into few programs in its entirety, and is found in isolated cases (i.e.
working with only a few selected farmers) within a projects' fiarnework (Biggs, 1989).
The collegial mode is the closest to the Farmer-first-and-last (FFL) approach prescribed by
Chambers (1989) as being more cost effective and suitable to resource-poor farmers.
While viewed as perhaps an extreme in participatory research, in FFL the main objective is
to empower farmers to learn, adapt and improve. Analysis is not conducted by outsiders
but rather by farmers or farmers assisted by outsiders. In the FFL approach, if outside
support weakens, either financially or by high turnovers of researchers, farmers may carry
on. Continuity is in the hands of the farmers and, it is argued, the sustainability of the
project becomes more likely (Chambers, 1989).

 

 

 

 

Table 2.
Types of Farmer Participation
Mode Objective

Contractual: Scientists contract with farmers to provide land or services

Consultative: Scientists consult farmers about their problems and then develop
solutions

Collaborative: Scientists and farmers collaborate as partners in the research process

Collegial: Scientists work to strengthen farmers' informal research and
development systems in rural areas

 

Source: Biggs, 1989.

Rocheleau's (1991) list is more expansive than Bigg's. She compiled the following
typology of difl‘erent types of collaborative arrangements for agroforestry trials between

26

informal (local) science and practice and "the scientific establishment" (Rocheleau 1991,
p. 122):

(1) Researcher-designed and managed trials, (usually on station or special
plots).

(2) Researcher-designed and managed trials, on site (farm), in local
peoples' work and production sites.

(3) Researcher-designed and user-managed trials, on site.

(4) Joint design and management of on-site trials by researcher and land
users.

(5) Trials designed and managed by land users, with outside researcher(s)
consulting.

(6) Trials designed and managed by land users. Outside researchers
observe and document existing trials, experiments and on-going innovation
and may also produce documents for local review, revision and use.

The choice of trial types above, and thus the amount of farmer participation,
depends on the type of research question, the variability of social, economic and ecological
conditions in the region, and the time, space and precision required to produce useful
answers to the questions at hand (Rocheleau, 1991).

The main debate around farmer's participation in research relates to the type of
technical changes that can be pursued by farmers and researchers together and how fast
technologies can diffuse (Borel and Romero, 1991). Borel and Romero (1991) claim that
any project is faced with two basic options in selecting research priorities. The first is for
researchers to propose fi'om the beginning (after some appraisal) some alternatives that
may appeal to the population. The second option is to start organizing people for
participatory research planning (including them in diagnosis, planning and design) without
giving prior reference to specific technical changes that could be pursued. Borel and
Romero (1991) used the first option in implementing a silvopastoral program in Costa

27

Rica primarily because the benefits of more decisive farmer participation were not clearly
understood at the time. This was partly a consequence of the project being managed by
foresters and agronomists, with little input fi'om socio-economic disciplines. Additionally,
project planning (i.e. grant) requirements in this instance demanded that the proponents
show evidence of working hypotheses, present a detailed time-table of the main activities
and present a rough description of the main experiments - all requirements that decrease
the possibility of farmer involvement at the design stage. The authors admit that it has
taken a long time for the project to reach its current level of farmer participation. To
increase farmer involvement at the design stage, it is suggested that a pre-project activity
is necessary where the principal activities are to organize farmers and design the main
project with farmer input before preparation of formal planning requirements (Borel and
Romero, 1991).

Ashby (1991) suggests that the greater degree to which an on-farm research effort
emphasizes the role of farmers as adapters of technology (versus adopters) the more
participatory the research style will be. A study conducted on the adoption of a diffuse
light storage technology in Peru found that farmers did not indiscriminately accept
technicians judgments, and few of them attempted to make all improvements at one time
(Rhoades gal, 1991). Farmers preferred to make alterations slowly, as they began to
understand the principles of the technology involved and as they saw the success of their
neighbors. Not dissimilar to the iterative process, this illustrates the need for an

evolutionary capacity to be built into technology development (Rhoades et al., 1991).

 

"Technology adaptation by farmers can be merely observed and tolerated as incidental to
the mainstream on-farm research effort, or it can be stimulated and integrated into this
research process" (Ashby, 1991 p.273).
nd r ’ nee
Shanna M (1992) note that in many geographic areas around the world, women
and children collect most of the household's fuelwood and fodder. Women are an

28

important source of information regarding forest products, the properties of plants that
produce them and traditional methods of forest management. As a result of the division of
labor by gender, women have difi‘erent needs and interests in forestry that have not been
adequately reflected in project planning (Sharma et__a_l, 1992).

The need to consider women's roles early in project planning and to actively
encourage their participation has been noted in several discussions elsewhere (Axinn, N.
1988; Shiva, 1988; Rocheleau, 1985; Fortrnann and Nabane, 1992). Nevertheless, Axinn
(N ., 1988) cites several assumptions and factors that, historically have led to the neglect of
incorporating women into agricultural development. First, women tend to be involved in
both agriculture and business and are generally noted as working more than men. They
may simply not have the time to formally participate in experiments. One factor that may
contribute to this is the arrangement of meetings or farm visits at an inappropriate time of
day for most women to attend. Second, where researchers are male, a male-on male bias
is often encountered. Third, even if a woman desires to participate in experiments it may
not be culturally acceptable to interact with male researchers. Thus, depending on the
circumstance, special consideration and procedures may be needed to identify and solicit

the cooperation of women farm managers for on-farm research.

2.7 Selection of On-farm Research Cooperators
As noted previously, Lightfoot and Barker (1988) found that common weaknesses

in the on-farm research process included unanticipated high variability between farmers'
management practices, insufi'rcient participation of the farmers and loss of farms from the
study. To enhance the participation of farmers in trials, Beer (1991) emphasizes the
importance of the selection criteria for farmer collaborators and sites on-farm. He
contends that the highest probability of success and impact of long-term collaborative
(agroforestry) trials is obtained by selecting innovative, experienced, motivated and locally
respected farmers. In implementing on-farm agroforestry research in Costa Rica (Beer,

29

1991), the project initially sought out the most favorable sites, selecting farmers (still
within the target group) with whom there was a high probability of establishing well
managed and reliable experiments. The project's rationale for this was that once the stafi‘
of the project gained experience with the technologies they wanted to promote, a more
representative group of farmers could then be selected and "the remaining problems that
occur when introducing a new technology (principally socio-economic) can be faced. "
Ashby (1986) concurs that cooperators should be innovative farmers, but she also adds
that an attempt should be made to represent a wide range of socio-economic resources.
The Costa Rican project noted above sought out innovative farmers However,
this same desire to work with innovative, dedicated, and interested farmers, led an
agroforestry project in Northern Zambia to utilize a self selection criterion/process for
farmer selection (Holden and Lawrence, 1990). Because selected villages in this project
were viewed as having a relatively egalitarian structure, in terms of cash and labor, the
farmers selected were considered to be fairly representative of other farmers in the area.
However, self selection is not without its own problems. In most instances this process
leads to a biased sample of more wealthy or powerful farmers, possibly also resulting in

the development of technologies more suited to their resources (Biggs, 1989).

2.8 Mb Tgrns gd Collgppration
The previous sections imply a need for a broad range of knowledge and

disciplinary backgrounds to successfirlly develop agroforestry technologies appropriate to
the communities or farming systems at hand. The need for multidisciplinary teams for
agroforestry research is well accepted, although it is not always a reality due to shortages
of trained personnel or financial constraints (Scherr, 1991). Most farming systems or on-
farm research programs claim to incorporate a multidisciplinary, efi‘ort. However, as
Conway (1985) notes, too often, while projects recognize the need for multidisciplinary
analyses of agricultural systems, social and biological scientists tend to work separately,

30

only coming together to write some final synthesis of their work. Yet experience suggests
that ”many, if not most, of the crucial questions for agricultural development lie not in one
province or the other, but in their intersection" (Conway, 1985). Gold and Tombaugh
(1987) argue that beyond a multidisciplinary approach of FSR, agroforestry research
requires an interdisciplinary’ approach as it depends on forestry, agronomy, horticulture,
and animal husbandry as major inputs along with a multitude of other inputs fi'om various
sciences. As analysis of the farming system requires intimate and integrated associations
of its various components, so should researchers studying those components integrate
their knowledge. Not only can this expand the researchers knowledge of his/her
colleagues respective fields, giving them a broader view, it enables a research team to
work to solve whole system problems rather than just addressing isolated pieces.

In addition to the use of multidisciplinary teams, Scherr (1991) stresses a need for
education and training of agroforestry (biophysical) researchers to expose them to a much
broader range of both social and teclmical on-farm research tools and paradigms. From
the same standpoint, efi‘ective participation of social scientists in a multidisciplinary setting
requires that they also have a working knowledge of pertinent technical disciplines (Tripp
and Byerlee, 1989; Horton 1986). Some relevant fields are noted: agronomy, soil
science, geography, agricultural economics, anthropology, rural sociology, ecology,
livestock husbandry, forestry and forest product utilization, rural extension, and
communication science (Scherr, 1991). Basic familiarity with methods used in some of
these various disciplines, and the scope of disciplinary knowledge, can help researchers
select appropriate methods for particular research problems or guide them to the necessary

expertise or resources for more in-depth training or advice (Scherr, 1991).

 

5 MW - group research whereby individuals fiom different disciplines work
togetheronacommonproblembutwithlimitedinteraction. W-groupresearch
wmwyiwwmrakfiommflermtdxiplinesmrkasammmmnnmflimeumflmwmcuonam
comptual synthesis (Roger and Boyd,l982 - p.88).

31

In instances when a potential shortage of trained personnel, finances, or time
exists, nongovernment organizations (N G05)6 and formal research and extension services
can benefit fi'om closer collaboration (F anington and Bebbington, 1992; F arrington,
1988b; Dasgupta and Mallick, 1992; Cook and Grut, 1989). The activities of NGOs are
community oriented and committed to the empowerment of local communities
(Farrington, 1988b; Sagar & F arrington, 1988). More importantly perhaps, NGOs take
great pains to remain accountable to their clients because participation of the latter is the
justifies their existence (Roling, 1988). This puts them in a position of being able to gain
valuable knowledge of local conditions, to assess local needs, and develop integrated
research and participatory methods of research. They also can identify strengths and local
potential for development and generate valuable feedback to formal research institutions
and promote new technologies. In addition, NGOs could benefit fiom greater access to
the technical knowledge at facilities of research institutions (F arrington, 1988; Farrington
and Bebbington, 1992). Indigenous NGOs are usually well equipped to mobilize local
resources. As such, they generally will be more cost-efi‘ective instruments for project
implementation. Finally, NGOs are more likely to be able to continue firnctioning when
external resources are withdrawn from a project (Cook and Grut, 1989).

Of course, some NGOs stray fi'om this romantic view, as a wide range of NGO
philosophies and approaches exist. Among other things, some NGOs are "top- down, "
while others can become narrowly tied to government contracts for service delivery. The
activities of some NGOs remain uncoordinated, and information exchange is poor,
especially among small NGOs where transaction costs are high (F arrington and
Bebbington, 1992). In arranging a research collaboration, care should be taken to make

sure that the goals of the institutions involved indeed complement one another.

 

5 Farrington (1988) uses the term Private Voluntary Organimtions (PVOs) as opposed to NGOs.
However, in later work (Farrington and Bebbington, 1992) the same checription ofPVOs fall under the
heading of NGOs. As such, and for simplicity, the term N00 is used here.

32

Borel and Romero (1991) support increasing collaboration in their review of a
CATIE Silvopastoral Project in Costa Rica. They suggest that greater collaboration
between development and research institutions could enhance the quality and relevance of
agroforestry research because the weaknesses of one could be strengthened by the other.
For example, just as development projects or institutions are frequently weak both in staff
and equipment, universities and research institutes have difficulties focusing on relevant

subjects to farmers and relating to development efi‘orts (Borel and Romero, 1991).

2.9 Evalugfl' g On-farm Research and Amforestly Interventions
Because of the relative scarcity of scientifically validated biophysical information in

agroforestry, monitoring and evaluating agroforestry technologies in development projects
plays an important role in the development and improvement of technical extension
recommendations (Scherr and Muller, 1991). Scherr and Muller (1991) conducted a
review of agroforestry technology monitoring and evaluating in 108 projects worldwide to
identify appropriate methodologies. Many of the projects were conducting research to
test technologies on-farm or on research plots. Technology evaluation focused on
biological performance of trees, often with inadequate consideration given to the
socioeconomic context in which the technology would be introduced. Few of the projects
appeared to use farmer assessment in evaluation.

The criteria necessary for an effective evaluation when dealing with complex
systems, such as agroforestry, or with interventions that are spread out over many years
are often contradictory in nature. A challenge faced by many agroforestry projects is how
to compromise between the sometimes contradictory requirements of rigorous scientific
research on component response and technical questions of immediate value to farmers
(Borel and Romero, 1992). It often becomes a matter of pitting "statistical" significance
against ”practical” significance and risk. The concept of "statistical significance" has been
reviewed by Borel and Romero(1991) and Sumberg and Okali (1988). According to

33

Borel and Romero (1991), when dealing with complex systems such as agroforestry, that
spread out the production of products and services over years, farmers' decisions to
change their systems is normally based on significant observations, practical difi‘erences
from the actual situation, and differences that remain fairly constant over time and over
various locations. Under these conditions, the standard test for statistical significance
becomes a "justification" that the research was well conducted. In other words, statistical
significance simply establishes whether or not observed difi‘erences between treatments
could have occurred even if true average treatment responses were the same. However,
the only valid evaluation of "practical" significance of treatment differences is the farmers'
acceptance, based on farmers' perceptions of technologies that may or may not coincide
with biophysical parameters used to conduct trials (Borel and Romero, 1991). Cook and
Grut (1989) agree that the best indicator of success in agroforestry is the extent to which
the recommended practices have been adopted by farmers. They conclude that project
evaluations should focus on this issue and that local people and local institutions should be
involved in all project evaluation efi‘orts (Cook and Grut, 1989).

To illustrate the practical significance of a technology, Sumberg and Okali (1989)
were initially discouraged by farmers who disregarded the recommended set of conditions
for tree establishment in an alley farming system in Nigeria and planted trees among
competitive crops, such as yarn and melon , instead of maize. To the surprise of the
researchers, trees sruvived the climbing vines, and although growth was not as rapid as
with maize, the trees survived within the local system. Sumberg and Okali realized that, as
their objective was to get the trees into the system, it did not matter if it took 6, 9, or 12
months for the trees to become established.

In evaluating agroecosystems in general, Marten (1988, p. 305) attests that "we
must also recognize that the ultimate purpose of evaluating agroecosystern performance is
to attain better agroecosystems, (where defining "better" is) a process squarely in the
domain of value judgments. " He argues that our judgment should be concerned with the

34

extent to which agroecosystems are meeting human objectives while avoiding the
presumption that one value or another of a system property is inherently good.
Researchers tend to assume that higher productivity, stability, and sustainability are better.
That may not always be the case. Many researchers also consider greater equitability,
autonomy, and solidarity among people to be better (Marten, 1988). The essentially
judgmental nature of the persons conducting much evaluation and irnpact7 assessment is
noted by Koppel (1990), Anderson and Herdt (1990) as well as Marten (1988) above.
To do a better job, they contend that investigators of research systems must seek advice
from the different actors in the research systems and must be aware of the fact that their
informants may "well be biased for all sorts of good human reasons" (Anderson and
Herdt, 1990).

Koppel (1990) lists several methods for visualizing linkages between activities and
accomplishments in agricultural research, noting that, just as no two research systems are
exactly alike, there is no single method of evaluation that fits all systems. The most
common type of evaluation, evaluating objectives and actions, starts fi'om the question,
"What happened in comparison to objectives?". \Vrthin this type of evaluation there are
actually many versions and depending on how the question is answered, it is possible to
make some judgments about whether a program was a success or not. It is important to
note the distinction between two types of objectives - explicit objectives and implicit
objectives. Explicit objectives are those that are openly and precisely stated at the outset.
Implicit objectives are often not stated openly and may be ambiguous and unclear.
Implicit objectives are not there because someone has something to hide; rather they are
there to remind us that programs evolve for a variety of reasons (Koppel, 1990). As a

result there may be many compromises and trade-offs in the implementation process.

 

7 ”Impact” here means "determining the significance, importance, value, or power of an event, idea, etc,
to produce changes” (Horton, 1990).

35

Implicit objectives behind these compromises and trade-offs are often carried into the
project, at times becoming more obvious as the project progresses (Koppel, 1990).

A second type of evaluation Koppel (1990) describes as "evaluating the decision-
making process." The emphasis is not on outputs and efi‘ects alone, but on processnhow
outputs and efl‘ects are achieved. Two different dimensions to this process are noted. The
first asks the question, "How similar or dissimilar are the formal or prescribed processes
from the informal or actual processes?". Formal and informal decision making usually
coexist in projects. In instances where a project contains decision-making arrangements
that cross several organizational settings, the project may represent an agreement about
how to proceed that can work in general, but will not work for all cases that arise during
implementation (Koppel, 1990). Two reasons are given for this. First not every detail and
problem of project implementation can be foreseen at the planning stage. Second,
agreements that were negotiated when the project started may no longer work because the
relationships between those that made the agreements have changed.

The second dimension to evaluating the decision-making process is participation,
which directs attention to the actors in programs and projects (Koppel, 1990). Several

questions/issues are covered:

1. Who made program decisions?
- about the allocation of input resources
- about the transformation of input resources
- about the distribution of outputs
- about the achievement of effects.

2. Who benefited?
- What benefits accrued to those who were the object of the program
outputs (according to planning)?
- To whom did benefits actually accrue?

3. Who paid?
-Who was supposed to pay? Who was supposed to bear what costs?

36

A third type of evaluation, evaluation of program levels, is based on the premise
that the kinds of activities being evaluated are not all the same. Here building blocks of
activities can be grouped into projects which, in turn, make up a program. Objectives are
applied to projects, which link together diverse activities. In turn, diverse projects can be
linked together to compose a program. In most cases, it will not be proper to apply the
same expectations for inputs, outputs, and efi‘ects to the program, project, or activity at
hand (Koppel, 1990). Which expectations are appropriate at what levels? What are the
implications (for a program) of potentially conflicting (project) objectives? For instance,
can a participatory activity be expected to operate without dificulty in a nonparticipatory
project? If not, what are the stresses on participatory objectives and processes? (Koppel,
1990)

Koppel notes that "this last approach of evaluating by levels begins from this body
of interacting objectives and seeks to explore the implications of decisions and actions on

some levels for decisions and actions on other levels" (Koppel, 1990).

2.10 Summag

Coupled with on-station research, on-farm research in particular, will conceivably
be more integral to the development of effective agroforestry technologies than in any
other FSR methodologies. The interaction between on-farm and on-station research
should be complementary. This interaction should be continuous, not sequential, to allow
for an iterative interplay of research to understand the potential of on-farm research.
Additionally, because of the combination of many unknown components in agroforestry,
the long-term nature of dealing with perennials in such systems, and the difiiculty in direct
monitoring and evaluation (i.e. biologically, economically), direct participation of farmers
in testing, monitoring and evaluation will be essential for the generation of needed

information.

37

Several factors may facilitate the on-farm research process. Some methods of
assessing local knowledge, constraints and needs for on-farm research were noted.
Researchers can choose fi'om a wide variety of methods that range fiom various forms of
Rapid Rural Appraisal to longer more detailed surveys. Long , detailed surveys have not
been found to be particularly useful to on-farm research, however, due to the large and
burdensome quantity of information usually generated and the resulting time its takes to
analyze that information. Regardless of the methods ultimately used in gathering
information, it is important that the information be made available in a timely manner that
allows it to become an integral part of the research program.

Single initial surveys or appraisals should not be utilized as the final word, so to
speak, in diagnosing the characteristics, needs and constraints of a community or farming
system. Rather, circumstances will change as time passes and as technologies are
deve10ped for the farming systems at hand. An iterative process of re-diagnosis and
redesign will become crucial. Researchers and farmers will have to work together on-farm
in exploratory trials, potentially for several years, to determine what components, sites,
and so forth, are suitable to the local system and merit further investment in long-term
formal experiments.

Researchers should be aware of the potential experimental concerns in on-farm
research: high variability between farmers' management practices and insuficient
participation of farmers. Trials should be flexible, simple and with few treatments.
Ultimately the trial design will depend upon the level of participation a research program
wants to elicit fi'om farmers and whether results fi'om the trials will by used primarily by
farmers or research stations and institutions.

Selection of on-farm cooperators is important. Collaborators should be

irmovative, experienced, motivated and locally respected farmers from the target area.

38

While researchers should work with enough farmers to allow for farmers dropping out for
unforeseen circumstances (i.e. sickness, death, outmigration), collaborators should also be
representative of the target community.

The use of interdisciplinary teams is important in understanding both the
complexity of agroforestry systems and the context (communities) in which they exist. It
is also important that researchers have a basic understanding of the scope of other
pertinent disciplines and methods used to guide them to the necessary expertise, should it
be needed. However, in instances where a shortage of trained personnel exists
collaboration with local institutions (N GOs) should be considered. As an addition to a
research program, NGOs often have a long-term commitment in communities that may
allow them to generate valuable feedback and promote new technologies, as well as to

continue the research of programs that have lost funding and research stafl‘.

2.11 Research Questions

This study reviews an on-farrn silvopastoral research program sponsored jointly by
Michigan State University and the J arnaican Agricultural Development Foundation. In
integrating the research conducted to date on indigenous knowledge, sociocultural and
farming systems, and biological and tree fodder systems, an attempt is made to answer

several questions:

1) Have the various stages of the MSU/JADF Agroforestry Program met their
initial objectives? Why or why not?

2) What was the process that led to selecting cooperators? Were innovative and
experienced cattle farmers selected?

3) Are cooperator households representative of the surrounding community with
respect to chosen socio-economic characteristics?

4) What was the extent of farmer involvement in trial implementation, i.e.
designing, planning, monitoring?

39

5) Did the various phases of the program build a base of knowledge sufi'rcient
enough to meet its major objective: the integration of tree fodder growth and
management knowledge with current small-scale farming systems to develop a sustainable
small-scale silvopastoral management system?

6) Based on what is already known about on-farm research for agroforestry
research, what insights may be gleaned fi'om the various phases of the program that may
contribute to improved implementation of similar agroforestry research programs in the
future?

Chapter 3

MATERIALS AND METHODS

3.1 Mgteggls’
Project research proposals, in addition to theses and presented papers from

researchers involved in various phases of the project, were used for information regarding
project objectives and implementation. This was supplemented by unstructured interviews
with on-farm researchers and persons supervising the project.

The information for descriptive statistical comparisons on the socioeconomic
characteristics of the Green Park community and of the on-farm cooperators were derived
from secondary data collected for the project fiom September 1990 - September 1991.
Survey responses were put into a spreadsheet package, which in turn was placed into five
SPSS statistical package files. This researcher had access to these five files, plus the hard
copy describing the associated variables. Data were analyzed using the SPSS statistical
package.

The following is a list of formal papers and project proposals fi'om which
information was obtained. Table 3 lists all the materials used, inclusive of the list below,

to answer the research questions listed on page 35.

Formal Papers

(1) Andreatta, S. 1993. A Stuoy of Community Formation and Resource Use in
a Caribbean Agrarian Resettlement: Green Park, Jamaica (1990-1991) From
and Ecological Anthropological Perspective. Ph.D. Dissertation, Michigan State
University, East Lansing, lWrchigan.

(2) Andreatta, S. and M. McDonough 1992. "Land, Trees and Labor:
Developing Alternative Agrosilvopastoral Systems Among J arnaican Small-Scale
Farmers. " Paper presented at the 4th North American Symposium on Society and
Resource Management. May 17-20, 1992 at University of Wisconsin,
Madison,WI.

41

(3) Krecik, Steve 1993. Fodder Tree Establishment and Production in
Seasonally Dry Areas of Jamaica. M. S. Thesis, Michigan State University, East

Lansing, MI.
(4) Krecik, 8., DO. Lantagne, MA Gold, and J.M. Roshetko 1994. "Cutting

Management of Leaucaena leucocephala, Calliandra calothyrsus, and Gliricidia
septum for Fodder Production." Nitrogen Fixing Tree Research Report Vol. 11.

(5) Morikawa, R. 1993. "Fodder Tree Establishment and Utilization in a
Seasonally Dry Area of Jamaica." Paper presented at the Fifih Annual Jamaican
Agricultural Seminar of the Jamaican Agricultural Research Program (JARP), July
1993, Kingston, Jamaica.

(6) Morrison, 8.]. 1991. Indigenous Knowledge Relating to Fodder Trees and
Silvopastoral Management Systems of Small-scale Farmers in Jamaica. M. S.
Thesis, Michigan State University, East Lansing, MI.

(7) Roshetko, J .M. 1991. Establishment and Nutritive Value of Native and
Exotic Fodder Tree Species in Jamaican Pasture Systems. M. S. Thesis, Michigan
State University, East Lansing, MI.

Project Proposals

(8) Phase 1 - Roshetko, J .M, M.A. Gold and DO. Lantagne "Fodder Tree
Establishment In Jamaican Pastures." Submitted October 1989.

(9) Phase 2 - Morrison, 8.]. and MA Gold "Indigenous Knowledge Relating to
Silvopastoral Management Systems of Small Farmers in Jamaica." Submitted
October 1989.

(10) Phase 3 - Andreatta, S. and M.H. McDonough "Energy, Trees, Land,
Labor and Capital: Developing Alternatives for Silvopastoral Management
Systems Among Jamaican Small Farmers." Submitted June 1990.

(10) Phase 4 & 5 - Krecik, 5., DO. Lantagne, and MA. Gold "Fodder Tree
Establishment and Production in a Seasonally Dry Area in Jamaica."

42
Table 3. Sources of Information for Research Questions

 

 

 

 

 

 

 

 

Research Question # Sources of Information

1 Research proposals, theses/papers and discussion with
researchers.

2 Analysis of data from socioeconomic study fiom phase
three. Discussions with the two on-fann researchers.

3 Analysis of socioeconomic data from phase three.
Discussion with on-farm researcher fi'om phase four.

4 Discussions with the two on-farm researchers.

5 Integration of project findings from the five phases.

6 Findings from theses/papers. Discussions with the two

on-farm researchers. Morikawa's letters/reports to Dr.
Gold and Dr. Lantagne from Jamaica.

 

 

 

3.2 Metho_(_ls
This section addresses the methods used to guide analysis. It is organized by the

research questions, and the methods used to answer each question are discussed.

Research Question 1: Have the various stages of the MSU/JADF Agroforem Program
met their initial objectives? Why or why not?

In reviewing project proposals and theses, the following questions guided the analysis:

1) Wlmt were the initial objectives of the difi‘erent phases of research? Explicit
objectives were listed in either the proposals or theses.

2) Were objectives altered as research progressed? Why or why not? Some
alterations could be found in written papers, some were found in both formal and
informal discussions with researchers.

Objectives are discussed, along with a summary of methods used to conduct
research in the various phases, under their respective phase headings in Chapter 4. For

43

more in depth coverage of methods used in each phase of research, consult Roshetko
(1991), Morrison (1991), Andreatta (1993), Krecik (1993) and Morikawa (1993).

Research Qpestion 2: What was the procgss that led to selection of coomtors? Were
innovative and gxperienced cattle farmers selected as coomstors?

The researcher implementing the first stage of on-farm research was asked what
process he used to recnrit cooperators. Determining whether cooperators might be
considered innovative or not was more problematic. Initially, analysis of the
socioeconomic data from phase three looks at a combination of whether a farmer
considers him/herself a hill-time farmer, years of experience raising livestock, and
willingness to utilize extension. Whether farmers seek out sources of new technology
information or participate actively in farm organizations might serve as potential indicators
for being innovative. Only a handfirl of farmers in Green Park, however, belonged to any
agricultural organization and only one cooperator did. Ultimately, researchers'
perceptions of farmers in the community were also included. Two individuals' (farmers)
names came up consistently when researchers were asked questions pertaining to
leadership or innovativeness among Green Park farmers, both of whom happened to be
cooperators. It is suggested that the individuals "might" be considered as innovative. This
should not be considered conclusive, however, as it is based on personal perceptions of

three individuals who were not members of the community.

Research Qpestion 3 - What are some of ths characteristics of the Green Park communig?
Are cooperator households representative of the surrounding community with rsspect to
chosen socioeconomic characteristics (i.e. access to resourfll?

Analysis of the socioeconomic data collected during phase three initially looks at

 

 

 

the Green Park community as a whole, to get a general perspective of where livestock
owners stand in terms of relative access to land and other resources. Because the focus

of the JADF/MSU Agroforestry Program is on silvopastoral systems, the main focus of

44

the analysis is on livestock (cattle and goat) owners. Comparisons are made between the
cooperators of the on-farm research phase and the general population of Green Park to
determine if these persons/households might be considered representative of the livestock
rearing community and/or whether they might be viewed as innovators in livestock or
agricultural production. This was possible because the names of respondents, and thus
cooperators, were included in the data set. Several variables were used as indicators.
Total number of acres used and their tenure arrangement (whether owned, leased, rented,
borrowed, squatted upon or family owned), access to irrigation and use of agricultural
inputs such as fertilizer, types of tools owned or used regularly (i.e. a tractor or plow), and
the number of cattle and goats owned were variables used to determine relative "wealth"
in terms of access to resources. Descriptive statistics generated from the data set are used
collectively to show possible characteristics of the Green Park community and farmer
cooperators. They are not meant to represent statistically significant findings.

During the analysis of socioeconomic data fiom phase three, it was decided that
having a list of those persons who attended the initial meeting to recruit cooperators for
on-farm research might be useful. The researcher from phase five felt that a particular
farmer, who arranged this meeting, would have a list of participants. On my behalf, the
researcher from phase five faxed a letter to a Peace Corps volunteer in the Green Park
area who he shared housing with while in Jamaica and who also knew the farmer. This
letter requested the volunteer to ask the farmer for the names of the meeting participants
and to send the list of names to the researcher. The list contained the names of 13
individuals all of whose names could be matched to survey responses. This is how the
number of women who attended this initial meeting was determined. It is possible that if
this list is based on recall alone it is inaccurate. Because the informant is known for

keeping track of such matters, however, it is felt that the source is reliable.

45

Research Quem'on 4 - What was the extent of farmer involvement in on-farm trig
implemfltatiop, is. desigm’pg, planning, monitoring, feedback?

Information was gathered from discussions with on-farm researchers fi'om phases

four and five. It was not possible to interview the farmers directly. Mthout feedback

 

from the farmers themselves, it is not possible to determine the farmers' perceptions of
how much input they had in the trial process or what they thought of the process at the

termination of the program.

Research Chreg'on 5 - mg the various phases pf the prom build a base of knowledge
suficient enmgh to m_eet its major objective: the integggtion of tree fodder growth a_n(_l

managementknowledge with current small-scale farming grstems to develop a sustainable
small-scale silvopastoral mgement system? Were phases complementm to one
Merl

 

In consolidating findings fi'om the various phases, the following questions steered
the inquiry:

1) What insights did the various researchers have that might either help or hinder
the continuation or expansion of the program (i.e. adding to the base of lmowledge
for consecutive phases)?

2) Were these insights or recommendations considered in subsequent phases?

3) What other factors might have helped or hindered the attainment of additional
knowledge to facilitate the on-farm research process (for instance timing of the
phases)?

Research Qupsp'on 6 - Based on what is already known About on-fjarm research for
aggoforestg in gmmwhst further insights may be gleaned fiom the various phases of
the program that may contribute to improved implemerfirtion of similar on-farm

sggoforestpy goggams in the future?

Interviews with the researchers involved in on-farm research focused on the

 

implementation process itself, seeking information that could not be gathered fi'om theses
or other papers. Perceptions of problems and constraints to the research experience were
sought out, as well as suggestions as to how researchers felt research could be facilitated
in the future. Questions focused on the researchers' perceptions of potential constraints to
farmers working with silvopastoral technologies and whether cooperators or other farmers

46

seemed receptive to the technologies. The interview with the researcher from phase five
took place within a month and a half of the researcher's return to Michigan State
University. A series of questions were used to guide the interview, and these questions
are listed in the Appendix. Discussion took place concurrently with a slide show
presented by the researcher of the Green Park community and the project. There were
points when the researcher answered a question, through a particular slide and the
description accompanying it, before the particular question was asked.

A greater emphasis was placed on interviewing the researcher involved in phase
five for several reasons. First, he was the most accessible; second because he was the last
researcher fi'om the program to work in Jamaica he had access to the most current
information on the situation in Green Park. This was important because it was not
possible for this study researcher to travel to Jamaica and visit the community or talk to
the people living there or participating in the program.

In reviewing the overall program, the approach used here combined aspects from
all three of Koppel's (1990) evaluation types. Explicit and implicit objectives were sought.
The review did not concern itself with inputs and outputs, as much as with the processes
that are used to achieve them. It looked at who made the decisions in project
implementation (e.g. design and management of on-farm trials) and who benefited.

By the same token, the various phases were not covered in equal detail. Instead an
approach was taken that is similar to Koppel's (1990) evaluation of program levels based
on the premise that the kinds of activities being evaluated are not all the same. As noted
earlier, activities can be grouped into projects (or phases) which, in turn, make up a
program. In most cases it would not be proper to apply the same expectations for inputs,
outputs, and effects to the program, project, or activity at hand (Koppel, 1990). Instead,
the various phases are covered according to how they contribute to the program as a
whole. That is whether they contributed to the base of knowledge that could facilitate

firrther advancement of the program at subsequent levels.

47
3-3 W

The order of Chapter 4 essentially follows the sequence of program phase
implementation In other words on-station trial and indigenous knowledge phases are
discussed first, followed by the socioeconomic phase and subsequently by the two on-farm
research phases. This sequence gives a perspective as to how the phases unfolded as the
program developed. Objectives and methods used to conduct research in the various
phases are summarized under their respective phase headings.

In labeling the various sections in Chapter 4, it should be noted that the section
titles represent the project phase of the program fiom which information was gathered.
For instance, the section including "Indigenous Knowledge" is not a section about
indigenous knowledge per se. Rather it covers the methods used and the information and
insights gathered fi'om that phase of the program.

In Chapter 5, the research questions for the study are answered. The chapter then
elaborates upon some of the other findings, as additional insights and concerns are raised

for on-farm research for agroforestry in Jamaica.

Chapter 4
FINDINGS

The initial proposal for the program addressed the first two, concurrent phases of
the project: seedling establishment and species trials, and indigenous knowledge. It was
anticipated that in establishing proposed baseline "on-station" trials, sites typical of
seasonally dry and drought prone areas in Jamaica should be represented. The project's
original intent was that more than one site representing a range of ecosystems within the
climate type would be established. In addition, it was expected that species identified in
the concurrent study on indigenous knowledge of Jamaican silvopastoral management also
be incorporated into the species trial (Roshetko et al. 1989; Morrison et al. 1989).

What transpired was different. The intentions for site selection were not met
entirely. As mentioned previously, Moneague was selected as the only station site by
JADF because of Alcan's Rio Hoe F arms' willingness to lend access to relatively
protected land (i.e. fi'om livestock) and resources, facilitating controlled experimental
conditions. This site represents an area that, although afi‘ected to some extent by seasonal
dry periods, has a relatively high amount of rainfall in comparison to Green Park. It is
unclear how much of a role climate played in the site selection process. Both Moneague
and Green Park were already selected at the time of preliminary visit to Jamaica (Dr.
Michael Gold, personal communication).

Table 4 illustrates that climate and biophysical characteristics of these two areas
are quite different. Moneague is classified as a subtropical moist forest zone (Holdridge,

' 1967), at an elevation of 500 meters. Green Park is classified as drought prone with
severely disturbed dry limestone vegetation (Kapos, 1986) ranging in elevation from 70 to
232 meters. Moneague has an average annual precipitation of 2,000 mm while Green
Park's is 1,140 mm. Seasonal dry spells efi‘ect both ofthese areas, but Green Park may go

through periods of up to eight months with less than 100 mm of precipitation.
48

49

Table 4. Biophysical Characteristics of Moneague, St Ann Parish and Green Park,
Trelawny Parish, Jamaica
(Source: Baker 1968, Holdridge 1967, Kapos 1986 and Roshetko 1991).

 

 

 

 

 

 

 

 

 

 

 

Moneagre Green Park
St. Ann Parish Trelawny Parish
Elevation 500 meters 70 - 232 meters
Average annual 15-25°C 18- 33°C (mean: 24-28°C)
temperature
"EEC
Average annual 2000m 1140m (w/ 8months receiving
precipitation less than 100mm).
Classification Subtropieal moist forest life zone Severely disturbed dry limestone
vegetation
Soil type St. Ann clay loam (well drained, med 95% either Lucky Hill (Ll-l) or Bonny
deep clay over rich deposits of bauxite) Gate (BG) clay loams - slightly acid
to alkaline
Soil nutrient P and K = Low BG - poor nutrient levels,
levels N = adequate low moisture holding capacity, highly
credible
. LH - poor nutrient levels, moderate
water holding capacity, slightly
erodible (ave. soil depth is 12cm. >
Bonnygate)
Soil analysis pH =7.48 None
Total N = 42%. P = 7 ppm
K = 95 ppm

 

 

50

In further support of the noted difi‘erences between Moneague and Green Park,
Andreatta (1993) later found that under Kaiser Bauxite's resettlement scheme, many
potential resettlers, given the choice of relocating from St. Ann Parish to Green Park in
the 1960s and 1970s, either chose compensation in lieu of property in Green Park, or
eventually sold off their lands in Green Park because of this difi‘erence in climate. The
people of St. Ann could not grow the crops in Green Park that they had become
accustomed to in the moister parish of St. Ann.

On-station trials started in March 1990 to test the performance of 22 potential
fodder trees under Jamaican conditions. Three trials were conducted. One studied the
efi'ects of weed control, fertilization, and planting pit size on survival and growth of direct
seeded and seedling transplants of three potential fodder trees (Calliandra calothyrus,
Gliricidia septum, and Leucaena leucocephala). The objective of the second trial was to
identify other fodder tree species, both native and exotic, that may be useful in Jamaica
and determine if they could be established by direct sowing. The trial established 15
fodder tree species by direct sowing in an improved grass pasture and evaluated by
survival and growth measures 10 months after planting. After the completion of on-
station trials, the third trial was conducted at Michigan State University in East Lansing,
Michigan where access to the necessary analytical equipment was readily available. This
last trial studied the estimated nutritive value of eight native and six exotic Jamaican tree
fodder species. Table 5 lists the tree species tested in the three different types of trials at
Moneague. Species' selections in trials one and two were initially based on a combination
of interest on the part of JADF to test the species, availability of the species in Jamaica,
and recommendations fiom the Nitrogen Fixing Tree Association in Hawaii as being
suitable for dry and drought prone areas.

Concurrently with the early phase of on-station trials in Moneague, farmers'
indigenous knowledge of fodder and cattle rearing was gathered in Green Park over a four
month period. The were three primary objectives of this study: I) explore small farmers'

51

Table 5. Fodder Tree SpeciesTested During Phase One (Source: Roshetko, 1991)

Seed/Cutting Source
Species Local Outside Jamaica

 

Albizia lebbeclr2 X
A. procera 2 X
A.saman2 X
Bauhinia variegataz,3
Brosimum alicastrum3

Bursera simaruba3

Cajanus cajan2

Calliandra calothyrsusl

C ecropia peltata3
Citharrerylum fruticosum3

C hamaecttisus palmensis3
Desmodium gyroides2

D. nicaraguense2

F lemingia rnacrophylla2
Gliricidia septum1

Guazuma ulmrfolia3
Haemaloxylum campechianum3
Ipomoea tiliacea3,4

><><>< ><><><><
><><

><><><><

><><><

Leucaena leucocephala 1, 2’ X X
Piscidia piscipula3 X

Sesbania grandrflora2 X
S. sesban2 X

 

l Weed control, fertilization, and planting pit size trial at Moneague, St. Ann, Jamaica.

2 Direct seeding trial at Moneague, St. Ann, Jamaica.

3 Nutritive value trial. These species were compared to Jamaican pasture grass species: Brachian'a
mutica, Cynodon plectostachus, Panicum maximum, Pennisetum purpurem as noted in JLA (1983). Tests
were conducted at Michigan State Univesity.

4 An herbaceous vine.

52

indigenous knowledge relating to fodder trees and silvopastoral management systems; 2)
conduct a preliminary investigation of the sociocultural environment of Green Park; 3)
examine pastoral land-use systems and technologies in current use. Data gathering
combined informal observation and informal topic-focused interviews using a three page
interview guide. Through "snowball sampling," where one farmer refers the interviewer to
anotherfarmeruntil allfarmersinastudyareahavebeencontacted, 40 GreenPark
resource users were interviewed over a four-month period. Thirty-five of these resource
users owned cattle. Findings showed that cattle farmers used several fodder sources in a
number of difi‘erent management systems in their attempts to feed their cattle. Besides
improved grasses, tree species mentioned most often were Guarunra ulmrflolia (Bacedar),
Samanea saman (Guango), Brosimun alicastrum (Breadnut) and Gliricidia sepium
(Quickstick). Except for one farmer who planted Gliricidia, however, the deliberate
planting and use of fodder trees was not apparent. Direct seeding trials on-station
included all of these species, but only one (Gliricidia) was included in the field trials on-
station (refer to Table 5). There were some other species mentioned during this phase
that could not be incorporated into the field trials because the research phases ran
concurrently and as such, the researcher from phase one did not have the information on
these species at the initiation of field trials. Instead, samples of these additional species
were gathered in or around Green Park and included in the nutritive (estimating crude
protein and mineral content) trial conducted later at Michigan State University (Table 5).

Along with fodder utilization, this study revealed several other things about the
community that could efi‘ect project implementation. First, Morrison (1991) observed an
apparent tension and distrust between some farmers, although he did not pinpoint the
exact source of this tension“.

One possible source of the animosity identified by Morrison was an animosity
between some of the farmers and the charcoal burners who, some farmers felt, were the

 

3 Themetnumbercffarrnerswasnotnoted

53

cause of much of the degradation taking place in the forests and on hillsides. Charcoal
burners were also farmers' primary competitors for a highly regarded fodder tree,
Brosimum alicastrum (Breadnut). Morrison (1991) suggested that this be further
investigated. Andreatta (1993) later revealed that while a certain amount of status and
prestige accompanies those that raise cattle in the community, conflicts around cattle
entering farmers' fields and destroying crops are not uncommon. Disgruntled farmers have
filed formal complaints in some instances. This same type of conflict was noted by
Morikawa (personal communication) during his research in phase four.

Morrison (1991) noted, and Andreatta (1993) later concurred, that in addition to
certain tensions, community cohesion is lacking in Green Park. Andreatta relates this to
the newness of the community as a resettlement area and the lack of developed extended
family and formal institutions in the area. Nevertheless, both Morrison (1991) and
Morikawa (personal communication) commented that many Green Park farmers who have
been in the area long enough to remember, spoke with fondness of the tradition of
"morning sport" 9 and of times when farmers in Green Park helped each other. This
suggests that some form of fiiendly and supportive association once may have existed
within the community's relatively short history.

Other considerations for project implementation are noted by Morrison (1991).
Farmers face dificulties in securing a tractor that may be used to plow rocky, hard soil
and make other improvements in cattle paddocks. Generally either farmers must wait to
secure a tractor fi'om the Agriculture Department or must rent one fi'om private operators
a high rate. Eight farmers had either used or heard of someone else using animal traction.
All but two of these eight felt that it was old-fashioned to use animal traction and were
uncertain about its viability for their areas. Labor also can be a constraint. Farmers claim
that even if they have money to hire, it is often dificult to find people who are reliable and

 

9 MorningsportistheJamaieantermusedforammualexchangeoflaborandserviceswherebyahost
farmerfeedsthosewhoassisthimwithfarmactivitiesworrison 1991).

54

hard working. Beyond simply maintaining cattle and paddocks, these factors make it
dificult to improve the quality of cattle rearing.

4.2 Ph 3: Soci nomic F e rs F in S ems

Identifying social indicators that might improve or hinder the introduction and
adoption of tree fodder production among livestock owners was the purpose of this phase
of the program. It was designed to follow-up and augment the study on indigenous
knowledge conducted in Green Park. Ideally, information on indigenous knowledge
gathered earlier would be integrated with this study's information on livestock production
practices, market economies and decision-making practices. Ultimately, the social
information gathered would serve as a foundation for an agroecosystem model to develop
alternative and sustainable silvopastoral systems (Project proposal for phase three, 1990).

With this in mind research three primary objectives were set out for research:

1. Understanding historic and current land use patterns and social factors that describe the
system: agricultural crops and livestock; land holdings and tenure; family structure;
migration patterns; local organizations; education; technology use.

2. Identifying labor requirements (year round and seasonal) associated with
agricultural/livestock management - both family and hired labor, as well as assessing
economic ability to hire.

3. Identifying economic factors associated with silvopastoral systems: decision-making
processes and markets.

Data were collected in the Green Park community from September 1990-
September 1991. Several methods were initially proposed to obtain the needed
information. First, the communities would be mapped, locating residences, farm lands,
forested/wooded regions, roads, and markets. The maps might include information on
livestock practices and orientation to market and subsistence production. This map never

materialized. The Alcan Company provided the researcher with a sectional map of the

55

Green Park community but the researcher did not provide illustration or interpretation of
this map that could be incorporated into the rest of the program.

A formal survey would be conducted to identify the characteristics of farmers (man
or woman, full-time or part-time), indigenous knowledge pertaining to livestock
production, and the farm households' relation to and interaction with markets.

The researcher attempted to interview the entire community using a formal survey
questionnaire. Of ninety households identified in Green Park, seventy-one took part in the
interviews. The researcher also spent time observing and helping farmers in the fields, and
accompanying the local veterinarian on rounds.

Secondary information obtained fi'om various libraries and data banks examined
historical agricultural practices in the area. In selected cases, oral histories were collected
from selected farmers that examined the agricultural history of the individual.

The purpose of the information gathered during this phase was to facilitate the
selection of farmers and identify issues that should be addressed in on-farm research and in
the development of sustainable silvopastoral systems (Krecik et al., 1991; Andreatta and
McDonough, 1990).

This phase produced a large amount of both quantitative and qualitative data.
Some of this data could be added to a baseline of information ifor use in firture studies,
particularly in Green Park or other Jamaican resettlement communities. Due to the timing
of this phase and the next phase, however, most of the information from this study was not
available to the on-farm researcher in phase four. Instead, preliminary findings in the form
of urrinterpreted frequencies, mostly confirming earlier findings from the study on
indigenous knowledge, were prepared for the people involved in overseeing the project at
MSU.

4.3 Ph Th n'u :An 3' of i onomiDa
This section looks at some of the data collected fi'om Phase Three and sets out to

address several questions:

1) What are some general characteristics of Green Park farmers? Of cattle
farmers in particular?

2) What resources are available to farmers (i.e. water, fertilizer, land)? What are
constraints (as perceived by the farmer) to livestock production that might be
considered in developing technologies for silvopastoral systems?

3) Information was not available on the particulars of the Green Park community
prior to the selection of cooperators for on-farrn research. Without this
information in advance, how did the project fare in its selection of farmer
cooperators relative to the community as a whole?

4) How do farmers gather new information on farming and cattle rearing? Do
they belong to farming associations?

5) Are cooperators representative of the surrounding community? Are women
included as cooperators?

6) Are cooperators representative of other cattle farmers in Green Park?
7) Are cooperators experienced cattle farmers?

8) Can cooperators be considered innovative?

For reasons discussed in the following section phase four did not incorporate
socioeconomic data from phase three into its agenda. The result was that cooperators for

on-farm research were selected without background information on community members.

Table 6 summarizes a few selected characteristics of the Green Park community.
Comparisons are made between the entire community. Individuals are also characterized
as either female or male respondents, full-time farmers or other primary occupations,
cattle farmers, and farmer cooperators.

Table 6. Summary of Selected Characteristics of the Green Park Community

57

 

 

 

 

 

 

 

 

 

 

 

 

 

Age Total Acreage Aaeage Aaeage Number II of Years Years Living
(Min/Max) Usedl Owned Leased ofCattle Rear'mg orFarmingin
(Mir/Max) (Min/Max) (Mm/Max) (Min/Max) Cattle or
(Mm/Max)
Entire 58 9.6 5.1 2.3 5.52 16.9 14.9
(11-=71) (24 I77) (.1 /74.5) (0 / 35) (0 / 39.5) (0 / 30) (0 / 29L
Male 59.3 10.9 5.3 2.9 6.6 19.5 15.4
(n= 52) (25 m) (0/308) (0 / 39.5) (0/70)
n=49
Farulenelpardan 54.2 6.0 4.6 .63 1.9 8.94 13.5
p1=19)~‘ (24/72) (0/35) (0/8) (0/41) (1 /27)
Fulltime4 farmer 63 9.7 4.6 3.2 7.8 17.2 16.1
(n=32) 65 (cane? 13.1 (Cattle) 6.3 (Cattle) 4.9 (Cattle) 7.4 (Cattle) 24.6 (Cattle) 16.5 (Cattle)
0111a primary 53.8 13.6 5.6 1.5 9.7 16.7 17.3
occupation (n=39) 55 (Cattle) 12.6 (Cattle) 7.9 (Cattle) 2 (Cattle) 9.2 (Cattle) 21.3 (Cattle) 15.6 (Cattle)
Malefillllime(n=23) 64.2 12 5.6 4.3 5.9 19.2 16.28
(25 /77) (.7 / 74.5) (o / 18) (0 / 39.5) (0 / 28) (0/60) (.4 / 29)
Female fulltime (n=9) 59.9 4.1 2.1 .44 2.1 1 1.63 15.6
(26/72) (.3 / 9.5) (0/ 8.5) (0/2) m6) (0/41) (2 / 21)
Cattle owner 59.3 12.9 7.2 3.3 8.6 22.8 16
(n=46) (25 / 77) (.5 / 74.5)6 (0/35) (0 / 39.5) (0/30) (1 / 70) (0/29)
Malecattle(n=40) 59.5 13.1 6.79 3.8 8.9 23.2 15.6
(25 /77) (2 / 74.5) (0/308) (0 / 39.5) (0/30) (1 / 70) (0 / 29)
Femalec‘tle(n=6) 58 10.9 9.3 .57 5.8 20.3 19.6
(49/72) (.5 /35) (.5 /35) (o / 2) (3 / 11) (11 /41) (14/22)
Farmeroocperatcrs 59 32.4 11.8 16.5 15.4 15.6 or 19 19.4 or23.5
@=5) (L6 / 77) (11 /74.5) (2 / 30.8) (4 / 39.5) (4 / 28) (2 / 21) (3 / 28)

 

 

 

 

 

 

 

 

 

' Total acreage is the summation of all land in use by farmers including areas defined as owned, leased,
rented, family landborrowedor squatted in Green Park. ltdoes not include lands heldorusedoutside
this area, which for at least five households, exceeds 100 acres.
2 This average includes the five farmer cooperators. If we exclude these households, the number of eattle
owned (n=66) drops to 4.7.
3 Number of female resondents varied depending on the file used The range was 16-19.
4 Fufltimefarmersarethosewhodescribefarmingastheirprimaryoccupation Theymayormaynot
haveothersorucesofincomethroughextra seasonal employment, spouseemployment, pensionsor

children.

5 This second number indicates the average ofonly fulltime/other primary occupation farmers who also

raise (mile (n=21).

5 'I'heminimumacreagelistedherereflectsonlythelandusedinGreenPark. Inaleastoneinstance,a
farmerislistedasusingatotalonacresofpropertyinGreenParkandyetraisesZSheadofcattle.This
samefarrnerisknowrrtohaveaccesstol3lacresof"familyland"outside,butnearbyGrcenPark
(Andreatta, 1993). AsacreageownedorusedmnsideomeenParkisnotincludedinthedataseLuse
caufionininterrn'efingnnnimumsmardmumsmroveranaccesstoresoumes.

 

58

Table 6 does not include land used (owned, leased or otherwise) outside Green
Park. At least five farmers are known to have access to over 100 acres in other areas,
including a non-cattle owner who raises sugar cane, and a cattle owner (Morrison, 1991).
In addition, there is a customary system of fieehold tenure called "family land,"
widespread throughout the Caribbean, that is a form of joint tenure, for all consanguine
children, legitimate and illegitimate, of a particular farmer (Blustain, 1981). The data set
noted this term, but the term was assigned inconsistently within tenure designations. One
farmer in Green Park designated seventeen acres as family land, and another farmer who
indicated having access to substantial land holdings (more than 100 acres) outside Green
Park designated that same land as family land. However, no questions pertaining to tree
tenure or planting on family land are noted, and as such it will not be discussed further
than saying that the circumstances around family land should be investigated in the future.

Cattle owners tend to own or have access to more land than non-cattle owners
(Table 6). However, cattle farmers who consider farming as their primary occupation
have slightly less total acreage, owned land, and head of cattle than cattle farmers who list
another primary occupation (referred to as part-time farmers). The exception is leased
land. Full-time farmers tend to have more leased land than part-time farmers. A farmer's
plots may be adjacent to one another, or spread throughout Green Park. Among cattle
farmers specifically, 70% (32) list their plots as spread out, while 30% (14) have plots that
are contiguous.

Both full-time and part-time farming households may have additional income
generated fi'om a spouse's or child(rcn)'s occupation, pensions fiom Jamaica or abroad, or
fi'om children living outside Jamaica. Forty-one percent (19) of cattle farming
respondent's have at least one child living abroad (usually in either Canada, England, or
the USA). Three have pensions coming from outside Jamaica, and one still has a wife
working in the U.S.A Thus, 50% of cattle farmers have access income generated by
family who have lived or are living abroad.

59

On average, cattle farmers in Green Park have been rearing cattle for a relatively
long time - more than twenty years with a range fi‘om l to 70 years. The manner in which
cattle farmers acquire knowledge about farming and livestock also varies.

Seven (22%) cattle farmers belong to a farm organization, while one farmer used
to belong. Six of these are members of the Jamaican Agricultural Society (J AS), while
one belongs to both JAS and the Coconut Industry. Activity appears to fluctuate,
however. Three farmers said that they are inactive, and only two farmers claim to learn of
new farming information fi'om their respective organizations. As Table 7
indicates, overall cattle farmers learn of new technologies fi'om some combination of other
Green Park farmers (through conversation or observation) and media (radio, TV., books)
- 46% (21 ); from some combination of self-reliance and/or media - 24% (1 1); from
extension agents (agricultural ofiicer) - 6% (3 ); or fi'om the farm store - 2% (1). The rest
- 22% (10) either claim they do not know how they acquire new information or feel that
things (technologies) do not change.

For livestock, the majority - 51% (26) of cattle farmers consult with the local
Green Park veterinarian; another 19% (9) consult with Green Park farmers and/or the
local veterinarian. The rest, 17% (8), go outside Green Park to either Orange Park or
F almouth.

Most cattle farmers raise some type of crop, whether for market or home or local
consumption, with 87% (40) planting in the fall and 62% (29) also planting in the spring.
Several farmers raise crops both in and out of Green Park, 20% (9). Thirty-nine (85%)
cattle farmers perceive drought as a major problem. While 41% (19) add water to crops,
only 4 farmers claim that water rates for irrigation were too high for crops. Since few
irrigation systems exist in Green Park (Morikawa, personal communication 1993 ), it is

not entirely clear how farmer‘s bring water to crops or if this addition is effective.

60

Table 7. Sources of New Technology Information Cited by Green Park Cattle Farmers

 

 

 

 

 

 

Source of Information Number of Cattle Farmers Using
Source (n=46)
Farm organization (e. g. JAS, the Coconut
Industry) 4% (2)
Combination of Green Park farmers and

media (radio, TV., books) 46% (21)

Combination of self-reliance and/or media 24% (l 1)
Extension agents (agricultural officer) 6% (3)
Farm store 2% (1)

 

Claim they don't know how new
information is acquired or feel that things

(technolog’cs) do not change. 22% (10)
Consult only with local Green Park
veterinarian regarding livestock problems 51% (26)

 

Consult with other Green Park farmers
and/or local veterinarian 19% (9)

 

Go outside Green Park to either Orange
Park or F almouth 17% (8)

 

 

 

 

61

A large number, 59% (27), of farmers also mentioned insect damage as a problem.
Farmers noted several other problems in crop production as well. Twelve cattle farmers
noted that crop stealing was a problem, but not more than six farmers mentioned any one
of these other problems: rats, stones, too much work, and cows invading fields. While
only six cattle farmers mentioned cows as a problem in raising crops, this was noted as a
major source of fiiction between cattle and non-cattle farmers (Morrison, 1991 and
Morikawa, personal communication 1993).

Only one cattle farmer buys fertilizer. A large number however, 74% (34), do add
cow manure or both chicken and cow manure to the soil, with two of these farmers
purchasing manure.

Farmers raise cattle for several reasons. The majority, 67% (31), use cattle as an
additional source of livelihood, while 23% (10) raise cattle to occupy their time. Two
farmers view cattle like money in the bank earning interest-«it improves itself over time.
Another farmer likes cattle rearing because it is a year-round activity.

Many cattle farmers, 61% (27) feel they could manage more cattle. The major
constraint to this is land, with 57% (25) claiming that current pasture land is insuflicicnt,
while 41% (18) claim that pasture land is suflicient to meet their current situation.
Regardless of the land situation, 78% (36) of farmers view a shortage of grass in the dry
season as a problem. The majority (80% (3 7)) of cattle farmers claim they utilize fodder
trees as an additional feed source. However, although most farmers have planted fi'uit
trees either for sale or home consumption at some point, the deliberate inclusion of fodder
trees in the farnring system is not apparent (also Morrison, 1991).

Tenure is an additional consideration in selecting cooperators and sites for on-farm
research. All cattle farmers said they would plant trees on land that they owned.
Seventeen (3 7%) said they would plant trees on leased land, while 10 said they would
consult the owner for permission before planting and 24% (11) said they would not plant
trees on leased land. On rented land, 19% (9) of cattle farmers said that they would plant

62

trees and that they have planted trees, 11% (5) said that planting trees depends on the
terms of the individual agreement, but the majority - 61% (28) said that they would not
plant trees on rented land.

Lack of labor does not appear to be a mq'or constraint in rearing cattle according
to the data, with only one farmer mentioning this as a problem. Only 13 % (6) cattle
farmers hire wage labor on a weekly basis. Seventeen percent (8) hire on an occasional
basis or irregularly, while the rest, 70% (32) do not hire. Seven (15%) farmers claim they
cannot afford to hire labor while 30% (14) indicate that if they do want to hire labor they
have difliculty finding workers and that laborers are not hard working. For the most part,
farmers rely on themselves and other members of the family for raising cattle.

Seven cattle farmers said they could not afford to hire labor. However, only 11%
(5) said that money (the lack of) is a problem in rearing cattle. It was not clear whether
this impeded the management at current levels or the expansion of herds. Morrison
(1991) alluded to a potential lack of access to credit as a problem for some farmers. This
should be investigated firrther.

W

Gender comparisons indicate that women are not generally involved in cattle rearing. Of
19 female respondents, 31% (6) of households own cattle. Fifty percent (3) of these
respondents are widows, and the other three have husbands. One female from the married
group is a retired school teacher whose spouse works on a tour ship outside of Jamaica.
The household also has the largest holding of owned land in Green Park. While she makes
the management decisions for the household, family members are charged with livestock
care. Fifty percent (3) of the female respondents owning cattle decide when to sell cattle,
as well as how lands are used. Joint decisions between spouses regarding farm
management are also made. Deciding how lands are managed and when to sell cattle lies
with both spouses in five and eight of total cattle households, respectively. Three women

63

were among the thirteen farmers attending the initial informational meeting held by Krecik
(Morikawa; personal communication fi'om Rupert Brown, 1994)”. However, only one
woman came to the water trough with her cattle and conversed with other cattle farmers
on a regular basis (Krecik, personal communication 1993). The two on-farm researchers
agreed with the findings of the data that cattle rearing does not appear to be in the regular
domain of women (Krecik and Morikawa, personal communication). Nevertheless, the
roles of women can play a major part in the household enterprise. For instance, in one
household (a cooperators) where farming is the major enterprise for both spouses, it is the
woman's chicken operation that brings in the most regular income for the family
(Morikawa, conversation with the Rupert Brown family regarding the Brown family,
1993)

Farmer Qoomrators:

The five farmer cooperators tend to have, on average, more access to total acreage
in Green Park, both in owned and leased land, than other cattle farmers (Table 5). They
also own more cattle. Two of these farmers each own 28 head of cattle - also the highest
number of cattle owned by an individual in Green Park. Except for one farmer,
cooperators have raised cattle for 19 years, slightly less than the average, but have lived or
farmed in Green Park for more than twenty years - well over the average for cattle owners
or other residents. The exception to this is one farmer who is a circular migrant recently
(1988) returned from the US. While he considers himselfa full-time farmer, he had only
lived in Green Park for three years at the time of the survey and had experience raising
cattle for only two years. All but two of his 14 acres are leased.

Two other cooperators consider themselves part-time farmers. They have primary
jobs as a mason and a carpenter, respectively, and neither reside in Green Park. They live
5 to 10 miles away but come into Green Park on a very regular basis. One of these

 

‘0 Mr.Brownarrangedthisinitialmeeting Inthisinstancc,Morikawafaxeer.BrcwninJamaiea,at
thisresearchcr'srequest Mr.Brownrespondedwithalistofallthcparticipantswhoattendedthat
meeting

64

farmers hires labor for cattle rearing on a several day per week basis. The other
cooperators do not hire labor.

Three raise cattle as a source of livelihood, while two see cattle rearing as an
investment that improves itself over time. All five claim they can manage more cattle.
They also feel that pasture land is suficicnt for their current levels of cattle. Three
farmers have additional land outside Green Park. Four cooperators say that planting trees
on leased land depends on the contract, while one says that his contract does allow for
planting. Three cooperators do not receive money fi'om outside Jamaica, one still has a
wife in the U. S. and also receives a pension from the U. S., and another receives money
fi'om a child(ren) living abroad.

Four farmers claim they use fodder trees as a food supplement in feeding cattle,
adding that either they collect fodder or are helped by a son. Cooperators saw initial
disadvantages of planting fodder trees quite difl‘crently, with two seeing no disadvantage,
one claiming that it would take too much additional time, one feeling that the land must be
cleared for planting, and one claiming that cattle would eat the trees down before they
matured.

Only one cooperator claimed membership in a farnring organization and he
belonged to two - the Jamaican Agricultural Society (J AS) and the coconut industry. Well
respected in the community, this particular farmer is perceived as a leader in the
community who experiments in farming (Krecik and Morikawa, personal communications
1993; Andreatta, statement at dissertation defense 1993). He has the largest total acreage
in Green Park, although almost 40 acres of this is leased, and he is one of the two farmers
who own 28 head of cattle. Both he and his wife are involved in the farm firll-time, with
his wife's chicken operation bringing in a substantial portion of income to help support the
education of six children who also help in the operation (Morikawa, personal
communication 1993). After Hurricane Gilbert caused heavy damage throughout Jamaica

65

in 1988, this farmer was one of the first in Green Park to take advantage of the
government's program to encourage coconut planting .

One other cooperator is perceived as being both knowledgeable and innovative in
farming and cattle raising (Krecik and Morikawa, personal communications 1993). He
resides outside of Green Park and is a part-time farmer hiring labor on a regular basis.
One cooperator is relatively inexperienced with farming or livestock and is new to the
community, having no reputation or linkages built up fully in the Green Park community.
Another might be considered "average" among cattle farmers in every way previously
discussed (i.e. land holdings, head of cattle, etc.), except age. At 77 years of age at the

time of the survey, he was the oldest person interviewed in Green Park.

4.4 Phase Fopr: On-farm resggpgh

In conducting on-farm research, the researcher intended to work with farms and
farmers that were represented the broad cross-section of interests and needs of farmers in
the Green Park area. The economically disadvantaged sections of the community would
receive emphasis. Information forthcoming from the socioeconomic study would be used
to help select the farmers meeting these criteria (Proposal for phase three). This,
however, did not happen.

About the time of the researcher's departure for Jamaica to begin implementing
phase four, some preliminary social data in the form of fi'equencies of results from the
survey were provided to advisors at MSU. No interpretation of this data however, was
included at the time. Much of the problem with acquiring more information in the form
of advice or recommendations for continuing research and selecting potential cooperators
from the anthropologist was due to the timing of these two phases of research. The
reseachcrs' schedules barely overlapped (approximately one week in Michigan), giving the
two researchers involved no time to interact personally for a usefirl exchange of
information and feedback. Dr. Maureen McDonough (personal communication, 1993)

66

indicated that no other information regarding the social data was requested fi'om the
researcher or program at this time.

Instead of attempting to select representative members fi'om a broad cross-section
of the community or to target farmers from economically disadvantaged sections of the
community, a meeting was held at the local community water trough with 14 (3 women)
of 46 potential cattle farmer cooperators. This general call to meeting resulted due to a
combination of a lack of new information about the community and a sense of urgency felt
by the researcher involved to get the trials underway in preparation for the upcoming rains
(Krecik, personal communication 1993). By the time of the meeting the researcher
already decided what species would be tested based on information fiom the research
conducted during phases one (on-station) in Moneague and two (indigenous knowledge)
in Green Park. Subsequently, the researcher described the types of trials to be conducted
and asked if any farmers were interested and willing to volunteer some of their land and
time to assist in the trials. Six of the farmers responded favorably, although one dropped
out early in the process. Thus, farmer "self-selection" was the method of acquiring
cooperators (Krecik, personal communication 1993).

With the selection of farmer cooperators in place, on-farm research was structured
initially around a combination of the contractual and consultative mode described by Biggs
(1989). The contractual mode was especially evident during phase four as the researcher
predetermined the trials he would conduct, and the cooperators' farms served mostly as a
source for land and labor.

The fourth phase consisted of two trials. The first trial established seven fodder
tree species, in fodder tree banks, at five planting sites in Green Park (Krecik, 1993). The
researcher selected trial species based on a combination of performance in initial trials on
station (the exception being Erythrina poepiggiana), the earlier report from the study on
indigenous knowledge relating to tree fodder use (Morrison, 1991), a literature search,
and communications with the Nitrogen Fixing Tree Association (NFTA), the Oxford

67

Forestry Institute, and the International Crops Research Institute for the Semi-Arid
Tropics (ICRISAT). The study objectives were threefold: 1) to demonstrate the
biological feasibility of establishing six selected exotic fodder tree species by direct seeding
in Green Park; 2) to demonstrate alternative establishment methods for two of those
species, and 3) to demonstrate establishment of an indigenous fodder tree species by
seedlings. Table 8 lists the species tested in this phase of on-farm research as well as the
types of trials conducted.

All trial sites were previously used for pasture, with four sites dominated by
Seymour grass (Andropogan pertusus) and one by Guinea grass (Panicum maximum).
Planting site characteristics were heterogeneous across sites, varying in soil depth, slope,
and aspect in comparison to each other. These characteristics showed little variation
within sites (Krecik, 1993).

Performance of direct seeding trials in Green Park was evaluated based on tree
height and survival, taken at 6 week intervals over 24 weeks, using a completely
randomized design. Trials were replicated on each farm.

Person-hours required for establishing and maintaining each site also were
recorded (Krecik, personal communication 1993 ), although only person-hours for trial
establishment are noted. General labor activities included herbicide application, plot and
planting pit location, pit establishment, seed pretreatment and sowing, and the acquisition,
movement and planting of seedlings and cuttings.

Krecik (1993) notes that survival and height growth in this trial was less than
reported for the earlier phase one trial on station in Moneague where the area receives
nearly twice the average annual precipitation of Green Park. In this previous trial
Gliricidia, Leucaena, Sesbania, and Callianrha grew two to seven times taller in six
months.

Table 8. Fodder Tree SpeciesTested On-farm in Green Park, Trelawny

 

 

 

Species Variety Establishment Tested in Phase
method One establishment
trials on-station
(Moneague)
Brosimum alicastrum local seedling
Cajanus Cajan ICP 88040 direct seeding X
Calliandra calothyrsus1 NFTA 896 direct seeding X
local seedling
Erythrina poepiggiana BLSF 2510 direct seeding
Gliricidia sepiuml NFTA 1 direct seeding X
local cutting
Leucaena leucocephala 1 NFTA direct seeding X
K636
Sesbania sesban NFTA 874 direct seeding X

 

Source: Krecik, 1993

1 Also a continuation of on-station trials at Moneague in a second
experiment conducted during this fourth phase.

 

69

Overall, growth and survival were found to vary considerably between sites.
Greatest overall tree growth was attributed to areas having greater soil fertility and less
exposure to the elements, i.e. excessive sun and wind exposure. While the poorest overall
growth was attributed to poor soils and high exposure to the elements, the particular
cooperator involved in the experiment at this site was also the most prone to neglect the
trials. Unlike the other cooperators, he often let his cattle in to the fenced-ofl‘ trial site to
graze (Krecik, personal communication 1993). As the literature suggests, difl‘erences in
management styles among farmers, such as lack of weeding and protection fi'om cattle,
have an equally damaging influence on trial outcome and should be noted. Morikawa
(personal communication 1993) had similar problems with this cooperator in the
continuation of the trials.

A special note on Cajanus cajan is in order. This species showed the best initial
survival and growth in the trials, but it also had a high mortality ten months after
establishment, due to insect and disease problems. Nevertheless, it is known as a short
lived species, and Krecik (1993) suggests that it might be useful for short-term fodder
production while other long-lived species are becoming established. From a cultural
standpoint, it is unlikely that farmers in Jamaica will utilize this species for fodder.
Cajanus cajan is a highly valued human food in Jamaica and is a major component in one
of Jamaica's national dishes - rice and peas (Morikawa, personal communication 1993).
While farmers may plant it for personal consumption, feeding it to livestock, at least in
Green Park, might take a lot of persuasion.

As suggested by Andreatta (1992) and Morrison (1991), Krecik found that the
introduction of Calliandra and Sesbania to Green Park is constrained by both the lack of
local seed sources and the unfamiliarity of farmers with these species. Given the limited
knowledge of these species' performances in Jamaica, the very difl‘ererrt conditions under
which they were tested previously in Moneague, and the fact that farmers had no idea of
the species' potential because they were introduced to it sight unseen, it is questionable

70

whether they should have been introduced during this relatively early and introductory
stage in project (on-farm) implementation. As noted earlier, before agroforestry
technologies can be taken to the farmer, even on an experimental basis, questions
pertaining to appropriate species and reliable, economical establishment methods must be
addressed on-station. In the future it may be more fi'uitful to initially work with fewer
species that farmers are familiar with, to accustom them to the idea of systematically
including fodder trees into their systems. In the meantime, introduced species could be
firrther tested in a controlled area more representative of Green Park's climate, and
farmers could then observe these trials to see if they warrant introduction into their own
systems.

The second experiment was conducted in Moneague, as a continuation of the trials
established during phase one. This experiment looked at the effects of two cutting
fi'equencies on Leucaena and of two cutting heights and frequencies on Calliandra and
Gliricidia. Problems with trial design existed in this project (Krecik, personal
communication 1993). These centered on statistical design and a lack of randomization of
treatments, and a general lack of experience of the researcher in establishing field trials
under a wide range of variable conditions. Some of these problems were overcome
through the analysis of data using non-parametric statistical techniques. This illustrates a
difficulty that can be encountered in overseas research programs where researchers
receiving "hands-on training" do not have ready access to advisors/experts and there is an
urgency to get the trial into the ground (for climatic or other reasons) without having any
prelinrinary exposure to the sites (Borel and Romero, 1991). Mistakes are caught afier the
fact and beyond the point of reestablishing trials. While all three species responded
favorably to management in Moneague (Krecik pt_a_l, 1993 ), the influence of cutting
frequency on Calliandra should be viewed as a guideline for management applications and
further study that takes into account other factors, including plant spacing, rainfall,
temperature, and solar irradiation.

71

It is diflicult to determine how much interaction existed between the researcher
and cooperators in establishing these trials. Morikawa (personal communication) noticed
during his later stay in Green Park, that a few of the trials previously established and
surrounded by fencing around them were no longer maintained by the farmcr(s). ASking
one cooperator why he was no longer caring for the plot the farmer replied that the
experiments did not belong to him. Instead the plots were viewed as belonging to
JADF/MSU. In part, this illustrates that if a farmer has no vested interest in a project, and
does not share fully in the management of the trials, it is unlikely that he/she will take the
initiative to continue experimentation once the project is gone. Nevertheless, if chance
permits, a follow-up should be conducted to investigate the ultimate course of the farmer's

action for this plot.

4.6 Pb Five: Further Ven u in n-Farm R h

In the last phase, an attempt was made to build on the work of the past two and a
half years. The two primary objectives of research were to evaluate the utilization of
fodder trees by beef cattle and to investigate and improve establishment techniques for
specific fodder tree species (Morikawa, 1993). The selection of species for trials were
narrowed down to two - Leucaena leucocephala (both a local and a hybrid variety) and
Brosimum alicastrum (Breadnut). In addition, observation and trials were continued in
Moneague, expanding upon the previous research conducted there during phases one and
four. In the previous phase, Brosimum alicastrum was included initially in on-farm trials
but then soon was abandoned because transplants did not survive on any of the trial plots.
Despite this result, its local popularity and interest among farmers in this species as a
fodder source warranted its firrther study in the mind of the researcher (Morrison, 1991;
Morikawa, personal communication). The researcher gave several reasons for an
emphasis on Leucaena in the trials. First, the cooperators unanimously preferred it
primarily due to animal preference. Cooperators felt that the value of other species as a

72

protein supplement, i.e. Sesbania, Cajanus, and Glirr'cidia, is undermined by its lack of
palatability. Second, Leucaena is a locally familiar species. Third, it is available locally
(Morikawa, personal communication with Dr. Gold 1992).

From a participation standpoint, this phase operated on a consultative mode much
more so than in the previous phase. Discussion with c00pcrators and their interest in
Leucaena and Brosimum were what ultimately led to the researcher's decision to focus on
these two species. In addition, where phase three focused primarily on establishment in
fodder banks, this phase integrated trials into farmers' working systems, integrating trees
with local food crops in many instances.

Four trials were conducted“: a feed trial comparing a 100% grass only diet and a
10% Leucaena/90% grass combination; a Leucaena establishment trial comparing direct
seeding with transplanted seedlings; a Leucaena seed source trial; and a Brosimum
alicastrum seed viability trial (in containers) (Morikawa, 1993). Establishment trials were
planted in cooperators' fields, allowing the farmers to intercrop local food crops. In
addition, informal trials involving breadnut in direct seeding in the field as well as nursery
production were carried out (Morikawa, personal communication 1993). Randomized
complete block designs were used in all trials.

To prepare for the field trials, the researcher secured the use of a tractor and disc
plow from a local fi'uit company, something he notes is a rare item to find during that time
of year (September).

Briefly summarized, the establishment and seed trials yielded mixed results.
Brosimum alicastrran showed high rates of germination, with rates being higher for
fresher seed sources. However, Brosimum does not provide seed on a regular basis and
large quantities of seed can be difficult to find. Leucaena showed potential, with a hybrid
(K636) exhibiting superior performance to the local variety (Morikawa, 1993). In

 

” Initially, theestablishementuialencompassedsixscpcratcficldtrialconfigumuons. Twoofthesesix
uialshowever, hadalmostmsurvivalmonkawapersonalcommunication toDr. Gold, l992)andwere

droppedfromtheinvestigation

73

addition, the local varieties exhibited an infestation of psyllids and an unidentified
caterpillar (Morikawa, personal communication 1993). Moisture availability at planting
was a major factor in the success of direct seedings of Leucaena (Morikawa, 1993).
Transplanted seedlings performed better than direct seedings, but are also more labor
intensive (Morikawa, 1993). Additional requirements for labor may be a deterrent to
farmers utilizing transplants in the field.

Results from the feed trials were not favorable and alone have the potential of
negating positive observations of farmers regarding Leucaena. Having little experience
with cattle or animal husbandry, the researcher consulted an animal husbandry specialist
(Dr. Eric Rugscgger, University of Florida at Gainesville), also conducting research in
Jamaica, to gain a better understanding of how to conduct the experiment. Eight beef
cattle were donated by three of the cooperators for the trials. The animals were fed two
different diets, one of grass only and the other of a 10% Leucaena (local) / 90% feed mix.
It is not clear why such a small percentage of local Leaucaena was included in the mix.
All animals were stall fed on one of these two diets for 72 days. Seven of the eight
animals lost weight, but it was determined that cattle were fed less than daily calorie
requirements. However weight loss was significantly greater in those cattle fed the
Leucaena mix than those fed grass only. The researcher attributes several possible factors
to the results. First, none of the animals were accustomed to stall feeding or to being
penned (Morikawa, personal communication). Second, animals fed the mix had only
encountered local Leucaena at random prior to the experiment and were not used to it as
a regular part of their diet. Third, rrrimosine12 contents were unknown and may have
interfered with animal metabolism. Fourth, the researcher was inexperienced in animal
nutrition (Morikawa, 1993). In a later personal communication with Dr. Rucggscgger

 

‘2 In L. leucocephala the toxic amino acid mimosine interferes with nutrient absorption, decreasing dry
matter, protein, fiber, and mineral digestibility (Jones 1979).

74

regarding a similar trial with which he was involved, a negative response by cattle to
Leucaena was indicated as well (Morikawa, personal communication 1993).

In addition to the above factors, the use of a local variety of Leucaena for trials
is problematic. Successful feed trials elsewhere utilizing Leucaena have focused on hybrid
varieties. The primary reasons for using hybrids have been a higher protein content in
hybrid varieties as Opposed to local varieties, the lower mimosinc content generally found
in hybrids, and hybrid resistance to psyllid infestations (Dr. Michael Gold and Dr. Doug
Lantagne, personal communications). At the time of feed trial implementation,
communications fi'om the researcher to advisors at Michigan State University indicated the
use of Leucaena, but not the variety. Advisors assumed that a hybrid was being studied
when in fact the local variety was used, primarily because the researcher could access a
suflicient amount of forage from local varieties. Hybrid varieties were not used because
trial plots were not yet yielding enough forage to feed cattle and the researcher was not
aware of the potential negative results in utilizing a local variety (Dr. Doug Lantagne,
personal communication). Considering that hybrids also performed better in field trials,
hindsight suggests that it would have been better to postpone feed trials until enough
forage could be collected from hybrids to conduct the trial.

In most parts of the world, ruminants rarely have problems with the mimosinc
contained in the foliage because the microbes in the first stomach convert the mimosinc to
nontoxic substances (Hensleigh and Holaway, 1988). In places where this microbe is not
naturally present (i.e. parts of Australia, Papua New Guinea, and a few other countries)
Leucaena should not be more than 30% of a ruminant's diet, but researchers are
introducing the necessary rumen nricrobc to these areas and Australia reportedly no longer
has this problem Runrinants elsewhere (India, Indonesia, Hawaii) have done well on
100% Leucaena diets (Hensleigh and Holaway, 1987). More research should be
conducted to determine whether cattle in this area of Jamaica carry the necessary microbe

for conversion of mimosinc and to deterrrrinc whether a relationship exists between

75

varying levels of hybrid Leucaena in cattle diets and a positive or negative response in
cattle.

There were other incidents that cfl‘ected the overall flow of this final phase of the
program. First, J ADF hesitated in releasing money to hire labor to build fences for
protecting crops and for corrals used for the feed trials. The reason given by JADF,
among other things, was that they felt the researcher paid more than the going rate for
labor in the area and could make due with less money. The researcher also had trouble
securing money to buy equipment and supplies for implementing the experiment. In
addition, J ADF promised to supply a truck to use on a regular basis for transporting
fodder fiom the field to the stalls, but this was never forthcoming. As a result, the
researcher spent a large proportion of time trying to secure regular use of a truck for
transporting fodder (Morikawa, personal communication 1993).

Despite the difliculties encountered, the researcher did observe several indications
of farmers' interest in incorporating fodder trees as an additional feed source for livestock.
Three of the cooperators pointed out that for years they cut Leucaena out of their
pastures and now they intentionally are encouraging its growth as part of the pasture
system (Morikawa, personal communication). Other farmers (non-cooperators) requested
seed to try in their own fields and deliberately have let Leucaena and Bacedar grow in
their fields As the researcher harvested Leucaena from the roadsidcs for the feeding
trials, a few people asked why he was taking the tree. He explained that he was
conducting a trial and the trees served as a source of fodder for cattle. Aficrword he
noted several occasions of these and other people actively harvesting Leucaena
(Morikawa, letter to Drs. Gold and Lantagne 1993). This harvesting of Leucaena can be
looked at from both a positive and a negative standpoint. On the positive side, harvesting
indicates that farmers are willing to experiment with new methods of improving cattle
rearing. On the negative side, if the feed trials conducted during this phase indicate a
negative efl‘ect on cattle well-being, similar observations of farmers testing local Leucaena

76

on their own could discourage further experimentation with hybrid varieties or other
species.

Green Park farmers also inquired about the usefulness of Leucaena for feeding
pigs and chickens (Morikawa, letter to Drs. Gold and Lantagne 1993). Dried for meal or
pellets, Leucaena is known as a highly palatable, digestible and nutritious source of food
for poultry (Hensleigh and Holaway, 1988). For poultry, a 4-6% Leucaena content is
recorrrrnended. Nonruminants (e.g. sheep or pigs) should not be fed Leucaena as major
portion of their diet (less than 10%) (Hensleigh and Holaway, 1988). Thus, if enough
interest exists in a community, future research might explore the potential for fodder trees
as feed sources for other farm animals in Jamaica.

A Note On flags: The program has only focused on cattle because JADF has a strong
preference for working with cattle only (Gold, 1993 personal communication). Although
24 cattle farmers mentioned raising goats (out of 32 total goat herders in Green Park) they
also listed praedial larceny as a major constraint to goat rearing (70%). In fact for eight
farmers praedial larceny was the primary reason that they quit raising goats. Dogs killing
goats, as well as a perception of the dry season as being "hard on goats" were listed by
eight farmers as additional constraints. It is not clear whether the difficulty during the dry
season may be due to a lack of food, water, or disease endemic to the time of year.
Farmers raise goats because, among other things, they receive a quicker return on
their investment than from raising cattle. The constraints listed above, in part, are due to
the inability to keep goats corralled in a manner that allows farmers to watch over them
more carefully. A few farmers suggested that they would be interested in finding a way to
feed goats in a corralled situation. Goats also are generally known to eat and assimilate
more types of fodder than cattle. With these things in mind, it may be fruitful to consider

goats in the development of a silvopastoral program in the future.

77

At this point the MSU/JADF Agroforestry Program ended due to funding
problems encountered by JADF. Pending available funding, a desire existed to continue
and expand on-farm research, as well as conduct further feed trials in order to better
understand the reasons for cattle weight loss in the earlier trial and to test the effects of

difl‘ercnt fodder species on cattle.

4.7 Summary
The JADF/MSU Agroforestry Research Program conducted five phases of

research from March 1990 to July 1993. The concurrent on-station and indigenous
knowledge phases yielded information on the potential for various species of fodder trees
in drought prone areas of Jamaica and elicited information from Green Park resource users
regarding their knowledge and use of local sources of fodder. The indigenous knowledge
study also provided some initial insights into potential constraints to livestock
improvement for Green Park cattle farmers as well as raising some concerns that may have
an affect on the outcome of program implementation.

The socioeconomic study gathered a substantial amount of information on the
history, structure and characteristics of the Green Park community. This study may be
used as a baseline for firrther study either in Green Park or in other resettlement
communities in Jamaica. But it did not provide additional useful information in a timely
manner to be incorporated into the subsequent on-farm research phase. Most of this had
to do with the timing of the different research phases. Nevertheless, the extensive amount
of time needed to process and analyze the amount and type of information generated
during this phase and the lack of communication between researchers also played a role.

Using a "self selection process," on-farm research worked with five cooperators.
Charatcrsitics of the Green Park cattle farming community and of cooperators are
discussed.

78

The first phase of on-farm research tested seven fodder species, with selection
based on the results from the previous on-station and indigenous knowledge research.
Research also was continued on the previously established trials from phase one, in
Moneague. In on-farm trials, trees did not perform as well as in Moneague, where they
receive two times the precipitation as Green Park, and tree growth and survival varied
considerably across sites. In addition, the researcher found that the use of species
unfamiliar to farmers was constrained due to a lack of a local seed source.

The second phase of on-farm research concentrated on two species of fodder trees
for further trials. Working with the same five farmers, feeding trials were initiated along
with new trials utilizing Leucaena leucocephala and Brosimum alicastrum. The planting
trials of Leucaena yielded mixed results; the negative results incurred during the feeding
trials could be harmful to the adaption of Leucaena as a fodder species by farmers.

Chapter 5
DISCUSSION

Because of temporal variations, agroforestry research (studies of existing systems
and experiments) must be implemented over many years before conclusive results on
component or system performance can be established (Scherr, 1991). While the program
did not go as far as MSU/JADF might have liked temporally, there are still some
observations and lessons that may be gleaned fi'om the experiences encountered thus far.
One thing that almost anyone involved with development projects overseas states is that
things rarely, if ever, turn out as planned. The MSU/JADF Agroforestry Program is no
exception. There were many unforeseen circumstances that hindered the smooth
transition of research between phases and that may have been counterproductive to the

advancement of the program had it continued.

5.1 A revig pf {he Respgrch Questions

The following summarizes the findings in relation to the research questions set

forth earlier.

Rgsearch Qpestion 1: Have the various gages of the MSU/JADF Wow
met their initial objectives? WM pr why not?

For the most part, the various phases accomplished what their individual

 

objectives set forth explicitly. On-station research established a baseline of knowledge on
fodder tree establishment for a drought prone area of J arnaica. The study on indigenous
knowledge collected information on farmer's indigenous knowledge systems, examined
pastoral land-use systems and technologies in use, and conducted a preliminary
investigation of the sociocultural environment in Green Park Since these first two phases
ran concurrently, however, there was only enough time to incorporate findings fiom the
indigenous knowledge study into a nutritive trial outside of Jamaica and not into
establishment trials for the first phase. Nevertheless, information fiom the study on

79

80

indigenous knowledge was a basis for information and species selection for on-farm
research (phase four).

Phase three was problematic. Data and observations were collected to accomplish
all the objectives, but the information was not interpreted and made available to the
program in a useful manner. Some of this was due to the timing of phases, but the amount
and type of data collected also required several months of analysis. As the literature
suggests, use of extensive formal surveys for diagnoses in on-farm reseach often encounter
the problem of gathering so much data that analysis becomes time consuming and also that
much of the data gathered are not pertinent to the project. As a result information
gathered cannot be incorporated into project planning and design. Thus, the data from
phase three were never used to facilitate the adoption of new silvopastoral technologies in
Green Park.

Phase four served as a basis for on-farm research in setting up establishment trials
under farmers' conditions in Green Park. As the obj ectivcs set forth, on-farm research
demonstrated the biological feasibility of establishing selected fodder tree species by direct
seeding and alternative establishment methods and demonstrated the establishment of an
indigenous fodder tree species. Species' performance was variable. Ultimately, this phase
did not seek out farmers with a broad cross-section of interests and needs with which to
work, as the data necessary to facilitate the selection process were not available.

Phase five investigated and attempted to improve establishment techniques for
specific fodder trees. The utilization of fodder trees by cattle was also evaluated. This
last objective was not met without difliculty, and the questionable success of the trials (i.e.
a negative response of cattle to local Leucaena) did not help promote fodder tree use
among farmers in Green Park.

R h 'n2:Whatw ther tl tothe l in foo trs?
Wers innpvsp'vg and szgpgjmged gattlg farmers glgsd gs grammars?

81

As noted, cooperators were "self selected" after an informational meeting
conducted by the on-farm researcher from phase four. Whether innovative farmers were
selected as a result is arguable. One of the cooperators claims membership in two farming
organizations and is the only farmer in Green Park who makes such a claim. This, of
course is not enough to determine innovativencss. In addition, much reliance has been
placed on the views of the various researchers involved in the program. When researchers
were asked about farmers who might be considered leaders or innovators in the
community, two of the cooperators names have come up consistently. Although these
two individuals indeed might be considered innovators, this is by no means absolute
because researchers were from outside the community and id not possess enough

knowledge of local growers to make such an assessment.

Research ngipn 3: Ar: cooperator households rsprcsentative pf ths surrounding
Erratum with rpm 19 chosen gem-economic characteristics

According to the data, cooperators tend to have greater access to resources (i.e.
land, cattle holdings) and have been raising cattle for a greater number of years than other

cattle farmers in Green Park.

Resgch ststion 4: What was the extent of farmer involvement in trial implsmentatjon,
i.e. desigm g, planning, mpnitoring?

In terms of designing, planning, and monitoring trials, it appears that little input
from cooperators was included, particularly in phase three. The researcher in phase three
had decided what species would be included in the trials before talking with farmers,
alhtough some of this decision was based on findings fiom the indigenous knowledge
study. The researcher designed the trials and trials were primarily set up in the form of
fodder banks. Farmers helped in the selection of trial locations, labor and maintenance.
Phase five operated in a consultative mode more so than in the previous phase.
Discussions with cooperators and their interest in Leucaena and Brosimum, ultimately led

to the researcher's decision to focus on these two species. In addition, where phase three

82

focused primarily on establishment of fodder banks, phase five integrated trials into
farmers' working systems, integrating trees with local food crops in many instances.

Remh ngp'pn 5: Did the vm'ous pm of the progpafl build a base of knowledge

suflicient enou to meet its ma'or oal: the inte tion of tree fodder owth and

management knowledge with current small-scalp Qgpmg' systems to develop a sum' abla

small-s c silvo astoral mana ement stem?

The program got off to a good start in establishing some baseline information for
further studies of silvopastoral systems for small-scale farming systems in drought prone
areas of Jamaica. But there were some problems. Continuing trials in Green Park based
on results of trials in Moneague with its difl‘erent climate and biophysical features is
problematic and is discussed firrther below. The timing of phases did not always facilitate
a transfer of information to subsequent phases, particularly between the socioeconomic
study (phase three) and the on-farm study (phase four). Fortunately the study on
indigenous knowledge did provide some valuable information for the on-farm research
component. The problems encountered in phase five could be a disincentive to
cooperators or other farmers in Green Park to systematically include fodder trees in their
own systems. It is possible, however, that farmers may be willing to allow the fodder
trees already growing in their systems to continue growing, instead of cutting them out as
they had done previously. Because it is still too early to tell, firrthcr investigations should

be made.

Researph Qacstion 6: Based on wh_at is already known about on-Larm research for

aggoforestg research, what insights may be gleaned from the various pMS of me
progpam that may contribute tp improvfl implementation of similar aggoforegry research

ro in the re?
The following sections discuss other insights from the MSU/JADF Agroforestry
Program that should be taken into account and which may contribute to improved
implementation of similar agroforestry programs in the future.

 

The first of these circumstances was the difi'erence in climate and biophysical
characteristics between on-station and on-farrn research sites. Although MSU had little to
say in the initial selection of these sites, it is important to recognize that if an iterative and
continuous interchange between the on-station and on-farm research areas is an objective,
the on-station site must be representative of the target areas for on-farm trials. This may
mean having more than one location for on-station (controlled) trials. In addition, it
would be beneficial if the on-station site were proximate to potential on-farm sites. This
would facilitate visitations by cooperators and potential cooperators alike, permitting them
in advance to see what the prospects are for some of the proposed teclmologics. This is
particularly essential when attempting to introduce into an area species with which farmers
are not familiar. In instances where only one researcher is responsible for overseeing both
areas, as was the case in phase five of this project, a great deal of time and resources spent
in traveling back and forth from the two areas could be better used by expanding on-farrn
trials and increasing farmer awareness. Indeed, with such limited time and resources the
researcher in phase five noted his frustration in working both areas, given the distance of
the Moneague site from Green Park (Morikawa, personal communication 1993).

On-station trials were expanded in Moneague, moving from establishment trials in
phase one to fodder production trials in phase five, using trials established during phase
one as a base with which to work. These trials will be usefirl in determining the potential
performance of fodder trees for areas resembling Moneague's climate and biophysical
characterisitcs, but it is doubtfirl that the information will be pertinent for areas similar to
Green Park.

It would be ideal to have two researchers, one working at Moneague and the other
at Green Park, with constant collaboration between the two, as well as with farmers. The
interaction between on-farm and on-station research should be continuous and dynamic,

not sequential (Scherr 1991). Initially it may be wise to assign relatively inexperienced

84

researchers to station trials, where they can (in the case of university students): 1) meet
the rigorous requirements of the scientific community through controlled experimental
design and replication; 2) familiarize themselves with the technologies at hand; and 3)
through collaborations with other researchers and farmers, acquire a better understanding
of on-farrn research and its constraints as well as methodologies that attempt to deal with
those constraints. "Biophysical academic research (e. g. for a higher degree) within
existing on-farm technology trials can create management conflicts between students,
farmer and project and therefore should be attempted with caution. The involvement of
students can be complementary to the main line of work but the activities should be kept
separate in most cases" (Beer, 1991 p.239).

The researcher whose focus is on-farm research and farmer collaboration should
have a solid foundation in experimental design and know the various limitations in
working with a variety of farmers and farming systems. He/she nwds to be willing to
venture beyond the randomized complete block, split plot, or use of analysis of variance in
evaluating technologies initially. Scherr (1991, p. 101) notes that "the added variability
introduced by farmer management of trials may draw into question the use or value of
statistics, even though that same variability may be one of the most important factors
requiring investigation." To reiterate, she suggests that researchers often need to work
together with farmers in exploratory or "diagnostic" trials, with new components, sites,
configurations or management systems, to determine jointly which are suitable for firrther
research efi‘orts. " Several years of on-farm diagnostic research may be needed before
investment in long-term formal experiments can be justified, or the specifics of
experimental designs such as selection of treatment factors, non-treatment factors, control

plots, and assessment criteria be appropriately selected" (Scherr, 1991).

85

5-3 fleas-amuse
Had the project continued, it is likely that problems would have been encountered

with limiting the number of cooperators to only five farmers. Projects need to plan for the
need to "discard" farmers and trials due to unexpected factors such as sickness, death, out
migration, etc. (Holden and Lawrence, 1990). A farmer drop-out rate of 50% during the
life of an on-farm trial is not unexpected, especially in agroforestry trials run for several
years (Pinney 1991). At the outset of the trial a contingency for such a drop-out rate
should be planned so that enough farms remain to ensure the precision of the treatment
estimates and allow for uscfirl inferences to be made. 13 As noted, both on-farm
researchers had some dificultics with one of the cooperators for the project. For instance,
he tended to neglect the plots by allowing cattle into the research area for grazing. While
all the cooperators were self-selected, looking at some of the characteristics of this farmer
(a circular migrant, new to the community/area and new to livestock raising and farming in
general) it is possible that itrritially he is not the best farmer in the community from which
to elicit feedback (in terms of livestock raising) initially.

This is not to say that farmers who do not have years of experience in
farming/livestock and innovation should not be included if they desire to participate. On-
farm research should include a wide range of farmer types fi'om the target group (Ashby,
1991). However, the selecting motivated, innovative, experienced, and respected farmers
is vital during the initiation of on-farm agroforestry research projects (Beer, 1991) because
it is important to work with enough farmers that have a background of knowledge and
experience with which they (the farmers) can make solid comparisons. As experience
suggests, it is important to remember that variability in farmer management will be high
when working with a wide range of farmer types, and this may limit the efl‘ective use of

 

‘3 AccordingtoPinney(l991-p27l),ifhalfthcfarmersdropoutwithnostrucurre,thcnancxpected
increaseworddocaninthestandardcrmrformunemsbyafactorofabmfiV2.

86

Cooperators should be actively sought beyond simply meeting with a group,
describing the research technologies and asking for volunteers. Such self selection can
lead to an unrepresentative group of farmers with which to work (Ewell, 1989). To
facilitate this search for innovative farmers, researchers need access to pertinent
information to help identify farmers that researchers may want to approach. There needs
to be more integration between social and biological researchers to identify potential
cooperators. Potential reasons for the a lack of farmer willingness or constraints to
participate should be sought out early on, before the initiation of on-farm research. For
instance, as the data indicated and Morikawa (personal communication) mentioned, some
farmers in Green Park who lease land are unwilling to plant trees for fear that the leasor
might see the improvements and reclaim their land when it comes time to renew the
contract. Armed with information on which farmers have access to enough land and also
indicate a willingness to plant trees on it, a researcher overseeing on-farm trials might
directly seek potentially interested farmer cooperators instead of relying on a general call
to meeting. In addition, the source of tension exhibited in the community noted by
Morrison was never fully pursued. It is still not known whether this hampered the
willingness of some farmers to participate in trials.

While secure land tenure is a definite consideration in choosing farmer
cooperators to work with in Jamaica, experience in Green Park suggests that proximity of
tenured plots in relation to either the homestead or worked farm land should also be taken
into account. Three of the five farmers involved in the on-farm trials deliberately placed
trials close to currently farmed lots because of proximity to the land that they already
frequented and because they had a vested interest in protecting the area fiom livestock
and other external forces. As such, these trials tended to run into fewer maintenance
problems than those placed on the other two farmers' lots (Morikawa, personal

communication 1993).

87

Andreatta (1993) suggests that, due to Green Park's nature as a relatively young
community without formal social institutions firlly developed or a history of collective
action, it is atypical of other Jamaican communities. Because Green Park lacks some of
the social institutions of traditional communities, i.e. a village school, a church that the
majority of community members attend, a ball court, etc., as well as a social fiamework
based on collectivism and extended family ties within Green Park; it does not constitute a
"community." To confirm these findings a wider range and variety of communities (e. g.
resettlement and non-resettlement) within seasonally dry and drought prone areas should
be surveyed. Defining the nature of the community per se, would not be the major
objective. Rather, such a study might explore whether cattle farmers in similar or difi‘erent
communities within dry and drought prone areas face similar or different concerns and
constraints as farmers in Green Park.

A potential problem of the Green Park survey is that it was administered once to
individual residents and resource users throughout the year. In other words, not all
individuals responded to the survey during the same time of year. It is difficult to
ascertain at this point, but it is possible that people have different perceptions of
constraints dependent upon the time of year (i.e. rainy or dry season). As Chambers, 51
g, (1981) note, the perceptions of, as well as the constraints and difficulties farmers face
will vary depending on the season. It may be wise to administer surveys separately, or
repeat the survey, over two difi‘erent seasons in order to measure more efl‘ectively peoples'
perceptions of concerns or constraints at any given time of year.

According to the Statistical Yearbook of Jamaica for 1991, the parish of Trelawny
has the lowest population density of all the parishes in Jamaica (SIOJ, 1991). It may be
that introducing similar technologies into areas with a higher population density will be
more (less) beneficial when coupled with difi‘erent pressures on the resource base (both
natural and human). As Nair (1992) notes, the type of agroforestry system found in a
particular area is determined, to some extent, by agroecological factors. He adds,

88

however, that several socioeconomic factors, such as human population pressure,
availability of labor and proximity to markets, also come into play, resulting in
considerable variation among agroforestry systems operating in similar or identical
agroclimatic conditions. Otherwise it may be problematic extrapolating results of the
community survey and on-farm trials based on an area with isolated constraints and

circumstances to other communities in seasonally dry areas of Jamaica.

5.4 Tactor [Jae and LaMr Inputs
The use of a tractor to clear land for planting in on-farm plots should be reviewed.

While farmers will attempt to secure a tractor for field work when possible, it is usually
quite difficult. Only two farmers mentioned using the tractor as a tool for farming.
Because farmers often must work within the constraints of physical labor available to
them, the researcher involved in these trials should make a point to work within these
constraints as well. Otherwise the number of person-hours it takes to establish and
maintain a technology will be misrepresented. In addition, when the time comes to expand
trials to other cooperators' farms, interested farmers may expect the availability of such
resources for their participation. Indeed, it may be the sole impetus for their participation
as one farmer (not a cooperator) showed in phase five. In this instance, the farmer
expressed an interest in planting trees when the researcher initially arrived. His farm plot
was adjacent to one of the cooperators sites that was in the process of being plowed with
the tractor. The researcher had the farmer‘s field plowed as well, as it was small. He also
gave the farmer some seed, although the farmer was never seen planting the seed
(Morikawa, personal communication). To be successful, incentives must be consistent.

Both Morrison (1991) and Andreatta (1993) note farmers' reluctance in using
animal traction as a potential method of clearing land because it is perceived as ”old
fashioned. " Overcoming the cultural barriers of using animal traction may be diflicult, or
even impossible, in Green Park. However, Green Park has a history of using animal

89

power, and several farmers mentioned their familiarity with it and with farmers using it in
other parts of Jamaica. Re-introducing animal traction in Green Park should be attempted,
especially for facilitating the clearing of fields for resource poor farmers who have an even
lower chance of accessing a tractor than the relatively resource rich and who may be
constrained be labor in improving their sites. JADF has indicated an interest in the
utilization of animal traction, and there is also a site in Jamaica for such training (Gold and
Lantagne, personal communication 1993). To further encourage its use, field trips to
the Jamaican training site might be arranged with local farmers to allow them to see first
hand the methods and benefits of animal traction.

Potential restrictions to improving pasture systems imposed by labor (or a lack
thereof) should be firrther investigated. Morrison (1991) indicated that labor constraints
are a potential problem, although it was difficult to confirm or negate this finding fi'om the
phase three data. Questions regarding labor in the phase three survey focused on
perceived constraints in farmers’ present systems. It was also noted that survey inquiries
were not made regarding whether farmers felt that labor might be a potential problem to
systematically incorporating fodder trees into their systems. As Cook and Grut (1989)
note, it is dificult for farmers to adopt agroforestry packages that require them to
organize and supervise labor to perform new tasks in particular sequences or time
intervals. Acquiring, transporting, planting, and caring for seedlings may require more
resources and more management skills than the average farmer can spare from subsistence
activities (Cook and Grut, 1989). In a survey of agroforestry project experiences in 21
countries in Africa, Kerkhof (1990) found that growing trees is more diflicult in dryland
farming areas. Growth rates are poorer, survival rates are poorer, and protection of
seedlings is more of a problem. The projects found that while farmers are willing to plant
some trees, the uptake of seedlings from nurseries was often disappointing. Some projects
found that encouraging natural regeneration was more acceptable to farmers than tree
planting because it is cheaper and less risky than esatablisheing seddling nurseries

90

(Kerkhof, 1990). Morikawa (1993) felt that the labor concern might be an issue for Green
Park farmers in utilizing transplants of Leucaena. Future projects should investigate this
firrther, including a study of econonric benefits and costs to systematic inclusic'rn of trees in

silvopastoral systems.

5.5 Di ei lin Trainin nd mm ni ti n

Finally, there is a need for better communication between researchers involved in
the project. An interdisciplinary versus what turned out to be a multidisciplinary
approach would facilitate this. Thus several pertinent disciplines should be involved in
studying the community and diagnosing potential constraints or problems in the
livestock/fanning system because no one person, or discipline, can be expected to know or
notice all the intricacies involved in a farming system. Ideally, at least one of the on-farm
researchers would be involved in a part of this study of the community in order to
familiarize himself/herself with the area and the people, within context (as opposed to
trying to visualize unseen faces and locations through discussions with previous
researchers stateside). Because this is a silvopastoral program, at least one person
involved in the diagnosis should have a solid background in livestock husbandry/nutrition,
particularly related to cattle.

As noted earlier by Scherr (1991) a need exists for educating and training
agroforestry researchers to expose them to a broader range of on-farm research tools.
Researchers familiar with a variety of disciplines can both take responsibility for
enlightening other disciplines to issues that they feel are important and can seek out the
appropriate sources of information as questions arise. The on-farm researcher must be able
to readily identify farmers who might be good cooperators and be able to select farmers
from a diversity of socioeconomic backgrounds. As in phases three and four, an on-farm
researcher may feel that time is of the essence, especially when working against seasonal

constraints. The on-farm researcher also must arm him/herself with enough background

91

to ask the right questions of his/her social science counterpart. Ultimately however, only
the person/people involved in collecting and analyzing social data can know the extent of
what data or other pertinent information exists and, based on their expertise, can interpret
the data and identify what issues are important to the success of project implementation.
Another objective of the J ADP/MSU Agroforestry Program, besides studying the
potential for silvopastoral systems, was the training of graduate students as future
researchers of academic, government, and non-government institutions alike. It is felt that
students must be given a chance to go out and learn fiom their mistakes and successes.
Because the use of interdisciplinary teams in research institutions is an increasing
phenomenon (i.e. ICRAF, CATIE, IDS, CIMMYT), students wishing to work for these
types of institutions must be able to work with people from other disciplines to efl‘ectively
assess a situation, diagnose potential problems and come up with potential solutions when

"solutions" are in order.

5.6 Fflgggk to and from the Commungy'
The JADF/MSU Agroforestry Program's goal was the improvement of small-scale

cattle farming systems. Green Park residents permitted researchers to study the
community in order to understand constraints to improving both cattle production and
farming systems. In addition, a few farmers donated land and labor for the study of
potential agroforestry technologies that might address some of these constraints.

The conunurrity was never given feedback, however, on the findings from the
various studies. For instance, would community members agree with the findings from the
indigenous knowledge study or the socioeconomic study? By the same token, did Green
Park farmers find the process used for studying technologies on-farm useful? This lack of
feedback, especially from the program to the community, draws into question the sincerity
of the program in addressing the concerns of the small-scale cattle farmer in Green Park.
Programs like this one need to incorporate the feedback of farmers and community

92

members into project planning and development if the program's ultimate goal is to benefit

farmers.

5.7 Conclusions

Flora (1992) looked at projects funded by a variety of sources, including
international research centers, major foundations and a number of governments. Findings
indicated that the longer the trajectory of funding, the more adaptations that the farnring
systems could make. Because of temporal variations, agroforestry research (both studies
of existing systems and experiments) must be irnplemented over many years before
conclusive results on component or system performance can be established (Scherr, 1991).
This is especially so when dealing with species and systems that are not only new to
communities, but to a country as well.

In this case, the MSU/JADF Agroforestry Program did not have the luxury of
time. In three and a halfyears of research it is difficult to assess the potential impact that
this project may have on the community involved or on the development of silvopastoral
systems for dry and drought prone areas of Jamaica. Many questions still remain. Will
encouraging and providing Brosimum alicastrum planting materials to cattle farmers to
actively plant, and not distributing the same to charcoal burners, cause increased
resentment between these two groups? Will a loss in cattle weight from feed trials
utilizing local Leucaena cause farmers to abandon Leucaena altogether?

Since the program lasted a relatively short period of time, there was no opportunity to
further test technologies and refine or revise research. Ashby (1991) suggests, however,
that a few farmers will probably continue to experiment on their own, adapting the
technologies to their own systems and needs (but perhaps not for cattle at all).

Over the course of three-and-a-half years, five researchers were given the
opportunity to expand their research skills and knowledge. A variety of fodder trees were

tested, and several were determined to have potential as an alternative food source for

93

cattle in Jamaica (Roshetko, 1991 and Krecik, 1992). Nevertheless, all of the researchers
concur that more research is necessary before making wider recommendations.

A time horizon was initially set at between three and five years, yet it was not
possible to foresee the end coming quite so abruptly. Nevertheless, an end was assumed.
To utilize resources more eficiently and efi‘ectively in the future, projects like this one
must plan for an end at the beginning, so to speak. Much of the problem with the
JADF/MSU Program centered around the timing of phases. Phases did not coincide in a
manner in which the transfer or integration of useful information to subsequent phases
could be made in a timely and usefirl manner. To avoid this in the future, consideration
must be given to the timing of difi‘erent phases of research in the firture.

Similar projects should look into integrating the efforts of local institutions (e.g.
collaboration with NGOs) into the research process so that these institutions also have an
understanding for the basis under which technologies are created. Thus, when a project
comes to an end and researchers return to their respective countries or institutions, there is

a better chance that someone will be left in the area to carry on.

Chapter 6
RECOMMENDATIONS

6-1 mm
It is important to get farmers' feedback on the program and technologies, and to

incorporate their suggestions into planning similar projects in the future. Further, other
farmers who were given seed or planting stock to experiment with on their own should be
sought out for their feedback and suggestions. If possible these interviews should be
conducted soon after planting, before details have faded from farmers' minds. Because a
review of an agroforestry program like this is not complete without the fwdback of the
farmers whose problems it has set out to address. As such, this study can not be
considered complete.

At no point in the J ADP/MSU Agroforestry Program is it indicated that the
farmers or community were given a summary, in writing or verbally, of the results of the
various phases of research in which they had taken part. To encourage the participation of
this, or any other community in future projects, a certain amount of accountability to the
people of the community by researchers is in order. Formally recognizing the cooperation
of the community involved by presenting the results of their cooperative research

endeavor is one way of doing this.

6.2 Further Stud:

Other communities in seasonally dry and drought prone areas of Jamaica should be
studied to see if farmers, especially cattle owners, face similar or different constraints as
farmers in Green Park. Research suggests that settlement communities in Jamaica difi‘er
fiom other communities in Jamaica. A sufiicient number of communities of both types
(i.e. settlement and non-settlement) could be studied to verify if difi‘erences do exist, and
what, if any, are these difi‘erences. Ideally an interdisciplinary team, including on-farm

94

95

researchers, should be included in the assessment of target communities so that each has a
better concept of the others' disciplines and approaches and also an understanding of
limitations.

Three-and-a-half years is a short period of time for observing agroforestry trials.
Clearly more research is needed, both on-station and on-farm, to test the performance of
various fodder tree species for dry and drought prone areas in Jamaica. Further study by
an animal husbandry/nutrition expert is necessary to better understand the reasons for the
negative response of cattle to Leaucaena, in addition to testing the palatability of other
species to, and responses of, cattle. One cannot expect farmers to invest the time and
money into further planting trials for silvopastoral systems if they feel cattle will not or
cannot eat the fodder.

Silvopastoral research in Jamaica should consider incorporating goats and possibly
other farm animals, as well as cattle into trials. Farmers have expressed an interest in this.
What is not palatable to cattle, may be so to goats (or other animals). It is also possible
that by penning goats and feeding them on a cut and carry basis, more palatable and
nutritious fodder will be freed up for cattle in pastures. It may be that the major endeavor
in such research will be finding an effective way to keep goats penned.

More effort should be made to incorporate the input of farmers into all phases of
the research program—diagnosing problems and constraints, designing trials, monitoring,
and evaluating. Farmers should feel a vested interest in and have an understanding of the
trial process, in order to continue research on their own as well as provide useful feedback

to researchers.

APPENDIX

APPENDIX
f r her

1. How representative of socioeconomic conditions (do you feel) are the participants of the on-farm trials
in comparison to the rest of the community?

2. Have any extension agents (i.e. fi'om the J. Dept ong) visitedorbeen involvedinthe research
process? Ifso, whenwerethey incorporated?

3. (Based on Bruce's and Susan's findings) Were you aware of any possible factionalism in the
community, especiallyamongfarmers/livestockowners? Ifso, doyouhaveanyideasastowhythis
animosity exits?

4. In 'recmiting" cooperators, what season did Steve have his meeting with community members? How
many of the potential farmers attended? Do you feel it possible that some potential particimnts did not
attend due to seasonal factors? Is it possible that some farmers did not attend due to community
factionalisrn?

5. What has your involvement been for the last year?

6. Whathavefarmers' responsesbeentothetechnologies? Alsodidtheprojectactuallypaycooperators
(i.e. R Brown) to participate in the on-farm research or, in Brown's ease, was he paid beeause he helped
the researchers integrate into the community, etc, etc?

7. How much of farmers' own inputs have been utilized in the on-farm trials (i.e. did the project suwly
all materials and money for labor, etc)? If so, will this allow a realistic environment on which to base

adoptability/adaptibility?

9. Have other farmers (outside on-farm participants) shown an interest in planting experimental trees?
Have you noticed actual planting in other farmers' lots? Has anyone? If so,how have farmers'
experimented with or adapted. the new technologies to their own situation(s)? (Steve mentioned he had
given some seeds to other farmers, although he did not mention to whom) Did you also distribute
seeds/cuttings to other (non trial ) farmers?

10. Steve mentioned that he took participants on ”field trips” or rather ”a" field trip, to see
"established”(‘n projects/technologies. Were these trips just for participants? Did you also attempt to do
this? Iknowthatwhen l askedyouthisonapreviousoceasionyou repliedthatyouplannedto.but
transportation arrangements fell through. Where did you plan to take farmers and why? Do you feel that
there should be money budgeted for transportation to facilitate such trips/exchanges of information in the
future?

11. Were there any plans to incorporate farmer-designed trials (-Vs- researcher designed)? Was
experimentation by farmers encouraged?

12. Theproject focusedononlyeattle. (Andreatta's surveyshowed thatwhile eight cattlefarnrers
stoppedraisinggoatsduetolarceny,manyeattlefarmeraaswellasnon-eattleowners,didraisegoatsand
wereinterestedinfindingwaysofkeepingthemfencedinandfeedingonacutandcanybasis). Doyou
thinkgoatsshoddhavebeenmmmomtedasweu-espedaflysincegoamwndwrespondfavomUyma
widervarietyoffodcbrsourcesthancattle? (Gutfeeling-Iknowthatyoudonothaveanerdensive
backgroundinanimalhusbandry.)

l3. Howdid (iftheydid)yourotjectiveslgoals changeastheprojectprogressed? Why/why not?

96

97

14. Whatelsewouldyouliketohaveseenaccomplished? Wouldyoudothingsdiflerentlyinthefirture?
Whatadditional informationaboutthecommunitydoyou think wouldhavefacilitated your research! or
increased farmers' interest in the technologies?

15. Arethereotherfarmersthatyouthoughtwotudhawbeengoodcooperators? Why?

16. Didyoukeeptrackoflaborrequirements i.e.thenumberofhoursittooktoestablishplotsand
maintainthem?

l7. Nowthattheprojectisover,doyouthinkanyofthetrialfarmerswillcontinuetoexperimentwithor
maintaintheplotsthattheprojectestablished? Afierthreeyears,doyoufeeltheprojectlefianylasting
impression?

BIBLIOGRAPHY

BIBLIOGRAPHY

Anderson, J .R and RW. Herdt 1990. ”Reflections on Impact Assessment." In
RG. Echeverria (ed) Methods for Diagnosing Research System Constraints and
Assessing the Impact of Agricultural Research Volume II: Assessing the Impact of
Agricultural Research. Proceedings of the ISNAR/Rutgers Agricultural Technology
Management Workshop, 6-8 July 1988, Rutgers University, New Jersey, USA.

Andreatta, S. 1993. " A Study of Community Formation and Resource Use in a
Caribbean Agrarian Resettlement: Green Park, Jamaica (1990-1991) From and Ecological
Anthropological Perspective." Ph.D. Dissertation, Michigan State University, East
Lansing, Michigan.

Andreatta, S. 1990. "Agrarian Land Use Transformations and Implications for Small
Farmers in Jamaica: A social Forestry Perspective.” A proposal for research
(unpublished).

Andreatta, S. and M. McDonough 1992. "Land, Trees and Labor: Developing
Alternative Agrosilvopastoral Systems Among Jamaican Small-Scale Farmers.” Paper
presented at the 4th North American Symposium on Society and Resource Management.
May 17-20, 1992 at University of Msconsin, Madison, WI.

Ashby, J .A. 1991. "Adopters and Adapters: The Participation of Farmers in On-F arm
Research.” In R Tripp (ed.) Planned Change in Farming Systems Research: Progress in
On-Farm Research. New York: John \Vrley & Sons.

Ashby, J .A 1986. "Methodology for the Participation of Small Farmers in the Design of
On-Farm Trials." Agricultural Administration 22: l-l9.

Axinn, G.H. 1988. Guide on Alternative Extension Approaches. Rome: FAO.

Axinn, N. 1988. "Gender Related Issues in International Development Assistance for
Agriculture and Rural Life.” Agriculture and Human Values. Winter-Spring pp. 69-76.

Baker, D. and B. Spence 1988. "Afro-Caribbean Agriculture: A Jamaican Maroon
Community in Transition.” The Geographical Journal 154(2): 198-208.

Beers, J. 1991. "Implementing on-farm agroforestry research: lessons learned in
Talanranca, Costa Rica.” Agroforestry Systems 15: 229-243.

Biggs, SD. 1989. Resource-Poor Farmer Participation in Research: A Synthesis of

Experiences from Nine National Agricultural Research Systems. OF COR Comparative
Study Paper No. 3. The Hague: ISNAR

98

99

Blustain, HS. 1981. "Customary Land Tenure in Rural Jamaica: Implications for
Development.” In H.S. Blustain and E. LeFranc eds. Strategies for Organization of
Small-Farm Agriculture in Jamaica. Mona, Jamaica; ICER, Univ. of West Indies.
Borel, R and F. Romero 1991. ”On-farm Research in a Silvopastoral Project: A Case
Study". Agroforestry Systems 15: 245-257.

Buck, L. 1990. "Planning Agroforestry Extension Projects: The Care International
Approach in Kenya." In W.W. Budd, I. Duchhart, L.H. Hardestry and F. Steiner (eds)
Planning for Agroforestry: Developments in Landscape Management. New York:
Elsevier Science Publishing Co., Inc.

Butler, L.M. and R0. Butler 1987. "Needs Assessment in International Development."
In DE. Johnson, et al., (eds.) Needs Assessment: Theory and Practice. Ames , Iowa:
Iowa State University Press.

Byerlee, D. and R. Tripp 1988. "Strengthening Linkages in Agricultural Research
Through a Farming Systems Perspective: The Role of the Social Scientist”. Experimental
Agriculture 24: 137-151.

Casley, DJ. and DA Lury 1987. Data Collection in Developing Countries. Oxford:
Clarendon Press.

Casley, DJ. and K. Kumar 1988. The Collection, Analysis, and Use of Monitoring and
Evaluation Data. Baltimore: The John Hopkins University Press.

Cernea, M.M., J .K. Coulter, F.A. Russell (eds) 1985. Research - Extension - Farmer:
A T wo— Way Continuum for Agricultural Development. A World Bank and UNDP
Symposium. Washington DC: The World Bank.

Chambers, R, R Longhurst, D. Bradley and R. Feachem 1981. "Conclusions and
Practical Implications." In R. Chambers, R. Longhurst, and A. Pacey (eds) Seasonal
Dimensions to Rural Poverty. London: Francis Pinter (Pub.) Ltd.

Collinson, MP. 1987. "Farming Systems Research: Procedures for Technology
Development.” Experimental Agriculture 23: 365-3 86.

Cook, CC. and M. Grut 1989. Agroforestry in Sub-Saharan Africa: A Farmer's
Perspective. World Bank Technical Paper Number 112. Washington, DC: The World
Bank.

Dasgupta, S. and RN. Mallick 1992. "FSR and NGO: Partners for Sustainable
Development of Resource-Poor Farmer." Paper presented at the International Farming
Systems Research/Extension Symposium, Michigan State University, 14-18 September
1992.

100

Echeverria, RG. (ed.) 1990. Methods for Diagnosing Research System Constraints and
Assessing the Impact of Agricultural Research Volume II: Assessing the Impact of
Agricultural Research. Proceedings of the ISNAR/Rutgers Agricultural Technology
Management Workshop, 6-8 July 1988, Rutgers University, New Jersey, USA

Etling, AW. and RB. Smith 1994. "Participatory Needs Assessment: A Key to F SRE."
Journal for F arming Systems Research-Extension 4(2): 45-56.

Evans, PT. 1988. "Designing Agroforestry Innovations to Increase Their Adoptability: a
Case Study from Paraguay." Journal of Rural Studies 4(1): 45-55.

Ewell, RT. 1989. Linkages Between On-Farm Research and Extension in Nine
Countries, OFCOR Comparative Study No. 4. The Hague: International Service for
National Agricultural Research (ISNAR).

Farrington, J. 1988a. "Farmer Participatory Research: Editorial Introduction."
Experimental Agriculture 24:269-279.

F arrington, J. 1988b. ”Private Voluntary Organizations as a Link Between Resource-Poor
Farmers and Formal Agricultural Research and Extension Services - Unexploited
Potential? A Proposal for Research." Workshop on "Operational approaches for
participative technology development in sustainable agriculture", ILEIA, Leusden, 11- 12
April, 1988.

Farrington, J. and AJ. Bebbington 1992. "From Research to Innovation: Getting the
Most From Interaction With NGOs in F SR/E. " Paper presented at the International
Farming Systems Research/Extension Symposium, Michigan State University, 14-18
September 1992.

Flora, C .B. 1992. "Evaluation and Impact of the Global Synthesis and Networking."
Paper presented at the 12th Annual Association of Farming Systems Research/Extension
Symposium, Michigan State University, 13-18 September 1992, East Lansing.

Ford, R. 1989. An Introduction to Participatory Rural Appraisal for Rural Resource
Management. Program for International Development, Clark University, Worcester,
Massachusetts.

Fortmann, L. and N. Nabane 1992. "The Fruits of Their Labours: Gender, Property and
Trees in erondoro District." Centre for Applied Social Sciences Occasional Paper Series
- NRM ; 6/1992. University of Zimbabwe, Harare.

Francis, PA, and AN. Atta-Krah 1989. "Sociological and Ecological Factors in
Technology Adoption: Fodder Trees in Southeast Nigeria. " Experimental Agriculture
25: 1-10.

101

Gibbs, C.J.N. 1985. "Rapid Rural Appraisal: An Overview of Concepts and
Applications. " In Grandstafi, Grandstafi and Lovelace, Rapid Rural Appraisal:
Proceedings of the 1985 RRA Conference.

Gold, MA and L.W. Tombaugh 1987. "Understanding Interdisciplinary Research and its
Application to Agroforestry Systems Research." In How Systems Work - Proceedings of
Farming Systems Symposium 1987. University of Arkansas, Fayetteville, AR.

Hensleigh, TE. and BK. Holaway (eds) 1988. Agroforestry Species for the Philippines.
Washington, DC: United States Peace Corps.

Hildebrand, PE. 1986. "The Sondeo: A Team Rapid Survey Approach." In P.E.
Hildebrand (ed) Perspectives on Farming Systems Research and Extension. Boulder:
Lynne Rienner Publishers.

Hildebrand, PE. 1981. "Combining disciplines in rapid appraisal: The Sondeo
approach. " Agricultural Administration 8: 423-432.

Holden, ST. and 0.]. Lawrence (1991). "Farmer Participatory Research and
Agroforestry Development - A Case Study from Northern Zambia." Agricultural Systems
36: 173-189.

Horten, D. 1986. "Farming Systems Research: Twelve Lessons from the Mantaro
Project." Agricultural Administration 23: 93-107.

Horton, DE. 1990. "Assessing the Impact of International Research: Concepts and
Challenges. " In Methods for Diagnosing Research System Constraints and Assessing the
Impact of Agricultural Research. Volume II: Assessing the Impact of Agricultural
Research. Proceedings of the ISNAR/Rutgers Agricultural Technology Management
Workshop, 6-8 July 1988, Rutgers University, New Jersey, USA.

Hunter, G. (ed.) 1982. "Enlisting the Small Farmer: The Range of Requirements."
Agricultural Administration Unit 5 Occasional Paper No. 4. London: Overseas
Development Institute (ODI).

Horowitz, D. (ed.) 1990. Making the Link: Agricultural Research and Technology
T ransfer in Developing Countries. Boulder: Westview Press.

JLA 1983. Livestock Manual for the Tropics. Kingston, Jamaica: The Jamaican
Livestock Association.

Jones, R]. 1979. ”The Value of Leucaena leucocephala as a Feed for Runrinants in the
Tropics". World Animal Review 31:13-23.

102

Kerkhof, P. 1990. Agroforestry in A fiica: A Survey of Project Experience. London:
Panos Publications, Ltd.

Kidd, CV. and D. Pirnentel. (eds) 1992. Integrated Resource Management:
Agroforestry for Development. Academic Press, Inc.

Koppel, B. 1990. "The Need to Know: Monitoring and Evaluating Agricultural
Technology Management. " In RG. Echeverr'ia (ed.) Methods for Diagnosing Research
System Constraints and Assessing the Impact of Agricultural Research. Volume II:
Assessing the Impact of Agricultural Research. Proceedings of the ISNAR/Rutgers
Agricultural Technology Management Workshop, 6-8 July 1988, Rutgers University,
New Jersey, USA

Krecik, Steve 1993a. Fodder Tree Establishment and Production in Seasonally Dry
Areas of Jamaica. MS. Thesis, Michigan State University, East Lansing, MI.

Krecik, S., D.O. Lantagne, M.A. Gold, and J.M. Roshetko 1994. "Cutting Management
of Leucaena leucocephala, Calliandra calothyrsus, and Gliricidia sepium for Fodder
Production." Nitrogen Fixing Tree Research Report Vol. 11.

Le Houe'rou , N.H. (ed.) 1980. Browse in Africa: The Current State of Knowledge.
Addis Ababa: ILCA 491 p.

Lightfoot, C., 0. De Guia Jr., and F. Ocado 1988. "A Participatory Method for
Systems-Problem Research: Rehabilitating Marginal Uplands in the Philippines."
Experimental Agriculture 24: 301 -309.

Little, PD. 1985. "Adding a Regional Perspective to Farming Systems Analysis:
Concepts and Analysis." Human Organization 44(4) Winter: 331-338.

MacDicken, K.G., G.V. Wolf, and CB. Briscoe (eds) 1991. Standard Research
Methods for Multipurpose Trees and Shrubs. Winrock International Institute for
Agricultural Development.

Malla, Y.B. 1988. "The Case for Placing More Emphasis on Private Tree Planting
Programmes." Forestry Research Paper Series No. 13. HMG-USAID—GTZ-FORD-
WINROCK Project. Kathmandu, Nepal.

Maurya, DA, A Bottrall, and J. F arrington 1988. "Improved Livelihoods, Genetic
Diversity and Farmer Participation: A Strategy for Rice Breeding in Rainfed Areas of
India." Experimental Agriculture 24: 311-320.

Merrill-Sands, D. and J. McAllister 1988. Strengthening the Integration of On-Farm
C Iient-Oriented Research and Experiment Station Research in National Agricultural
Research Systems (NARS): Management Lessons from Nine Country Case Studies.

103

OFCOR Comparative Study Paper No. 1. The Hague: International Service for National
Agricultural Research (ISNAR).

Morikawa, R 1993. "Fodder Tree Establishment and Utilization in a Seasonally Dry Area
of Jamaica." Paper presented at the Fifth Annual Jamaican Agricultural Seminar of the
Jamaican Agricultural Research Program (JARP), July 1993, Kingston Jamaica

Morrison, B]. 1991. Indigenous Knowledge Relating to Fodder Trees and Silvo-
pastoral Management Systems of Small-Scale Farmers in Jamaica. M. S. Thesis,
Michigan State University, East Lansing, MI.

Nair, P.K.R. 1993. An Introduction to Agroforestry. The Netherlands: Kluwer
Academic Publishers.

Nair, P.K.R. 1992. "Agroforestry Systems: An Ecosystem Approach." In N.P. Sharma
(ed) Managing the World's Forests. Iowa: Kendall/Hall Publishing Co.

Ndufa, J .K., E. Ohlsson, and KB. Shepherd 1992. "Participatory Research Methods for
Agroforestry Technology Development in Western Kenya. " Paper presented at the 12th
Annual Association for Farming Systems Research Symposium, 13-18 September 1992,
Michigan State University, East Lansing.

Noronha, R and J .S. Spears 1985. "Sociological Variables in Forestry Project Design."
In M.M. Cemea Putting People First. Sociological Variables in Rural Development.
World Bank.

Pandy, K. 1982. Fodder Trees and Tree Fodder in Nepal. Swiss Development
Cooperation. Berne, Switzerland.

Pinners, E. and V. Balasubramanian 1991. "Use of the Iterative Diagnosis and Design
Approach in the Development of Suitable Agroforestry Systems for a Target Area. "
Agroforestry Systems 15: 183-201.

Pinney, A. 1991. "Farmer Augmented Designs for Participatory Agroforestry Research."
Agroforestry Systems 15: 259-274.

Raintree, J .B. 1983. "Strategies for Enhancing the Adoptability of Agroforestry
Innovations." Agroforestry Systems 12173-187.

Raintree, J .B. 1987. D&D User's Manual: An Introduction to Agroforestry Diagnosis
and Design Nairobi, Kenya: International Council for Research in Agroforestry.

Raintree, J .B. 1990. "Agroforestry Diagnosis and Design: Overview and Update." In
W.W. Budd, I. Duchhart, L.H. Hardestry, and F. Steiner (eds) Planning for

104

Agroforestry: Developments in Landscape Management. New York: Elsevier Sciences
Publishing Co., Inc..

Raintree, J .B. and F. Torres 1986. Agroforestry Research in Farnring Systems
Perspective: The ICRAF Approach. ICRAF Working Paper No. 39. ICRAF, Nairobi,
Kenya.

Rajasekaran, B. 1992. "Farmer Participatory Approaches to Integrate Indigenous
Knowledge Systems and Research Station Technologies Toward Sustained Food
Production and Resource Conservation in India." Paper presented at the 12th Annual
Association for Farming Systems Research/Extension, 13-18 September 1992, Michigan
State University, East Lansing.

Reynolds, L., C. di Domenico, AN. Atta-Krah, and J. Cobbina 1991. "Alley Farnring in
South-Western Nigeria: The Role of F arming Systems Research in Technology
Development. " In R. Tripp (ed.) Planned Change in Farming Systems. New York: John
Wiley & Sons

Rhodes, R, R. Booth, E. Schmidt, and O. Cuyubamba 1991. "Post-Harvest Technology
Development with Farmers: The Difiirsed Light Storage Case." In R Tripp (ed.)
Planned Change in Farming Systems. New York: John Wiley & Sons.

Rocheleau, DE. 1991. "Participatory Research in Agroforestry: Learning From
Experience and Expanding Our Repertoire." Agroforestry Systems 15: 111-137.

Rocheleau, DE. 1985. "Criteria for Re-appraisal and Re—desigrr: Intra-household and
Between-household Aspects of FSRE in Three Kenyan Agroforestry Projects." Paper
presented at the Annual Farming Systems Research and Extension Symposium, Kansas
State University, Manhattan Kansas, 7-10 October 1984. ICRAF Working Paper No. 37.

Rocheleau, DE. 1984. "Land Use Planning With Rural Farm Households and
Communities: Participatory Agroforestry Research. " Paper presented to the Networkshop
on the Role of the Sociologists and Anthropologists in F arnring Systems Research.
Sponsored by ARPT Zambia and CIMMYT East Afiica, 27-29 November 1984, Lusaka
Zambia.

Rocheleau, D.E., F. Weber, and A. Field-Juma 1988. Agroforestry in DrylandAfi'ica.
Nairobi: ICRAF.

Roshetko, J .M. 1991. Establishment and Nutritive Value of Native and Exotic Fodder
Tree Species in Jamaican Pasture Systems. M. S. Thesis, Michigan State University, East
Lansing, MI.

105

Rusten, ER 1989. An Investigation of an Indigenous Knowledge System and
Management Practices of Tree Fodder Resources in the Middle Hills of Central Nepal.
Ph.D. Dissertation, Michigan State University, East Lansing, MI.

Sagar, D. and J. F arrington 1988. "Participatory Approaches to Technology Generation:
From the Development of Methodology to Wider-Scale Implementation. " Agricultural
Administration Network Paper No.2 Agricultural Administration Network. London:
Overseas Development Institute.

Scherr, SJ. 1991. "On-Farm Research: The Challenges of Agroforestry. " Agroforestry
Systems 15: 95-110.

Shaner, W.W., P.F. Philipp, W.R Schmehl 1982. Farming Systems Research and
Development. Boulder”. Westview Press.

Shanna, N.P, R, Rowe, M. Grut, R Kramer and H. Gregersen (1992). "Conditions for
Sustainable Development." In N.P. Sharma (ed) Managing the World's Forests. Iowa:
Kendall/Hall Publishing Co.

SIOJ 1991. Statistical Yearbook of Jamaica - 1991. Kingston: The Statistical Institute
of Jamaica.

Sumberg, J. and C. Okali 1988. "Farmers, On-Farm Research and the Development of
New Technology." Experimental Agriculture 24: 333-342.

Torres, F. 1983. "Role of Woody Perennials in Animal Agroforestry." Agroforestry
Systems 1:131-163.

Tripp, R. 1991a. "An Overview of the Cases of On-Farm Research." In R. Tripp (ed.)
Planned Change in Farming Systems. New York: John Wiley and Sons

Tripp, R. 1991b. "The Limitations of On-Farm Research." In R. Tripp (ed) Planned
Change in Farming Systems. New York: John Wiley and Sons

Tripp, R. and J. Woolley 1989. The Planning Stage of On-Farm Research: Identifying
Factors for Experimentation. Colombia: CIMMYT and CIAT.

 

'rrrrrrrlrllrrr