|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
ESTABLISHMENT OF SILVOPASTORAL SYSTEMS IN DEGRADED, GRAZED
PASTURES: TREE SEEDLING SURVIVAL AND FORAGE PRODUCTION UNDER TREES
ALYSON B. K. DAGANG
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Alyson B.K. Dagang
To my Mother and Father, whose boundless love gives me life
There are many individuals and organizations who contributed to this study and my
doctoral program to whom I am indebted and grateful. I thank my chair, Dr. P.K. Nair for his
dedication and guidance throughout this process, and my committee, Dr. Peter Hildebrand, Dr.
Kaoru Kitajima, Dr. Tim Martin, Dr. Lynn Sollenberger, and Dr. Marilyn Swisher, for their faith
and confidence especially through rocky times.
I would like to recognize and express my sincere gratitude to the individuals and their
institutions that supported me during my doctoral studies, including the School of Forest
Resources and Conservation (Cherie Arias, Sherry Tucker, Dr. George Blakeslee, Dr. Wayne
Smith), the Institute of Food and Agricultural Sciences, the Center for Tropical Conservation and
Development, the College of Agriculture and Life Sciences, the National Security and Education
Program, the Southeast Alliance for Graduate Education (NSF-SEAGEP), the Department of
Energy FLAS program, the University of Florida Alumni Association, and the School for
International Training (SIT).
This study would not have been possible without the constant support I received from the
farmers, families, and other collaborators in Panama. Thank you to Mr. Severito Martinez, Dr.
Juan Jean, Viodelda de Suarez, Antonio Suarez, Famila Jaen, Familia Suarez, Familia Martinez,
Familia Graj ales, Familia Villareal, Familia Aguilar, Lic. Jose Villareal, personnel from the
Laboratorio de Suelos del Instituto de Investigacion Agricola de Panama (IDIAP), Dr. Rodrigo
Velarde, and Dr. Jaime Velarde.
Over the years, I have greatly benefited from and been enriched by the presence of the
members of the UF Agroforestry lab. To Andrea Albertin, Shinjiro Sato, Matt Langholtz, Paul
Thangata, Jimmy Knowles, Bocary Kaya, Robert Miller, Eddie Ellis, John Bellow, Brian Becker,
Abiud Mwale, Asako Takimoto, Solomon Haile, Soumya Mohan, Alain Michel, David Howlett,
Joyce Lepetu, Subrajit Saha, Mark Drew, and Wendy Francesconi, thank you for your support,
friendship, humor, and tremendous spirit.
To my treasured compafieros and sisters who have been an integral part of the many years
of this process, thank you Sharene Esias, Molly Rhodes, Deb Sparadeo, Yvie Fabella, Mikilin
Esposito, Leilani Pedro, Steve Taranto, Osvaldo Jordan, Luis Dominguez, Juan Nuques, Cynthia
Gomez, Alicia Peon, Leonardo Martinez, Jennie Saqui, and Pio Saqui.
I express my profound gratitude to the Dagang family for their love and support. And,
most importantly and profoundly, to my mother, Catherine Henig, without whom this endeavor
would have never come to fruition.
TABLE OF CONTENTS
ACKNOWLEDGMENT S ................. ...............4....___ ......
LIST OF TABLES ............ ............ ...............10...
LIST OF FIGURES ............ ............ ...............11...
AB S TRAC T ............._. .......... ..............._ 13...
1 INTRODUCTION ................. ...............15.......... ......
2 AGROFORESTRY AND LAND USE IN PANAMA AND A GENERAL
DESCRIPTION OF THE STUDY SITE............... ...............18..
Agroforestry ................. .. ............ ...............18.......
Benefits of Agroforestry Systems ................ ................ ......... ........ ...._18
Relevance of Agroforestry in Panama ...._. ......_._._ .......__. ............1
Silvopastoral Systems............... ...............20
Choice of Tree Component .............. ...............21....
M icroclim ate............. ...... .. .. .. .... ...................2
Forage component Recent Studies on Forage Vegetation in Silvopastoral Systems ...24
Sum m ary ................ ..... ....... ........ ... ..... .. ..... .. .. .......2
Land Use and Land Use Change in Panama: A Background to the Impetus for the
Presented Research............... ...............26
Introduction ............... ...............26....
Emergence of the Isthmus ................ ........... ........................ ..............27
Development of Human Land Use in Panama ........._.._. ......_. ......._.. .......2
Introduction of Cattle and Land Use Change ........._.._. ......__ ......_.. ........2
Frontier Expansion and Green Revolution in Panama .............. ...............30....
Impacts of the Green Revolution. ............_. ...._.. ...._... ...........3
Land Use in Panama Today ............._. ...._... ...............31...
Cattle Ranching in Panama .............. ...............33....
Ranching Importance and Benefits .............. ...............33....
Economic Importance of Cattle ............._. ...._... ...............34...
Pasture Proliferation ........._..._.._ ...............35.._.._._ .....
Changing Nature of Ranching ........._..._.._ ...............35....._._ ....
Conclusion ........._..._.._ ...._._. ...............37.....
Research Site Description............... ..............3
Location ........._..._.._ ...._._. ...............37.....
E col ogy ........._..._.._ ...._._. ...............3_ 8...
Climate .................. ...............3 8..
Local Farming Systems .............. ...............39....
Species Descriptions ........._..._.._ ...._._. ...............39.....
Tectona grandis .................. ............ ....._ ..........4
Origin, Natural Habitat, and Environment ....__ ................ ........._ ......40
U ses .............. ...............41....
Botany ............... ........... ... .....__ .............4
Germination and Establishment .............. ...............41....
Adaptability and Performance ............. ...... ._ ...............42....
Rooting and Competition .............. ...............43....
Burning ......................... ..._ ...... ...............43
Potential benefits of teak plantations ............._..... ..._ ....._ ..........4
Bonabacopsis quinata (syn. Pochota quinata, Bonabacopsis quinatunt) .........................45
Anacardium occidentale ........._..... ...._... ...............46.....
Botanical description............... ..............4
Cultivation ........._..... ...._... ...............47.....
U ses ................. ...............48...
Planting Configuration .............. ...............50....
3 TREE SEEDLING SURVIVAL AND IMPACT OF HERBIVORY ON
SILVOPASTORAL SYSTEM ESTABLISHMENT .............. ...............61....
Literature Review .............. ...............62....
Tree Seedling Survival .............. ...............62....
Effects of Cattle Grazing ............. ...._._. ...............64....
Herbivory ........._..... ......_. ...............66....
Obj ectives and Hypothesis ................. ...............70.......... ....
Methods and Materials .............. ...............71....
Study Site............... ...............71..
Experimental Design .............. ...............71....
M material s ................ ...............71................
Establishment .............. ...............72....
Measurements ................. ...............72.................
Data Analy si s............... ...............72
R e sults............... .. .. ........ ...............73.......
Seedling Survival .................. ...............73.................
Ob served Causes of Mortality ................. ...............74.......... ....
Herbivory ........._..... ......_. ...............74....
Sources of Herbivory............... ..............7
D iscussion............... ..............7
Seedling Survival ........._.._... ........._..... ...............75.....
Observed Causes of Seedling Mortality ........._..... ....__. ....._. ...........7
Herbivory ........._..... ......_. ...............79....
Sources of Herbivory............... ..............8
Conclusion ........._..... ...............82._._.........
4 EFFECTS OF SCATTERED LARGE TREES INT PASTURES ON A Hyparrhenia
rufa-DOMINATED MIXED SWARD ................. ...............89........... ....
Literature Review .............. ...............89....
L ight .................. ...............89..
Biomass Allocation .............. ...............91....
Below ground Factors............... ...............91
Obj ective and Hypothesis ................ ...............93................
Methods and Materials .............. ...............93....
Study Site............... ...............93..
Experimental Design .............. ...............93....
Measurements ................. ...............94.................
Data Analysis............... ...............95
Re sults ................ ...............95.................
Forage Mass............... ...............95..
Forage Digestibility ................. ...............96.......... .....
Forage Composition .............. ...............96....
Forage Mass............... ...............97..
Forage Composition .............. ...............101....
Conclusion ................ ...............102................
5 INTERACTIONS BETWEEN TREE SEEDLINGS AND UNDERSTORY
VEGETATION DURING THE EARLY PHASE OF SILVOPASTORAL SYSTEM
ESTABLISHMENT ........._... ...... ...............113...
Literature Review ............. ...... ...............113...
Competitive Ability ............. ...... ...............114...
Competition for Soil Moisture. ............. ...... ...............116..
Root Biomass Allocation .................. ........_ ...............118..
Competition for Nutrients (Fertilization Studies)............... ...............12
M icroclimate Effects .............. ...............121....
Trenching Effects .............. ...............121....
Obj ectives and Hypothesis ................. ...............124......... .....
Methods and Materials .............. ...............124....
Study Site............... ...............124.
Experimental Design .............. ...............124....
M material s ................. ...............125................
Establishm ent .............. ...............125....
Measurements ................. ...............126................
Data Analysis............... ...............12
Re sults ................ .......... ...............126......
Herbage Removal ............... ... ..... .............12
Effects of the Species Treatment on Biomass ................ ............ ........ .........127
Stem Biomass ............. ...... __ ...............128...
Root Biomass ................. ...............128...............
D iscussion................. .............12
Seedling Growth ................. ...............129................
Stem and Root Biomass ................. ...............131...............
Conclusion ................ ...............132................
6 SUMMARY AND CONCLUSIONS ................ ...............140...............
Experimental Findings............... ...............14
Seedling Survival and Herbivory .............. ...............140....
Effects of Large Trees on Understory Forage ............ ..... .__ ........._......14
Interactions between Seedlings and Vegetation ......____ ..... ... ._ ..........._....142
Implications for Implementation .............. ...............143....
Options for Grazing ............ ..... ._ ...............143...
Manipulating Forage with Trees .............. ...............144....
Tree E stabli shment ............ ..... ._ ...............144...
Future Research ............ ..... ._ ...............144...
A PLANTING CONFIGURATIONS OF THE THREE TREE-SPECIES SEEDLINGS
FOR ESTABLISHMENT OF A SILVOPASTORAL SYSTEM IN RIO, GRANDE,
COCLE, PANAMA .........._.... ......... ...............146....
B COMPARISONS OF MEANS OF INCIDENCE OF TREE SEEDLING HERBIVORY
ACROSS TREE SPECIES AND PLANTING CONFIGURATION ................. ...............147
C FORAGE SAMPLING SCHEMATIC OF HERBAGE MASS HARVESTED AT
THREE DISTANCES FROM TREE STEM IN THE FOUR CARDINAL
DIRECTIONS CARRIED OUT UNDER SCATTERED TREES IN PASTURES IN
RIO GRANDE, COCLE, PANAMA .............. ...............148....
LIST OF REFERENCES .........._._ ...._.... ...............149.....
BIOGRAPHICAL SKETCH ..............._ ...............160......_ ......
LIST OF TABLES
2-1 Results of effects of Ziziphus joazeiro and Prosopis juliflora trees on buffelgrass
pasture in Northeast Brazil............... ...............52.
2-2 Total farm land, farms with cattle, and area under pasture in Panama, 2000. ...................57
2-3 Economic importance of cattle in Panama by province, 2000 ................. .............. .....58
3-1 Comparison of effects of planting configuration and species on survival of 675
seedlings planted in five blocks in degraded pastures on-farm over two years in
Cocle, Panama............... ...............84.
4-1 Analysis of variance for polynomial orthogonal contrasts of sample mean forage
mass comparing the effects of distance and season under dispersed Anacardium
occidentale trees in Rio Grande, Cocle, Panama. ........._._.._ ....... ........_.._.....104
4-2 Analysis of variance for polynomial orthogonal contrasts of sample mean forage
mass comparing the effects of distance and season under dispersed Tectona grandis
trees in Rio Grande, Cocle, Panama. ............. ...............105....
4-3 Post hoc comparisons of mean forage mass at three distances from dispersed 7
grandis tree stems in grazed, degraded pastures in Rio Grande, Cocle, Panama. ...........106
4-4 Post hoc analysis of forage digestibility across three distances from dispersed
Cashew trees (A. occidentale) and by two seasons in grazed pastures of Rio Grande,
Cocle, Panama............... ...............107
5-1 Analysis of the effects of the repeated measures herbage removal, tree species, and
time on biomass accumulation of tree seedlings planted on-farm in a non-grazed
pasture in Rio Grande, Cocle, Panama. ............. ...............133....
5-2 Comparisons of the within-subj ect effects of the repeated measure herbage removal
on biomass accumulation of tree seedlings planted on-farm in a non-grazed pasture
and observed over two years in Rio Grande, Cocle, Panama. ............. ....................13
5-3 Effects of the interactions of three seedling species with harvest time (6, 12, and 24
months after planting) on biomass accumulation of tree seedlings planted on-farm in
a non-grazed pasture in Rio Grande, Cocle, Panama. ........._..._.._ ...._._. ...............135
LIST OF FIGURES
2-1 Topographic map of the Panamanian isthmus. .........._...._ ....._. ......._._ ........5
2-2 Panama forest cover and areas of deforestation in 1947. ....._____ ........._ ..............54
2-3 Changes in land use and human population in Panama 1961-2003 ............... ..............55
2-4 Farm sizes and areas in Panama 2000. ........._._. ....____ ...............56.
2-5 Proportion of pasture area to total land area by corregimiento in Panama, 2003 ..............59
2-6 Research study site location, Rio Grande corregimiento, Cocle province, Republic of
Panam a. .............. ...............60....
3-1 Comparison of the survival curves of three tree seedling species (Anacardium
occidentale, Bombacopsis quinata, and Tectona grandis) (N = 675) planted in three
planting configurations (diagonal, fence, and line) during 900 days in pastures of Rio
Grande, Cocle province, Panama. ........._.._.._ ...._.._....._._ ...........8
3-2 Incidence of mortality among Anacardium occidentale, Bombacopsis quinata, and
Tectona grandis seedlings planted in three planting configurations for silvopastoral
system establishment in farmers' fields in Rio Grande, Cocle, Panama. ..........................86
3-3 Incidence of herbivory of three species of tree seedlings (N = 225 seedlings per
species) browsed by cattle, leaf-cutter ants, or other observed sources during a two-
year experiment in grazed on-farm pastures in Rio Grande, Cocle, Panama.. ........._.._......87
3-4 Incidence of cattle, leaf-cutter ant, and other sources of herbivory of tree seedlings
(Anacardium occidentale, Bombacopsis quinata, Tectona grandis) planted in three
planting configurations in grazed pastures in Rio Grande, Cocle, Panama.............._.._.. ...88
4-1 Forage mass under two species (Anacardium occidentale and Tectona grandis) of
isolated, large trees in a Hyparrhenia rufa-dominated mixed sward during two
seasons in Rio Grande, Cocle, Panama. ......___ .... ....._. ....._._............0
4-2 In vitro organic matter digestibility of forage from Hyparrhenia rufa mixed swards
under two species (Anacardium occidentale and Tectona grandis) of large, isolated
trees in pastures during two seasons, in Rio Grande, Cocle, Panama. ............................109
4-3 Proportional botanical composition of Hyparrhenia rufa mixed swards at three
distances from two species (Anacardium occidentale and Tectona grandis) of large,
isolated trees in pastures at the end of the wet season in Rio Grande, Cocle, Panama. ..1 10
4-4 Composition of forage categorized by weeds, grass, legume, and necromass across
three distances (0.5 (close to tree stem), 1.0 (drip line), 2.0 (open pasture)) from
Cashew (A. occidentale) tree stems in grazed pastures in Rio Grande, Cocle,
Panama. ........._..._.._ ...............111._.._._ ......
4-5 Composition of forage categorized by weeds, grass, legume, and necromass across
three distances (0.5 (close to tree stem), 1.0 (drip line), 2.0 (open pasture)) from Teak
(T. grandis) tree stems in grazed pastures in Rio Grande, Cocle, Panama. ........._.._........112
5-1 Responses of three species of tree seedlings to three understory-herbage- removal
treatments during the first two years after tree planting in a field site in Rio Grande,
Cocle, Panama............... ...............136
5-2 Biomass accumulation of stems and roots of three species of tree seedlings planted
for the establishment of silvopastoral systems in a field site in Rio Grande, Cocle,
Panama. ........._.._.. ...._... ...............137....
5-3 Changes in seedling biomass accumulation in stems and roots, and root:shoot ratio
(numbers above bars) changes during the two-year establishment of silvopastoral
systems in pastures in Cocle, Panama............... ...............138
5-4 Root: shoot ratios of three species of seedlings across grass removal treatments
during the two-year establishment phase of silvopastoral systems planted in pastures
in Rio Grande, Cocle, Panama. ........._.._.. ...._... ...............139..
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ESTABLISHMENT OF SILVOPASTORAL SYSTEMS IN DEGRADED, GRAZED
PASTURES: TREE SEEDLING SURVIVAL AND FORAGE PRODUCTION UNDER TREES
Alyson B. K. Dagang
Chair: P.K.R. Nair
Maj or Department: School of Forest Resources and Conservation
Silvopastoral systems that integrate trees on animal production units are reported to be a
promising land-use activity. Research on methods of integrating trees into smallholder pasture
systems for development of such systems in the tropics has, however, received little attention. In
Panama, smallholder pastures are abundant across the landscape, but they are often extensive,
degraded, overgrazed, and of low productivity. Based on the premise that integration of
silvopastoral systems on degraded pastures might be an effective technology that is accessible
and affordable for small-scale producers, this research was carried out on-farm for two years in
central Panama to help devise best management practices for optimizing tree-seedling survival,
reducing competition between seedlings and herbaceous vegetation, and managing effects of
large trees on forage.
Three experiments were conducted. The first one examined seedling survivorship and
herbivory of three tree species (Anacardium occidentale, Bombacopsis quinata, and Tectona
grandis) planted in three configurations (grouped in diagonals, in lines, and along fences). The
second experiment examined the effects of herbaceous vegetation on the establishment of tree
seedlings. Seedling growth and biomass distribution to shoots and roots were evaluated in
relation to four herbaceous removal regimes, which included removal of surrounding vegetation
both above- and belowground. In the third experiment that focused on the effects of large,
dispersed trees on forage characteristics, two tree species, Anacardium occidentale and Tectona
grandis, were evaluated for their effects in terms of mass, digestibility, and botanical
composition of the forage underneath.
Research results revealed that Anacardium occidentale seedlings survived best in grazed
pastures and the fence planting configuration resulted in the lowest seedling survival. Seedling
herbivory was greatest for Bombacopsis quinata, and cattle and leaf-cutter ants (Atta spp.) were
the herbivores that browsed seedlings most. Tree seedlings performed differently under the
different herbaceous vegetation removal regimes. Bombacopsis quinata grew best overall and
maintained a consistent root: shoot ratio during the two years of study However, Anacardium
occidentale performed better than the other species in terms of biomass allocation to shoots.
Similarly, the effects of large trees on understory forage varied with tree species. Forage mass
under 7 grandis was suppressed in comparison to A. occidentale. Conversely, forage
digestibility was lower under A. occidentale than under T. grandis. Finally, while forage
botanical composition was uniform (with a greater proportion of grass) under 7 grandis across
distances from tree stem, under A. occidentale, proportions of botanical composition were more
varied and comprised more legume than grass.
These results can be used for development of recommendations and guidelines on tree
species selection, planting configuration, grazing, weeding, and forage management for
successfully integrating silvopastoral systems into smallholder pastures in Panama.
In Panama, pastureland covers about 1.3 million hectares, constituting more than 20% of
the landscape. Existing pastures are extensive, low in productivity, commonly under some
degree of degradation, and practically devoid of trees. Although high-intensity technologies and
management technologies such as use of feed lots and supplemental, processed feeds exist to
augment productivity, these are untenable for most producers. New production strategies that
can be easily accessed, implemented, and afforded by producers must be sought. Silvopastoral
systems the integration of trees into livestock systems are considered to be one such approach
with the potential to address the problem of increasing degradation of existing pastures in
Panama. Based on the premise that tree integration on degraded pastures can augment soil
health, forage production, and environmental services, silvopastoral systems, might be an
effective technology that is accessible and affordable for producers. Several management
aspects of silvopastoral systems have, however, not been researched and therefore remain
unknown. It was in this context that the present study was undertaken.
The study, exploratory in nature, involved applied, on-farm research to devise appropriate
means of establishing silvopastoral systems on degraded pastures and to investigate how best to
integrate tree seedlings into grazed, degraded pastures in Panama. Maj or areas of investigation
included appropriate tree species and their optimal planting configuration in terms of seedling
survival and seedling herbivory. Consequences of large trees on pastures in terms of effects on
forage mass, forage digestibility, and forage botanical composition; and interactions between
herbaceous vegetation and establishing seedlings as they pertain to removal of herbaceous
material around seedlings were also investigated.
Three species, chosen by participating farmers, were used in the study: Anacardium
occidentale (cashew), Tectona grandis (teak), and Bombacopsis quinata (tropical cedar).
Anacardium occidentale is a locally abundant species that is valued for the marketable, well-
priced nut it produces and for its fruit, which is consumed by farm families and livestock.
Tectona grandis is arguably the most valuable tropical hardwood species that has been heavily
promoted throughout Panama in reforestation efforts and as a plantation species. Producers
perceive T. grandis as a commodity species that can provide added income from the pasture to
the household. Bombacopsis quinata is a multi-purpose, native hardwood species that is used
locally in live and dead fences, furniture making, and in construction.
The overall obj ective of this research was to gain knowledge of some of the bases of
silvopastoral system establishment in degraded, grazed pastures. Through monitoring seedlings
for survival and herbivory over two years, manipulating herbaceous vegetation and tree seedlings
above- and belowground, and testing forage characteristics close to and far from isolated trees,
the study was also aimed at understanding some of the interactions that occur in silvopastoral
systems in extensive pastures in Panama.
The study sought to examine particular assumptions regarding the use and performance of
A. occidentale, T. grandis, and B. quinata in silvopastoral systems as well as the impact of these
species on pasture. Specifically, the following general hypotheses were tested:
* The pattern in which tree seedlings are planted in pasture (planting configuration) impacts
the survival and herbivory of seedlings.
* Differences exist among tree species in terms of their performance under different planting
configurations in silvopastoral systems.
* Removal of herbaceous vegetation around establishing seedlings has positive effects on
* Isolated, large trees impact mass, digestibility, and botanical composition of the understory
This dissertation is presented in six chapters. Following this introductory chapter, Chapter
2 expands upon the problem statement providing an in-depth discussion and background to the
drivers behind land use in Panama today and presenting the overall context for the motivation
behind the research presented. Chapter 2 also includes a review of the relevant silvopastoral
system literature as well as tree species and research site descriptions. Chapters 3, 4, and 5
present the experiments conducted in this research. Chapter 3 comprises the presentation of the
experiment and its results that examined seedling survival and herbivory of three tree species on
five farms in extensive pastures in Central Panama. Chapter 4 provides the results from the
study that examined the consequences of dispersed, large trees on forage characteristics in
pasture. Chapter 5 presents the results from the experiment that studied the effects of above- and
belowground vegetation removal on tree seedling growth in a controlled field site. Each of the
three chapters includes an explanation of the experimental methodology, a review of the
pertinent literature, a description of the study results, and a discussion of the findings. Finally,
Chapter 6 provides a synthesis of the results of the experiments, implications for the on-farm
integration of trees into extensive pastures, and recommendations for future research based on
the outcomes of the research.
AGROFORESTRY AND LAND USE INT PANAMA AND A GENERAL DESCRIPTION OF
THE STUDY SITE
Agroforestry entails the deliberate growing of woody perennials on the same unit of land
as agricultural crops and/or animals in some form of special mixture or sequence that results in a
significant interaction of woody and non-woody components (Nair, 1993). There is evidence of
the implementation of agroforestry systems dating 10,000 years before present (Miller and Nair,
2006; Gakis et al., 2004). Widespread study of these traditional practices has grown during the
20th century. Researchers who seek appropriate technologies to respond to growing food needs,
diminishing global ecological health, and the rise in land degradation have embraced
agroforestry practices as a suite of systems with the potential to meet some of these demands
(Huxley, 1999). Some of these systems include alley cropping for soil improvement, fodder
production for livestock and dispersed trees in pasture for enhancing animal production, fallow
enhancement for soil enrichment, home gardens for food and nutritional security, and others
(Nair, 1993). Silvopastoral systems, a type of agroforestry, involve the interaction of woody
perennials, forages, and livestock. The three components in the system are intentionally
managed for optimal interactions aimed at augmenting agricultural production and
environmental services (Sharrow, 1999). Silvopastoral systems will be discussed further in this
Benefits of Agroforestry Systems
Agroforestry systems such as improved fallows, alley cropping, and silvopasture offer
benefits for agricultural production and environmental enhancement. Benefits from improved
fallows involve the augmentation of soil physical and chemical properties through the short-term
planting of soil-improving tree species. These can be an answer to exhausted soils or degraded
lands (Nair et al., 1999; Sanchez, 1999). Alley cropping is the combining of woody perennials
and annual crops in fields with the aim of enhancing crop production through enriched nutrient
cycling (Jordan, 2004).
Improvement in agricultural production through agroforestry systems is based in part on
the contribution of woody species to enhanced nutrient cycling. The woody perennial
component of the systems may provide multiple services to crops and/or forage by accessing
belowground resources in lower soil columns through deep roots. Likewise, increased capture of
light can enrich the overall production of the system (Ong et al., 1996). In some cases, the
woody component may provide needed soil moisture to neighboring vegetation by excising
moisture from deep soil sources and redistributing it near the soil surface, a debated phenomenon
known as hydraulic lift (Burgess et al., 1998; Emerman and Dawson, 1996).
Relevance of Agroforestry in Panama
Currently well-known and implemented agroforestry systems in Panama include home
gardens, live fences, dispersed trees in pastures and crop fields as well as to a lesser extent coffee
(Coffea spp.) and cocoa (Theobroma cacao) shaded perennial systems. Although certain systems
such as live fences are extensively used in Panama, agroforestry systems have not been
holistically embraced by Panamanian land managers as an alternative for improving agricultural
production. However, the existing multitude of agroforestry systems are in fact relevant to
Panama in that they have the capacity to address important challenges that the agricultural and
environmental sectors face today, including issues of burning, deforestation, and land
Three current deleterious situations include 1.) burning for plot clearing and short-term soil
enhancement, 2.) deforestation for pasture creation, and 3.) pasture degradation. These
situations are highly detrimental to the natural resource base and agroecological conditions in the
short-term and in the long-term. Pasture degradation and creation are among the leading causes
of deforestation. As such, integration of silvopastoral systems into existing agricultural
enterprises can potentially enable farmers to reduce the degradation of their farms (Serrao and
Toledo, 1990). Benefits and characteristics of silvopastoral systems will be discussed in detail in
the next section.
As noted above, silvopastoral systems, a form of agroforestry, include land-use practices
that involve woody perennials, forage plants, and livestock simultaneously during a period of
time to enhance production and/or the environment. One type of silvopastoral system, cut-and-
carry fodder banks entails the growing of forages in a confined space. Forages are harvested and
taken to livestock as opposed to being directly grazed. Another type of silvopastoral system
includes grazed systems. These may involve the establishment of high quality fodder banks
which are protected from herbivory at most times but are periodically grazed by cattle. Another
grazed system includes dispersed tree systems in which trees grow on pasture at different stand
densities but trees are not directly grazed. However, depending on the tree species, livestock
commonly graze fallen fruits, seeds, nuts, and foliage. Each of these systems offers different
advantages and benefits for agricultural production.
From improved microclimate to increased productivity, there is a multiplicity of
production and conservation benefits reported by researchers that occur in silvopastoral systems.
Garret et al. (2004) suggest multiple obj ectives are achievable through the implementation of
silvopastoral systems. They postulate that social, environmental, and economic benefits can be
obtained through improving forage quality, increasing timber production, sequestering carbon,
reducing contaminant run-off, enriching wildlife habitat, and improving landowner income. For
example, studies in semiarid northeastern Brazil conclude that maintaining 30% of tree cover
when converting forest vegetation to pasture increased forage and beef production in comparison
to areas with no remaining trees (Araujo Filho 1990 as cited by Menezes et al. 2002). Although
researchers agree on the benefits offered by silvopastoral systems, there is a great deal of
research that needs to be carried out in order to make appropriate recommendations for
silvopastoral systems in terms of tree density, forage cultivars, and animal stocking rates.
Although several aspects of agroforestry systems in general and silvopastoral systems in
particular have been studied, the following brief review of literature will highlight general topics
of silvopastoral system research which are included in this particular study. In the following
chapters, specific reviews of literature address the topics in greater detail.
Choice of Tree Component
Species selection for the tree component in a silvopastoral system is vital in that the unique
characteristics of each species including rooting habit, litter quality, canopy architecture,
allelopathy, radiation interception, and other traits can have decisive impacts on the nature and
outcome of the system and its parts. Research has yet to identify and ubiquitously recommend
appropriate tree species to be used in temperate or tropical pasture systems. However, Garret et
al. (2004) agree that properties such as canopy density, species phenology, vigor, and growth
habit are crucial characteristics to be identified for the integration of a tree component into
silvopastoral systems. Likewise, Cajas-Giron and Sinclair (2001) suggest that the canopy strata
which trees occupy as well as the products they offer in terms of leaf forage, fruits, and other
products are key determinants for the choice of tree species in silvopastoral systems.
Some studies have been conducted testing pine species (Pinus spp.). For example, in a
modeling study by Ares et al. (2003) based on data from long-term silvopastoral studies in the
southern U.S.A., it was found that growth of southern pines (Pinus spp.), was sensitive to
understory composition. Also, differences in grazing, fertilization, and tree population density
significantly affected the growth of the studied pine stands. Similarly, in New Zealand, Chang
and Mead (2003) in an eight-year study found radiata pine (Pinus radiata) diameter growth to be
sensitive to understory forage composition although tree height was not significantly affected at
the end of the experiment. Moreover, in a study looking at broad-leaved species, Teklehaimanot
et al. (2002) found significant differences in growth between sycamore (Acer psuedopla~tanus)
and alder (Alnus rubra) in a study in North Wales. They attributed these differences to species
amenability to spacing and/or different levels of nitrogen availability in the soil. However,
neither species had a significant effect on sheep and lamb stocking rates in terms of productive
Within a silvopastoral system, the multiple effects of microclimate created by the tree
component and the understory vegetation can have positive and negative impacts on production
as a whole as well as on the individual parts of the system. Microclimate characteristics and
potential consequences were studied by Menezes et al. (2002) in semiarid Brazil using two
unique tree species (Ziziphus joazeiro and Prosopis juliflora) and buffel grass (Cenchrus ciliaris)
as the primary understory vegetation. They found that microclimate effects on pasture soil
differed by tree species. The results of their study provide an excellent example of the
microclimatic effects of trees on pasture and highlights how these can differ by species (Table
As seen in the Menezes et al. (2002) experiment, canopy radiation interception and
therefore canopy architecture can play an important role in the effects of the tree component on
understory vegetation. In West Virginia, Feldhake (2001) studied the effects of black locust
(Robinia pseudoacacia) canopy on a tall fescue (Festuca arundinacea) pasture. He studied
photosynthetically active radiation (PAR), red/far-red ratio, and soil temperatures and found that
under increasingly cloudy conditions (25% PAR), % PAR under black locust canopy relative to
open field PAR doubled. Moreover, the author posited that the presence of the black locust
canopy reduced the extent of extreme conditions that the understory vegetation had to endure and
therefore to which it must adapt which he asserted may be beneficial. He concluded that
increased radiation use efficiency of the forage under diffuse light conditions as opposed to
direct sun increased forage production. Feldhake (2001) also found a significant difference in
soil temperature when comparing open-field and under-canopy temperatures. During a mid-day
reading, there was a difference of 6.5oC in soil temperature under the two scenarios with
equivalent soil moisture. In response to a 10% decrease in soil moisture, soil temperature in the
open field increased 12oC while under the black locust canopy soil surface temperature increased
2oC. According to Feldhake (2001), temperature conditions under the black locust canopy were
consistently within the appropriate range for tall fescue. Feldhake (2002) also found significant
differences in night temperatures in an on-farm silvopastoral system. His research results
showed that average below canopy nighttime temperatures in a southern West Virginia 35-yr-
old, 17-m-tall mixed conifer site with orchardgrass (Dactylis glomerata) understory was 11.5oC
higher than open field temperatures. Results from the Feldhake experiments demonstrate the
potential for the use of trees to moderate extreme temperatures that can be disadvantageous for
forage plants in pasture systems.
Contrary to the findings of Feldhake (2001; 2002), Dulormne et al. (2004) found no
significant differences between air temperatures or humidity under the tree canopy of a
Gliricidia sepium-Dichanthium aristatum silvopastoral system and Dichanthium aristatum open
field in Guadeloupe. However, there was a significant difference in wind speed between the two
system types. On the other hand, grass growth in the wet season was significantly greater in the
open field. However, during the dry season, there was no significant difference observed for
grass dry matter production between the two field types. Likewise, in the dry season no
significant difference was found between treatments in terms of soil porosity among the three
tested soil. However, interestingly, Dulormne et al. reported that in a previous study
(Tournebize, 1994) carried out on the same study site, it was observed that air temperature and
humidity were in fact higher under the Gliricidia sepium canopy. Nevertheless, the authors note
that in the previous study, the canopy of G. sepium was far larger (covering the entire interrow)
than the current canopy studied and therefore may have resulted in these different findings. The
comparison of these two studies illustrates how different management schemes can affect the
interactions among silvopastoral system components. They also highlight the importance for
research to address how different management types can result in distinct agronomic and
Forage component Recent Studies on Forage Vegetation in Silvopastoral Systems
As mentioned in the microclimate section, the varied characteristics of tree species can
influence the overall productive outcome of a silvopastoral system. Positive and negative effects
can occur belowground between the forage plant and tree component as well as aboveground
through shading and fallen leaf litter.
A vivid example of the dynamic effects of tree-forage interactions was found in an
experiment carried out in Australia studying the raintree Sanzanea sanzan in a dispersed tree
silvopastoral system. Durr and Rangel (2002) looked at forage growth proximate to the S. sanzan
canopy. The authors sampled biomass accumulation under the canopy, at the drip line, and in
open field. They found no significant difference in aboveground biomass accumulation between
the drip line and open field samples. However, under the canopy, aboveground biomass
averaged 90% more than the drip line and open field samples (found to be significantly
different). Another part of this experiment examined the botanical composition of the forage
species in the different canopy regions and found important contrasts that could explain the
sizable differences in aboveground accumulation in the different canopy zones. The below
canopy zone which was found to have overwhelmingly greater abundance of aboveground
biomass was dominated by PanicuntPPP~~~~PPP~~~PPP nzaxinzun, an important tropical forage species. The drip
line was populated by a mix ofP. nzaxinaun and Urochloa nzosa~nbicensis and the open field was
dominated by U. nzosa~nbicensis. This species specialization by canopy region was generally
static most of the year except during the dry season when there was an increase in U.
nzosa~nbicensis at the drip line. This study illustrates how understory forage species can differ in
preferences for proximity to tree crowns, another important element in the design and research of
Kallenbach et al. (2006) addressed a similar issue in Missouri, USA, looking at forage
growth, nutritional quality, and livestock performance under young mixed stands of pitch pine
(Pinus rigida, loblolly pine (Pinus taeda), and black walnut (Juglans nigra). Their experiment
produced diverse results. Using pasture blocks with and without trees, they measured forage
abundance over two years and found that pasture without trees consistently produced more
forage than the pasture with trees. Yet, there were apparent seasonal differences of less forage
abundance in the treeless pastures which the authors speculate can be attributed to the buffering
of temperature and wind in the treed pastures.
Forage is a principal component of silvopastoral systems. Its abundance or scarcity can be
the determining factor in the productivity of a farming system. Forage species that demonstrate
shade tolerance and effective rooting abilities may provide greater advantages when used in
silvopastoral systems. Likewise, tree species without highly competitive tendencies that are not
especially sensitive to effects of understory competition may be preferential for silvopastoral
systems. It is plausible that, given the appropriate companion components and management,
forage productivity can be enhanced through the integration of silvopastoral systems in livestock
farming systems. Considering the need to develop alternatives to present day, traditional
agriculture in the interest of ecosystem health and farm productivity and survival, agroforestry is
one option for farmers. Silvopastoral systems in particular offer viable options for agricultural
improvement and ecosystem health through the integration of woody perennials into farming
systems. Specific, specialized research is needed on silvopastoral systems in the tropics due to
the importance of synergy among system components and that these be optimal for the success
of the systems.
Land Use and Land Use Change in Panama: A Background to the Impetus for the
This section discusses historical, human, ecological, and social drivers behind present day
land use. The aim of the discussion is to illustrate the motivations behind the research reported
in this dissertation, which was devised in response to contemporary Panamanian realities of land
use change, degradation, and indications of declining agricultural productivity. Factors
contributing to land use change are multifaceted, not only made up of modern agro-ecological
realities but are also a result of the natural history of the isthmus and the land use practices
applied by pre-colonial populations, Spanish colonists, and 20th century homesteaders. Such
historical factors coupled with current socioeconomic conditions transcend and shape today's
land use issues. In order to understand these situations and thereby shed light on the conception
of this research, this section will convey the development of the Panamanian isthmus, pre-
historic land use, the legacies of fire and savanna crops left by pre-colonial populations and
colonists, consequences of the introduction of cattle on to the landscape, and the nature of land
Emergence of the Isthmus
Three million years ago, the Panamanian isthmus emerged connecting Central America
and South America. The occurrence had profound impacts on regional terrestrial and marine
ecology including the definitive separation of the Atlantic and Pacific Oceans (Coates, 1997).
The connection of the Americas through the emergence of the isthmus also gave way to the
Great American Faunal Exchange (Webb, 1997).
With the rise of the isthmus, a mountain range was formed, a feature that creates one of the
central pieces of Panama' s topography (Figure 2.1i). The resulting cordillera central is the
central mountain range that moves through Mesoamerica and continues into Panama creating
two prominent and distinct climatic and ecological zones. These include what are known as the
Pacific seasonal region and the wet Atlantic region. Historically, this geographic and climatic
distinction has had a decisive impact on the ecological, agricultural, and human development of
Panama. The unique eco-climatic regions created by the central range continue to influence land
Two unique precipitation zones are created in part by the predominant directions of trade
winds. These generally blow from northeast to southwest causing areas north and east of
mountain ranges to be wet, and those south and west of mountain systems to be drier. This
occurs in Panama consequent to the presence of the central mountain range. The phenomenon is
also known as an orographic rain shadow. Murphy and Lugo (1995) site Panama as a primary
example of this geographical contrast in precipitation patterns. They state, "The Pacific coast of
Panama, supporting semideciduous forest, receives about 1780 mm of annual rainfall whereas
the evergreen forest of the Caribbean coast receives over 3300 mm. On the Caribbean side,
minimum monthly rainfall is normally > 38 mm while the Pacific coast receives < 13 mm during
the cooler months of February and March." This situation results in the northern part of the
country being subj ect to continuous, very humid conditions throughout the year (3000 to 4000
mm) while the southern plains and mountains of the country are seasonally dry during five to six
months of the year (Murphy and Lugo, 1995).
Contrasting precipitation and topography have brought about the development of unique
ecological zones (Piperno and Pearsall, 1998). On the north side (Atlantic), there are steep
slopes, dense forest canopy, abundant fast-moving rivers, few mangroves, and extensive
wetlands. On the south side, there are dry, wide plains; moderate mountain slopes; extensive
rivers; mangrove forests; and varied seasonal forest types (ANAM, 2000).
Development of Human Land Use in Panama
Today, land use is a product of land occupation, manipulation, and cultivation by human
civilizations over millennia coupled with the demands of political and economic changes
experienced during the 20th century. To understand what is going on today in terms of land use,
food production, and conservation, it is crucial that one become familiar with the history of the
Panama' s topographical and ecological contrasts play a key role in the nature of the natural
and human transformation of the landscape and development of land use on the isthmus. The
unique ecological zones were fundamental to the development of human civilization during the
pre-colonial period in Panama. The flatter, drier southern side of the country with more
abundant river systems was favored by pre-colonial populations for farming, fishing, hunting,
and general existence. The very wet inhospitable, adverse conditions of the northern side of the
country presented greater challenges to survival than the southern coast (Linares, 1980).
Although the wet north coast presented challenges, some populations did live there. However,
their agricultural practices were profoundly distinct in that very small plots were slashed, soon
abandoned, and left for long fallow periods whereas southern populations developed expansive
crop savannas (Cooke, 1997).
Research reveals that pre-colonial populations in Panama began to use fire to manipulate
forests and augment abundance of desirable forest products during the period of 1 1,000 yr BP
Panamanian agriculture commenced in the period of 7,000 yr BP coinciding with the
introduction of maize (Zea mays) to the isthmus (Pipemo and Pearsall, 1998) and was rapidly
widespread by 2000 BP. In fact, it is reported that at the time of the Spanish arrival to the
isthmus (early 16th century), much of the southern flatlands was void of forest cover as a result of
the widespread use of fire and agricultural practices by pre-colonial populations as anthropogenic
savannas dominated the landscape (Jaen, 1985). However, the arrival of the Spanish in the 16th
century changed land use and land cover dramatically. Notably, the Spanish conquest provoked
a significant decrease in the pre-colonial population and a concomitant recovery of forests on the
landscape (Cooke, 1997).
Introduction of Cattle and Land Use Change
In 1521, Spanish merchants began to import cattle (Heckadon-Moreno, 1997) to graze
Panama' s former savannas and recovering forests. Introduction of cattle to the isthmus marked a
crucial turning point for the landscape as cattle counteracted forest recovery and impeded fallow
regrowth. Limiting forest regrowth was important to Spanish colonists for two reasons: it
facilitated the creation of extensive haciendas and controlled the invasive natural landscape
Following initial colonial settlement, the northern region was comparatively unpopulated
and became densely forested with a marked recovery of forests along the alluvial coastal plain.
The mountainous region, populated by descendants of indigenous groups escaped from slavery,
was cultivated in the traditional indigenous slash-and-burn system. The southern plains were
dominated by European settlers engaged in agriculture and cattle raising. The eastern region of
the country was sparsely populated by communities of escaped slave populations. However, by
the 18th century demographic changes spurred amplification of the anthropogenic savannas.
Settlers used cattle, fire, and traditional agricultural practices in tandem to increase space for land
settlement. The combined use of these was fundamental to population expansion and land
incursion. Agricultural area doubled between the beginning of the 17th century and the end of
the 19th century in the central provinces (Jaen, 1985). Characteristics of the rural Panamanian
landscape changed little from the 19th century through the early 20th century (Figure 2.2). Today,
of Panama' s 7.5M ha of land area, approximately 2.25M ha are covered by forest, 1.5M ha are
covered by pasture, and 0.5M ha are devoted to crops (Figure 2.3).
Frontier Expansion and Green Revolution in Panama
Today, Panama' s rural human and ecological landscape resembles in some ways that of the
early 20th century. However, certain developments have modified this situation. Firstly,
provision of basic medical care during the 20th century augmented the expansion of the human
population base (Heckadon-Moreno, 1997). In response to the new, growing population,
forested areas of the southern region neighboring the principle areas of commerce and cultivation
were expanded into including the southern portion of the Azuero Peninsula and the province of
Chiriqui (Heckadon, 1983; Jaen, 1985). Also, the population boom provoked an important rural-
to-rural migration that vastly expanded the agricultural frontier into 400 yr old forests (Herrera,
Impacts of the Green Revolution
The time at which rural-to-rural migration and large-scale expansion began (beginning in
the late 1950s) coincided with the initiation of the green revolution and heightened concurrently
with the spread of green revolution practices. For the rural sector in Panama, widespread rural
migration and the green revolution worked in concert as each circumstance mutually fueled the
other (Priestley, 1982). These conditions, coupled with a fervent State-sponsored campaign (the
"Conquer the Atlantic" campaign set forth by ruling General Omar Torrij os) to facilitate the
relocation of peoples from areas of burgeoning population growth and land scarcity into the
hinterlands, spawned a massive migration into forests (Dagang et al., 2003). Government-
sponsored migration into forest lands initiated multiple new agricultural frontiers, opening new
lands for cultivation and pasture creation, and in some cases, application of green revolution
practices (mechanization, synthetic inputs, new crop varieties, etc.). Green revolution practices
enabled farmers to increase agricultural production capacity and concomitantly continue
expansion into forests (Priestley, 1982). For example, between 1950 and 1970, the area devoted
to pasture production doubled (Jaen, 1985). Increased food production facilitated an increase in
family size thus provoking greater population growth and consequent further migration and
expansion of the frontiers (Figure 1.1).
Land Use in Panama Today
Panama' s landowners and occupiers consist of peoples who own or occupy small, medium,
or large parcels of land. In Panama, generally a small parcel can comprise 0.5 30 ha; a
medium-sized farm may be considered 31 120 ha; and a large farm may comprise more than
approximately 120 ha. Due to a multitude of global and national social, economic, and political
issues, the Panamanian agricultural economy has suffered in the past ten years which has
provoked important changes in land use and a transformation of the landscape (Dagang, 2004).
During this period, many small farms have been sold to medium and large farmers permitting the
consolidation of large landholdings (Figure 2.4). While these small farms traditionally
maintained a diversity of crops and livestock, larger owners generally choose to cultivate
monocrops and/or engage in single-species livestock raising. Some smallholders who have sold
their land have moved to urban areas to seek wage labor opportunities while others move to an
agricultural frontier area to continue traditional farming (Rudolf, 1999) and pasture creation,
among these are the agricultural frontiers initiated in the 1960s during the green revolution and
the campaign to "Conquer the Atlantic." Changes in farmer populations and parcel size resulting
from socioeconomic and political transformations have resulted in important alterations of the
landscape and land use patterns. The new, changed (and still transforming) Panamanian
agricultural landscape comprises medium and large-scale farming on the southern plains and into
the piedmont, dwindling small farmer population in the southern mountains, small farming on
degraded lands in the northwest, aggressive frontier expansion into the wet north and into the
east, abutting protected areas and indigenous reserves.
Today, smallholder farmers and some large landowners on the frontiers are moving into
the less populated areas of the wet north and extreme eastern regions of the country. However,
the newly migrating farmers have met different challenges than their predecessors. Frontier
expansion has become more tenuous due to a diminishing supply of unclaimed land and
increased demand for it by a larger population. Expansion is being limited by the preservation
status of protected areas and by the country's autonomous indigenous reserves. Conflicts among
populations for rights to land occupation and use have arisen and ignited social discord
(Benjamin and Quintero, 2005). Such diminishing supply of available land and the consistent
outward migration to agricultural frontiers are gradually prompting some producers to think
about how to reap greater production from their land but in a manner that will not damage their
limited commodity. This research was conceived and conducted to respond to this need.
Cattle Ranching in Panama
The changing landscape is dominated by cattle ranching and pasture proliferation.
Increasing cattle population and concomitant expansion of pastureland calls for a greater focus
and increased emphasis of research on pasture productivity within the context of growing land
scarcity as mentioned above. To embrace this situation optimally, it is vital that the dynamics of
today's land use, dominated by pasture and cattle, be understood. The following sections discuss
Ranching Importance and Benefits
Cattle ranching is pervasive throughout Panama and plays a strong cultural and ecological
role on the isthmus. Cattle and pasture are dominating features throughout the landscape (Table
2.1). Generally, cattle are highly valued within Panamanian society and ranching is an activity
that symbolizes wealth. Ranchers are generally regarded as influential community members and
important stakeholders (Dagang et al., 2003).
In addition to its cultural relevance, there are multiple incentives for raising cattle. Firstly,
raising cattle has traditionally been a more profitable and stable investment than many banking
ventures, providing salaried sectors of society with a steady, low-risk investment. According to
the National Bank of Panama, a 6-month investment in cattle can produce as much as 20% in
earnings on initial investment as compared to typical certificate of deposit interest rates (Banco
Nacional, 2003). Secondly, raising cattle is commonly embraced by city dwellers, who choose
to maintain strong ties with the countryside. To strengthen these connections, contribute to
kinship welfare, and simultaneously earn income on a stable investment, salaried city dwellers
will invest in cattle to achieve these multiple objectives (Dagang and Nair, 2003). Thirdly, cattle
provide emergency funds for farmers during moments of critical need such as family illness,
school initiation, and other expenses. Fourthly, in contrast to other agricultural endeavors, cattle
can provide immediate liquidity at key moments for their owners as opposed to crops which can
only be cashed in at harvest time. Fifthly, cattle provide the means for farmers without land title
or other property to attain credit. When producers do not hold title to their farm, cattle can be
used as collateral to obtain bank loans. Sixthly, cattle are used as a vehicle for land claim. For
instance, in regions of unoccupied land, forests are cleared for crop planting and after harvest are
seeded to grass (Joly, 1989). Then, cattle are introduced onto these lands for land claim. The
law recognizes use of land for cattle, not forest, as justification for land claim. As such clearing
forest and creating pastures for cattle fulfill two general obj ectives for squatters and
homesteaders: to inhibit the regrowth of forest, and to demonstrate to government land title
inspectors that requirements have been met for legal land claim (Villalobos, 2003).
Economic Importance of Cattle
Cattle represent an important part of the national economy particularly for the rural sector
which constitutes half of the Panamanian population. Cattle sales contributed more than $11 1
million to the national economy in 2000 (Table 2.2). This amount comprised 19% of the
agricultural contribution to the GDP, more than any other agricultural activity. These figures do
not include the contributions of the dairy industry to the economy in which annual milk sales
averaged approximately $30M. In addition, of the 503 corregimientosl surveyed in the 2000
agricultural census, 268 corregimientos produced more than $100,000 each in cattle activities
and 14 corregimientos produced more than $1M in cattle sales during the year 2000 marking
significant contributions to the rural economy.
SA "corregimiento" is the smallest political division recognized by the State. For example, corregimientos comprise
towns, districts comprise corregimientos, and provinces comprise districts.
The prominence and importance of cattle ranching is reflected in the vast areas occupied
by cattle in Panama. Of the 7.5M ha that constitute the country of Panama, approximately 1.5M
ha are cattle pasture. These 1.5M ha makeup approximately 20% of Panama's total land mass
and 71% of all agricultural land in Panama (Censo, 2001). Approximately half of the
corregimientos nationwide are covered by 40% or more with pasture and 1 12 of these
corregimientos are covered by more than 70% of pasture (Figure 2.5). Traditionally pastures are
extensive, maintain less than one head of cattle per hectare, are often degraded, covered by
naturalized grasses, managed non-intensively, and may have both flat and sloped topography.
Changing Nature of Ranching
Raising cattle has traditionally been a low-input activity. However, certain sectors of the
cattle industry are changing due to changes in economic globalization and a future that speaks of
the need to have to compete with imports. The agricultural sector has received incentives to
intensify cattle production. Laws 24 and 25 of 2001, including the "Programa para la
Reconversi6n Agropecuaria (Agricultural Conversion Program)," provide low interest loans,
reimbursements, and other assistance for farmers interested in improving their production
techniques. This program is sponsored by the Inter-American Development Bank and part of the
effort reimburses farmers on their investments in advanced agricultural technology. These
programs are geared toward large farming enterprises.
The goal of these laws is to equip and prepare farmers to compete with their counterparts
in other parts of the world in light of the imminent reduction of tariffs and assorted free trade
agreements Panama has pending (Gordon, 2001). In addition, recent law that mandates grading
of meat quality is slowly catalyzing changes in the meat industry particularly in terms of animal
genetics, nutrition, management, and investment. These changes have the potential to bear
significant effects on the ecological consequences of cattle ranching particularly in the reduction
of the use of extensive pastures. One of the emphases of these changes has been the reduction of
space in which cattle are raised i.e. the promotion of feed lots and stabling of cattle for fattening
in shorter time periods as opposed to the traditional system of grazing cattle during 3 5 years
on extensive pastures. However, the programs designed to encourage farmers toward confined
fattening (feed lots) programs have not been fruitful. Purchasing of feed which is unsubsidized
has not proven cost effective for farmers. In many cases, producers who originally tried these
techniques have reverted to extensive pasture fattening or semi-pastured feedlots.
In the past Hyve years farming conditions have begun to change as a result of the oscillating
economic situation and government programs geared toward improving agricultural productivity
nationwide and activity-wide. On some farms, pastures are beginning to be managed more
intensively through improvement in animal genetics, feed supplementation, and pasture
improvement (17% of pastureland has been planted with improved grasses and 97% of
corregimientos report having some type of improved grasses). However, these types of changes
require costly monetary investments. As a result, small-scale ranchers who raise cattle in an
extensive nature have been obliged in many cases to withdraw from the ranching business. It has
become more difficult economically to raise cattle extensively, due to declining productivity and
the increased cost of living. This implies that large areas of land are used that are costly to
maintain and that because cattle are fattened on pasture as opposed to feedlots, the cattle are
older when they are sold and thus the quality of the meat is low and money earned is less.
Hence, the traditional system requires more time for production and, today, renders fewer
earnings. It is proj ected that the change in technology use and intensification may render a
marked reduction in small-scale cattle farmers and only those farmers able to access credit and
invest in technology for farm improvement will prevail (Name, 2002). Due to the inaccessibility
of advanced technologies for some farmers and in other cases the inability to expand
landholdings, coupled with the existing need to improve traditional farming practices both for
land health and income, it is necessary to seek alternatives to agricultural practices employed
today. Agroforestry systems may be an alternative to traditional farming practices; silvopastoral
systems may be particularly important in the context of improving traditional cattle and pasture
Pre-historic peoples have left a vivid, indelible legacy of fire and savanna-like crop fields
on the Panamanian landscape. Introduction of cattle by the Spanish solidified the perpetuation of
the pre-historic legacies and added cattle to these to become an established trio of legacy land
use practices which have been embraced in their entirety by land use managers of the 20th and
21st centuries. The nature of land use today pillared by deforestation, pasture creation, and cattle
insertion has begun to confront its limits in that the supply of remaining unclaimed forest for
deforestation is diminishing and the existing pastures which in some cases have been worked for
centuries and in other cases during millennia exist in various stages of degradation. The research
presented in this dissertation was carried out in response to this land use crisis in Panama and
seeks to take a closer look at the potential of silvopastoral systems as an alternative for land
managers and their farms.
Research Site Description
Panama lies between Costa Rica and Colombia on the Central American isthmus. The
study site lies in the center of the country on the southern coast and is located in the
corregimiento of Rio Grande, in the Penonome district of the province of Cocle (08.3 1"T,
80.21oW)(Figure 2.6). The corregimiento of Rio Grande consists of extensive flatlands with a
landscape dominated by rice fields and cattle pastures. These lands are known to have been
inhabited and cultivated prior to colonial settlement, by pre-Columbian peoples, and were among
the first cultivated and grazed during the arrival of Spanish settlers (Jaen, 1985).
Rio Grande forms part of the dry tropical forest life zone (as described by Holdridge, 1967)
that characterizes Panama' s central Pacific flatlands. Dry forest zones are primarily climatically
determined and occur on a range of soil types. As depicted by Murphy and Lugo (1995), Central
American tropical dry forest occurs in the lowlands and temperature varies little throughout the
year. Seasons, therefore, are noted by changes in precipitation regimes. In the case of Rio
Grande, centuries and perhaps millennia of anthropogenic land use has eliminated the native
landscape. The corregimiento of Rio Grande lies approximately between 0 and 25 masl. Local
soil types are classified as chromic luvisols and dystric nitosols (ultisols and alfisols) (FAO,
1972; Nair, 1993). Specifically, Matthews and Guzman (1955) classify soils in the study site
area as pertaining to "Chumico sandy clay loam." Soil pH ranges from 4.3 to 5.9 and percentage
of soil organic matter ranges from 1.61 to 4.02.
There are two well-defined climatic seasons on Panama' s southern coast the wet season
and the dry season. In the last ten years in Rio Grande, the dry season has extended from
January to June and the wet season from July to December (observations from farmers). During
the wet season, 93% of the annual precipitation occurs. However, the corregimiento of Rio
Grande is situated in a well known microclimate called the Arco Seco or dry arc of Panama' s
central provinces in which a semi-circular area of the country's central plains receives less
precipitation than the surrounding areas just a short distance away. Rio Grande receives between
900 tol200 mm precipitation annually. Temperature ranges from 25 to 31oC.
Local Farming Systems
In Rio Grande, the dominant agricultural activities include growing rice (Oryza sativa) and
corn (Zea mays), and raising beef and dual purpose cattle. Most producers are semi-subsistence
in which they produce for household sustenance as well as market a portion of their products.
Although the community is relatively small, there are a wide range and diversity of producer
* day laborers who rent out their labor to farmers for a wage,
* day laborers who also cultivate small parcels for home consumption,
* smallholder farmers of crops who are almost entirely of a subsistence nature,
* smallholder farmers of crops and cattle who are almost entirely subsistence farmers,
* medium-scale farmers with crops for home and market,
* medium-scale farmers with crops for home and cattle for market,
* medium-scale farmers with crops for market and cattle for market, and
* large-scale farmers with rice and cattle for market.
Cattle include dairy, beef, and dual-purpose. Market crops include corn, rice, and some
seasonal peppers. The studies reported in this dissertation were undertaken within the context of
the local farming systems in Rio Grande. The five farms where the trials took place
encompassed a range of production types.
In the experiments presented in this dissertation, three species of woody perennials were
studied. These include Tectona grandis, Bombacopsis quinata, and Anacardium occidentale.
These species were selected by the farmers who were involved in the study. Of the three species,
Tectona grandis is the only non-native species and was chosen by the farmers on the basis of the
high price of its timber. Bombacopsis quinata was chosen based on the strong wood it produces
and its versatile utility on-farm. Anacardium occidentale was chosen for two of the products it
bares, its fruit and nut. The following information presented here provides a broad background
of the characteristics of these species. Because these species are studied closely throughout this
work, it is important to have a complete understanding of their defining characteristics. The
information available on each of the species is disparate. According to the available literature, it
is apparent that Tectona grandis has been studied and probed more extensively than either
Anacardium occidentale or Bombacopsis quinata, as such the length of each species review is
Origin, Natural Habitat, and Environment
Teak (Tectona grandis L.) is native to Southeast Asia and parts of the Indian sub-
continent. In the Philippines, it is also regarded as a naturalized species. Teak occurs naturally
as part of an assemblage of mixed forest species in its natural habitat. Although teak occurs
naturally in diverse ecological settings, moist deciduous forest is regarded as being its original,
native habitat (Kadambi, 1972) and develops best on fertile, well drained soils. In Thailand, teak
is found at altitudes between 100 and 1000 mas1 while in Indonesia, teak occurs in rainfall ranges
of 1500 to 2500 mm. However, rainfall for optimum growth is regarded to range from 1500 to
2000 mm yet trees will tolerate minimum precipitation of 500 mm with a maximum of 5000 mm
and temperatures between 2o and 48oC. Due to the species' plasticity in a range of conditions
and proven adaptability, it has been planted throughout tropical Africa, the Americas, and other
parts of Asia. Likewise, it is known to have been planted in plantations on the Indian
subcontinent and in Burma since the middle of the 19th century. Kadambi (1972) notes that
experimentation with teak planting began in Panama in the 1920s.
Teak gained its worldwide reputation initially as a prized wood due to its excellent
performance as a material for shipbuilding. Its hue, texture, and durability make it a desired
wood throughout the world (Bailey and Harj anto, 2005; Husen and Pal, 2006), for furniture
making, cabinetry, wharf construction, and for railcars. The qualities that make teak a
formidable wood species for these crafts include termite resistance, strength, appearance, water
resistance, and workability. Teak wood has been known to last intact for more than five
centuries (Kadambi, 1972).
Part of the Verbenaceae family, teak leaves are large, elliptical, and obovate with tapering
petioles. They produce abundant white flowers and the fruit takes the form of a hard berry-nut.
Seeds have four inner cells with an additional central cavity. Generally regarded as hardy, teak
trees are light-demanding, deciduous, and when mature become quite large, some known to
reach more than 40 m in height. Mature teak trees in favorable conditions can be generally
characterized by a tall, straight, cylindrical bole. The phenological cycle of the species consists
of the initiation of leaf senescence commensurate with the onset of the local dry season (in the
case of Panama this occurs in January). Leaves emerge in May while flowering initiates in
September in Panama. Numerous white flowers abound during the dry season in Panama as teak
trees defoliate entirely during this period.
Germination and Establishment
Seed germination is epigeous. One fruit can produce up to 4 seedlings resulting from the
multi-cavity fruit as mentioned above. Leaves are small during the initial growing season while
the seedling taproot can elongate up to 30 cm during this period. The taproot is known to reach
60 to 90 cm during the second and third growing seasons. Lack of light, drought, overhead drip,
excessive grazing, and resource competition from weeds are the known leading barriers to
germination and establishment of teak seedlings (Kadambi, 1972).
Adaptability and Performance
Abundant fruit production and a multi-cavity fruit enable teak to proliferate throughout the
landscape. Likewise, teak's well-documented plasticity and adaptability to diverse and, in some
cases, adverse conditions have also enabled its expansion throughout the tropics. In a study by
Piotto et al. (2003), teak was one of two exotic species compared with seven native species for
performance factors in Costa Rica. Teak was among the highest performing species in terms of
mean annual increment, a key growth marker. Both in height and DBH, teak was among the
highest producers. However, teak exhibited higher variability across plantations and
management strategies than its native counterparts. It also demonstrated a comparatively high
rate of bifurcation. The authors concluded that exotic species were promising; but, for optimal
timber production, they required more intensive management schemes compared to native
In a similar study, Piotto et al. (2004) compared the survival and growth of 13 native
species in mixed and single-species plantations with teak under dry forest conditions on the
Costa Rican Pacific coast. The native species were equally divided into slow-growth species and
fast-growth species. In the slow-growing category, teak rated second to Dalbergia retusa in a
single-species plantation with a survival rate of 90%. Similarly, compared to the species in the
fast-growing category, teak was second to Pseudosama~nea guachapele (92%) in a mixed species
plantation in terms of survival percentage. After 58 months of growth, teak surpassed all slow-
growth species in height and DBH. In comparison with the fast-growing species, teak was
second to S. pazrahyba in height and DBH. However, despite these promising characteristics
demonstrated in multiple research studies, Perez and Kanninen (2005) claim that in Costa Rica
and in several other Central American countries, teak plantations have not reached anticipated
levels of productivity.
Rooting and Competition
Teak in its juvenile stage exhibits aggressive rooting habits characterized by one or two
well-developed tap roots and extensive lateral roots located just below the soil surface. The
taproot is known to develop into a series of vertical roots. Root competition from neighboring
vegetation and other teak trees in plantation conditions markedly hampers teak growth
(Kadambi, 1972). Teak' s sensitivity to root competition presents considerable problems at the
plantation level as numerous population density studies have shown the superiority of planting
teak plantations sparsely. In a root distribution study, Divakara et al. (2001) tested interspecific
root competition between bamboo (Bambusa arundinacea) and teak by tracing 32P uptake. They
found that when 32P was applied at 25 cm depth, teak uptake of P increased exponentially as
lateral distance to bamboo increased. However, when 32P was applied at 50 cm depth, teak P
uptake declined as lateral distance to bamboo clumps increased. Although these two species are
well-known for being highly competitive belowground, this study may indicate teak' s ability to
specialize in upper soil horizon P uptake when faced with a fierce competitor such as bamboo.
Similarly, Shankar et al. (1998) note that in a 35-yr-old taungya field, the competitive presence
of introduced teak may have inhibited the invasion of the site by nonnative and weedy surface-
One way in which plantation owners have sought to ameliorate teak' s sensitivity to
surrounding vegetation is through burning. Teak is known to benefit from fire. Burning
provokes a rejuvenation of tree vigor, increased growth (height and diameter), and in plantation
situations a renewed uniformity within the plantation (Kadambi, 1972). Ultimately, teak' s fire
hardiness allows it to prevail over its neighbors for survival.
Potential benefits of teak plantations
There is much controversy over the introduction of exotic species into foreign landscapes
and the consequences for the environment and wildlife. Studies and cases of negative impacts of
the effects of exotic species abound. In Panama, there are numerous testimonies based on
empirical evidence to the negative effects of teak plantations there. Some of these impacts
include erosion on slopes due to the large, slow decomposing leaf litter left following the dry
season and teak' s ability to inhibit the growth of understory vegetation to a certain degree
particularly under a closed canopy. There are also claims in Panama that teak plantations do not
provide wildlife habitat. For example, in their work on comparisons of wildlife habitat in
Tanzania, Hinde et al. (2001) found teak plantations to be favorable for 'gleaner' wildlife
species. Also in Tanzania, Jenkins et al. (2003) found wildlife use of teak plantations to depend
on plantation age, distance to food sources, and animal type. Younger plantations maintained
wildlife communities similar to those of native opened woodland. However, the authors stress
the need for these plantations to have direct connectivity with natural areas for wildlife to
As teak plantations were shown to provide habitat for some large mammals, Saha (2001)
found no significant difference in plant diversity in a comparison study of vegetation
composition in a secondary forest (30 to 35 yr) and in a teak plantation (16 to 18 yr). Overall,
for the two land-use types, species richness was similar as were seedling density and the
abundance of animal dispersed species. However, Saha indicates that the plantations tested
possessed dissimilar composition and structure in comparison to the secondary forest.
An alternative use of teak plantations may be for carbon sequestration and storage. In
Panama, Kraenzel et al. (2003) found 20 yr teak plantations could sequester and store 85% as
much carbon as did local mature forest. Similarly, litterfall abundance in the teak plantations
was similar to that of local forest whereas litter quantity on nearby pasture was 25 to 30% less
than that of surrounding forest and the studied teak plantations.
Bonabacopsis quinata (syn. Pochota quinata, Bonabacopsis quinzatum)
Bombacopsis quinata Jacq. (bombacopsis) is a deciduous species native to the Americas
ranging from southern Honduras through Columbia and Venezuela. It is a large tree known to
reach 30 to 35 m in height and 1 to 2 m in diameter. Bombacopsis requires a defined dry season
and occurs in areas of annual precipitation ranging from 800 to 3000 mm (Cordero et al., 2003).
It grows from 0 to 900 mas1 and is more commonly found on flat land than on hillsides.
Bombacopsis prospers in well-drained, neutral or acidic soils and is characterized by a main stem
lined with large stems and a fluted base.
Leaves of bombacopsis trees are compound and usually possess 3 to 7 leaflets. Seeds are
wind-dispersed. Flowers are pinkish-white and the encapsulated fruits are 4-10 cm long. One of
the defining characteristics ofbombacopsis is its ability to thrive during extended dry seasons.
During 5 to 6 months of the year, bombacopsis is completely deciduous; this period usually
coincides with the local dry season. However, precipitation plays an important role in the
production capacity and specific gravity ofbombacopsis timber (Cordero and Kanninen, 2002).
Timber from this species is prized for its ability to maintain its shape and form during
moisture loss. The heartwood is reddish and the sapwood is white in color. It is generally used
for exterior and interior construction, furniture, and general carpentry. It is also a highly valued
reforestation species for its survival capacity, pest and disease resistance, and proven growth
rate. Bombacopsis has also become a desirable species due to its easy propagation using stumps,
bareroots, and by seeding.
In Venezuela and in Costa Rica in the past twenty years, bombacopsis has been planted
widely for timber production (Cordero and Kanninen, 2002). In Venezuela, bombacopsis is one
of the most important commercial forest species. In the moist deciduous forests of the western
plains, it is prominent in the standing stock volume and occurs naturally in prolific stands. In
this region, Kammesheidt (1998) found that bombacopsis recovered poorly after being logged
which lead to the near disappearance of the species in the studied forests even after more than 19
yr following the logging event. The author attributes this to the small bombacopsis seeds' need
for gap conditions and litter-free soil to germinate. Consequently, Kammesheidt suggests that
the often-prescribed timber harvest cycle of 30 yr will be inadequate for the regeneration of the
species. In fact, Cordero et al. (2003) recommend a rotation cycle of 50 yr for plantation-grown
bombacopsis (to maximize heartwood content).
Cashew (Anacardium occidentale L.), a member of the Anacardiaceae family, is a small to
medium-sized tree averaging a maximum of 20 m height and 1 m diameter. Cashew is known to
grow in regions generally from 0 to 1000 mas1 with mean annual rainfall between 600 to 1200
mm. Trees can withstand dry periods of up to 9 months and can tolerate infertile, shallow soils
Cashew leaves are oval, average 10 to 20 cm in length, and can measure up to 20 cm in
width. Young leaves are reddish or light green and mature into dark green. Flowers are
yellowish pink and usually emerge during the middle of the dry season on newly developed
shoots. Following pollination, nut growth is vigorous and reaches its maximum size 30 days
after initiation while the fruit pedunclee) develops at a slower rate. Fruit generally requires 70
days to reach maturity (Behrens, 1996).
Widespread cashew planting is prevalent in India, Brazil, Indonesia, and Tanzania.
Cashew is also frequently found on farms throughout Mesoamerica. In Tanzania, cashew trees
are prevalent on small farms whereas in India they are a popular species used in wasteland
reclamation. Maj or et al. (2005) found cashew to be among the most abundant food species in
eastern Amazonian homegardens. The prevalence of cashew can be attributed to its hardiness
under adverse environmental conditions.
Cashew' s hardiness has been shown to be a product of its ability to capture resources and
withstand drought. For example, in Ghana cashew tree planting and production is known today
to be expanding rapidly and concern exists over the potential instability that extensive cash crop
lands can cause in terms of the consumption of important water and nutrient resources which is
thought to be particularly acute in the case of cashew due to its drought hardy nature and its
frequent placement on barren lands in this case in forest-savanna transition zones. In response to
this concern, Oguntunde and van de Giesen (2005) investigated cashew water use. Their
research addressed the amplification of cashew plantations in West Africa. They found that
cashew responded sensitively to certain climatic conditions. For example, under conditions of
high radiation and high vapor pressure deficit, stomata were shown to close despite non-limiting
soil moisture availability. Therefore, when sensing environmental moisture deficiency cashew
restricted its water uptake instead of accessing soil moisture to counter the moisture deficit. The
authors concluded that cashew soil water uptake was directly related to climatic conditions rather
than soil moisture availability. We may deduce that this apparent mechanism of cashew's, to
conserve water reserves during periods of moisture deficit, may aid in cashew' s noted ability to
Studies have also been done in Brazil to investigate the physiological drivers behind
cashew' s ability to thrive in resource poor conditions. In a study that looked at various
physiological characteristics of cashew gas exchange, de Souza et al. (2005), like Oguntunde and
van de Giesen (2005), found cashew stomata behavior to be highly influenced by changes in
vapor pressure deficit. Prompt stomata closure in response to high vapor pressure deficit was
effective in restricting transpiration. The authors concluded that cashew' s ability to quickly and
effectively provoke stomata closure lends to cashew's ability to prosper on drier soils.
Cashew products and byproducts have a multiplicity of uses and values. From the world
market to rural homegardens (Isaac and Nair, 2005; Maj or, 2005), cashew is grown for the sale
of its kernel, for its fruits in industrially produced beverages, and for the nut shell liquid which is
used in a ranges of industries. The nut shell liquid is used abundantly and in a variety of
scenarios, including as a substitute for asbestos, in the car industry, as a wood sealant, germicide,
and others (Behrens, 1996).
In addition to providing multiple products for the global market, cashew has been shown to
provide services for biodiversity restoration as well. In a comparison of single and mixed-
species plantation types in Thailand, Kaewkrom et al. (2005) found that the combination of teak,
tamarind (Tama~rindus indica), and cashew was superior in providing habitat for establishment of
species from adj acent forests. They found that the diverse nature and multi-strata shading in the
tri-species canopy resulted in a reduction in weeds and pioneer species abundance giving way to
an acceleration of succession in the understory. The plantation, with the combination of teak,
tamarind, and cashew, housed the largest number of native forest tree species compared to other
plantation types. Kaewkrom et al. (2005) also found that the plantations with three species (as
opposed to the others having only two or single species) had scaled litter decomposition rates
providing a continuous release of nutrients to the soil nutrient pool. Finally, the authors noted
that the presence of cashew in the plantations may have played an important role in attracting
frugivores thereby potentially enhancing and diversifying the seed bank via the deposition of
other forest species seeds by these animals.
In the aforementioned study, Kaewkrom et al. (2005) allude to the role of leaf litter playing
an important role in nutrient storage and release. Building on this, Isaac and Nair (2005) carried
out one of the few studies that examined the dynamics of cashew leaf litter. They compared
cashew, mango (Mangifera indica), and j ackfruit (Artocarpus heterophyllus) leaf litters. Initial
characteristics of the cashew litter were different from the others. They found cashew litter to
have high nitrogen and cellulose concentrations coupled with intermediate quantities of phenols
and low amounts of lignin, relative to the other species. Likewise, of the three species, cashew
litter was the fastest to reach 95% decomposition (in 6 months). Soil under cashew litter also
held the largest quantities of actinomycetes, bacteria, and fungi relative to the other species in the
experiment. Nutrient release (N, P, K) from cashew litter was gradual throughout the 6 months
of its decay in which cashew litter released 97% of its N and K nutrients and 94% of P. With
these results, Isaac and Nair (2005) conclude that the cashew species can make an excellent
component in agroforestry systems due to its ability to provide a steady stream of soil nutrients
important to crops.
Researchers are also looking to cashew for use with livestock. In Brazil, Ferreira et al.
(2004) tested the use of cashew bagasse (fruit mass and fiber that remains following processing)
as an additive to grass silage for livestock feed. The study results showed that the cashew
bagasse had a positive effect on the nutritive composition of the silo and a positive effect on silo
conservation quality. In addition to the use of cashew for agricultural purposes, researchers in
Cuba are testing cashew for its ability to improve conditions of soils from abandoned mining
regions. In one study in Cuba, Izquierdo et al. (2005) tested cashew for its soil reclamation
capacity. They found cashew trees rapidly improved the targeted soil physical and biochemical
properties, including the improvement of soil electrical conductivity, total organic C
concentration, total N, and the reactivation of certain microbial processes in the mined soil.
However, while in the above study cashew played an important role in soil amelioration,
Ngatunga et al. (2003) found in Tanzania that cashew cultural practices acidified soil. According
to Ngatunga et al. (2003), due to the overwhelming infestation of powdery mildew disease in
cashew trees, Tanzanian farmers apply large quantities of sulfur to Eight this crop killing disease.
The abundance of deposited sulfur in the last decade has resulted in the acidiaication of soils in
Tanzania' s cashew producing region. This situation, lowering of pH of farm soils, can have dire
consequences as cashew is often intercropped with annual crops which, in the long run, will
unlikely be able to withstand the imminent acidiaication of these soils. Finally, one important
new use of cashew under investigation concerns its medicinal properties. In Brazil, Medonga et
al. (2005) studied a range of plant species for their ability to kill mosquito larvae. They found
that among a range of native species studied, cashew was the most effective at killing the larvae
of the dengue-spreading mosquito Aedes aegypti.
In Chapter 3, seedlings of the three species described above were planted in three different
planting configurations, which included plantings in lines, grouped on a diagonal, and along
fences. Investigation of different planting configurations was based on the premise that cattle
browse and treading of tree seedlings may occur differently depending on the organization of
seedlings in the pasture. Prior to the establishment of the experiment, participating farmers noted
that cattle tended to congregate along fences and may have an impact on planted tree seedlings.
On the other hand, farmers suggested that planting in lines would create alleyways for cattle to
move through. They also proposed that the diagonal configuration would create a greater
shading effect on the pasture that could benefit cattle during high temperatures. In addition,
Teklehaimanot et al. (2002) noted that trees planted in different configurations can impact tree
architecture and shading, and can create "micro-woodland" habitat for the benefit of wildlife.
Table 2-1 Results of effects of Ziziphus joazeiro and Prosopis juliflora trees on buffelgrass
pasture in Northeast Brazil.
Results by treespce
Test as compared to open pasture Ziziphus joazeiro Prosopis jirllotrar
Soil moisture No effect Less soil moisture than pasture (early season)
Maximum soil temperatures Lower No significant effect
Maximum air temperatures Lower Little effect
Loss of P from litter under crown Lower NA
Mineralized net N Greater Greater than pasture and Z. joazeiro
Crown radiation interception 65-70% 20-30%
Source: Menezes et al., 2002.
Figure 2-1 Topographic map of the Panamanian isthmus.
Source: NASA-SERVIR (Mesoamerican Regional Visualization and Monitoring
System), http://servir.nsstc.nasa.gov/, 2006.
figure 2-2 Panama forest cover and areas of detorestation In 194 I.
Source: ANAM, 1999.
1961 1986 1994 2003
+ Forest cover (ha) + Permanent pasture (ha) Total agricultural land (ha) Human Population (people)
Figure 2-3 Changes in land use and human population in Panama 1961-2003.
Source: Pagiola et al., 2004; FAOSTAT, 2006.
Total farm area (ha)
No. of farms
19.9 200 -
Farm size categories
Figure 2-4 Farm sizes and areas in Panama 2000.
Source: Censo, 2001.
Table 2-2 Total farm land, farms with cattle, and area under pasture in Panama, 20 30.
agricultural Total area agricultural
No. of total No. of % cattle area in pasture land in
Province farms (1000) cattle farms farms (10,000 ha) (10,000 ha) pasture (%)
Bocas del Toro 4.72 1,282 27 9.74 3.68 38
Chiriqui 48.50 7,305 15 42.79 24.60 57
Cocle 31.22 4,347 14 25.24 10. 15 40
Colon 10.95 2,136 20 16.99 7.63 45
Darien 5.31 1,543 29 23.23 7.00 30
Herrera 18.84 4,590 24 19.01 11.64 61
Los Santos 17.31 5,795 34 30.76 23.20 75
Panama 65.86 4,526 7 48.62 20. 17 41
Veraguas 33.72 7,615 23 60.16 30.17 50
Source: Censo, 2001.
Table 2-3 Economic importance of catt e in Panama by province, 2000.
Earnings from monthly Farmstead
Province cattle ($ IM) income ($) poulation
Bocas del Toro 2.6328.03,2
Cocle 6.44 220.60 113,764
Chirqul 6.02302.10 140,909
Darie 5.28116.50 21,016
Herrer 9.56249.80 55,743
Los Santos 26.4623.04,8
Verauas 6.85166.90 125,562
Source: Censo, 2001.
3 100 200 Kilometers
Figure 2-5 Proportion of pasture area to total land area by corregimiento in Panama, 2003.
Source: Dagang, 2004.
Proportion of Pasture areas to total area in each Corregimiento
Pasture area (%")
6~r2 .. I~ Corregirniento boundary
Figure 2-6 Research study site location, Rio Grande corregimiento, Cocle province, Republic of
S ource: www.Iib.utexas.edulmap s/cia00.html
40I 80 km
*, La Palma *
Research study site
COL OMBI A
TREE SEEDLING SURVIVAL AND IMPACT OF HERBIVORY ON SILVOPASTORAL
Finding a balance among food production, income generation, and environmental
preservation is a growing challenge. Likewise, an increasing world population requires greater
products and services from the land base. In light of these realities, it is vital that land use and
land management be carried out optimally and efficiently to maximize production of food,
income, and environmental integrity. The study presented in this chapter sought to test one
aspect, seedling survival and herbivory, of a land management strategy that intends to increase
the productive capacity of the land unit, diversify its products, and potentially increase the
environmental services it offers.
Considering that more than 20% of Panama is covered by pastures and most of these are
degraded and of low productivity, it seems both logical and necessary to focus on improving the
services pastures can provide. Being that cattle production in extensive pastures is the most
dominant land use system in the country, and considering the growing needs of the human
population coupled with the diminishing natural resource base, I focused on testing the
integration of fruit and hardwood trees into extensive, degraded pastures. When designing a
study to further develop an existing land use system, it is vital that the land strategies already
employed be included in the new design. For this reason, this study included the existing system
of cattle grazing in extensive, degraded pastures in its structure. Therefore, the experiment was
carried out in pastures that were actively grazed by cattle. The inclusion of cattle in
experimental pastures was made due to farmer interest, as farmers in Panama are generally not
willing to remove cattle from their pastures for the establishment of trees.
Tree Seedling Survival
Some researchers suggest a relationship exists between seedling survival and particular
seedling characteristics. Through their research of seedling survival under distinct
microenvironments with variation in competition, trenching, light, and soil nutrient availability
in the US Southeast, Beckage and Clark (2003) proposed that seed size may be an important
factor in seedling survival. In their experiment, small-seeded yellow poplar seedlings
(Liriodendron tulipifera) exhibited far greater growth than larger seeded species. Also, in a
study in Costa Rica examining the effects of light gradients on seedlings, Balderrama and
Chazdon (2005) relate the importance of size to seedling survival and growth, although in this
case seedling size, rather than seed size, was proposed to have had a positive impact upon
seedling survival. Balderrama and Chazdon (2005) also suggest that within the importance of
seedling size and more importantly seedling height, seedling architecture may play a role in
survival within light-compromised environments. However, Benitez-Malvido et al. (2005) found
in the Central Amazon that seedlings of Pouteria caimito demonstrated an inverse relationship
between survivorship and initial seedling height. Also, seedlings of Chrysophyllum pomiferum
demonstrated a negative relationship between seedling size and height relative growth rate.
Factors contributing to survival and growth of seedlings can be difficult to generalize and
seedling responses in terms of survival and growth can be species specific (Benitez-Malvido et
al., 2005). Beckage and Clark (2003) found species performed distinctly under different
resource situations. In the study, Liriodendron tulipifera flourished in high resource
environments but did not do well in competitive environments. Quercus rubra responded little
to competitive environments and responded similarly across treatments. However, Balderrama
and Chazdon (2005) found that responses from different tropical species varied more in survival
than in growth across different light availability treatments. Hyeronima alchorneoides and
Virola koschnyi survived under low light situations; however, they did not respond as well as
other species in terms of growth in high light conditions. The often studied Dipteryxpna~naensis
and Vochysia guatemalensis did not exhibit this tradeoff in that they had high survival rates
under low light conditions coupled with high growth rates in high light conditions.
Griscom et al. (2005) also found different species to respond distinctly in the field. When
comparing Cedrela odorata, Enterolobium cyclocarpum, and Copaifera aromatica, herbicide
application had a greater, significantly positive effect on survival of C. odorata seedlings than on
other species in the study. Ramirez-Marcial (2003) also assessed survival of different species in
anthropogenic environments and found that species growth and response to grazing differed.
She found relative height and diameter growth rates ofLiquidamnbar styraciflora seedlings were
significantly associated with cattle grazing while growth rates of Cornus disciflora and
Oreopanax xalapensis were not.
Another factor that can have differential effects on seedling species is habitat. In fact,
Benitez-Malvido et al. (2005) found that the pasture conditions (temperature, humidity, and soil
moisture) in their study, unique to the native forest habitat of the seedling species that were
studied, may have impeded acclimation of certain species, specifically Chrysophyllum
pomiferum and M~icropholis venulosa, to the area. The authors contend that the dramatically
different habitat conditions in which the seedlings attempted to establish brought about higher
rates of seedling mortality for certain species while other species such as Pouteria caimito
thrived in pasture conditions but not in forest.
Another relevant factor when considering seedling survival is the effect of existing
vegetation on seedling establishment. In a Hawaiian forest, Denslow et al. (2006) found that
existing vegetation severely constrained woody seedling establishment. Presence of grasses
impeded growth of the species Acacia koa, Sophora chrysophilla, and Dodonea viscosa.
Sanchez and Peco (2004) also suggest that presence of grasses during seedling establishment of
Lavandula stoechas in Spain negatively impacted seedling growth. They also concluded that
grass roots form a belowground layer that functions as a barrier to seedling roots and prevents
their penetration into deeper soil layers.
More specifically, Posada et al. (2000) found that different grass types impacted
establishing seedlings differently in an abandoned pasture in Colombia. Molassesgrass (M~elinis
minutiflora) permitted significantly greater colonization and growth of woody individuals than
kikuyugrass (Pennisetum clan2destinum). The authors suggest that the stoloniferous growth habit
of P. clan2destinum created a physical barrier that inhibited seed germination and seedling
establishment of woody perennials. Similarly, seedlings that achieved germination within the
stolon mat suffered due to low light and mechanical damage by fast growing P. clan2destinum
grass shoots. In contrast, the bunch grasss M. minutiflora allocated less biomass to stolons and
had more open surface area between plants which they found to be more conducive to woody
perennial seedling establishment.
Effects of Cattle Grazing
Effects of cattle grazing such as browsing and treading can have negative impacts on
seedling survival. Stammel et al. (2006) studied the emergence and establishment of six tree
species under different land management strategies including grazing, and they found that
treading effects from cattle tended to have a negative impact on seedling emergence. Moreover,
treading caused vegetation removal, soil disturbance, puddling, and desiccation. Likewise,
seedlings in a study carried out in a Panamanian pasture by Griscom et al. (2005) encountered
negative effects of cattle on seedlings, in which cattle impacted seedling growth and survival by
trampling and browsing seedlings. They found that negative effects from cattle grazing could be
species specific. In their study, exclusion of cattle from seedlings had a greater, significantly
positive effect on Enterolobium cyclocarpum when compared with other species. Also for
Cedrela odorata seedlings, presence of cattle significantly reduced dry mass across the species.
Overall, presence of cattle and absence of herbicides caused the greatest mortality among all
seedling treatment combinations in the study. Evans et al. (2004) also found species-specific
effects of cattle on seedlings in which cattle avoided grazing Salix spp. and only when other
forage was scarce would cattle browse this species. Ganskopp and Bohnert (2006) also suggest
that cattle will select for high quality forage and that cattle in their study traveled longer
distances to access higher quality forage. They make the point that cattle return year after year to
the same grazing areas presenting a problem for range managers and causing large areas of
pasture to not be used. However, the non-use of some pasture areas by cattle may provide a
window of opportunity for the establishment of woody perennials.
Although the research discussed above indicates potential negative effects of cattle grazing
on woody perennial establishment, some studies suggest that the presence of cattle can in fact
benefit seedling survival. Posada et al. (2000) propose the notion that grazing can serve as a tool
for the regeneration of forests on abandoned pastures. They suggest that cattle browse can
reduce aggressive grass species in pastures. In addition, they put forth the notion that initial
colonization of tropical grasslands is dominated by wind-dispersed species consisting of woody
shrubs or small trees that frequently occur in disturbed areas. Establishment of such species,
they note, can lead to the shading out of grasses and the creation of suitable microclimates for
forest species establishment. Other studies conclude similarly. For example, in a study in
Argentina de Villalobos et al. (2005) found that grazing may benefit woody perennial seedling
survival. They contend that grazing caused a reduction in grass biomass above- and
belowground, potentially increasing surface soil moisture and thereby enhancing woody seedling
establishment and growth. In contrast with Stammel et al. (2006), de Villalobos et al. (2005)
contend the creation of gaps by cattle treading may induce periodic woody perennial seedling
establishment. Despite finding negative impacts on seedlings from cattle, Griscom et al. (2005)
also suggest that seedling survival and growth may benefit from cattle through the removal of
competing biomass, which has the potential to increase seedling access to light, water, and
Leaf-cutter ants (Atta spp.) are an abundant invertebrate species in tropical ecosystems
(Jaffe and Vilela, 1989) and they function as important selective herbivores throughout the
Neotropics (Rao et al., 2001). These herbivores can have a tremendous impact on the landscape.
Leaf-cutter ant herbivory can reduce plant reproductive potential through decreased seed
production and result in reduced seedling survivorship (Vasconcelos and Cherrett, 1997). In
addition, leaf-cutter ants prefer young leaves over mature leaves thereby hindering regeneration.
Leaf-cutter ants manifest preference for particular species. Rao et al. (2001) found decreased
density of adult trees of preferred species in ant-foraging zones in comparison with ant-free
areas. They suggest that repeated exposure to ant defoliation may induce mortality and trigger a
reduction of species diversity.
Similarly, anthropogenic intervention into natural tropical landscapes has been shown to
increase the density of leaf-cutter ant nests (Jaffe and Vilela, 1989). Impact by Atta spp. has
been observed to heighten within human-influenced natural systems. Jaffe and Vilela (1989)
suggest two reasons for the increase in Atta populations in human-intervened natural systems.
First, they propose that due to species diversity, abundance of palatable vegetation free of
defense mechanisms is low and may be highly dispersed in forests in comparison to human-
affected environments. They argue that the diversity of forest vegetation makes ants susceptible
to poisonous plants and consequently may subdue the Atta population. Secondly, the authors
contend that Atta nests require exposure to sunlight. This requirement is often a rarity on the
tropical forest floor. However, because human interference is often coupled with the removal of
tree cover and a consequent increase in sunlight, these conditions may be advantageous for
increases in nest density. Therefore, they propose that proliferation of human-affected
landscapes decreases non-desirable plant abundance and concomitantly increases leaf-cutter ant
nest density. For example, Terborgh et al. (2006) also examined leaf-cutter and plant presence in
a comparison of Atta populations on different sized islands and mainland Venezuela. They
found that leaf-cutter ants browsed less selectively at high population densities, and were able to
generate wide impacts on plant communities. In addition, Atta population density was greater on
smaller islands resulting in a greater impact on the landscapes of the islands. Rao (2000)
attributed this occurrence in part due to an absence of Atta predators on small islands, which
were too small to maintain populations of predators such as armadillos (Dasypus novemcinctus).
As noted above, the effects of herbivory on a landscape can be cross-cutting and intense.
Detrimental consequences due to herbivory can occur for different plant species as well as for
cohorts of different age classes (Terborgh et al., 2006). However, species responses to herbivory
can vary (Midoko-Iponga et al., 2005). Variables such as habitat, seedling height, herbivory
intensity, pathogens, competition, and seedling non-structural carbohydrate reserves can
influence seedling response to herbivory (Benitez-Malvido et al., 2005). For example, according
to Allcock and Hik (2004), habitat played a pivotal role in the response of seedlings to
mammalian herbivory in an Australian grassland. In their study, seedlings exposed to herbivores
in grassland were similar in size to seedlings grown in herbivore exclosures in woodlands after
three years of observation. The authors deduced that rapid seedling growth in the grassland
habitat counterbalanced the negative impacts of herbivory. Seedlings were able to recover from
herbivory more quickly due to the potentially higher resource habitat in the grassland, especially
regarding light availability. On the other hand, the slower growth rates and recovery time of
seedlings in the woodland habitat placed seedlings at greater risk to repeated herbivory and
mortality. As it took longer for seedlings to grow their apical meristems beyond the reach of
herbivores, their risk to herbivory was observed to be greater and prolonged.
Vasconcelos and Cherrett (1997) also found in their research that taller seedlings
experienced less mortality than others. To compound the risk of repeated herbivory and eventual
mortality, Haukioja and Koricheva (2000) note that the breaking of apical dominance due to
herbivory can result in vigorous vegetative growth leading to higher susceptibility of plants to
herbivores. Such induced susceptibility (young leaf growth coupled with shorter seedling
stature) can cause seedlings to be more attractive to herbivores.
Hester et al. (2004) also found seedling height to play an important role in response to
herbivory, particularly in the case of Pinus sylvestris in a simulated browse greenhouse
experiment. They found that slow height growth of browsed P. sylvestris seedlings caused them
to remain in a size range susceptible to herbivores in comparison to non-browsed seedlings.
However, they concluded slow growth response of P. sylvestris seedlings, including fewer
shoots, may make seedlings less desirable to herbivores. Hester et al. (2004) found that Betula
pend'ula and Sorbis aucuparia seedlings responded better to simulated browsing than P.
sylvestris with increased biomass above- and belowground. Likewise, Allcock and Hik (2004)
found that Eucalyptus albens seedlings experienced greater survival than that of Callitris
glaucophylla due to Eucalyptus' ability to rebound from herbivory through hastened re-
sprouting. The authors suggested a decline in the C. glaucophylla population would occur if
sustained grazing were to occur in the study site.
Herbivory intensity and energy reserves may also play an important role in seedling
response to herbivory. In an experiment using seedlings species (Acer rubrum, Acer saccharum,
Quercus rubra, and Prunus serotina) from the US Northeast, Canham et al. (1999) examined the
effects of different degrees of manual defoliation on the survival and biomass allocation of
seedlings. They found that in response to complete leaf removal, survival declined sharply.
They suggested survival was closely tied to total carbohydrate reserves and concentrations of
carbohydrate reserves. Although effects of defoliation on carbohydrate reserves were consistent
across species, consequences for survival differed by species. Rao et al. (2001) concurred in
their conclusions that if seedlings are able to persist through the sapling stage, their survival may
likely be due to the accumulation of energy reserves which may better equip them to survive and
recover from a defoliation event. Similarly, Haukioja and Koricheva (2000) in their comparison
of woody perennials and herbs concurred that plant regrowth following herbivory is dependent
on energy and nutrient storage; however, they emphasize that such storage must occur in
unthreatened plant organs when herbivory is a factor. Being that mature woody perennials store
a small proportion of their biomass in leaves (in comparison with herbs), Haukioj a and
Koricheva (2000) concluded that woody plants were better suited than herbs to withstand
Just as response to herbivory by seedlings can be species-specific, so may herbivores
maintain preferences for particular species (as noted to be the case with Atta spp.). Hester et al.
(2004) contend that herbivore choice can be affected by a multitude of factors, including
individual location, plant morphology, plant chemical composition, and neighboring species.
The authors also distinguish preferences among different herbivores. They note that
morphological differences among saplings are more important to mammalian herbivores than
plant chemistry. However, they propose that secondary chemical composition and morphology
may interact to influence herbivore choice.
Tree seedling survival, herbivory, and recovery from herbivory are intricate processes
which, according to the research, seem to be impacted by a range of tree species, herbivore, and
habitat characteristics. Species characteristics such as seed size, seedling height, and architecture
seem to play important roles in a species' ability to survive. These characteristics coupled with
variations in habitat including light availability, moisture, and existing vegetation can result in
differences in seedling survival. Similarly, seedling herbivory can also have important impacts
on survival. Herbivore preferences can have particularly negative impacts on seedling survival
and ability to persist. Also, seedling response to herbivory can be sensitive to species-specific
characteristics such as seedling architecture and biomass allocation particularly in the case of
storage of non-structural carbohydrates, as well as habitat conditions and herbivory intensity.
Considering that research suggests seedling survival, herbivory, and response to herbivory can be
species-specific and taking into account that seedling survival is vital to the establishment of a
silvopastoral system (the larger focus of this study), the following research was undertaken to
investigate the survival and herbivory of three important tree species used in agricultural systems
Objectives and Hypothesis
The obj ective of this study was to assess the potential for the integration of Anacardium
occidentale, Bombacopsis quinata, and Tectona grandis seedlings into actively grazed pastures.
I hypothesized that cattle herbivory (the grazing or browsing of seedlings by cattle) and treading
would play an important role in seedling survival and that seedling species would be a
determining factor for survival and herbivory.
Methods and Materials
This study was conducted on five farms in the Rio Grande corregimiento of Cocle
province, Republic of Panama (see Chapter 2 for specific local and regional characteristics).
Each on-farm study site consisted of a 2 ha pasture dominated by the naturalized grass
Hyparrhenia rufa. Pasture stocking rate averaged approximately 0.5 to 1.0 animal unit per ha
(one animal unit = ~ 270 kg).
A randomized complete block design was used. There were five blocks; one block on each
farm. Each block contained a complete set of treatment combinations which comprised a total of
135 seedlings. There were nine treatment combinations consisting of three species and three
planting configurations. The species were Anacardium occidentale, Bombacopsis quinata, and
Tectona grandis. The planting configurations included seedlings planted in pastures on a
diagonal, in lines, and along fences (APPENDIX A). There were fifteen seedlings planted for
each treatment combination. Each experimental unit was planted in random locations throughout
The three tree species were chosen by farmers participating in the study. The seedlings
were acquired through local nurseries. A. occidentale andB. quinata seedlings were
approximately 180 days in age and measured approximately 30 cm height at the time of planting.
In accordance with local and regional planting technique, T. grandis was planted using bareroot
stalks approximately 180 to 220 days in age.
On each farm, a circular area of 1 m diameter was cleared of vegetation manually for each
seedling. Holes were dug 30 cm deep and 30 cm in diameter. Seedling nursery bags were
removed and seedlings were placed in holes as they were backfilled with the excavated soil.
Within each experimental unit, seedlings were planted 3 m apart.
Seedlings were surveyed weekly for two years. They were observed for mortality and
herbivory. We recorded mortality, potential cause of mortality, sign of herbivory, and source of
herbivory. Seedlings were considered dead when their stems had dried and/or when their stems
and leaves had disappeared. Cause of mortality was categorized into cattle, leaf-cutter ant (Atta
spp.), natural, and other. Cattle and leaf-cutter ant effects were distinguished visually.
Herbivory was determined when a portion of a seedling had been removed. Source of herbivory
was also categorized into cattle, leaf-cutter ant, and other and were also distinguished visually.
Statistical analyses were performed using SPSS. A survival analysis was conducted to
analyze the seedling mortality and cause of mortality data. The Kaplan-Meier survival
probability via the Log Rank test was used to compare the survival curves and source of
mortality curves for species and planting configuration. SAS JMP was used to analyze the
interaction factors of species and configuration through Cox regression analysis. SAS was used
to analyze herbivory data. A two-way analysis of variance was conducted. Tukey's Honestly
Significant Difference test was used to determine mean separations at the .05 significance level.
A chi-square analysis and Goodman and Kruskal Tau test were used to analyze source of
During the two years of the experiment, the survival analysis revealed 250 of a total of 675
planted seedlings survived, a survival rate of 37%. Survivorship was significantly affected by
the planting configuration, species, and planting configuration x species interaction treatments.
The Log Rank test revealed significant differences in the survival curves across configuration (p
< 0.001), species (p < 0.001), and planting configuration x species interaction (p < 0.001) (Table
The survival analysis for species reveals some insight into species performance. For
example, much of the total mortality (70%) across species that occurred over the life of the
experiment occurred by day 300 (73% ofA. occidentale, 65% ofB. quinata, and 73% of T.
grandis) (Figure 3-1). Likewise, the three species experienced mortality in a similar pattern, in
two large events during the first third of the experiment and in smaller increments toward the end
of the experiment (Figure 3-2). Also, across species, of those seedlings that died, 27% were A.
occidentale, 35% were 7: grandis, and 38% were B. quinata. Within species, mortality rates
were 51% for A. occidentale seedlings, 67% for T. grandis, and 71% for B. quinata, amounting
to seedling survival rates of 49%, 33%, and 29% for A. occidentale, T. grandis, and B. quinata,
When examining the results of the interactions between species and planting configuration,
the survival analysis reveals that in the diagonal configuration, species performed significantly
different (p < 0.001). There were a total of 127 seedling deaths in the diagonal configuration
which included 19 mortality cases for cashew, 64 for tropical cedar, and 44 for teak. Mean
survival time for seedlings planted in the diagonal configuration was 500 days. Similarly, 170
seedling deaths occurred in the fence configuration consisting of 57 mortality cases for cashew,
62 cases for tropical cedar, and 51 cases for teak. Mean survival time for the fence configuration
was 451 days. However, the Log Rank test revealed that within the fence configuration there
was not a significant effect on survival for species (p = 0.069). Within the line configuration,
there were a total of 128 seedling deaths made up of 34 cases for cashew, 39 cases of tropical
cedar, and 55 cases for teak. The mean survival time for seedlings planted in the line
configuration was 572 days. The line configuration had a significant effect on survival (p =
0.003). The different patterns in which seedling species mortality and risk to mortality occurred
over time are illustrated in the survival curves in Figure 3-1.
Observed Causes of Mortality
Browsing and treading by cattle were the dominant observed causes of seedling mortality.
Of the total 425 seedlings that died, 345 (81.1% of the total) died due to effects from cattle.
Other observed causes of mortality included effects from leaf-cutter ants, natural causes, and
from machinery. Using the Log Rank test there were significant differences in the survival
curves across the 'causes of mortality' factor, p = 0.005. Survival curves reveal that the mortality
cases that occurred due to "natural causes" occurred sooner after planting than the other
mortality cases, and 46.5% of the cases that occurred due to cattle effects expired during the
period of 210 287 days.
The effects of species and planting configuration on herbivory were tested. Of the species,
overall B. quinata was browsed most frequently while A. occidentale was browsed least
frequently. A significant main effect was captured for species, p < 0.0001. A significant two-
way interaction was obtained when examining the configuration x species interaction, p <
0.0001. However, contrary to survival, the main effect for configuration was not significant.
Using the Tukey hsd test, significant differences in herbivory were observed between B.
quinata and A. occidentale. In the diagonal configuration, B. quinata experienced significantly
greater herbivory than did A. occidentale. In the fence configuration, 7: grandis was browsed
significantly more than the other two species. Finally, in the line configuration, B. quinata
experienced significantly more herbivory than the other two species (APPENDIX B).
Sources of Herbivory
Three categories of sources of herbivory were recorded, including cattle, leaf-cutter ants,
and other. Overall 68.1% of the herbivory cases occurred due to cattle, 30.5% due to leaf-cutter
ants, and 1.5% due to other causes. Among the species, B. quinata had the largest proportion of
cases of herbivory due to cattle grazing and due to leaf-cutter ants with a total of 57.2% and
56.4%, respectively (Figure 3-3). A. occidentale had the largest number of cases for the third
category of "other" sources of herbivory. In addition, when using the chi-square test, there was a
significant effect for species on source of herbivory, p < 0.05. Also, the Goodman and Kruskal
Tau test was significant for the species effect on source of herbivory, z = .009, p < 0.05.
The effect of configuration on source of herbivory was significant at p < 0.05. In addition,
the Goodman and Kruskal Tau test was significant for configuration at z = .01, p < 0.05.
Relative to source of herbivory as an outcome, line had the highest proportion of cases for cattle
(37.4%) and leaf-cutter ants (39.5%), whereas diagonal and fence were highest for 'other'
(37.0%) (Figure 3-4).
The overall seedling survival rate of 37% can be regarded as an adequate yield for a field
planting trial considering the continuous grazing of cattle and the long-term nature of the study.
Mortality occurred at distinct times over the life of the study. High seedling mortality took place
relatively early (1-60 day) while moderately high mortality occurred toward the end of the
experiment (Figure 3.2). This pattern occurred similarly across species. The period right after
transplanting is expected to be a bottleneck for survival due to difficulty of establishment into
existing vegetation (Sanchez and Peco, 2004). The second mortality period occurred between
day 200 and day 320. This period coincided precisely with the local dry season when rainfall
can drop below 13 mm per month (Murphy and Lugo, 1995). Consequently, it is likely that
moisture scarcity played an important role in the persistence of seedlings and their ability to
establish early on. Overall, median seedling mortality occurred at day 286 (in the third month of
the 5-6 month dry season). It is important to note that during the dry season, seedlings
experienced increased threat to survival as during this period moisture stress typically can lead to
seedling mortality; concomitantly, forage scarcity is typical of the dry season period, which can
lead to increased grazing of seedlings by cattle. Thus, during the dry season seedlings were
likely subj ect to the typical moisture deficits of this period that are reportedly experienced in
natural settings, in addition to the added burden of likely forage-deprived cattle. However, it is
relevant to note that these conditions were not directly measured in this study.
The species treatment was significantly different across the seedling mortality survival
curves and, overall A. occidentale experienced the greatest survivorship among the species
followed by T. grandis and B. quinata, respectively. A. occidentale's perseverance in the
pastures is reflective of its local abundance. Its ability to withstand prolonged drought
conditions may have aided its survival. Also, its ability to persist and eventually thrive in the
seedling stage was seen in the experiment discussed in Chapter 5. In that study, A. occidentale
seedlings did not experience notable growth in the first year of the experiment but flourished
during the second year. Similarly, T. grandis also suffered less mortality than B. quinata. T.
grandis leaves are less brittle but seemingly equally unpalatable as A. occidentale leaves. These
characteristics may have provided T. grandis with an added benefit for survival.
Spatial placement of the planted seedlings may have been key to their survival in terms of
planting configuration. This was reflected in the significant effect planting configuration had on
survival. It is likely that seedlings were subjected to strong neighboring competition by already
existing vegetation in the pasture both above- and belowground. Although seedlings were
spaced at equal distances throughout the configuration treatments (3 m x 3 m), seedlings in the
fence treatment suffered most such that there were no significant differences among species
planted along fences. It is likely that seedlings in the fence treatment were subj ect to more
frequent cattle presence and treading due to the more abundant shade (where cattle tend to
congregate) that occurred along fences in comparison to open pasture. Also, competition may
have been more intense along fences than in open pasture (in lines and diagonals) as most fences
comprised mature, live tree posts and trees with established roots systems and canopies which
likely had an advantage over seedlings in acquiring resources, particularly during the dry season.
However, it is important to note that competition between large trees and seedlings was not
directly measured in this study.
Despite lower total survival in the fence treatment, seedlings planted along fences may
have benefited from periodic weeding of fences, which entails the cutting away of all vegetation
surrounding live and dead fence posts just prior to and during the dry season. This practice is
carried out to avoid the spreading of local human-induced fires into pastures. The elimination of
competing grasses and forage vegetation along fences in itself may have provided an advantage
to seedlings already negatively affected by on-going cattle presence, shade, and competitive
effects of nearby large trees. Likewise, the removal of competing vegetation during a critical
period such as the dry season when available soil moisture is reduced may have had an even
more dramatic, important effect on seedling survival in the fence treatment.
Line and diagonal treatments may have been subj ect to competition as well due to their
having a greater abundance of surrounding vegetation as well as having the presence of
neighboring seedlings surrounding them in comparison to the fence treatment. However, their
greater survival indicates that these configurations provided an advantage for seedling survival.
The design of each of these configurations formed alleyways between seedling rows which may
have facilitated cattle movement through the configurations and potentially reduced cattle
treading and consequent seedling damage. In addition, other studies have proposed that planting
seedlings in small groups can reduce cattle damage due to a clustered, island effect that is formed
when seedlings are grouped together; creating conditions where cattle may be less apt to graze in
contrast to the fence treatment which consisted of one long, accessible row of seedlings.
Observed Causes of Seedling Mortality
According to the data, cattle treading and grazing were the primary observed cause of
seedling mortality in this experiment. Taken as a whole, 81.1% of seedling mortality was caused
by cattle. As reflected in the survival curve, seedling mortality due to "other" circumstances
occurred largely during the same brief periods, i.e. the maj ority of these cases occurred at three
particular times. Being that the "other" category included causes of mortality such as those due
to accidental cutting by a machete during weeding and being run over by a machine, it seems
presumable that the "other" mortality cases would have occurred more or less during the same
time period as weeding and presence of machines took place only at specific moments.
Almost half of the seedlings that died due to effects from cattle died between day 210 and
day 287 after planting during the first four months of the lengthy dry season. There may have
been two different dynamics behind the seedling mortality during this period. First, available
forage for cattle is scarce during the dry season particularly late in the season when drought is
often prolonged. For sustenance, cattle are known to browse any type of living plant during this
period; even those plants that are not customarily browsed during the wet season will be
consumed when scarcity occurs. Therefore, it seems logical that cattle would act most
vigorously upon seedlings precisely at a time when customary forage is unobtainable. However,
it is likely that an additional factor influenced seedling mortality during this period. That is,
during the dry season period, seedlings were weakened due to moisture scarcity. Effects of cattle
such as browsing and treading (which seedlings would normally be able to effectively rebound
from in the wet season) may have been too severe in the dry season, and, therefore, led to
mortality. This situation is further intensified as seedlings may not have developed an adequate
root structure to capture dwindling soil moisture particularly while competing with long-
established pasture grasses. Therefore, given the presumably weakened status of seedlings
during the dry season coupled with often amplified cattle effects such as grazing and treading, it
is not unexpected that mortality would heighten particularly due to cattle during this period.
In contrast to seedling survival, seedling herbivory was significantly affected by species
but was not significantly affected by planting configuration of seedlings. Additionally, the
interaction effect of species and configuration was significant as has been noted in other
agroforestry communities (Teklehaimanot et al., 2002). It is interesting to note that species
played a significant role in herbivory. This result may provide some insight into the importance
of tree species as a driver or determining factor of herbivory in grazed pastures and,
consequently, establishment of silvopastoral systems in grazed pastures. At the same time,
considering insight gained from the results and in terms of drivers, it could then perhaps be
conceived that species (as well as other factors) is a more relevant driver of seedling herbivory
than is configuration. These broad, potential insights will bear upon the ultimate purpose of this
research, i.e., to aid farmers in decisionmaking regarding the establishment of dispersed trees in
pasture and the creation of appropriate silvopastoral systems.
Across species, B. quinata was indeed browsed most among all of the species. This is not
surprising given that B. quinata seedlings possess succulent green leaves. However, it is
noteworthy that herbivory of B. quinata occurred most given that the seedlings in the study
experienced leaf senescence and, generally, this species is known to defoliate completely during
seasonally dry periods. Hence, although it seems appropriate that B. quinata leaves were
browsed more often than others given their better palatability, their leaves were not present
during at least half of the experimental period. This leads us to believe that B. quinata leaves
were, in fact, likely browsed quite intensely while they were present.
B. quinata was more heavily impacted by herbivory than A. occidentale. In contrast to B.
quinata, A. occidentale's fibrous, brittle leaves were less appetizing to the observed herbivores.
This condition was likely a deterrent to the browsing of A. occidentale and may have enhanced
the herbivory of B. quinata. As noted above, this situation may have served as an added
advantage for the survival ofA. occidentale. Furthermore, in the case of T. grandis, the texture
and herbivores' lack of preference for T. grandis leaves were similar to those of A. occidentale
which may have lead to those seedlings being browsed less than B. quinata.
When examining the results of the post hoc test of the interaction of seedling species and
planting configuration on herbivory, significant results varied. A. occidentale was shown to be
the least desirable to herbivores overall, across configuration treatment interactions. This result
was to be expected given the significant main effect of A. occidentale. However, the surprising
result was that 7: grandis herbivory was significantly greater in the fence configuration than the
other species. Although leaf growth data were not recorded, it is possible that T. grandis
benefited from the shade from the live fence in the fence configuration which may have provided
an increase in soil moisture along the fence treatment area. Given T. grandis' documented
aggressive character and ability to readily dominate available resources in comparison to other
species, it is possible that T. grandis was able to capture shade-induced moisture increases better
than the other species on the fence and consequently increase its leaf growth. Increased leaf
growth could have then lead to increased herbivory due to greater leaf presence in comparison to
the other species particularly during periods of moisture scarcity. However, it is important to
note that soil moisture and leaf growth parameters were not measured in this study.
Sources of Herbivory
Similar to the survival study, it was evident that cattle were the observed herbivore that
grazed seedlings the most. Cattle are known to graze palatable woody perennials when given the
opportunity in both pasture and forest environments (Ramirez-Marcial, 2003). In the case of
pasture, cattle herbivory can lead to the local elimination of certain woody perennials in
pasturelands. However, prior to the initiation of the experiment, there was the expectation that
leaf-cutter ants (Atta spp.) would play a more dominant role in the herbivory of seedlings, given
the abundance of these in the study site and past farmer experience, particularly in the case of B.
quinata. It is not surprising though that cattle and leaf-cutter ants browsed B. quinata seedlings
most often, for the same reasons mentioned above palatability and texture. In contrast, the
undesirability ofA. occidentale by the leading herbivores (cattle and leaf-cutter ants) led it to be
the most browsed by "other" sources. Hence, the results which clearly show significant
differences among sources of herbivory demonstrate that species was a main factor that shaped
the way in which source of herbivory occurred. Like the survival data, cattle were the greatest
overall browsers of seedlings, particularly in the line configuration. It is not understood why
planting configuration may have had a significant effect on herbivory. In fact, prior to the
installation of the experiment, it was assumed that fence would have the greatest amount of
herbivory being that shade abounds along fences and it is in this area where cattle tend to
Tree-seedling survival is shown to be highly responsive to changes in season, herbivore
(cattle) presence, tree species characteristics, configuration, and possibly proximity to large trees
(in the case of the fence configuration). Each of these factors played a determining role in the
survival and mortality of the seedlings studied in this experiment. The greatest amount of
mortality occurred during the first dry season, indicating that if producers can find the means to
support the survival of planted seedlings through this period, the total proportion of surviving
seedlings could be greater in the long-term. Cattle were the overwhelming predators of seedlings
and, if seedling survival is a farmer priority, then cattle should be removed during seedling
establishment. However, if cattle are the farmer priority, then seedlings can be grazed and will
rebound with a satisfactory survival percentage (37%). As will be manifested in the subsequent
chapters, it was found here that tree species is key to seedling survival and herbivory. In all four
analyses, species had a significant effect on the outcomes. As noted, characteristics such as
aggressive growth type, leaf palatability, shade tolerance, and regrowth ability are a few of the
considerations that should be made when selecting appropriate tree species for grazed
silvopastoral system establishment. Configuration also played an important role, particularly in
terms of seedling mortality where in the fence treatment the most mortality occurred and
seedling lifespan was shortest; however, the fence configuration experienced the least amount of
The varied results of this experiment are indicators of the delicate balance that occurs in
natural systems. Although human-induced systems are often characterized as being less
biologically diverse and complex than naturally occurring systems, it has become evident
through this study that the integration of silvopasture into pasture systems is in fact complex.
The complexity lies in the many factors the system comprises: trees, grasses, and livestock;
however, complexity is heightened by competition among the system components, presence of
other herbivores, and local conditions. These must also be combined with farmer preferences
and land management goals. Given these considerations coupled with the present need to
augment the production capability and environmental integrity of agricultural systems, it is
important that research on silvopastoral systems be intensified.
Table 3-1 Comparison of effects of planting configuration and species on survival of 675
seedlings planted in five blocks in degraded pastures on-farm over two years in
Source Nparm IDF L-R Chi Square Prob > Chi Square
Species 2 2 40.13 0.00
Configuration 2 2 19.99 0.00
Species x Configuration 4 4 60.91 0.00
Block 4 4 63.35 0.00
0 100 200 300 400 500 600 700 800 900
0 100 200 300 400 500 600 700 800 900
0 100 200 300 400 500 600 700 800 900
- Anacardium occidental (cashew)
- Bombacopsis quinata (tropical cedar)
- Tectona grandis (teak)
Figure 3-1 Comparison of the survival curves of three tree seedling species (Anacardium
occidental, Bombacopsis quinata, and Tectona grandis) (N = 675) planted in three
planting configurations (diagonal, fence, and line) during 900 days in pastures ofRio
Grande, Cocle province, Panama.
-x- 4-. occidentale
Terminal events (#)
Time (days after planting)
Figure 3-2 Incidence of mortality among Anacardium occidentale, Bombacopsis quinata, and
Tectona grandis seedlings planted in three planting configurations for silvopastoral
system establishment in farmers' fields in Rio Grande, Cocle, Panama.
Incidence of herbnvory 400
Figure 3-3 Incidence of herbivory of three species of tree seedlings (N = 225 seedlings per
species) browsed by cattle, leaf-cutter ants, or other observed sources during a two-
year experiment in grazed on-farm pastures in Rio Grande, Cocle, Panama. The y-
axis (Incidence of herbivory) refers to the number of events when seedlings were
impacted by herbivores.
Incidence of herblvory 250
Figure 3-4 Incidence of cattle, leaf-cutter ant, and other sources of herbivory of tree seedlings
(Anacardium occidentale, Bombacopsis quinata, Tectona grandis) planted in three
planting configurations in grazed pastures in Rio Grande, Cocle, Panama. The y-axis
(Incidence of herbivory) refers to the number of events when seedlings were impacted
EFFECTS OF SCATTERED LARGE TREES IN PASTURES ON A Hyparrhenia rufa-
DOMINATED MIXED SWARD
To be able to promote the implementation and use of silvopastoral systems with certainty,
it is imperative that the dynamics of the systems and their parts be understood. Garnering
knowledge of interactions in silvopastoral systems is of particular importance due to their
complexity, as they comprise multiple, multi-dimensional components including trees, crops, and
livestock. Within the context of seeking to understand diverse biophysical interactions of
silvopastoral systems as a means to work toward the promotion and wider implementation of
silvopastoral systems in Panama, this research studied the effects of mature, dispersed trees on
forage in extensive degraded pastures. Effects of two species of trees (Anacardium occidentale
and Tectona grandis) were assessed on pastures dominated by the naturalized African grass,
Hyparrhenia rufa. Analyses included the testing of forage mass, digestibility, and composition
along a gradient of distances from mature trees.
A debate abounds concerning the effects of light on forage growth in tree-pasture systems.
Belsky (1994) proposed that light is not a primary factor in the growth of perennial species under
trees. She found that the environmental conditions under tree canopies were more prominent
than the potential effects of competition for light between trees and perennials. Clason (1999)
also suggested that canopy shading did not play a role in his research on subtropical forage
growth under a mixed pine plantation (Pinus taeda and Pinus echinata) in Louisiana, USA.
Rather, he found competition for soil moisture between trees and forage to be a greater
determining factor in reduction of forage yields under trees. Ares et al. (2006) also contended
that overstory shade was not a prominent factor affecting forage production under large native
pecans (Carya illinoinensis) in Kansas, USA. Rather, they attributed fluctuations in forage yield
to changes in local climatic conditions. Likewise, in Argentina Fernandez et al. (2006) studied
the interactions between Festuca pallescens and Pinus ponderosa. They found that at a stand
density of 350 trees per ha, light levels under the pine canopy and in areas between canopies
However, disparity exists in this debate. Some researchers conclude that light does in fact
have an important effect on forage growth under trees. In fact, in a study in Appalachia, USA
testing the performance of orchardgrass (Dactylis glomerata) in open pasture, woodlands, and
woodland-grass edge sites, Belesky (2005) found a significant relationship between grass dry
matter and light availability to grass. Grass dry matter was greatest as leaf of grass growing in
transition zone edge sites, suggesting that availability of light in edge sites facilitated grass
growth. Similarly, in their research on a mixed forage pasture with dispersed poplar (Populus
spp.) trees, Douglas et al. (2006) found forage growth was reduced 23% under trees when
compared to open pasture. The authors attributed the differences in treatment effects,
particularly in terms of season, to differences in light reception below trees and in open pasture.
However, other research results (Peri et al., 2002) show that effects of changes in light may vary
by forage species. For example, in the study carried out by Douglas et al. (2006), white clover
(Trifolium repens) was significantly more abundant in open pasture than under trees. On the
other hand, orchardgrass composition in pasture was twofold greater under trees than in open
pasture while differences were not found in perennial ryegrass (Lolium perenne) growth under
trees and open pasture. Similarly, Fernandez et al. (2002), studying the effect of overstory Pinus
ponderosa canopy on the tussock grass Stipa speciosa in Argentina concluded that S. speciosa
growth was limited as a result of the interception of light by the overstory canopy. They found
that as pine stocking rate increased, grass growth decreased.
Consistent with the differing results of the effects of light on tree-pasture systems, some
research has looked closer at plant responses to diminished light availability in silvopastoral
systems. Specifically, changes in grass allocation to above- versus belowground biomass
consequent to changes in available light have been examined. Fernandez et al. (2004) examined
the changes in biomass allocation of the forage species, Festuca pallescens, relative to different
shade intensities in Argentina. They deduced that changes in allocation of biomass resulted in
increases in leaf production. Under a stand density of 500 pruned pine trees per ha, radiation
was reduced by 75%. They proposed that the forage species changed its biomass allocation
pattern in response to shading: allocation to storage roots was reduced while allocation to leaves
increased. The authors asserted that this change may affect species susceptibility to herbivory.
A shift in biomass allocation, from storage organs to leaves, can leave a plant less equipped to
respond to herbivory with new growth.
Belesky (2005) concurs that leaf production should not be achieved at the expense of
structures contributing to plant persistence. Reduced allocation to roots can also result in
reduced drought tolerance due to decreased soil foraging and water uptake by roots, particularly
when in competition with tree roots. Moreover, both Belesky (2005) and Fernandez et al. (2002)
found shading reduced tiller production in forage grasses.
Considering the potential effects of reduced light availability on pasture grasses under
trees, Rietkerk et al. (1998) suggest that a tradeoff exists between light availability and soil
nutrient availability in that although light in the understory often becomes reduced due to
shading by the overstory canopy, trees may confer beneficial effects on understory conditions
and vegetation. Silva-Pando et al. (2002) proposed that a relationship existed between shade
intensity and soil nutrient availability. Moreover, as suggested by Belsky (1994) and others,
factors other than changes in light availability may impact forage growth in tree-pasture systems.
Such factors include soil water use (Clason, 1999) and belowground competition for nutrients
and space (Ares et al., 2006). In fact, Rietkerk et al. (1998) suggested that tree roots' zone of
influence extended beyond the tree crown implying that tree root systems can have a strong,
extensive effect on understory vegetation belowground.
Silva-Pando et al. (2002) also proposed the existence of mechanisms other than light, such
as physiological aspects of trees and forage in the understory and overstory, that may affect
forage growth. Indeed, Douglas et al. (2006) and Fernandez et al. (2006) found soil water
availability to be less under trees than in open pasture. They both suggest that rainfall was
captured by trees in the overstory thereby limiting soil moisture content, and consequently,
moisture availability to understory vegetation. Also, uptake of water by tree roots might play an
important role in limiting the availability of moisture belowground. However, Fernandez et al.
(2004) only found a disparity in soil moisture availability between open pasture and under trees
during periods of high moisture availability, at which time grasses under trees had better water
status than grasses in open pasture. The authors attributed this to lower evaporative demand
under the tree canopy.
There is a range and diversity of research and opinions concerning large tree effects on
understory forages. There seems to be much debate on which aspects of tree-forage interactions
ultimately determine outcomes: light may or may not be a factor, climate, soil moisture, species-
specific traits, and tradeoffs of light reduction and buffering of extreme conditions are considered
to play some type of role in impacting characteristics of understory forage.
Objective and Hypothesis
The obj ective of this study was to evaluate and compare the impacts and consequences of
large, dispersed trees in pasture on the characteristics ofHyparrhenia rufa-dominated forage
growing in mixed swards in degraded pastures. Characteristics included forage growth, in vitro
organic matter digestibility, and forage composition as characterized by proportions of grass,
legumes, weeds, and necromass on the pasture. I hypothesized that along a range of distances
relative to stems of trees, influence and impacts of trees on pasture components and
characteristics would become reduced with increasing distance from the tree stems.
Methods and Materials
This study was conducted in the sectors of La Calendaria and Los Olivos, Rio Grande
corregimiento, Cocle province, Panama (see Chapter 2 for specific local and regional
characteristics). Data were gathered from pastures on one farm in each sector. The pasture is
dominated by the naturalized African grass Hyparrhenia rufa with few naturally occurring
legume species. Field burning is a common practice in the area; however, broadleaf herbicide
application is rare. Pastures had been grazed by cattle consistently during at least two decades.
Mature trees were dispersed throughout the pastures. In the wet season, cattle stocking rate
averaged 0.5 to 1.0 AU per ha.
The study consisted of two similar experiments. These experiments were structured as
randomized complete block designs. Each experiment was alike except for the tree species that
was used; one experiment used Anacardium occidentale and the other experiment used Tectona
grandis. All experimental design aspects of the study were similar for both experiments. There
were three blocks for each species and each block contained all of the treatment combinations.
Forage was harvested on a gradient of three distances from tree stems in the four cardinal
directions. Distances were formulated according to the crown size of each tree. The radius of
each canopy was measured and distances were gauged based on the space pertaining to 50%,
100%, and 200% (identified as 0.5, 1.0, and 2.0 distances) of the radius of each tree canopy
(APPENDIX C). Forage samples were harvested randomly within the context of corresponding
direction and distance from the tree stem, yielding twelve destructive samples per tree, for both
experiments. Sampling of forage mass, digestibility, and botanical composition occurred in May
and September of 2002, in May and December of 2001, and in December of 2001, respectively.
Sample sites were chosen at each distance in each cardinal direction. A metal wire ring,
0.5 m in diameter, was placed in the selected sites and all herbage within the ring was harvested
manually (by machete and hand clippers) to ground level. The forage fresh weight was recorded.
To evaluate in vitro organic matter digestion (IVOMD), herbage was bagged and oven-dried at
600 C. Dried samples were ground and milled through a 1 mm screen. In vitro organic matter
digestion was performed by a modification of the two-stage technique (Moore and Mott, 1974).
To assess composition, fresh samples were air dried and separated by hand into pre-established
categories of grass, weeds, legume, and necromass. "Grass" was categorized as all green
biomass pertaining to the species Hyparrhenia rufa. "Weeds" were plants that participating
farmers identified as being undesirable or harmful to cattle, and/or not beneficial to or
contributing to good pasture and cattle production. These included a variety of plant types,
including forbs and shrubs. "Legumes" were categorized as those plants with characteristics that
resembled the Fabaceae family. "Necromass" was all biomass identified as dead material. After
forage categorization, samples were bagged and weighed.
Statistical analyses were performed using SAS and SPSS. Dependent variables (forage
mass, IVOMD, and forage botanical composition) were analyzed using the ANOVA procedure.
When main effects were significant, Tukey hsd post-hoc test was used to compare means.
Orthogonal polynomial contrasts were used to describe the effect of location.
When analyzing the distance by season interaction for A. occidentale, there was no
significant effect on forage mass (p = 0.641), nor was there a significant main effect for the
distance variable (p = 0.76) in the case ofA. occidentale. There was no significant linear or
quadratic effect of distance on mass or its interaction with the season variable (Table 4-1). There
was a main effect of season on forage mass (p < 0.001) with wet season obtaining an overall
higher mean than dry season. In the post hoc test, we observed that there was a significant
seasonal effect within each distance, 50% (p = 0.015), 100% (p = 0.002), and 200% (p < 0.001).
Wet season marginal means were greater than dry season marginal means at each distance.
In the analysis of forage mass under Tectona grandis, there was no significant two-way
interaction between distance and season (p = 0.368). There was a significant linear effect (p =
0.001) of distance, but the quadratic effect only approached significance (p = 0.097) (Table 4-2).
In the post hoc test, distance 2.0 mean forage mass was significantly greater than distance 1.0 (p
= 0.018) and distance 0.5 (p = 0.004) (Table 4-3). However, there was no significant main effect
for season (p = 0.926) under 7: grandis.
Forage digestibility under A. occidentale was affected by distance and season (p = 0.042
and p < 0.001, respectively) but there were no interactions. The post hoc test revealed that
forage digestibility was significantly greater at the farthest distance from the tree stem (2.0) than
at the 0.5 distance (close to the tree stem) while the drip line (1.0) and 0.5 distances were not
significantly different. In addition, in the post hoc analysis of the season variable, wet season
digestibility was significantly greater than dry season digestibility at the 0.5 and 2.0 distances
from the A. occidentale tree stems (Table 4-4).
However, results were different for T. grandis forage digestibility. There was no distance
effect for T. grandis (p = 0.746). The season variable was significant at p < 0.001 under T.
grandis. Wet season digestibility was significantly greater than that of the dry season at
distances 2.0 (p = 0.001) and 0.5 (p < 0.001) (Table 4-4).
Under A. occidentale, there were no treatment effects on forage botanical composition.
Likewise, under T. grandis, the effect of distance on weeds, grass, and legume was not
significant. However, results for necromass under 7 grandis were different from the other
forage components in that the effect of distance on necromass was significant (p = 0.035). When
examining further the comparisons of means of necromass by distance, there was a significant
difference between distances 0.5 and 1.0, where necromass at the drip line (distance 1.0) was
significantly greater than necromass close to the stem (distance 0.5) (p = 0.049). No significant
difference was observed in necromass abundance between distances 1.0 and 2.0 (p = 0.982) or
0.5 and 2.0 (p = 0.314).
Effects of trees on understory forage can vary by season, climate, and soil conditions. In
this research, forage mass was affected by distance and season; however, these effects were
dependent on tree species. Distance of forage from the tree stem did not have a significant effect
on forage mass below A. occidentale but did play a role below T. grandis. Forage mass was
significantly greater at the 2.0 distance than at the 0.5 and 1.0 distances below T. grandis. At the
same time, seasonal effects influenced forage mass under A. occidentale but did not have an
effect on T. grandis forage. The difference found for forage mass under A. occidentale in the dry
season and the wet season touches upon the importance of seasonal effects on herbage
abundance in tropical pastures. This result was to be expected given the seasonal contrast in
moisture availability. Although accurate rainfall data for the study site could not be obtained,
records at the nearby recording site show the annual rainfall as about ~ 900-1100 mm, 90% of
which is received in eight months during May to December, the remaining 4 months being quite
dry. However, results of forage mass under A. occidentale should not be generalized across
species because although forage mass was significantly higher under A. occidentale during the
wet season than in the dry season, forage mass did not differ significantly under 7 grandis
between seasons. In fact, forage mass was lower under T. grandis in the wet season than in the
dry season. Thus, season did not have the same affect on forage mass under the two tree species.
The consistency of forage mass abundance under 7 grandis across seasons contrasted with the
sizable increase in forage abundance under A. occidentale from the dry season to the wet season;
forage mass under T. grandis experienced a decrease during the same period (Figure 4-1). These
results suggest: 1) dry season conditions augmented forage mass under 7 grandis while wet
season conditions induced a suppressive effect on forage growth under 7 grandis; or 2) based on
the consistency of forage mass abundance across season, 7 grandis maintained a steady,
suppressive effect on forage throughout the year, regardless of season; and 3) growth
performance of forage was different under different tree species.
Increased forage mass under T. grandis in the dry season may have been related to two
traits pertaining to T. grandis: deciduousness and aggressive growth habit. During the dry
season, 7 grandis was completely deciduous. At this time, the entire stem and branches of T.
grandis individuals are leafless indicating that T. grandis may enter a type of dormancy during
this period. If such dormancy occurs, an attenuation of T. grandis' aggressive growth type,
including a temporary reduction in belowground resource use, may occur as part of the
dormancy process. Relief from T. grandis' highly aggressive growth complemented by
increased availability of belowground resources and light may have provided the forage under 7
grandis with increased access to resources, leading to increased growth and accumulation of
forage mass during this period.
However, it is also plausible that the consistency of overall low forage mass abundance
under 7 grandis across seasons and distances may be the consequence of a consistent
suppressive effect of the tree species. In this case, the decrease in forage mass in the wet season
could have been the result of the intensification of T. grandis' suppressive effect due to an
increase in soil moisture, reduced stress, and consequent increase in resource availability to the
tree. However, it is important to note that these parameters were not directly measured in this
The contrasting results of forage growth under A. occidentale and T. grandis emphasize the
difference in effects of individual tree species on forage. Also emphasizing the importance of
tree species effect on pasture, forage mass was notably less under T. grandis in comparison to
herbage under A. occidentale in both the wet and dry seasons. Higher yielding forage
performance under A. occidentale and the apparent suppression of forage growth under T.
grandis further accentuates the distinct effects tree species can have on forage.
Species-specific effects were also evident when comparisons were made of results within
distances. Like season, distance played a different role in the results by species. Unlike season,
distance was not a relevant factor for forage mass under A. occidentale; however, under T.
grandis distance from the tree stem played a role in determining forage abundance. Forage mass
at the farthest distance (2.0) was significantly greater than forage mass at the drip line (1.0) and
close to the tree stem (0.5) under T. grandis. There was no difference between the 0.5 and 1.0
distances, suggesting that the tree had some effect on nearby forage. However, when examining
the absolute values of forage mass at different distances under T. grandis, the differences are
seemingly slight. Nevertheless, decreased forage abundance closer to the T. grandis tree stem
broadens the argument regarding the aggressive character of this tree species. This is also
emphasized by the lack of distance effect ofA. occidentale on forage.
Differences in distance can be influenced by seasonal effects as well; for example, during
the dry season forage mass at the drip line can be buffered from high temperatures and
evapotranspiration rates while in the wet season moisture at the drip line is captured by the tree
crown. In comparison, open pasture during these periods is exposed to temperature,
evapotranspiration, and moisture fluxes. These effects are related to and can be impacted by tree
species type. For example, canopy architecture and leaf type can determine the degree of light
availability, temperature buffering, and evapotranspiration at the drip line. Also, root systems
and belowground performance can differ by species. Rooting ability, root length, root
architecture, and biomass allocation to roots can determine species effectiveness at acquiring and
outcompeting grasses for resources at both distances. In fact, Behrens (1996) notes that roots of
mature A. occidentale trees are known to extend beyond the drip line as much as twice the length
of the tree canopy. Species with more effective root systems may be better equipped to
outcompete grasses at the drip line and potentially in open pasture.
For a better understanding of the difference in effects of particular tree species on forage,
we may consider the impacts of cattle, tree canopy, leaf type, and allelopathy on conditions
around trees, and how these can differ by tree species and thereby impact forage. In the case of
this experiment, in which forage mass under A. occidentale was markedly greater than that under
T. grandis, it is worthwhile to consider how cattle may impact forage around these species. A.
occidentale is an abundant producer of large, nutritious fruit which attracts cattle to its
immediate surroundings. Also, A. occidentale commonly possesses a globular, densely-leafed
canopy which casts cool shade, frequently pursued by cattle in extensive, denuded pastures. As
such, cattle are lured by shade and fruit to A. occidentale trees and thus can often be observed
congregating close to these. Such presence of cattle brings the benefits of deposition of dung
and urine to trees and surrounding areas. Dung and urine can add organic material and nutrients
to the environment thereby benefiting soil and forage under the tree and as well as the tree itself.
Conversely, T. grandis does not produce fruits relished by cattle. Also, T. grandis does not
tend to attract cattle (in this experiment). In this experiment, T. grandis trees possessed a conical
canopy shape which did not produce shade that was attractive to cattle. In addition, leaf
characteristics of the two species are unique. T. grandis grows a very large, thick leaf that, when
added to the ground following leaf-fall, requires prolonged periods of time to decompose.