<%BANNER%>

Establishment of Silvopastoral Systems in Degraded, Grazed Pastures: Tree Seedling Survival and Forage Production under ...


PAGE 1

ESTABLISHMENT OF SILVOPASTORAL SYSTEMS IN DEGRADED, GRAZED PASTURES: TREE SEEDLING SURVIVAL AND FORAGE PRODUCTION UNDER TREES IN PANAMA By ALYSON B. K. DAGANG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007 1

PAGE 2

Copyright 2007 by Alyson B.K. Dagang 2

PAGE 3

To my Mother and Father, w hose boundless love gives me life 3

PAGE 4

ACKNOWLEDGMENTS There are many individuals and organizations who contributed to this study and my doctoral program to whom I am indebted and gratef ul. 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 Sollenber ger, and Dr. Marilyn Swisher, for their faith and confidence especially through rocky times. I would like to recognize and express my sin cere gratitude to the individuals and their institutions that supported me during my doc toral 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 Sc iences, 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 Flor ida Alumni Associati on, and the School for International Training (SIT). This study would not have been possible withou t the constant support I received from the farmers, families, and other collaborators in Pa nama. Thank you to Mr. Severito Martinez, Dr. Juan Jean, Viodelda de Suarez, Antonio Suarez, Famila Jaen, Familia Suarez, Familia Martinez, Familia Grajales, Familia Villareal, Familia Agui lar, Lic. Jose Villareal, personnel from the Laboratorio de Suelos del Institu to 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 Andr ea 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, 4

PAGE 5

Joyce Lepetu, Subrajit Saha, Mark Drew, and Wendy Francesconi, thank you for your support, friendship, humor, and tremendous spirit. To my treasured compaeros and sisters who ha ve 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 Ta ranto, Osvaldo Jordan, Luis Do minguez, Juan Nuques, Cynthia Gomez, Alicia Peon, Leonardo Mart inez, 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. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES .........................................................................................................................10 LIST OF FIGURES .......................................................................................................................11 ABSTRACT ...................................................................................................................................13 CHAPTER 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 .............................................................................19 Silvopastoral Systems ......................................................................................................20 Choice of Tree Component .............................................................................................21 Microclimate ....................................................................................................................22 Forage component Recent Studies on Fora ge Vegetation in Silvopastoral Systems ...24 Summary ..........................................................................................................................25 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 ................................................................28 Introduction of Cattle and Land Use Change ..................................................................29 Frontier Expansion and Green Revolution in Panama ....................................................30 Impacts of the Green Revolution .....................................................................................30 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 .......................................................................................................37 Location ...........................................................................................................................37 Ecology ............................................................................................................................38 Climate ............................................................................................................................38 Local Farming Systems ...................................................................................................39 Species Descriptions ...............................................................................................................39 6

PAGE 7

Tectona grandis ...............................................................................................................40 Origin, Natural Habitat, and Environment ...............................................................40 Uses ..........................................................................................................................41 Botany ..............................................................................................................................41 Germination and Establishment ......................................................................................41 Adaptability and Performance .........................................................................................42 Rooting and Competition ................................................................................................43 Burning ............................................................................................................................43 Potential benefits of teak plantations ...............................................................................44 Bombacopsis quinata (syn. Pochota quinata, Bombacopsis quinatum ).........................45 Anacardium occidentale ..................................................................................................46 Botanical description ................................................................................................46 Cultivation ................................................................................................................47 Uses ..........................................................................................................................48 Planting Configuration ....................................................................................................50 3 TREE SEEDLING SURVIVAL A ND IMPACT OF HERBIVORY ON SILVOPASTORAL SYSTEM ESTABLISHMENT.............................................................61 Introduction .............................................................................................................................61 Literature Review ...................................................................................................................62 Tree Seedling Survival ....................................................................................................62 Effects of Cattle Grazing .................................................................................................64 Herbivory .........................................................................................................................66 Objectives and Hypothesis .....................................................................................................70 Methods and Materials ...........................................................................................................71 Study Site .........................................................................................................................71 Experimental Design .......................................................................................................71 Materials ..........................................................................................................................71 Establishment ..................................................................................................................72 Measurements ..................................................................................................................72 Data Analysis ...................................................................................................................72 Results .....................................................................................................................................73 Seedling Survival .............................................................................................................73 Observed Causes of Mortality .........................................................................................74 Herbivory .........................................................................................................................74 Sources of Herbivory .......................................................................................................75 Discussion ...............................................................................................................................75 Seedling Survival .............................................................................................................75 Observed Causes of Seedling Mortality ..........................................................................78 Herbivory .........................................................................................................................79 Sources of Herbivory .......................................................................................................81 Conclusion ..............................................................................................................................82 7

PAGE 8

4 EFFECTS OF SCATTERED LARGE TREES IN PASTURES ON A Hyparrhenia rufaDOMINATED MIXED SWARD...................................................................................89 Introduction .............................................................................................................................89 Literature Review ...................................................................................................................89 Light ................................................................................................................................89 Biomass Allocation .........................................................................................................91 Belowground Factors .......................................................................................................91 Objective and Hypothesis.......................................................................................................93 Methods and Materials ...........................................................................................................93 Study Site .........................................................................................................................93 Experimental Design .......................................................................................................93 Measurements ..................................................................................................................94 Data Analysis ...................................................................................................................95 Results .....................................................................................................................................95 Forage Mass .....................................................................................................................95 Forage Digestibility .........................................................................................................96 Forage Composition ........................................................................................................96 Discussion ...............................................................................................................................97 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 Introduction ...........................................................................................................................113 Literature Review .................................................................................................................113 Competitive Ability .......................................................................................................114 Competition for Soil Moisture .......................................................................................116 Root Biomass Allocation ...............................................................................................118 Competition for Nutrients (Fertilization Studies) ..........................................................120 Microclimate Effects .....................................................................................................121 Trenching Effects ..........................................................................................................121 Objectives and Hypothesis ...................................................................................................124 Methods and Materials .........................................................................................................124 Study Site .......................................................................................................................124 Experimental Design .....................................................................................................124 Materials ........................................................................................................................125 Establishment ................................................................................................................125 Measurements ................................................................................................................126 Data Analysis .................................................................................................................126 Results ...................................................................................................................................126 Herbage Removal ..........................................................................................................126 Effects of the Species Treatment on Biomass ...............................................................127 Stem Biomass ................................................................................................................128 8

PAGE 9

Root Biomass.................................................................................................................128 Discussion .............................................................................................................................129 Seedling Growth ............................................................................................................129 Stem and Root Biomass .................................................................................................131 Conclusion ............................................................................................................................132 6 SUMMARY AND CONCLUSIONS...................................................................................140 Experimental Findings ..........................................................................................................140 Seedling Survival and Herbivory ..................................................................................140 Effects of Large Trees on Understory Forage ...............................................................141 Interactions between Seedlings and Vegetation ............................................................142 Implications for Implementation ..........................................................................................143 Options for Grazing .......................................................................................................143 Manipulating Forage with Trees ...................................................................................144 Tree Establishment ........................................................................................................144 Future Research ....................................................................................................................144 APPENDIX A PLANTING CONFIGURATIONS OF TH E THREE TREE-SPECIES SEEDLINGS FOR ESTABLISHMENT OF A SILVOPAS TORAL SYSTEM IN RIO, GRANDE, COCL, PANAMA..............................................................................................................146 B COMPARISONS OF MEANS OF INCIDENCE OF TREE SEEDLING HERBIVORY ACROSS TREE SPECIES AND PL ANTING CONFIGURATION...................................147 C FORAGE SAMPLING SCHEMATIC OF HERBAGE MASS HARVESTED AT THREE DISTANCES FROM TREE S TEM IN THE FOUR CARDINAL DIRECTIONS CARRIED OU T UNDER SCATTERED TREES IN PASTURES IN RIO GRANDE, COCL, PANAMA...................................................................................148 LIST OF REFERENCES .............................................................................................................149 BIOGRAPHICAL SKETCH .......................................................................................................160 9

PAGE 10

LIST OF TABLES Table page 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 conf iguration and species on survival of 675 seedlings planted in five blocks in degraded pastures on-farm over two years in Cocl, Panama. ...................................................................................................................84 4-1 Analysis of variance for polynomial ort hogonal contrasts of sample mean forage mass comparing the effects of distance and season under dispersed Anacardium occidentale trees in Rio Grande, Cocl, Panama. ............................................................104 4-2 Analysis of variance for polynomial ort hogonal contrasts of sample mean forage mass comparing the effects of distance and season under dispersed Tectona grandis trees in Rio Grande, Cocl, Panama. ...............................................................................105 4-3 Post hoc comparisons of mean forage mass at three distances1 from dispersed T. grandis tree stems in grazed, degraded past ures in Rio Grande, Cocl, Panama. ...........106 4-4 Post hoc analysis of forage digestib ility across three distances from dispersed Cashew trees ( A. occidentale ) and by two seasons in graz ed pastures of Rio Grande, Cocl, Panama. .................................................................................................................107 5-1 Analysis of the effects of the repeated measures herbage removal, tree species, and time on biomass accumulation of tree seedli ngs planted on-farm in a non-grazed pasture in Rio Grande, Cocl, Panama. ...........................................................................133 5-2 Comparisons of the within -subject effects of the repeated measure herbage removal on biomass accumulation of tree seedlings pl anted on-farm in a non-grazed pasture and observed over two years in Rio Grande, Cocl, Panama. .........................................134 5-3 Effects of the interactions of three se edling species with harvest time (6, 12, and 24 months after planting) on biomass accumula tion of tree seedlings planted on-farm in a non-grazed pasture in Ri o Grande, Cocl, Panama.......................................................135 10

PAGE 11

LIST OF FIGURES Figure page 2-1 Topographic map of the Panamanian isthmus. ..................................................................53 2-2 Panama forest cover and areas of deforestation in 1947. ...................................................54 2-3 Changes in land use and hu man 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, Cocl province, Republic of Panama. ..............................................................................................................................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, Cocl province, Panama. ......................................................................................85 3-2 Incidence of mortality among Anacardium occidentale, Bombacopsis quinata, and Tectona grandis seedlings planted in three planti ng configurations for silvopastoral system establishment in farmers fields in Rio Grande, Cocl, Panama. ..........................86 3-3 Incidence of herbivory of three specie s of tree seedlings (N = 225 seedlings per species) browsed by cattle, leaf-cutter ants or other observed sources during a twoyear experiment in grazed on-farm past ures in Rio Grande, Cocl, 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 past ures in Rio Grande, Cocl, 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, Cocl, Panama. ...........................................................................108 4-2 In vitro organic matter di gestibility 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, Cocl, 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, Cocl, Panama. ..110 11

PAGE 12

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, Cocl, 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, Cocl, Panama. .....................112 5-1 Responses of three species of tree seed lings to three understo ry-herbageremoval treatments during the first two years after tr ee planting in a field site in Rio Grande, Cocl, 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, Cocl, Panama. ............................................................................................................................137 5-3 Changes in seedling biomass accumulation in stems and roots, and root:shoot ratio (numbers above bars) changes during the tw o-year establishment of silvopastoral systems in pastures in Cocl, Panama. .............................................................................138 5-4 Root:shoot ratios of thre e species of seedlings acro ss grass removal treatments during the two-year establishm ent phase of silvopastoral systems planted in pastures in Rio Grande, Cocl, Panama. ........................................................................................139 12

PAGE 13

Abstract of Dissertation Pres ented 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 IN PANAMA By Alyson B. K. Dagang May 2007 Chair: P.K.R. Nair Major Department: School of Fo rest 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 tropi cs has, however, received little attention. In Panama, smallholder pastures ar e abundant across the landscape, but they are often extensive, degraded, overgrazed, and of low productivity. Based on the premise that integr ation of silvopastoral systems on degraded pastures might be an effective technol ogy that is accessible and affordable for small-scale producers, this re search was carried out onfarm for two years in central Panama to help devise best management practices for optimizing tree-seedling survival, reducing competition between seedlings and he rbaceous vegetation, and managing effects of large trees on forage. Three experiments were conducted. The firs t 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 herb aceous vegetation on the establishment of tree 13

PAGE 14

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 aboveand belowground. In the third experi ment 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 te rms of mass, digestib ility, and botanical composition of the forage underneath. Research results revealed that Anacardium occidentale seedlings survived best in grazed pastures and the fence planting configuration resu lted 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 ot her species in terms of biom ass allocation to shoots. Similarly, the effects of large tr ees on understory forage varied with tree species. Forage mass under T. 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 T. 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 developmen t of recommendations and guidelines on tree species selection, planting c onfiguration, grazing, weeding, and forage management for successfully integrating silvopa storal systems into small holder pastures in Panama. 14

PAGE 15

CHAPTER 1 INTRODUCTION In Panama, pastureland covers about 1.3 milli on 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-inte nsity technologies and management technologies such as use of feed lo ts and supplemental, processed feeds exist to augment productivity, these are untenable for most producers. New produc tion strategies that can be easily accessed, implemented, and afforded by producers must be sought. Silvopastoral systems the integration of trees into livestock sy stems 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 integr ation on degraded pastures can augment soil health, forage production, and environmental se rvices, silvopastoral systems, might be an effective technology that is accessible and affo rdable 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 appl ied, on-farm research to devise appropriate means of establishing silvopastoral systems on degr aded pastures and to investigate how best to integrate tree seedlings into graz ed, degraded pastures in Panama Major 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 re moval of herbaceous material around seedlings were also investigated. 15

PAGE 16

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, wellpriced 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 ef forts and as a plantation species. Producers perceive T. grandis as a commodity species that can provi de added income from the pasture to the household. Bombacopsis quinata is a multi-purpose, native ha rdwood species that is used locally in live and dead fences, furniture making, and in construction. The overall objective of this research was to gain knowledge of some of the bases of silvopastoral system establishment in degrade d, grazed pastures. Through monitoring seedlings for survival and herbivory over two years, manipul ating herbaceous vegetation and tree seedlings aboveand 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 assump tions 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 fo llowing general hypothes es were tested: The pattern in which tree seedlings are planted in pasture (planting c onfiguration) impacts the survival and herbivory of seedlings. Differences exist among tree speci es in terms of their perfor mance under different planting configurations in silvopastoral systems. Removal of herbaceous vegetation around esta blishing seedlings has positive effects on seedling survival. Isolated, large trees impact mass, digestibil ity, and botanical compos ition of the understory forage. 16

PAGE 17

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 in cludes a review of the relevant silvopastoral system literature as well as tree species and re search site descriptions. Chapters 3, 4, and 5 present the experiments conducted in this researc h. 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 Pa nama. Chapter 4 provide s the results from the study that examined the consequences of disper sed, large trees on fora ge characteristics in pasture. Chapter 5 presents the results from the experiment that studied the effects of aboveand belowground vegetation removal on tree seedling grow th in a controlled field site. Each of the three chapters includes an e xplanation of the experimental methodology, a review of the pertinent literature, a descripti on of the study results, and a disc ussion 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. 17

PAGE 18

CHAPTER 2 AGROFORESTRY AND LAND USE IN PANA MA AND A GENERAL DESCRIPTION OF THE STUDY SITE Agroforestry 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 se quence that results in a significant interaction of woody and non-woody com ponents (Nair, 1993). There is evidence of the implementation of agroforestry systems da ting 10,000 years before present (Miller and Nair, 2006; Gakis et al., 2004). Widespread study of these traditional practices has grown during the 20 th century. Researchers who seek appropriate technologies to respond to growing food needs, diminishing global ecological health, and th e rise in land degradation have embraced agroforestry practices as a suit e 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 en hancing 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 thre e components in the syst em are intentionally managed for optimal interactions aimed at augmenting agricultural production and environmental services (Sharrow, 1999). Silvopasto ral systems will be discussed further in this chapter. Benefits of Agroforestry Systems Agroforestry systems such as improved fa llows, 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 speci es. These can be an answer to exhausted soils or degraded 18

PAGE 19

lands (Nair et al., 1999; Sanch ez, 1999). Alley cropping is th e combining of woody perennials and annual crops in fields with the aim of e nhancing crop production th rough enriched nutrient cycling (Jordan, 2004). Improvement in agricultural pr oduction through agroforestry systems is based in part on the contribution of woody speci es 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 thr ough 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 mois ture to neighboring vegetation by excising moisture from deep soil sources and redistribu ting 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 fo r improving agricultural production. However, the existing multitude of ag roforestry 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 degradation. 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 19

PAGE 20

short-term and in the long-term. Pasture degr adation and creation are among the leading causes of deforestation. As such, integration of s ilvopastoral systems into existing agricultural enterprises can potentially enable farmers to re duce the degradation of their farms (Serrao and Toledo, 1990). Benefits and characteristics of sil vopastoral systems will be discussed in detail in the next section. Silvopastoral Systems As noted above, silvopastoral systems, a form of agroforestry, include land-use practices that involve woody perennials, fo rage plants, and livestock simultaneously during a period of time to enhance production and/or the environment. One type of silvopastoral system, cut-andcarry 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 hi gh 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 past ure at different stand densities but trees are not directly grazed. However, depe nding on the tree species, livestock commonly graze fallen fruits, seeds, nuts, and foliag e. 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 objectives are achievable through the implementation of silvopastoral systems. They postulate that soci al, environmental, and eco nomic benefits can be obtained through improving forage quality, in creasing timber production, sequestering carbon, reducing contaminant run-off, en riching wildlife habitat, and improving landowner income. For example, studies in semiarid northeastern Brazil conclude that maintain ing 30% of tree cover 20

PAGE 21

when converting forest vegetati on 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 silv opastoral 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 system s in general and silvopastoral systems in particular have been studied, the following brief re view 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 comp onent in a silvopastoral system is vital in that the unique characteristics of each species including root ing habit, litter quali ty, canopy architecture, allelopathy, radiation interception, and other traits can have deci sive 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 pastur e systems. However, Garret et al. (2004) agree that properti es such as canopy density, spec ies phenology, vigor, and growth habit are crucial characteristics to be identified for the integr ation 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 conduc ted 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 21

PAGE 22

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 altho ugh tree height was not significantly affected at the end of the experiment. Moreover, in a st udy looking at broad-leaved species, Teklehaimanot et al. (2002) found signifi cant differences in growth between sycamore ( Acer psuedoplatanus) 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 capacity. Microclimate Within a silvopastoral system, the multiple effects of microclimate created by the tree component and the understory vegetation can ha ve 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 thei r study provide an excel lent example of the microclimatic effects of trees on pasture and hi ghlights how these can di ffer by species (Table 2.1). As seen in the Menezes et al. (2002) e xperiment, 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, Feldha ke (2001) studied the ef fects of black locust ( Robinia pseudoacacia ) canopy on a tall fescue ( Festuca arundinacea ) pasture. He studied photosynthetically active radiati on (PAR), red/far-red ratio, and soil temperatures and found that 22

PAGE 23

under increasingly cloudy conditions (25% PAR), % PAR under bl ack locust canopy relative to open field PAR doubled. Moreover, the author po sited that the presence of the black locust canopy reduced the extent of extr eme conditions that the understor y vegetation had to endure and therefore to which it must adapt which he asse rted may be beneficial He concluded that increased radiation use efficiency of the fora ge under diffuse light conditions as opposed to direct sun increased forage production. Feldha ke (2001) also found a significant difference in soil temperature when comparing open-field a nd under-canopy temperatures. During a mid-day reading, there was a difference of 6.5 o C in soil temperature under the two scenarios with equivalent soil moisture. In response to a 10% de crease in soil moisture, soil temperature in the open field increased 12 o C while under the black locust canop y soil surface temperature increased 2 o C. According to Feldhake (2001), temperatur e conditions under the black locust canopy were consistently within the appropria te range for tall fescue. Feld hake (2002) also found significant differences in night temperatures in an on-farm silvopastoral system. His research results showed that average below canopy nighttime temper atures in a southern West Virginia 35-yrold, 17-m-tall mixed conifer site with orchardgrass ( Dactylis glomerata ) understory was 11.5 o C higher than open field temperatur es. Results from the Feldhake experiments demonstrate the potential for the use of trees to moderate extrem e 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 temperat ures 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 signif icant difference in wind speed between the two system types. On the other hand, grass growth in the wet season was significantly greater in the 23

PAGE 24

open field. However, during the dry season, th ere was no significant difference observed for grass dry matter production between the two fiel d types. Likewise, in the dry season no significant difference was found between treatmen ts 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 s ite, 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 ha ve 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 differe nt management types can result in distinct agronomic and physiological outcomes. Forage component Recent Studies on For age Vegetation in Silvopastoral Systems As mentioned in the microclim ate section, the varied charac teristics of tree species can influence the overall productive outcome of a sil vopastoral system. Positive and negative effects can occur belowground between the forage plan t 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 Samanea saman in a dispersed tree silvopastoral system. Durr and Rangel (2002) looked at forage growth proximate to the S. saman canopy. The authors sampled biomass accumulation under the canopy, at the drip line, and in open field. They found no signifi cant 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 24

PAGE 25

different). Another part of th is experiment examined the bota nical 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 Panicum maximum, an important tropical forage species. The drip line was populated by a mix of P. maximum and Urochloa mosambicensis and the open field was dominated by U. mosambicensis This species sp ecialization by canopy region was generally static most of the year except during the dry season when there was an increase in U. mosambicensis 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 silvopastoral systems. Kallenbach et al. (2006) addressed a similar issue in Missouri, USA, looking at forage growth, nutritional quality, and livestock perfor mance 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 pastur e 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 author s speculate can be attr ibuted to the buffering of temperature and wind in the treed pastures. Summary Forage is a principal component of silvopastoral systems. Its abundance or scarcity can be the determining factor in the pr oductivity of a farming system. Forage species that demonstrate shade tolerance and effective r ooting abilities may provide greater advantages when used in silvopastoral systems. Likewise, tree species without highly competitive tendencies that are not 25

PAGE 26

especially sensitive to effects of understory competition may be preferential for silvopastoral systems. It is plausible th at, given the appropriate companion components and management, forage productivity can be enhan ced through the integration of sil vopastoral systems in livestock farming systems. Considering the need to de velop alternatives to present day, traditional agriculture in the intere st 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 need ed 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 Presented Research Introduction This section discusses histor ical, human, ecological, and soci al drivers behind present day land use. The aim of the discussion is to illust rate the motivations behind the research reported in this dissertation, which was devised in respons e 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 hi story of the isthmus a nd the land use practices applied by pre-colonial populati ons, Spanish colonists, and 20 th century homesteaders. Such historical factors coupled with current socioeconomic conditions transcend and shape todays land use issues. In order to understand these situations and ther eby shed light on the conception of this research, this section will convey the development of the Panamanian isthmus, prehistoric land use, the legacies of fire and savanna crops left by pre-colonial populations and 26

PAGE 27

colonists, consequences of the introduction of cattle on to the landscape, a nd the nature of land use today. Emergence of the Isthmus Three million years ago, the Panamanian isthmus emerged connecting Central America and South America. The occurrence had prof ound 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 Panamas topograp hy (Figure 2.1). The resulting c ordillera central is the central mountain range that moves through Meso america and continues into Panama creating two prominent and distinct clima tic and ecological zones. These include what are known as the Pacific seasonal region and the we t Atlantic region. Historicall y, this geographic and climatic distinction has had a decisive impact on the eco logical, agricultural, a nd human development of Panama. The unique eco-climatic regions created by the central range con tinue to influence land use today. Two unique precipitation zones ar e created in part by the pred ominant directions of trade winds. These generally blow from northeast to southwest causing ar eas north and east of mountain ranges to be wet, and those south and we st 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. Mu rphy and Lugo (1995) site Panama as a primary example of this geographical contra st in precipitation patterns. They state, The Pacific coast of Panama, supporting semideciduous forest, receiv es about 1780 mm of annual rainfall whereas the evergreen forest of the Caribbean coast r eceives over 3300 mm. On the Caribbean side, 27

PAGE 28

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 subject to continuous, very humid conditions th roughout the year (3000 to 4000 mm) while the southern plains a nd 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-m oving 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 occupa tion, manipulation, and cultivation by human civilizations over millennia coupled with th e demands of political and economic changes experienced during the 20 th century. To understand what is go ing on today in terms of land use, food production, and conservation, it is crucial that one become familiar with the history of the landscape. Panamas topographical and ecologi cal 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 funda mental 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-colo nial populations for farming, fishing, hunting, and general existence. The very wet inhospitable, adverse conditions of the northern side of the country presented greater challe nges to survival than the sout hern coast (Linares, 1980). Although the wet north coast presen ted challenges, some populations did live there. However, 28

PAGE 29

their agricultural practices were profoundly distin ct in that very small plots were slashed, soon abandoned, and left for long fallow periods wher eas 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 fo rest products during the period of 11,000 yr BP Panamanian agriculture commenced in the period of 7,000 yr BP coinciding with the introduction of maize ( Zea mays) to the isthmus (Piperno and P earsall, 1998) and was rapidly widespread by 2000 BP. In fact, it is reported th at at the time of the Spanish arrival to the isthmus (early 16 th century), much of the southern flatlands was void of forest cover as a result of the widespread use of fire and agricultural pract ices by pre-colonial popul ations as anthropogenic savannas dominated the landscape (Jaen, 1985). Howe ver, the arrival of the Spanish in the 16 th century changed land use and land cover dramat ically. Notably, the Spanish conquest provoked a significant decrease in the pr e-colonial population and a concomita nt recovery of forests on the landscape (Cooke, 1997). Introduction of Cattle and Land Use Change In 1521, Spanish merchants began to impor t cattle (Heckadon-Moreno, 1997) to graze Panamas former savannas and recovering forests. Introduction of cattle to the isthmus marked a crucial turning point for the land scape as cattle counter acted forest recovery and impeded fallow regrowth. Limiting forest regrowth was importa nt to Spanish colonists for two reasons: it facilitated the creation of extensive haciendas and controlled the invasi ve natural landscape (Jaen, 1985). Following initial colonial se ttlement, the northern region was comparatively unpopulated and became densely forested with a marked recove ry of forests along the alluvial coastal plain. The mountainous region, populated by descendants of indigenous groups escaped from slavery, 29

PAGE 30

was cultivated in the traditional indigenous slas h-and-burn system. The southern plains were dominated by European settlers e ngaged in agriculture and cattle raising. The eastern region of the country was sparsely populated by communitie s of escaped slave populations. However, by the 18 th century demographic changes spurred amplif ication of the anthropogenic savannas. Settlers used cattle, fire, and traditional agricultur al practices in tandem to increase space for land settlement. The combined use of these was fundamental to population expansion and land incursion. Agricultural area double d between the beginning of the 17 th century and the end of the 19 th century in the central provin ces (Jaen, 1985). Characterist ics of the rural Panamanian landscape changed little from the 19 th century through the early 20 th century (Figure 2.2). Today, of Panamas 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 ar e devoted to crops (Figure 2.3). Frontier Expansion and Green Revolution in Panama Today, Panamas rural human and ecological land scape resembles in some ways that of the early 20 th century. However, certain developments have modified this situation. Firstly, provision of basic medi cal care during the 20 th 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 portio n of the Azuero Peninsula and the province of Chiriqui (Heckadon, 1983; Jaen, 1985). Also, the population boo m provoked an important ruralto-rural migration that vastly e xpanded the agricultural frontier in to 400 yr old forests (Herrera, 1986). 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 he ightened concurrently 30

PAGE 31

with the spread of green revolu tion practices. For the rural sect or in Panama, widespread rural migration and the green revolution worked in con cert 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 Torrijos) to facilitate the relocation of peoples from areas of burgeoning population growth and la nd scarcity into the hinterlands, spawned a massive migration into forests (Dagang et al., 2003). Governmentsponsored migration into forest lands initiate d 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 cr op varieties, etc.). Gr een revolution practices enabled farmers to increase agricultural pr oduction capacity and c oncomitantly continue expansion into forests (Pries tley, 1982). For example, between 1950 and 1970, the area devoted to pasture production doubled (Jaen, 1985). Increased food production facil itated an increase in family size thus provoking greater population gr owth and consequent further migration and expansion of the frontiers (Figure 1.1). Land Use in Panama Today Panamas landowners and occupiers consist of peoples who own or occupy small, medium, or large parcels of land. In Panama, genera lly 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 gl obal and national social, economic, and political issues, the Panamanian agricultural economy ha s suffered in the past ten years which has provoked important changes in land use and a tran sformation 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 31

PAGE 32

monocrops and/or engage in single -species livestock rais ing. Some smallholders who have sold their land have moved to urban areas to seek wa ge labor opportunities while others move to an agricultural frontier area to continue traditio nal farming (Rudolf, 1999) and pasture creation, among these are the agricultural fr ontiers 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 import ant alterations of the landscape and land use patterns. The new, changed (and stil l 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 fr ontier expansion into the wet north and into the east, abutting protected areas and indigenous reserves. Today, smallholder farmers and some large la ndowners 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 be ing limited by the preservation status of protected areas and by the countrys autonomous indigenous reserves. Conflicts among populations for rights to land o ccupation 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 ar e 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 concei ved and conducted to respond to this need. 32

PAGE 33

Cattle Ranching in Panama The changing landscape is dominated by cattle ranching and past ure proliferation. Increasing cattle population and concomitant expans ion of pastureland calls for a greater focus and increased emphasis of research on pasture pr oductivity within the c ontext of growing land scarcity as mentioned above. To embrace this situ ation optimally, it is vital that the dynamics of todays land use, dominated by pasture and catt le, be understood. The following sections discuss these issues. 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 dom inating features throughout the landscape (Table 2.1). Generally, cattle are highly valued within Panamanian society and ranching is an activity that symbolizes wealth. Ranche rs are generally regarded as in fluential 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 stab le investment than many banking ventures, providing salaried sector s of society with a steady, low-ri sk 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 ce rtificate 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, sa laried 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 ot her agricultural e ndeavors, cattle 33

PAGE 34

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 cr edit. When producers do not hold title to their farm, cattle can be used as collateral to obtain bank loans. Sixthl y, cattle are used as a vehicle for land claim. For instance, in regions of unoccupi ed land, forests are cleared for cr op 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 cl aim. As such clearing forest and creating pastures for cattle fulfill two general objectives for squatters and homesteaders: to inhibit the regrowth of fore st, 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 natio nal economy particularly for the rural sector which constitutes half of the Panamanian populat ion. Cattle sales contributed more than $111 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 ac tivity. These figures do not include the contributions of the dairy indus try to the economy in which annual milk sales averaged approximately $30M. In addition, of the 503 corregimientos 1 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. 1 A corregimiento is the smallest politi cal division recognized by the State. For example, corregimientos comprise towns, districts comprise corregimientos, and provinces comprise districts. 34

PAGE 35

Pasture Proliferation The prominence and importance of cattle ranching is reflected in the vast areas occupied by cattle in Panama. Of the 7.5M ha that const itute the country of Panama, approximately 1.5M ha are cattle pasture. Thes e 1.5M ha makeup approximately 20% of Panamas total land mass and 71% of all agricu ltural land in Panama (Censo, 2001). Approximately half of the corregimientos nationwide are covered by 40% or more with pasture and 112 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, an d may have both flat and sloped topography. Changing Nature of Ranching Raising cattle has traditionally been a low-i nput activity. However, certain sectors of the cattle industry are changing due to changes in economic globalization and a fu ture that speaks of the need to have to compete w ith imports. The agricultural sector has received incentives to intensify cattle production. Laws 24 and 25 of 2001, including the Programa para la Reconversin Agropecuaria (Agric ultural Conversion Program), provide low interest loans, reimbursements, and other assistance for farm ers interested in im proving their production techniques. This program is sponsored by the Inter-American Developmen t Bank and part of the effort reimburses farmers on their investment s in advanced agricultural technology. These programs are geared toward large farming enterprises. The goal of these laws is to equip and prepar e farmers to compete with their counterparts in other parts of the world in light of the immi nent 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 th e meat industry particularly in terms of animal genetics, nutrition, management, and investment. These changes have the potential to bear 35

PAGE 36

significant effects on the ecological consequences of cattle ranching particularly in the reduction of the use of extensive pastures. One of the em phases 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 tradi tional system of grazing cattle during 3 5 years on extensive pastures. However, the programs de signed to encourage farmers toward confined fattening (feed lots) programs have not been fru itful. 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 se mi-pastured feedlots. In the past five years farming conditions have begun to change as a result of the oscillating economic situation and government programs g eared 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 improve d grasses). However, these types of changes require costly monetary investments. As a resu lt, 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 extens ively, due to declining productivity and the increased cost of living. This implies that large areas of la nd 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 t hus 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 projected that the change in technology use and intensification may render a marked reduction in small-scale cattle farmers a nd only those farmers able to access credit and 36

PAGE 37

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 traditiona l farming practices; silvopastoral systems may be particularly important in the context of improving tradi tional cattle and pasture management. Conclusion Pre-historic peoples have left a vivid, indelible legacy of fire and savanna-like crop fields on the Panamanian landscape. Introduction of catt le by the Spanish solidif ied 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 th eir entirety by land use managers of the 20 th and 21 st centuries. The nature of land use today pillared by deforestation, pastur e 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 pastur es which in some cases have been worked for centuries and in other cases during millennia exist in vari ous stages of degradation. The research presented in this dissertation wa s 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 Location 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 a nd is located in the corregimiento of Rio Grande, in the Penonom district of the province of Cocl (08.31 o N, 37

PAGE 38

80.21 o W)(Figure 2.6). The corregimiento of Rio Gra nde 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 ar rival of Spanish settlers (Jaen, 1985). Ecology Rio Grande forms part of the dry tropical fore st life zone (as described by Holdridge, 1967) that characterizes Panamas central Pacific flatland s. 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 lowl ands and temperature vari es 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 li es approximately between 0 and 25 masl. Local soil types are classified as chromic luvisols a nd dystric nitosols (ultis ols and alfisols) (FAO, 1972; Nair, 1993). Specifically, Ma tthews and Guzman (1955) clas sify soils in the study site area as pertaining to Chumico sa ndy clay loam. Soil pH ranges from 4.3 to 5.9 and percentage of soil organic matter ranges from 1.61 to 4.02. Climate There are two well-defined climatic seasons on Panamas 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 o ccurs. However, the corregimiento of Rio Grande is situated in a we ll known microclimate called the Arco Seco or dry arc of Panamas central provinces in which a se mi-circular area of the countrys central plains receives less 38

PAGE 39

precipitation than the surrounding area s just a short distance away. Rio Grande receives between 900 to1200 mm precipitation annually. Temperature ranges from 25 to 31 o C. Local Farming Systems In Rio Grande, the dominant agricultu ral activities include growing rice ( Oryza sativa ) and corn ( Zea mays), and raising beef and dual purpose cat tle. Most producers are semi-subsistence in which they produce for household sustenance as well as market a port ion of their products. Although the community is relatively small, ther e are a wide range and diversity of producer types, including: 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 almo st entirely of a subsistence nature, smallholder farmers of crops and cattle w ho 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. Species Descriptions In the experiments presented in this dissert ation, three species of woody perennials were studied. These include Tectona grandis, Bombacopsis quinata and Anacardium occidentale These species were selected by th e farmers who were involved in th e 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 39

PAGE 40

bares, its fruit and nut. The following inform ation 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 dispar ate. According to the available literature, it is apparent that Tectona grandis has been studied and probed mo re extensively than either Anacardium occidentale or Bombacopsis quinata as such the length of each species review is correspondingly unique. Tectona grandis Origin, Natural Habitat, and Environment Teak ( Tectona grandis L.) is native to Southeast Asia and parts of the Indian subcontinent. In the Philippines, it is also regard ed as a naturalized species. Teak occurs naturally as part of an assemblage of mixed forest spec ies in its natural habita t. Although teak occurs naturally in diverse ecological se ttings, moist deciduous forest is regarded as being its original, native habitat (Kadambi, 1972) and develops best on fertile, well drained so ils. In Thailand, teak is found at altitudes between 100 and 1000 masl while in Indonesia, teak occurs in rainfall ranges of 1500 to 2500 mm. However, rainfall for optimu m 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 2 o and 48 o C. Due to the species plasti city in a range of conditions and proven adaptability, it has been planted throughout tropical Af rica, 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 19 th century. Kadambi (1972) notes that experimentation with teak planti ng began in Panama in the 1920s. 40

PAGE 41

Uses Teak gained its worldwide re putation initially as a prized wood due to its excellent performance as a material for shipbuilding. Its hue, texture, and durab ility make it a desired wood throughout the world (Bailey and Harjanto, 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). Botany Part of the Verbenaceae family, teak leaves ar e 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 ma ture 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). Leav es 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, 41

PAGE 42

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 fru it enable teak to prol iferate throughout the landscape. Likewise, teaks well-documented plasti city and adaptability to diverse and, in some cases, adverse conditions have also enabled it s 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 performi ng 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 de monstrated a comparatively high rate of bifurcation. The authors concluded that exotic species were promising; but, for optimal timber production, they required more intensiv e management schemes compared to native species. In a similar study, Piotto et al. (2004) compared the surviv al and growth of 13 native species in mixed and single-spe cies plantations with teak un der dry forest conditions on the Costa Rican Pacific coast. The native species we re equally divided into slow-growth species and fast-growth species. In the slow-g rowing 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 Pseudosamanea guachapele (92%) in a mixed species plantation in terms of survival percentage. After 58 m onths of growth, teak surpassed all slowgrowth species in height and DBH. In comp arison with the fast-growing species, teak was second to S. parahyba in height and DBH. However, de spite these promising characteristics demonstrated in multiple research studies, Pe rez and Kanninen (2005) claim that in Costa Rica 42

PAGE 43

and in several other Central American countries, teak plantations have not reached anticipated levels of productivity. Rooting and Competition Teak in its juvenile stage e xhibits 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 ve rtical roots. Root competition from neighboring vegetation and other teak trees in plantati on conditions markedly hampers teak growth (Kadambi, 1972). Teaks sensitivity to root co mpetition 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 32 P uptake. They found that when 32 P was applied at 25 cm depth, teak uptake of P increased exponentially as lateral distance to bamboo increased. However, when 32 P was applied at 50 cm depth, teak P uptake declined as lateral distance to bamboo clumps increased. Althoug h these two species are well-known for being highly competitive belowground, this study may indicate teaks ability to specialize in upper soil horizon P uptake when f aced 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 surfacerooted species. Burning One way in which plantation owners have sought to ameliorate teaks sensitivity to surrounding vegetation is through burning. Teak is known to benefit from fire. Burning provokes a rejuvenation of tree vigo r, increased growth (height a nd diameter), and in plantation 43

PAGE 44

situations a renewed uniformity within the plantation (Kadambi, 1972). Ultimately, teaks fire hardiness allows it to prevail ove r 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 wild life. Studies and cases of negative impacts of the effects of exotic species abound. In Pana ma, 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 teaks 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 comp arisons of wildlife habitat in Tanzania, Hinde et al. (2001) found teak planta tions 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 natura l areas for wildlife to benefit. As teak plantations were shown to provide habitat for some large mammals, Saha (2001) found no significant difference in plant dive rsity in a comparison study of vegetation composition in a secondary forest (30 to 35 yr) and in a teak plan tation (16 to 18 yr). Overall, for the two land-use types, species richness wa s similar as were seedling density and the abundance of animal dispersed species. However, Saha indicates that the plantations tested possessed dissimilar composition a nd structure in comparison to the secondary forest. 44

PAGE 45

An alternative use of teak plantations may be for carbon sequestrati on and storage. In Panama, Kraenzel et al. (2003) f ound 20 yr teak plantations coul d 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 li tter quantity on nearby pasture was 25 to 30% less than that of surrounding forest and the studied teak plantations. Bombacopsis quinata (syn. Pochota quinata, Bombacopsis quinatum ) Bombacopsis quinata Jacq. (bombacopsis) is a deciduous species native to the Americas ranging from southern Honduras through Columbia a nd Venezuela. It is a large tree known to reach 30 to 35 m in height and 1 to 2 m in diam eter. Bombacopsis requires a defined dry season and occurs in areas of annual precipitation rangi ng from 800 to 3000 mm (Cordero et al., 2003). It grows from 0 to 900 masl 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 pink ish-white and the encapsulated fruits are 4-10 cm long. One of the defining characteristics of bo mbacopsis is its ability to thri ve 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 im portant role in the production capacity and specific gravity of bom bacopsis 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, a nd general carpentry. It is also a highly valued reforestation species for its survival capacity, pest and disease resist ance, and proven growth 45

PAGE 46

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 bomb acopsis 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. Consequen tly, 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 cycl e of 50 yr for plantation-grown bombacopsis (to maximize heartwood content). Anacardium occidentale Cashew ( Anacardium occidentale L.), a member of the Anacar diaceae 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 masl with mean annual rain fall between 600 to 1200 mm. Trees can withstand dry periods of up to 9 m onths and can tolerate infertile, shallow soils (Behrens, 1996). Botanical description 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 in to 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 46

PAGE 47

after initiation while the fruit (peduncle) develops at a slower rate. Fr uit generally requires 70 days to reach maturity (Behrens, 1996). Cultivation Widespread cashew planting is prevalent in India, Brazil, Indonesia, and Tanzania. Cashew is also frequently found on farms thr oughout Mesoamerica. In Tanzania, cashew trees are prevalent on small farms whereas in India they are a popular species used in wasteland reclamation. Major et al. (2005) found cashew to be among the most abundant food species in eastern Amazonian homegardens. The prevalence of cashew can be attrib uted to its hardiness under adverse environmental conditions. Cashews hardiness has been shown to be a pr oduct of its ability to capture resources and withstand drought. For example, in Ghana cashe w tree planting and prod uction 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 pl antations in West Afri ca. 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 se nsing environmental moisture deficiency cashew restricted its water uptake instead of accessing soil mo isture to counter the moisture deficit. The authors concluded that cashew soil water uptake wa s directly related to c limatic conditions rather than soil moisture availability. We may deduce that this apparent mechanism of cashews, to 47

PAGE 48

conserve water reserves during periods of moisture deficit, may aid in cashews noted ability to withstand drought. Studies have also been done in Brazil to investigate the physio logical drivers behind cashews ability to thrive in resource poor c onditions. In a study th at looked at various physiological characteristics of cashew gas exchange, de Souza et al. (2005), like Oguntunde and van de Giesen (2005), found cashew stomata behavi or to be highly influenced by changes in vapor pressure deficit. Prompt stomata closur e in response to high vapor pressure deficit was effective in restricting transpir ation. The authors concluded that cashews ability to quickly and effectively provoke stomata closure lends to cas hews ability to prosper on drier soils. Uses Cashew products and byproducts have a multiplic ity of uses and values. From the world market to rural homegardens (Isaac and Nair, 2005; Major, 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 industr y, 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 biodivers ity restoration as well. In a comparison of single and mixedspecies plantation types in Thai land, Kaewkrom et al. (2005) found that the combination of teak, tamarind ( Tamarindus indica ), and cashew was superior in provi ding habitat for establishment of species from adjacent forests. They found that th e diverse nature and multi-strata shading in the tri-species canopy resulted in a reduction in we eds 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 48

PAGE 49

plantation types. Kaewkrom et al. (2005) also found that the plan tations 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 ma y have played an important role in attracting frugivores thereby potentially enhancing and dive rsifying 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 st orage and release. Building on th is, Isaac and Nair (2005) carried out one of the few studies that examined the dyna mics of cashew leaf litter. They compared cashew, mango ( Mangifera indica ), and jackfruit ( 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 gradua l throughout the 6 months of its decay in which cashew litte r released 97% of its N and K nutrients and 94% of P. With these results, Isaac and Nair (2005) conclude th at 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 us e 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 f eed. The study results s howed that the cashew 49

PAGE 50

bagasse had a positive effect on the nutritive com position 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 ca shew for its soil reclamation capacity. They found cashew trees rapidly improve d the targeted soil physical and biochemical properties, including the improvement of so il electrical conductivity, total organic C conventration, total N, and the reactivation of cer tain 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 Tanz ania that cashew cultural practi ces 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 fight this crop killing disease. The abundance of deposited sulfur in the last decade has resulted in the acidification of soils in Tanzanias 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 acidi fication of these soils. Finally, one important new use of cashew under investigation concerns it s medicinal properties. In Brazil, Medona et al. (2005) studied a range of plant species for th eir ability to kill mosquito larvae. They found that among a range of native species studied, cashew was the most e ffective at killing the larvae of the dengue-spreading mosquito Aedes aegypti. Planting Configuration In Chapter 3, seedlings of the three species de scribed above were plan ted in three different planting configurations, which included plan tings in lines, grouped on a diagonal, and along fences. Investigation of differe nt planting configurations was based on the premise that cattle browse and treading of tree seedlings may occu r differently depending on the organization of 50

PAGE 51

seedlings in the pasture. Prior to the establishm ent of the experiment, participating farmers noted that cattle tended to congregate along fences an d 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 benef it 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-w oodland habitat for the benefit of wildlife. 51

PAGE 52

Table 2-1 Results of effects of Ziziphus joazeiro and Prosopis juliflora trees on buffelgrass pasture in Northeast Brazil. Test as compared to open pasture Ziziphus joazeiroProsopis juliflora 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 crownLower NA Mineralized net N Greater Greater than pasture and Z. joazeiro Crown radiation interception65-70% 20-30% Results by tree species Source: Menezes et al., 2002. 52

PAGE 53

N Figure 2-1 Topographic map of the Panamanian isthmus. Source: NASA-SERVIR (Mesoamerican Re gional Visualization and Monitoring System), http://servir.nsstc.nasa.gov/, 2006. 53

PAGE 54

Figure 2-2 Panama forest cover and areas of deforestation in 1947. Caribbean Sea Pacific Ocean N Forest cover Deforested land Source: ANAM, 1999. 54

PAGE 55

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 1961 1986 1994 2003 Years Millions Forest cover (ha) Permanent pasture (ha) Total agricultural land (ha) Human Population (people) Figure 2-3 Changes in land use and hu man population in Panama 1961-2003. Source: Pagiola et al., 2004; FAOSTAT, 2006. 55

PAGE 56

0.5 19.99 20 49.99 50 99.99 100 199.99 200 499.99 500 < No. of farms Total farm area (ha) 99,160 16,253 7,555 3,282 1,522 402 0 100,000 200,000 300,000 400,000 500,000Farm size categories Figure 2-4 Farm sizes and areas in Panama 2000. Source: Censo, 2001. 56

PAGE 57

Table 2-2 Total farm land, farms with cattl e, and area under pasture in Panama, 2000. Province No. of total farms (1000) No. of cattle farms % cattle farms Total agricultural area (10,000 ha) Total area in pasture (10,000 ha) Proportion of agricultural land in pasture (%) Bocas del Toro4.721,282279.743.6838 Chiriqui48.507,3051542.7924.6057 Cocl31.224,3471425.2410.1540 Colon10.952,1362016.997.6345 Darien5.311,5432923.237.0030 Herrera18.844,5902419.0111.6461 Los Santos17.315,7953430.7623.2075 Panam65.864,526748.6220.1741 Veraguas33.727,6152360.1630.1750 Source: Censo, 2001. 57

PAGE 58

Table 2-3 Economic importance of cattle in Panama by province, 2000. Province Earnings from cattle ($ 1M) Average household monthly income ($) Farmstead population Bocas del Toro 2.63 282.60 23,402 Cocl 6.44 220.60 113,764 Coln 4.41 377.60 36,830 Chiriqu 26.02 302.10 140,909 Darin 5.28 116.50 21,016 Herrera 9.56 249.80 55,743 Los Santos 26.46 235.70 43,684 Panam 13.67 540.40 486,201 Veraguas 16.85 166.90 125,562 Source: Censo, 2001. 58

PAGE 59

Figure 2-5 Proportion of pasture area to total land area by co rregimiento in Panama, 2003. Source: Dagang, 2004. 59

PAGE 60

Research study site Figure 2-6 Research study site location, Rio Grande corregimient o, Cocl province, Republic of Panama. Source: www.lib.utexas.edu/maps/cia00.html 60

PAGE 61

CHAPTER 3 TREE SEEDLING SURVIVAL AND IMPACT OF HERBIVORY ON SILVOPASTORAL SYSTEM ESTABLISHMENT Introduction Finding a balance among food production, in come generation, and environmental preservation is a growing chal lenge. 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 a nd efficiently to maximize production of food, income, and environmental integrity. The study pr esented in this chapter sought to test one aspect, seedling survival and herb ivory, of a land management strategy that intends to increase the productive capacity of the land unit, divers ify 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 catt le production in extensive pastures is the most dominant land use system in the country, and considering the growi ng needs of the human population coupled with the diminishing natu ral resource base, I focused on testing the integration of fruit and hardwood trees into ex tensive, degraded pastures. When designing a study to further develop an existi ng 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 gr azed by cattle. The inclusion of cattle in experimental pastures was made due to farmer in terest, as farmers in Panama are generally not willing to remove cattle fr om their pastures for the establishment of trees. 61

PAGE 62

Literature Review Tree Seedling Survival Some researchers suggest a re lationship exists between seed ling survival and particular seedling characteristics. Through their rese arch 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, smallseeded yellow poplar seedlings ( Liriodendron tulipifera ) exhibited far greater growth than larger seeded species. Also, in a study in Costa Rica examining the effects of li ght gradients on seedlings, Balderrama and Chazdon (2005) relate the importance of size to s eedling 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 w ithin the importance of seedling size and more importantly seedling heig ht, 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 s eedling height. Also, seedlings of Chrysophyllum pomiferum demonstrated a negative relati onship between seedling size and he ight relative growth rate. Factors contributing to survival and growth of seedlings can be diffi cult to generalize and seedling responses in terms of survival and grow th can be species specific (Benitez-Malvido et al., 2005). Beckage and Clark (2003) found sp ecies 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 simila rly across treatments. However, Balderrama and Chazdon (2005) found that responses from different tropical speci es varied more in survival 62

PAGE 63

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 Dipteryx panamensis 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 greate r, significantly positive effect on survival of C. odorata seedlings than on other species in the study. Ramir ez-Marcial (2003) also assessed su rvival of different species in anthropogenic environments and found that speci es growth and response to grazing differed. She found relative height and diameter growth rates of Liquidambar 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 effect s on seedling species is habitat. In fact, Benitez-Malvido et al. (2005) found that the pa sture conditions (temperature, humidity, and soil moisture) in their study, unique to the native forest ha bitat of the seedling species that were studied, may have impeded acclimation of certain species, specifically Chrysophyllum pomiferum and Micropholis venulosa to the area. The authors c ontend 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 conditi ons but not in forest. Another relevant factor when considering seedling survival is the effect of existing vegetation on seedling establishment. In a Ha waiian forest, Denslow et al. (2006) found that 63

PAGE 64

existing vegetation severely constrained woody s eedling establishment. Presence of grasses impeded growth of the species Acacia koa Sophora chrysophilla and Dodonea viscosa. Sanchez and Peco (2004) also suggest that presen ce 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 function s 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 (Melinis minutiflora ) permitted significantly greater colonizati on and growth of woody individuals than kikuyugrass ( Pennisetum clandestinum ). The authors suggest that the stoloniferous growth habit of P. clandestinum created a physical barrier that inhi bited seed germination and seedling establishment of woody perennials. Similarly, seedlings that ach ieved germination within the stolon mat suffered due to low light a nd mechanical damage by fast growing P. clandestinum grass shoots. In contrast, the bunch grasss M. minutiflora allocated less biomass to stolons and had more open surface area between plants wh ich 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 ma nagement 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 distur bance, 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 seedli ngs, in which cattle impacted seedling growth and survival by 64

PAGE 65

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 herb icides caused the greatest mortality among all seedling treatment combinations in the study. Ev ans 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 sp ecies. Ganskopp and Bohnert (2006) also suggest that cattle will select for hi gh quality forage and that cattl e 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 fo r 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 es tablishment of woody perennials. Although the research discussed above indicates potential negative effects of cattle grazing on woody perennial establishment, some studies sugge st that the presence of cattle can in fact benefit seedling survival. Posada et al. (2000) pr opose the notion that grazi ng can serve as a tool for the regeneration of forests on abandoned pastur es. 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 fre quently occur in disturbed areas. Establishment of such species, they note, can lead to the shading out of gra sses and the creation of su itable microclimates for forest species establishment. Other studies c onclude similarly. For ex ample, in a study in Argentina de Villalobos et al (2005) found that grazing may be nefit woody perennial seedling 65

PAGE 66

survival. They contend that grazing cause d a reduction in grass biomass aboveand belowground, potentially increasing surface soil moisture and thereby enhancing woody seedling establishment and growth. In contrast with St ammel et al. (2006), de Villalobos et al. (2005) contend the creation of gaps by cattle treading may induce peri odic 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 nutrients. Herbivory Leaf-cutter ants ( Atta spp.) are an abundant invertebra te species in tropical ecosystems (Jaffe and Vilela, 1989) and they function as important selective herb ivores throughout the Neotropics (Rao et al., 2001). These herbivores can have a trem endous impact on the landscape. Leaf-cutter ant herbi vory can reduce plant reproductive potential through decreased seed production and result in reduced seedling survivor ship (Vasconcelos and Cherrett, 1997). In addition, leaf-cutter ants prefer young leaves over mature leaves thereby hindering regeneration. Leaf-cutter ants manifest prefer ence for particular species. Ra o et al. (2001) found decreased density of adult trees of pref erred 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 natu ral tropical landscapes has been shown to increase the density of le af-cutter ant nests (Jaffe and Vilela, 1989). Impact by Atta spp. has been observed to heighten within human-influen ced 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 66

PAGE 67

defense mechanisms is low and may be highly dispersed in forests in comparison to humanaffected environments. They argue that the dive rsity of forest vegetation makes ants susceptible to poisonous plants and c onsequently may subdue the Atta population. Secondly, the authors contend that Atta nests require exposure to sunlight. Th is 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 conditi ons may be advantageous for increases in nest density. Therefore, they propose that proliferat ion of human-affected landscapes decreases non-desirable plant abundance and concomitan tly increases le af-cutter ant nest density. For example, Terbor gh et al. (2006) also examined l eaf-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 la ndscape can be cross-cutting and intense. Detrimental consequences due to herbivory can o ccur 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., 200 5). Variables such as habita t, seedling height, herbivory intensity, pathogens, competition, and seedli ng non-structural carbohydrate reserves can influence seedling response to herbivory (Benit ez-Malvido et al., 2005). For example, according to Allcock and Hik (2004), habitat played a pi votal role in the response of seedlings to mammalian herbivory in an Australi an grassland. In their study, s eedlings exposed to herbivores 67

PAGE 68

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 he rbivory. Seedlings were able to recover from herbivory more quickly due to the potentially highe r resource habitat in the grassland, especially regarding light availability. On the other hand, the slower grow th rates and recovery time of seedlings in the woodland habitat placed seedlings at greater ri sk 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 comp ound the risk of repeated herbivory and eventual mortality, Haukioja and Korichev a (2000) note that the breaking of apical dominance due to herbivory can result in vigorous vegetative growth leading to hi gher 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 sl ow height growth of browsed P. sylvestris seedlings caused them to remain in a size range susceptible to herb ivores 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 he rbivores. Hester et al. (2004) found that Betula pendula and Sorbis aucuparia seedlings responded better to simulated browsing than P. sylvestris with increased biomass aboveand belo wground. Likewise, Allcock and Hik (2004) found that Eucalyptus albens seedlings experienced greater survival than that of Callitris 68

PAGE 69

glaucophylla due to Eucalyptus ability to rebound from herbivory through hastened resprouting. The authors s uggested a decline in the C. glaucophylla population would occur if sustained grazing were to o ccur in the study site. Herbivory intensity and energy reserves may also play an important role in seedling response to herbivory. In an expe riment using seedlings species (Acer rubrum Acer saccharum Quercus rubra, and Prunus serotina) from the US Northeast, Canha m 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 comple te leaf removal, survival declined sharply. They suggested survival was closely tied to to tal carbohydrate reserves a nd concentrations of carbohydrate reserves. Although eff ects of defoliation on carbohydrate reserves were consistent across species, consequences for survival differe d by species. Rao et al. (2001) concurred in their conclusions that if seedlings are able to persist through the sa pling 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, Hauki oja and Koricheva (2000) in their comparison of woody perennials and herbs concurred that pl ant regrowth following herbivory is dependent on energy and nutrient storage; however, they emphasize that such storage must occur in unthreatened plant organs when he rbivory is a factor. Being th at mature woody perennials store a small proportion of their biomass in leaves (in comparison with herbs), Haukioja and Koricheva (2000) concluded that woody plants we re better suited than herbs to withstand herbivory. Just as response to herbivory by seedlings can be species-specific, so may herbivores maintain preferences for particular sp ecies (as noted to be the case with Atta spp.). Hester et al. (2004) contend that herbivore choice can be affected by a mu ltitude of factors, including 69

PAGE 70

individual location, plant morphology, plant chem ical 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, herbivor y, and recovery from herbi vory are intricate processes which, according to the research, seem to be imp acted by a range of tree species, herbivore, and habitat characteristics. Species ch aracteristics such as seed size, se edling height, and architecture seem to play important roles in a species ability to survive. These char acteristics coupled with variations in habitat including light availabilit y, moisture, and existing vegetation can result in differences in seedling survival. Similarly, seed ling herbivory can also have important impacts on survival. Herbivore preferences can have part icularly 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 carbohydr ates, as well as habitat conditi ons and herbivory intensity. Considering that research suggests seedling surviv al, 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 impor tant tree species used in agricultural systems in Panama. Objectives and Hypothesis The objective 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 s eedlings by cattle) and treading 70

PAGE 71

would play an important role in seedling su rvival and that seedling species would be a determining factor for survival and herbivory. Methods and Materials Study Site This study was conducted on five farms in the Rio Grande corregimiento of Cocl 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 appr oximately 0.5 to 1.0 animal unit per ha (one animal unit = ~ 270 kg). Experimental Design A randomized complete block design was used. There were five blocks; one block on each farm. Each block contained a complete set of tr eatment combinations which comprised a total of 135 seedlings. There were nine treatment combinations consis ting of three sp ecies 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 each pasture. Materials The three tree species were chosen by farm ers participating in the study. The seedlings were acquired through local nurseries. A. occidentale and B. quinata seedlings were approximately 180 days in age and measured approxi mately 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. 71

PAGE 72

Establishment 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. Measurements Seedlings were surveyed weekly for two years. They were observed for mortality and herbivory. We recorded mortalit y, 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. Ca ttle and leaf-cutter ant effects were distinguished visually. Herbivory was determined when a portion of a se edling had been removed. Source of herbivory was also categorized into cattle, leaf-cutter ant, and other and were also distinguished visually. Data Analysis Statistical analyses were performed using SPSS. A survival analysis was conducted to analyze the seedling mortality and cause of mo rtality 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 regr ession analysis. SAS was used to analyze herbivory data. A two-way analysis of variance was conducte d. Tukeys 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 herbivory data. 72

PAGE 73

Results Seedling Survival During the two years of the expe riment, the survival analysis revealed 25 0 of a total of 675 planted seedlings survive d, a survival rate of 37%. Survivor ship was significantly affected by the planting configuratio n, species, and planting c onfiguration x species in teraction treatments. The Log Rank test revealed signif icant differences in the survival curves across configuration (p < 0.001), species ( p < 0.001), and planting configur ation x species interaction ( p < 0.001) (Table 3-1). The survival analysis for species reveals so me insight into species performance. For example, much of the total mortality (70%) acro ss species that occurred over the life of the experiment occurred by day 300 (73% of A. occidentale, 65% of B. quinata and 73% of T. grandis) (Figure 3-1). Likewise, the three species experienced mortal ity in a similar pattern, in two large events during the first third of the expe riment and in smaller increments toward the end of the experiment (Figure 3-2). Also, across sp ecies, of those seedlings that died, 27% were A. occidentale 35% were T. 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, respectively. When examining the results of the interactions between spec ies and planting configuration, the survival analysis re veals that in the diagonal configurat ion, species performed significantly different ( p < 0.001). There were a total of 127 seedli ng deaths in the diagonal configuration which included 19 mortality cases for cashew, 64 for tropical cedar, a nd 44 for teak. Mean survival time for seedlings planted in the diagonal configuration was 500 days. Similarly, 170 seedling deaths occurred in the fence configura tion consisting of 57 mortality cases for cashew, 73

PAGE 74

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 reveal ed that within the fe nce 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 surv ival time for seedlings planted in the line configuration was 572 days. The line configura tion had a significant effect on survival ( p = 0.003). The different patterns in which seedling speci es 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 domin ant observed causes of seedling mortality. Of the total 425 seedlings that di ed, 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 o ccurred due to cattle e ffects expired during the period of 210 287 days. Herbivory The effects of species and plan ting 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 twoway interaction was obtained when examining the configuration x species interaction, p < 0.0001. However, contrary to survival, the main effect for configurati on was not significant. 74

PAGE 75

Using the Tukey hsd test, significant differe nces 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, T. grandis was browsed significantly more than the other two species. Finally, in the line configuration, B. quinata experienced significantly more herbivory than the othe r two species (APPENDIX B). Sources of Herbivory Three categories of sources of herbivory were recorded, including cat tle, 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 l eaf-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 add ition, when using the chi-s quare test, there was a significant effect for speci es on source of herbivory, p < 0.05. Also, the Goodman and Kruskal Tau test was significant for the spec ies effect on source of herbivory, = .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 wa s significant for configuration at = .01, p < 0.05. Relative to source of herbivory as an outcome, lin e 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). Discussion Seedling Survival 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 tim es over the life of the study. Hi gh seedling mortality took place 75

PAGE 76

relatively early (1-60 day) while moderately hi gh 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 precis ely with the local dry season when rainfall can drop below 13 mm per month (Murphy and L ugo, 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 subject to the typical moisture deficits of this period that are re portedly experienced in natural settings, in addition to th e added burden of likely forage-dep rived cattle. However, it is relevant to note that thes e conditions were not directly measured in this study. The species treatment was significantly differe nt across the seedling mortality survival curves and, overall A. occidentale experienced the gr eatest 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 surviv al. Also, its ability to persis t 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 th e first year of the experiment but flourished during the second year. Similarly, T. grandis also suffered less mortality than B. quinata T. 76

PAGE 77

grandis leaves are less brit tle 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 ha ve been key to their survival in terms of planting configuration. This was reflected in th e significant effect plan ting 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 configura tion treatments (3 m x 3 m), seedlings in the fence treatment suffered most such that there were no significant di fferences among species planted along fences. It is lik ely that seedlings in the fence treatment were subject to more frequent cattle presence and tr eading due to the more abundant shade (where cattle tend to congregate) that occurred along fences in comp arison 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 r oots 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 compe tition between large trees and seedlings was not directly measured in this study. Despite lower total survival in the fence tr eatment, 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 fen ces in itself may have provided an advantage to seedlings already negativel y affected by on-going cattle presence, shade, and competitive effects of nearby large trees. Likewise, the re moval of competing vegetation during a critical 77

PAGE 78

period such as the dry season when available so il moisture is reduced may have had an even more dramatic, important effect on seed ling survival in the fence treatment. Line and diagonal treatments may have been s ubject 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 form ed alleyways between seedling rows which may have facilitated cattle movement through the configurations and potentially reduced cattle treading and consequent seedling damage. In a ddition, 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 togeth er; 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 majority 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 weed ing and being run over by a machine, it seems presumable that the other mortality cases woul d have occurred more or less during the same time period as weeding and pres ence 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 dur ing the first four months of the lengthy dry season. There may have been two different dynamics behind the seedling mo rtality during this period. First, available 78

PAGE 79

forage for cattle is scarce during the dry season pa rticularly 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 scarci ty. 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 mois ture particularly while competing with longestablished pasture grasses. Therefore, given the presumably weakened status of seedlings during the dry season coupled with often amplifie d cattle effects such as grazing and treading, it is not unexpected that mortality would heighten particularly due to cattle during this period. Herbivory In contrast to seedling survival, seedling he rbivory was significantly affected by species but was not significantly affect ed 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 re sult 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 syst ems 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 releva nt driver of seedling herbivory 79

PAGE 80

than is configuration. These br oad, potential insights will bear upon the ultimate purpose of this research, i.e., to aid farmers in decisionmaking re garding the establishment of dispersed trees in pasture and the creation of appr opriate 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 gr een leaves. However, it is noteworthy that herbivory of B. quinata occurred most given that the seedlings in the study experienced leaf senescence and, ge nerally, 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 l eaves 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 a ppetizing to the obser ved herbivores. This condition was likely a de terrent 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 of A. 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 s eedlings being browsed less than B. quinata When examining the results of the post hoc te st 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 T. grandis herbivory was significantly greater in the fence configuration than the 80

PAGE 81

other species. Although leaf growth data were not r ecorded, 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 domin ate available resources in comparison to other species, it is possible that T. grandis was able to capture shade-induc ed 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 greate r leaf presence in comparison to the other species particularly duri ng periods of moisture scarcity. However, it is important to note that soil moisture and leaf growth pa rameters were not measured in this study. Sources of Herbivory Similar to the survival study, it was evident th at 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 fo rest environments (Ramirez-Marc ial, 2003). In the case of pasture, cattle herbivory can l ead 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-cut ter ants browsed B. quinata seedlings most often, for the same reasons mentioned above palatability and texture. In contrast, the undesirability of A. occidentale by the leading herbivor es (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. Li ke the survival data, cattle were the greatest overall browsers of seedlings, particularly in th e line configuration. It is not understood why 81

PAGE 82

planting configuration may have had a significant effect on herb ivory. 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 congregate. Conclusion Tree-seedling survival is shown to be highly responsive to changes in season, herbivore (cattle) presence, tree speci es characteristics, conf iguration, and possibly pr oximity to large trees (in the case of the fence configurat ion). 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, in dicating that if producers can find the means to support the survival of planted seedlings through this period, the total pr oportion of surviving seedlings could be greater in the long-term. Ca ttle were the overwhelmi ng predators of seedlings and, if seedling survival is a farmer priorit y, 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 tr ee species is key to seedling surviv al 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 fen ce treatment the most mortality occurred and seedling lifespan was shortest; however, the fence configuration experienced the least amount of herbivory. 82

PAGE 83

The varied results of this experiment are indi cators of the delicate ba lance that occurs in natural systems. Although human-induced syst ems are often characterized as being less biologically diverse and complex than naturally occurring systems, it has become evident through this study that the integrat ion of silvopasture into pasture systems is in fact complex. The complexity lies in the many factors the syst em comprises: trees, grasses, and livestock; however, complexity is heightened by competiti on 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 consid erations coupled with the present need to augment the production capability and environmenta l integrity of agricultural systems, it is important that research on sil vopastoral systems be intensified. 83

PAGE 84

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 Cocl, Panama. Source Nparm DF 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 84

PAGE 85

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Surviving 0 100 200 300 400 500 600 700 800 900days Diagonal 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Surviving 0 100 200 300 400 500 600 700 800 900days Fence 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Surviving 0 100 200 300 400 500 600 700 800 900days Line Anacardium occidentale (cashew) Bombacopsis quinata (tropical cedar) Tectona grandis (teak) Figure 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, Cocl province, Panama. 85

PAGE 86

0 10 20 30 40 50 600 60 120 180 240 300 360 420 480 540 600 660 720 780 840Time (days after planting)Terminal events (#) A. occidentale B. quinata T. grandis Figure 3-2 Incidence of mortality among Anacardium occidentale, Bombacopsis quinata, and Tectona grandis seedlings planted in three planti ng configurations for silvopastoral system establishment in farmers fields in Rio Grande, Cocl, Panama. 86

PAGE 87

0 100 200 300 400 500 600 700 800 A. occidentale B. quinata T. grandis Species Incidence of herbivory Cattle Leaf-cutter ants Other Figure 3-3 Incidence of herbivory of three sp ecies of tree seedlings (N = 225 seedlings per species) browsed by cattle, leaf-cutter ants or other observed sources during a twoyear experiment in grazed on-farm pastur es in Rio Grande, Cocl, Panama. The yaxis (Incidence of herbivory) refers to the number of events when seedlings were impacted by herbivores. 87

PAGE 88

0 50 100 150 200 250 300 350 400 450 500 Diagonal Fence Line Incidence of herbivory Cattle Leaf-cutter ants Other Figure 3-4 Incidence of cattle, l eaf-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, Cocl, Panama. The y-axis (Incidence of herbivory) refers to the number of events wh en seedlings were impacted by herbivores. 88

PAGE 89

CHAPTER 4 EFFECTS OF SCATTERED LARGE TREES IN PASTURES ON A Hyparrhenia rufaDOMINATED MIXED SWARD Introduction To be able to promote the implementation and use of silvopastoral systems with certainty, it is imperative that the dyna mics of the systems and their parts be understood. Garnering knowledge of interactions in si lvopastoral systems is of par ticular importance due to their complexity, as they comprise multiple, multi-di mensional components including trees, crops, and livestock. Within the context of seeking to understand diverse biophysic al interactions of silvopastoral systems as a means to work towa rd 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 fo rage mass, digestib ility, and composition along a gradient of distan ces from mature trees. Literature Review Light A debate abounds concerning the effects of light on forage growth in tr ee-pasture systems. Belsky (1994) proposed that light is not a primary factor in the gr owth of perennial species under trees. She found that the environmental condit ions 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. Ar es et al. (2006) also contended 89

PAGE 90

that overstory shade was not a prominent factor affecting forage produc tion under large native pecans ( Carya illinoinensis ) in Kansas, USA. Rather, they attr ibuted fluctuations in forage yield to changes in local climatic condi tions. 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 were similar. However, disparity exists in this debate. Some researchers conclude th at 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 availa bility of light in edge sites facilitated grass growth. Similarly, in their research on a mi xed 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 attri buted 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 effect s of changes in light may vary by forage species. For example, in the study ca rried out by Douglas et al. (2006), white clover ( Trifolium repens ) was significantly more abundant in ope n 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 90

PAGE 91

growth was limited as a result of the interception of light by the overstory canopy. They found that as pine stocking rate incr eased, grass growth decreased. Biomass Allocation Consistent with the differing results of the e ffects of light on tree-p asture systems, some research has looked closer at pl ant responses to diminished light availability in silvopastoral systems. Specifically, changes in grass al location to above vers us belowground biomass consequent to changes in availa ble light have been examined. Fernandez et al. (2004) examined the changes in biomass alloca tion 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 dens ity of 500 pruned pine trees per ha, radiation was reduced by 75%. They proposed that the fo rage species changed its biomass allocation pattern in response to shading: allocation to storage roots was re duced 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 l eaf 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 produ ction in forage grasses. Belowground Factors Considering the potential effects of reduced light availability on pasture grasses under trees, Rietkerk et al. (1998) suggest that a tr adeoff exists between light availability and soil nutrient availability in that although light in the understory often becomes reduced due to 91

PAGE 92

shading by the overstory canopy, trees may confer beneficial effects on understory conditions and vegetation. Silva-Pando et al. (2002) propose d 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 (Clas on, 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 implyi ng 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, Dougl as et al. (2006) and Fernande z et al. (2006) found soil water availability to be less under trees than in ope n pasture. They both s uggest that rainfall was captured by trees in the overstory thereby limiting soil moisture content, and consequently, moisture availability to understo ry vegetation. Also, uptake of wa ter by tree roots might play an important role in limiting the ava ilability of moisture belowground. However, Fernandez et al. (2004) only found a disparity in so il moisture availability between open pasture and under trees during periods of high moisture av ailability, at which time gra sses under trees had better water status than grasses in open pasture. The aut hors 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 deba te on which aspects of tree-forage interactions ultimately determine outcomes: light may or may not be a factor, climate, soil moisture, species92

PAGE 93

specific traits, and tradeoffs of light reduction and buffering of ex treme conditions are considered to play some type of role in impac ting characteristics of understory forage. Objective and Hypothesis The objective of this study wa s to evaluate and compare th e impacts and consequences of large, dispersed trees in pasture on the characteristics of Hyparrhenia rufa -dominated forage growing in mixed swards in degr aded pastures. Characteristics included forage growth, in vitro organic matter digestibility, and forage compos ition 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 Study Site This study was conducted in the sectors of La Calendaria and Los Olivos, Rio Grande corregimiento, Cocl province, Panama (see Chapter 2 for specific local and regional characteristics). Data were gathered from past ures 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 th roughout the pastures. In the wet season, cattle stocking rate averaged 0.5 to 1.0 AU per ha. Experimental Design The study consisted of two similar experiments. These experiments were structured as randomized complete block designs. Each experi ment was alike except for the tree species that was used; one experiment used Anacardium occidentale and the other experiment used Tectona 93

PAGE 94

grandis. All experimental design asp ects 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 di stances) of the radius of each tree canopy (APPENDIX C). Forage samples were harveste d randomly within the context of corresponding direction and distance from the tr ee stem, yielding twelve destruc tive samples per tree, for both experiments. Sampling of forage mass, diges tibility, and bota nical composition occurred in May and September of 2002, in May and December of 2001, and in December of 2001, respectively. Measurements 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 site s and all herbage within the ring was harvested manually (by machete and hand clippe rs) to ground level. The forage fresh weight was recorded. To evaluate in vitro organic matter digestion (IVOMD), herbage was bagged and oven-dried at 60 o C. Dried samples were ground and milled th rough a 1 mm screen. In vitro organic matter digestion was performed by a modification of th e two-stage technique (Moore and Mott, 1974). To assess composition, fresh samples were air drie d and separated by hand into pre-established categories of grass, weeds, legume, and necrom ass. 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 categor ized as those plants w ith characteristics that 94

PAGE 95

resembled the Fabaceae family. Necromass was a ll biomass identified as dead material. After forage categorization, sample s were bagged and weighed. Data Analysis 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 signifi cant, Tukey hsd post-hoc test wa s used to compare means. Orthogonal polynomial contrasts were used to describe the effect of location. Results Forage Mass 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 signi ficant main effect for the distance variable ( p = 0.76) in the case of A. occidentale. There was no significant linear or quadratic effect of distance on mass or its interaction with the seas on 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 te st, we observed that th ere 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 sign ificant linear effect ( p = 0.001) of distance, but the quadratic e ffect 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, th ere was no significant main effect for season ( p = 0.926) under T. grandis 95

PAGE 96

Forage Digestibility Forage digestibility under A. occidentale was affected by distance and season ( p = 0.042 and p < 0.001, respectively) but there were no intera ctions. The post hoc test revealed that forage digestibility was significantly greater at the farthe st distance from the tree stem (2.0) than at the 0.5 distance (close to the tree stem) while the drip line (1.0) a nd 0.5 distances were not significantly different. In addition, in the post ho c analysis of the seas on 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). Forage Composition Under A. occidentale, there were no treatment effects on forage botanical composition. Likewise, under T. grandis the effect of distance on weed s, grass, and legume was not significant. However, results for necromass under T. 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 necr omass 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). 96

PAGE 97

Discussion Forage Mass Effects of trees on understory forage can vary by season, climate, and soil conditions. In this research, forage mass was affected by di stance 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 th e 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 th e 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 ma ss was significantly higher under A. occidentale during the wet season than in the dry season, forage mass did not differ significantly under T. 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 a ffect on forage mass under the two tree species. The consistency of forage mass abundance under T. 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 T. grandis while wet season conditions induced a suppressive effect on forage growth under T. grandis; or 2) based on 97

PAGE 98

the consistency of forage mass abundance across season, T. 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 gr owth habit. During the dry season, T. grandis was completely deciduous. At this tim e, 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 resource s and light may have provided the forage under T. grandis with increased access to resources, leadin g to increased growth and accumulation of forage mass during this period. However, it is also plausible that the c onsistency of overall lo w forage mass abundance under T. 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 conse quent increase in resour ce availability to the tree. However, it is important to note that these parameters were not dir ectly measured in this investigation. The contrasting results of forage growth under A. occidentale and T. grandis emphasize the difference in effects of indivi dual 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 98

PAGE 99

herbage under A. occidentale in both the wet and dry seas ons. Higher yielding forage performance under A. occidentale and the apparent suppressi on of forage growth under T. grandis further accentuates the distinct eff ects 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 fo rage 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, decr eased forage abundance closer to the T. grandis tree stem broadens the argument regarding th e aggressive character of this tree species. This is also emphasized by the lack of distance effect of A. 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 pa sture during these periods is exposed to temperature, evapotranspiration, and moisture fluxes. These e ffects 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 dr ip line. Also, root systems and belowground performance can differ by spec ies. Rooting ability, root length, root architecture, and biomass allocation to roots can determine species effectiveness at acquiring and 99

PAGE 100

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 effec tive root systems may be better equipped to outcompete grasses at the drip line a nd 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 al lelopathy on conditions around trees, and how these can differ by tree specie s and thereby impact fo rage. 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 cons ider how cattle may impact forage around these species. A. occidentale is an abundant producer of large, nutri tious fruit which attracts cattle to its immediate surroundings. Also, A. occidentale commonly possesses a glo bular, densely-leafed canopy which casts cool shade, frequently pursue d 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 th e 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 expe riment). In this experiment, T. grandis trees possessed a conical canopy shape which did not produce shade that wa s 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. Forage Digestibility 100

PAGE 101

Forage digestibility proved to respond to oversto ry presence of trees di stinctly from forage mass. In contrast to forage mass, distance was relevant to A. occidentale but not to T. grandis while season played a role for both species, also unlike the case of forage mass. Overall, it is to be expected that forage digest ibility would decrease in the dr y season in comparison to the wet season due to desiccation. The season effect was consistent along the thr ee gradients of distance from the tree stem for both species, emphasizing the decisive effect seasonal conditions have on forage digestibility. For both seasons, digestibility of forage growing under T. grandis was consistently and notably greater than that under A. occidentale (Figure 4-2). It is possible to attribute the differences in digestibility of the two forages at the 0.5 and 1.0 distances to contrasts in light availability. This suggestion is based on the assumption that due to its dense evergreen canopy (Behrens, 1996), A. occidentale limited understory light av ailability more than T. grandis did. When light availability becomes restricted, seve ral changes can occur in plant characteristics such as decreases in plant non-structural car bohydrates, increases in cell wall content, and increases in internode length. Such shifts in pl ant characteristics, provoked by reductions in light availability, can result in decrease s in the digestibility values of forage (Lin et al., 2001). Forage Composition For each species, distance to tree was tested for their affects on the four forage components: grass, weed, legume, and necroma ss. Composition of the forage differed by tree species in which composition under A. occidentale varied notably in each of the categories. Forage under A. occidentale comprised more legume overall than did forage under T. grandis alluding to a benefit for cattle. However, the forage composition under T. grandis was highly uniform (Figure 4-3) across distances. Di stinct proportional composition of forage under A. occidentale and T. grandis suggest a potential effect of tree species on forage performance in the 101

PAGE 102

context of composition (Figure 4-4, Figure 4-5). Moreover, such marked consistency of the forage contents across parts under T. grandis alludes to the possible lack of relevance of distance from the tree to forage composition in the case of this species, particularly in the case of the proportions of weeds, grass, and legume as part of the total forage composition. However, distance had a signifi cant effect on necromass under T. grandis This contrast in results as compared to grass, legume, and weeds is quite noteworthy as it is an indicator of a clear forage response to tree pr esence. The differences in necromass by distance may be an indication of an interaction between forage a nd trees. Creation of a microclimate beneath and around trees on pasture, includi ng buffering of temperature and reduced evapotranspiration under the tree canopy may maintain forage vigor a nd prevent the generation of necromass. Conclusion This study has shown that effects of large tr ees on pasture can be species-specific and variable, and therefore s hould not be generalized. The most interesting element of the results presented here is the lack of cons istent impact of the distance vari able: distance to tree stem did not have a constant, clear main effect on mass, digestibility, or composition of the forage underneath. This may indeed indicate that the s imple presence of an isol ated tree on pasture is not the determining factor when considering the consequences for effects on forage. Rather, it was season and species that exerted a more prominent influence on variation of forage characteristics. The greater relevance of season and species is key as Panamanian producers tend to dislike the presence of trees in pastures because they believe that trees, rega rdless of species, have universal negative effects on forage. In light of the diverse results of th is study, it is imperative that research on effects of particular tree species on pasture be continued, in order to formulate a 102

PAGE 103

working framework of recommendations to supp ort farmer decisionmaking in silvopastoral establishment and management. 103

PAGE 104

Table 4-1 Analysis of variance for polynomial or thogonal contrasts of sample mean forage mass comparing the effects of dist ance and season under dispersed Anacardium occidentale trees in Rio Grande, Cocl, Panama. Source Distance Type III Sum of Squares df Mean Square F Sig. Partial Eta Squared distance Linear 10513.92 1 10513.92 3.444 0.077 0.135 Quadratic 2595.903 1 2595.903 1.494 0.235 0.064 distance x season Linear 1759.341 1 1759.341 0.576 0.456 0.026 Quadratic 389.404 1 389.404 0.224 0.641 0.01 Error (distance) Linear 67157.789 22 3052.627 Quadratic 38228.803 22 1737.673 104

PAGE 105

Table 4-2 Analysis of variance for polynomial or thogonal contrasts of sample mean forage mass comparing the effects of dist ance and season under dispersed Tectona grandis trees in Rio Grande, Cocl, Panama. Source Distance Type III Sum of Squares df Mean Square F Sig. Partial Eta Squared distance Linear 18840.317 1 18840.317 12.161 0.001 0.269 Quadratic 1479.516 1 1479.516 2.909 0.097 0.081 distance x season Linear 1768.478 2 884.239 0.571 0.571 0.033 Quadratic 2621.69 2 1310.845 2.578 0.091 0.135 Error (distance) Linear 51124.044 33 1549.213 Quadratic 16781.606 33 508.534 105

PAGE 106

Table 4-3 Post hoc comparisons of mean forage mass at three distances 1 from dispersed T. grandis tree stems in grazed, degraded pastur es in Rio Grande, Cocl, Panama. 95% Confidence Interval for Difference (i)Distance (j)Distance Mean difference (i-j) Std. Error Sig. Lower bound Upper bound 0.5 1 -8.325 4.332 0.178 -19.219 2.569 2 -32.353* 9.277 0.004 -55.685 -9.02 1 0.5 8.325 4.332 0.178 -2.569 19.219 2 -24.028* 8.164 0.018 -44.56 -3.495 2 0.5 32.353* 9.277 0.004 9.02 55.685 1 24.028* 8.164 0.018 3.495 44.56 1 The distances were: 0.5 = close to tree, 1.0 = drip line, 2.0 = open pasture. 106

PAGE 107

Table 4-4 Post hoc analysis of forage digestibil ity across three distances from dispersed Cashew trees (A. occidentale ) and by two seasons in grazed pa stures of Rio Grande, Cocl, Panama. 95% Confidence Interval for Difference Distance (i) Season (j)Season Mean difference (i-j) Std. Error Sig. Lower bound Upper bound 0.5 Dry Wet -6.201* 1.271 0.000 -8.836 -3.566 Wet Dry 6.201* 1.271 0.000 3.566 8.836 1 Dry Wet -3.004 1.616 0.076 -6.356 0.347 Wet Dry 3.004 1.616 0.076 -0.347 6.356 2 Dry Wet -6.898* 1.776 0.001 -10.582 -0.3213 Wet Dry 6.898* 1.776 0.001 3.213 10.582 107

PAGE 108

Sample mean forage mass (g)Season Distance Cashew Teak 72.88 142.98 89.43 181.51 90.37 184.69 0 50 100 150 200 DryWetDryWetDryWet 5050100100200200 78.08 88.90 65.59 74.68 56.33 59.20 0 50 100 150 200 DryWetDryWetDryWet 5050100100200200 Sample mean forage mass (g)Season Distance Cashew Teak 72.88 142.98 89.43 181.51 90.37 184.69 0 50 100 150 200 DryWetDryWetDryWet 5050100100200200 78.08 88.90 65.59 74.68 56.33 59.20 0 50 100 150 200 DryWetDryWetDryWet 5050100100200200 Figure 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, Cocl, Panama. 108

PAGE 109

TeakMean forage digestibility (%)Season Distance Cashew 26.26 27.82 28.38 35.28 30.82 32.46 0 5 10 15 20 25 30 35 40 Dry Wet Dry Wet Dry Wet 50 50 100 100 200 200 30.00 30.92 36.01 31.03 35.12 35.33 0 5 10 15 20 25 30 35 40 Dry Wet Dry Wet Dry Wet 50 50 100 100 200 200 TeakMean forage digestibility (%)Season Distance Cashew 26.26 27.82 28.38 35.28 30.82 32.46 0 5 10 15 20 25 30 35 40 Dry Wet Dry Wet Dry Wet 50 50 100 100 200 200 30.00 30.92 36.01 31.03 35.12 35.33 0 5 10 15 20 25 30 35 40 Dry Wet Dry Wet Dry Wet 50 50 100 100 200 200 Figure 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, Cocl, Panama. 109

PAGE 110

0.000.200.400.600.801.001.20 0.5 1 2 0.5 1 2 A occidentale T. grandis Proportion of forage composition Weeds Legume Grass NecromassDistance 0.000.200.400.600.801.001.20 0.5 1 2 0.5 1 2 A occidentale T. grandis Proportion of forage composition Weeds Legume Grass NecromassDistance Figure 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, Cocl, Panama. 110

PAGE 111

5.77 6.34 5.16 0 5 10 15 20 25 30 35 40 25.36 25.52 22.56 0 5 10 15 20 25 30 35 40 2.99 4.14 2.88 0 5 10 15 20 25 30 35 40 28.27 30.09 21.35 0 5 10 15 20 25 30 35 40 0.5 1 2Sample mean forage weight (g)Distance Weeds Grass Legume Necromass 5.77 6.34 5.16 0 5 10 15 20 25 30 35 40 25.36 25.52 22.56 0 5 10 15 20 25 30 35 40 2.99 4.14 2.88 0 5 10 15 20 25 30 35 40 28.27 30.09 21.35 0 5 10 15 20 25 30 35 40 0.5 1 2Sample mean forage weight (g)Distance Weeds Grass Legume Necromass Figure 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, Cocl, Panama. 111

PAGE 112

12.46 8.64 9.97 0 5 10 15 20 25 30 35 40 14.34 19.73 16.97 0 5 10 15 20 25 30 35 40 12.08 7.56 9.42 0 5 10 15 20 25 30 35 40 35.88 37.53 20.34 0 5 10 15 20 25 30 35 40 0.5 1 2Weeds Grass Legume NecromassSample mean forage weight (g)Distance 12.46 8.64 9.97 0 5 10 15 20 25 30 35 40 14.34 19.73 16.97 0 5 10 15 20 25 30 35 40 12.08 7.56 9.42 0 5 10 15 20 25 30 35 40 35.88 37.53 20.34 0 5 10 15 20 25 30 35 40 0.5 1 2Weeds Grass Legume NecromassSample mean forage weight (g)Distance Figure 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, Cocl, Panama. 112

PAGE 113

CHAPTER 5 INTERACTIONS BETWEEN TREE SEEDLINGS AND UNDERSTORY VEGETATION DURING THE EARLY PHASE OF SILVOP ASTORAL SYSTEM ESTABLISHMENT Introduction In Panama, there is national interest in im proving the welfare of fa rmers and balancing the costs of agriculture with greater attention toward environmental conservation. Part of this effort is to look at how to increase the environmental su stainability of existing farm fields including the thousands of hectares of extensive pastures that cover the landscape. There is interest on the part of some farmers in integrating trees in or around their pastures. Part of this interest is fueled by the constant, daily needs of farmers for produc ts obtained from trees including fence posts, construction materials, fodder for cattle, food, medicine, and fuelwood. Despite the national interest in improving agricultural sustainability, fa rmer interest, and farmer need for tree-derived products, practically few efforts are underway in fomenting or developing systems to integrate trees into pastures. Likewise, little res earch has been done in Panama on appropriate management of tree seedlings that are planted for establishing silvopastoral system of dispersed trees in extensive pastures. This research aims to examine potential manageme nt strategies of tree seedlings planted in pasture, focusing on the tree species: Anacardium occidentale Bombacopsis quinata and Tectona grandis. Effects of three herbage removal regimes on growth of field-grown tree seedlings form the focus of the two-year study reported here. Literature Review Several experiments have been carried out to assess the extent of in teractions between tree seedlings and grasses and other herbaceous vegetation into which tree seedlings have been 113

PAGE 114

planted, the mechanisms that determine such in teractions, and the res ponses of diverse woody species to manipulation of herbaceous vegetation and surrounding conditions. Competitive Ability Tree seedling and grass interactions are often characterized as competitive, and frequently result in the domination of one t ype of vegetation or species or ag e class over another. Research suggests that there are numerous ways in wh ich types of vegetation gain dominance or outcompete one another. For example, in an e xperiment in Australia that investigated the interactions between tree seedlings and grasses, Florenti ne and Fox (2003) found that Eucalyptus victrix seedlings did not effectiv ely compete with grasses as grasses overcame tree seedlings during the period of establishment. Likewise, in a pot experiment, Sanchez and Peco (2004) interplanted Lavandula stoechas subsp. pedunculata with Mediterranean perennial grasses. There was a significant difference in seedling survival for pots planted with and without perennial grasses. In pots without grass, 78.36% of tree seedlings survived ; in pots planted with grasses, only 7.36% of seedlings survived. It has been suggested that many different plant characteristics can confer a superior compe titive advantage to plants, including total plant biomass, development of an elongated tap root, high leaf area, root st orage function, and early germination (Casper et al., 2003; Rajaniemi et al, 2003; Harmer and R obertson, 2003; Espigares et al., 2004). Blair (2001) suggests that size is not paramount to effective ability to compete. Rather, individuals with greater compe titive ability are more likely to acquire greater belowground resources regardless of their size. The author tested whether belowground competition in soils with isolated pockets of nutrients is dependent on plant size. Blair concluded that competition for nutrient patches may occur in unique ways, and may depend more on resource patch size and root foraging ability than on plant size. However, Collet et al. (2006) contend that changes in the 114

PAGE 115

spatial dimensions of seedling root growth may be the consequence of direct effects of nearby herbaceous vegetation including alle lopathy and mechanical effects. In an experiment involving Quercus petraea seedlings, Collet et al (2006) studied interacti ons between seedlings and grasses. They tested seedling performance under grass-removal treatments. Experiment results revealed that seedling branch roots were sign ificantly shorter when grown with grass than without grass. Likewise, at va rious dates during the four year field experiment, differences in seedling tap root length were statistically significant. Seedli ng biomass distribution was also affected by the treatments: biomass distribution to roots was more for seedlings grown with grass, compared with those grown without grass. Also, smaller seedlings, as opposed to larger ones, were shown to have allocated more biom ass to roots. Howeve r, Cahill (2003) found no relationship between belowground competitive abili ty and root system size in a Canadian grassland experiment. In accordance with some of the research c ited above, Peltzer and Kochy (2001) suggest that the characteristics of a good competitor cons titute the ability to withstand suppression by neighbors, effectively exploit avai lable resources, hinder growth of other plants, and grow faster or survive longer at low resource levels. In ad dition, the authors found that competitive ability may not rely as much on total accumulated plant bi omass but rather on growth rate, in addition to the other characteristic s mentioned. Through their greenhouse experiment looking at effects of neighbor plants (grasses, shrubs, and intact vegetation), Peltzer and Kochy (2001) found less competition for resources between woody plants th an between grasses. Consequently, they suggest that some type of facilitation may occu r among woody plants, and this may be the reason behind the occurrence of concentrations of w oody plants on the landscape, particularly in savanna conditions. 115

PAGE 116

It has also been suggested that resource and spatial gaps (areas unoccupied by grass roots) in belowground soil layers may better induce tree seedling establishment. Coll et al. (2004) suggest that gaps in soil resour ce abundance and differences in so il resource distribution in areas of grass growth may be consequent to indirect effects from gra ss roots induced by high grass root density, creating zones of resource depletion which may induce or detour changes in spatial distribution of tree seedling root systems. Jurena and Archer ( 2003) propose that roots compete for space, not only resources, belowground. In a fi eld experiment, they tested the establishment of Prosopis glandulosa seedlings with Schizachyrium scoparium and Paspalum plicatulum grasses in areas with and without gaps in grass roots belowground, and with and without aboveground gaps of grass. No relationship was found between belowground biomass and aboveground gap size, although aboveground spatia l gaps had a positive impact on seedling survival. However, Lindh et al. (2003) found r oot biomass in aboveground gaps to be notably less than root biomass under closed canopy in a NW US coniferous forest. Yet, Jurena and Archer (2003) found seedling roots preferentially grew in unoccupied spatial gaps belowground. Within these gaps, vertical spatial gaps in the so il had greater impact than horizontal spatial gaps on seedling establishment. Overall, the authors concluded that spatial an d temporal differences in competitive intensity among vegetation may bring about diverse windows of opportunity for tree seedling establishment in grasslands. Their conclusion agrees with that of Cahill (2002) and Jose et al. (2004) who argue that competition is not made up of a suite of static interactions rather these can fluctuate in intensity and vary spatially and temporally. Competition for Soil Moisture As seedlings and grasses have been shown to compete for space, some researchers have found that the most intense interactions and co mpetition among vegetation occur for moisture. Such heightened competition can occur due to drought, increases in biomass of herbaceous 116

PAGE 117

vegetation that provoke decreases in soil moisture c ontent, and other factors that trigger moisture availability deficits (Benayas et al., 2003; Jose et al., 2004; Schenk, 2006). In an experiment to test th e interactions among existing native and invasive forbs, annual and perennial grass species, and blue oak ( Quercus douglasii ) seedlings in a greenhouse, Gordon and Rice (2000) found significant differences in soil water potential in th e different competitive neighborhoods over time. The authors proposed that higher soil water depletion rates in the competing neighborhoods might have been due to significantly higher bi omass production rates and longer root length s of the annual grass Bromus diandra (non-native) in comparison to the competitive neighborhoods created with the forb Erodium botrys (non-native) and with the perennial grass Nassella pulchra (native). Blue oak seedling leaf number and leaf area were highest when grown with N. pulchra Root biomass of oak seedli ngs was lower in treatments involving high density plantings of other species. Blue oak shoot emergence was significantly affected by neighborhood competition; 89% of oak seedlings emerged in a no-neighbors situation and in the case of nonnative neighbors planted in low densities 56% of oak seedlings emerged. Water potentials in a ll treatments had important impacts on oak seedling growth and elongation. Davis et al. (2005) found different results in an experiment in Minnesota,USA, in which they examined the effects of native and non-native species on oaks as well as effects of moisture on oak ( Quercus ellipsoidalis ) seedling establishment by mani pulating moisture and nutrient levels. Results indicated no important impacts we re made due to neighboring grass type and, in addition, soil moisture content had a positive, sign ificant effect on seedling growth. Benayas et al. (2003) also investigated th e effects of native grasses on oa k seedlings by testing, in a pot experiment consisting of native Mediterranean herbaceous vegetation and Quercus faginea 117

PAGE 118

seedlings, in which different herbaceous removal trea tments were used. Results revealed that in the treatment that excluded aboveground biomass (regarded as the belowground competition treatment), biomass of herbaceous vegetation corre lated negatively with soil moisture content. Similarly, soil moisture was lowest for treatm ents with no competition during the dry season. However, elimination of herb shoots did not affect seedling survival. Coll et al. (2004) combined beech (Fagus sylvatica ) seedlings and perennial grasses in a pot experiment, in which, beech seedlings grown al one, free of grasses, increased their initial height by 87% in contrast to seedlings grown with grasses in which height increased just 1%. After two growing seasons, diameter and height growth of beech seedlings grown with grasses were reduced. The authors correlated reduction in seedling diameter with reduced soil water content, as during both growing seasons treatmen ts involving interplant ed seedlings and grass experienced an important decrease in soil water content. They also noted that grasses were more efficient at absorbing nutrients than beech seedlings. It is apparent that there is a wide variety of results and opinions regarding tree seedlinggrass interactions and competition. It has been s hown that trees and grasses can utilize different strategies to outcompete one a nother for aboveand belowground resources and space. It has also been shown that competition and interactive relationships can change over time and space and that these will also vary with species and environmental conditions. Evidently, available soil moisture plays a key role in seedling establis hment and may determine species survivorship. Root Biomass Allocation One way in which plants respond to competiti on is through shifts in biomass allocation. To explore this idea, Harmer a nd Robertson (2003) studied change s in tree seedling root systems associated with intercropped grasses grown in nur sery beds. In their experiment, five of six seedling species had greater root:s hoot ratios when planted with grasses as compared to when 118

PAGE 119

planted without grass. Over time, biomass accumu lation increased in roots rather than shoots. Tap root lengths were overall longer for species planted with grasses; ho wever, mean length of the root systems were shorter in the grass treatments. Also in th e grass inclusion treatments, tap root made up a larger proportion of total root length for the final harvest, re sulting in an increase in root length relative to total root biomass. Yet, significan t differences were not found for most of the variables. Authors attributed lack of significance to the short-term nature of the study and the differing emergence and germination dates of th e seedlings, and attributed the differences in responses among tree seedlings to the unique re sponses among different species to grass presence. In an experiment exploring a similar topic, Nilsson and Orlander (1999) found comparable results when testing Norway spruce ( Picea abies ) seedlings and grasses using treatments of mounding and herbicide. They found spruce seedli ngs in a grass inclusio n treatment allocated greater biomass to roots than in grass exclus ion treatments. Additionally, the presence of neighboring grasses brought about increased evapotranspiration in spruce seedlings. Collett et al. (2006) also found increased allo cation of biomass to roots in Quercus petraea seedlings when seedlings were grown with grasses. However, in a study in New Zealand examining mountain beech ( Nothofagus solandri ) seedlings in forest understor y, Platt et al. (2004) found that belowground root trenching and trenching combin ed with fertilization significantly increased biomass allocation to roots. More literature exists with parallel as well as contrasting results: in some competitive situations, biomass allocation to roots increases and in other situations of competitive exclusion the same result occurs. 119

PAGE 120

Competition for Nutrients (Fertilization Studies) It seems apparent that competition for mois ture and space can create situations of exclusion in which herbaceous vegetation ca n outcompete woody pere nnial seedlings, as competition for soil nutrients can be intense fo r seedlings and herbaceous vegetation. Several studies have examined the effects of fertilization treatments on the interactions between tree seedlings and herbaceous vegetation, some of these include Hangs et al. (2002), Thevathasan et al. (2000), Ramsey et al. (2005), and Platt et al. (2004). For ex ample, in a Canadian boreal environment, Hangs et al. (2002) tested Populus tremuloides Epilobium angustifoilum and Clamagrostis canadensis (early succession species) in competition with Picea glauca and Pinus banksiana for N. In the study, the grass species, Calamagrostis canadensis outcompeted P. glauca and P. banksiana seedlings for N during the establis hment phase of tree seedlings and vegetation in a pot experiment. Hangs et al. (2 002) concluded that the ability of herbaceous vegetation to efficiently access NH 4 + and NO 3 more effectively than the other species ensured the grasses with a competitive edge over the tree seedlings. Th evathasan et al. (2005) also looked at the effects of C. canadensis E. angustifolium P. tremuloides together with other species on black spruce ( Picea mariana ) based on NO 3 accumulation rates. The early succession species (also deemed weed species in the literature) bene fited most from the accumulated NO 3 Weeds were able to outcompete black spruce seedlings for resources. In treatments of low weed density, black spru ce seedling performance improved. Ramsey et al. (2002) also tested fertilizer treatments although coupled with an herbicide regime to assess interacti ons between longleaf pine ( Pinus palustris ) and varied naturalized grasses in Florida, USA. They found fertiliz ation to have a negligible effect on seedling development. The authors postulated that fertilizati on of seedlings in an old field site potentially reduced seedling survival due to the possi ble stimulation of surrounding weed growth by 120

PAGE 121

fertilizer. Concurrently, herbaceous weed cont rol significantly increased pine seedling height and root collar diameter. Herbace ous weed control also resulted in higher seedling survival than that of the control after the first growing seas on. Furthermore, weeding treatments increased the growth of seedlings out of the characteristic P. palustris grass stage. Platt et al. (2004) combined fertilization with vegetation removal treatments and added root trenching to control herbaceous vegetation growing w ith mountain beech seedlings ( Nothofagus solandri ). They found that the combination of root trenching a nd fertilization significan tly increased seedling stem diameter 231% and height growth 167% und er the same treatment regime in comparison with the seedling control group. However, seedling growth in fertilizer-o nly treatments (in the absence of trenching and herb aceous vegetation removal) was not significantly higher. Microclimate Effects An indirect form of interaction between gra sses and tree seedlings was tested by Ball et al. (1997; 2002) in New South Wales, Australia, through assessing the effects of grasses on groundlevel microclimate. They found that grasses ar ound seedlings caused a change in temperature near to the ground creating a microclimate around seedlings. The microclimate significantly lowered minimum temperatures above the grass surface and consequently lowered tree seedling leaf temperature 13 o C, leading to a significant decrease in seedling growth. Concurrently, seedling leaf temperatures increased linearly w ith increased bare ground area (removal of grass) surrounding the seedling. Trenching Effects The strategy used by Pratt et al (2004), root trench ing, has been used for research in some cases to form a physical barrier between root sy stems, unlike treatments such as clipping and herbicide applications, which in most cases only eliminate vegetation inte raction superficially. In experiments with mature trees, Ludwig et al. (2004) and Harringt on et al. (2003) used 121

PAGE 122

trenching to assess competition be tween trees and neighboring herbaceous vegetation. In an East African savanna, Ludwig et al. (2004) trenched mature Acacia tortilis trees in an experiment to examine plant-tree interactions within the context of hydraulic lift. They found live grass biomass to be significantly higher in trenched plots. Overall, grass growth benefited from trenching and the reduction in competition. In addition, total aboveground biomass was significantly higher in trenched pl ots than in control plots. In a unique study, Harrington et al. (2003) a ssessed mature woody species ability to outcompete native herbaceous vegetation. Matu re long leaf pines demonstrated belowground intraspecific competition by limiting longleaf seedling growth nearby through root competition. The authors also suggested that pine needle litter could play a role in curbing the growth of herbaceous vegetation as litter can diminish the penetration of s unlight and moisture to lower layers. During the first two years of the study, th ere was effective separation of herbaceous and pine roots using trenching. The authors deem ed that trenching was effective at reducing competitive interactions among the species. Howeve r, in the third year of the study, trenching effects were reduced by pines ca pacity to access soil moisture beyond the trenched areas. At the same time, stand basal area for pine increased substantially and the increased absorption of moisture occurred in proportion to the increase in basal area. The authors found that the presence of pine in the overstory and relativ e buffering of high surface temperatures did not provide an overall benefit to herbaceous vegetati on as increases in pine population caused an increase in competitive interactions aboveand belowground. In contrast to these, Holl (1998) found in an experiment on the effect of trenching on seedling performance in a pasture and a forest in Costa Rica that trenching had a significant effect on root biomass of grasse s and shrubs but not on tree seedli ngs. Grass fine root biomass 122

PAGE 123

was greater than shrub fine root biomass in nontrenched treatments; however, fine root biomass for grasses and shrubs in the trenched treatments were similar. Seedling growth was also greater with grass than with shrubs across treatments. Also in a tropical setting, Barberis and Tanner (2005) examined the effects of trenching on seedling performance in a semi-evergreen forest in Panama. They evaluated seedling survival and growth in forest gaps and in forest unders tory in relation to soil moisture, soil nutrient availability, and light availabil ity in both trenched and non-trench ed plots. In the study, soil moisture was significantly affected by trenchi ng, by 40% in the dry season and by 2% in the wet season. Also in the dry season, s eedlings in trenched plots had gr eater leaf area th an those in non-trenched plots. In terms of overall effects of trenching on soil moisture, the authors generalized that the increase in seedling growth in trenched pl ots was a function in part of improved soil moisture. However, the authors fo und soil nutrient availability also played an important role in seedling performance. In f act, trenching was less effective in the understory when seedlings became less limited by nutrients. Like soil moisture and nutrient availability, light availability also played a significant role in seedli ng performance in trenched and nontrenched plots. According to the results, light gaps were more effective at increasing seedling growth and survival than trenching. The author s suggested that, the im portance of belowground competition in limiting the growth of tropical tree seedlings decreases as soil fertility increases and decreases as drought decreases. We can also generalize that the increases in growth due to gaps are greater than increases due to trench ing in wetter and more fertile sites. It is evident that interactions between tree seedlings and grasses are extensive, somewhat complex, and not of any consistent pattern. Th ere are numerous variables and scenarios that affect belowground interactions and competiti on, including soil conditions, species, surrounding 123

PAGE 124

vegetation, climate, regions, ecology, and many othe rs. Existing research alludes to the possible relationships between some of th ese variables; however, concrete determinations are few. Objectives and Hypothesis The objective of this study was twofold. One element of the experiment sought to assess whether vegetation surrounding seed lings affected the development of the studied seedlings. The other objective was to use information fr om this study to provide recommendations to producers regarding species selection and require d weeding regimes for trees integrated into pasture systems. I hypothesized that increasing herbage removal would lead to increasing seedling growth, trenched seedli ngs would prosper over seedlings in the other herbage removal treatments, and A. occidentale would be the hardiest specie s among the three species tested. Methods and Materials Study Site The study was conducted on La Cabimosa Fa rm, La Candelaria sector, Rio Grande corregimiento, Cocl province, Panama (see Chapter 2 for specific local and regional characteristics). Experimental Design A completely randomized design was used with three tree species (listed in the next section) and four levels of herbage removal around seedlings, thus a total of 12 treatment combinations. Tree seedlings were planted in ro ws of thirty, and treatment combinations were randomly assigned to each row. Treatments included zero removal of herbage around the seedling (control), removal of herbage with in a 50 cm diameter around the seedling stem, removal of herbage within a 100 cm diameter ar ound the seedling stem, and removal of herbage within a 100 cm diameter around the seedling stem combined with a backfilled trench around the seedling. There were 10 repetitions for each treatment combination and each season, totaling 124

PAGE 125

360 seedlings. To maintain the herbage remova l treatments, weeds were clipped to ground level once monthly using a machete. Materials Three tree species were chosen by participa ting farmers to be used in the study. The species were Anacardium occidentale Bombacopsis quinata and Tectona grandis. All of the seedlings were acquired th rough a local nursery. The A. occidentale and B. quinata seedlings were approximately 180 days in age and 30 cm in height at the time of planting. In accordance with local and regiona l planting technique, T. grandis seedlings were planted using bareroot stalks and were approximately 200 days in age. Establishment On the Cabimosa farm in a fenced field pr eviously used for past ure and seasonal rice production, soil was tilled by tracto r in preparation for planting. Most of the standing herbage was removed by tilling; remaining weeds were removed manually using a machete. Holes were dug 3 m apart and measured 30 cm deep by appr oximately 30 cm wide in rows 3 m apart, resulting in a planting configura tion of 3 m x 3 m. Circular tr enches were excavated at 100 cm diameter around the seedlings, which were rando mly chosen to correspond to the trenching treatment. Within each trench, a single layer of thin black plastic was placed to line the trench and the trench was backfilled. At the time of planting, the nursery bags of A. occidentale and B. quinata were removed and the root ball with its orig inal soil was placed inside the hole with the previously removed soil which had been loos ened and rocks removed prior to planting. T. grandis bareroots were planted similarly into the hol es and backfilled. Seedlings were planted in June, 2000. 125

PAGE 126

Measurements Seedlings were harvested 6, 12, and 24 months af ter planting. Harvesting consisted of the complete uprooting of the seedling. Seedlings were removed from the ground using knives and fingers. Roots were washed to disperse soil particles. Seedli ngs were weighed fresh and then divided into roots, stem, and leav es. Observations were recorded for the total dry mass weight of each seedling as well as its roots, stems, and leaves separately. It is notable that data for T. grandis in month 24 was unavailable due to seedling mortality. In addition, due to the phenology of T. grandis and B. quinata, leaf data for these species were often unavailable due to the absc ission of its leaves prior to the time of harvest and observation. Data Analysis Logarithmic transformations of the data were applied to improve normality of the distribution. Statistical analyses were performed using SAS. Analysis of variance (ANOVA) was conducted. When the ANOVA results indicated a significant effect ( = 0.05), a Scheff test was conducted to carry out multiple comparisons of means. The data used in the analysis consisted of weights per s eedling or seedling part. Results Herbage Removal Effects of herbage removal on tree seedlings varied by species and over time as there was an overall significant main effect of the herbage removal treatment variable on seedling biomass (Table 5-1). In the Scheff comparison of m eans, the Control and Trench treatments were significantly different while the 50 cm and 100 cm treatments were not (Table 5-2). Overall, the lowest growth occurred in the Control group (Figure 5-1). 126

PAGE 127

Effects of the Species Treatment on Biomass The three tree species differed significantly in biomass production over time; also there was a significant interaction betw een species and the repeated measure time Each species was significantly different at each time, with the exception of teak at 24 months (Table 5-3). In relationship to the weeding regimes, A. occidentale showed the largest fluctuation in responses to weeding regimes, with its highest mean biomass in the 100 cm treatment and the lowest in the Control treatment. Results for B. quinata were somewhat similar for the 50 cm and 100 cm treatments. However, B. quinata biomass was highest overall in the Trench treatment. At the same time, B. quinata had the highest mean biomass of all species in all of the herbage removal treatments except in the 100 cm treatment while T. grandis had the lowest mean in all of the treatments. Between the 6 and 12 month harvests, there was a general decreas ing trend in mean biomass for the Control, 50 cm, and Trench treat ments. However, in the 100 cm treatment, growth was stagnant overall. Conversely, in the 24 month harvest, overall growth increased sharply (Figure 5.2). It should be noted that in the 24 month harvest, no data was available for T. grandis growth due to mortality of T. grandis seedlings during this period. In the ANOVA analysis, there was a significant inte raction between spec ies and harvest ( p < 0.0001). When a comparison of means was conducted, at 6 months 12 months, and 24 months, all of the species were significantly different except for the absence of T. grandis in the 24 month harvest. At the 6 and 12 month harvests, B. quinata had the highest mean biomass values. However, in month 24, A. occidentale had the highest mean of the two speci es, which was almost twice that of B. quinata 127

PAGE 128

Stem Biomass There was a significant main effect for herbage removal on stem biomass ( p < 0.0446). However, there was no significant interaction of herbage removal with th e other variables. Similar to the results of total s eedling biomass stated above, in a comparison of means test there was a significant difference ( p < 0.0137) between the Control and the Trench treatments in the stem data. However, the 50 cm and 100 cm treatments were not significantly different. At the same time, there was a significant inte raction effect between species and harvest variables. All of the species were significantly different at month 6 a nd at month 12, but were not significantly different at month 24. In terms of differences in growth patterns, A. occidentale stem growth made little progress during the fi rst year. However, between 12 and 24 months, there was a substantial increase in its stem biom ass (Figure 5-3): while the mean stem weight was 62.75 g per seedling at 12 months, it was 836.12 g at 24 months. Furthermore, the stem biomass of A. occidentale was about three times the amount of its root biomass during the study. Both B. quinata and T. grandis stem weight decreased from m onth 6 to month 12. However, in month 24, B. quinata stems rebounded in growth, from a me an weight of 240.5 g in month 6 to 340.02 g in month 24. Root Biomass Similar to the stem biomass data, there was a significant interaction between harvest and species in the root biomass data ( p < 0.001). Root biomass yields we re significantly different for each measurement interval (harvest) across species (except T. grandis in month 24), unlike the stem data but similar to th e overall growth data. For A. occidentale, root:stem ratios were 0.405, 1.185, and 0.29 at months 6, 12, and 24, respectively. B. quinata root:stem ratios were somewhat consistent over time at 1.185, 1.465, and 1.21 at 6, 12, and 24 months respectively. In contrast, for the first two harv est dates, root biomass of T. grandis decreased over time, its 128

PAGE 129

root:stem ratio at the 6 month harvest interval was 3.31 and at the 12 month interval it was 2.0 (Figure 5-4). Discussion To promote the integration of trees into pastur elands, it is imperative that the dynamics of the interactions between grasses and trees be understood. Of part icular importance is the period of seedling establishment to ensure the growth of healt hy seedlings within a nascent silvopastoral system. In addition, it is vital that land managers be aware of the management practices required during the tree seedling establishm ent phase of the system to en sure its longevity and vitality. In Panamas southern plains, pastures with few dispersed trees dominate the landscape. Trees remain in pastures for myriad reasons ; however, both trees and emerging seedlings are seldom cared for or managed. This study was established primarily to examine the dynamics of seedling establishment in pastures to help land ma nagers interested in successfully establishing a greater number of trees in their pastures. Seedling Growth In the study, observations were made of total seedling growth (including roots, stems, and leaves when intact) in response to experime ntal treatments. Rem oval of herbage surrounding seedlings had a significant effect on seedling gr owth. The hypothesis was that the absence of herbage competitors aboveand belowground would have beneficial effects on seedling growth; thus, increasing herbage control was expected to have positiv e effects on seedling growth. However, only the Trench treatment had a sign ificant impact on seedling growth. This may indicate that although herbage was removed aboveground, for example in the 50 cm and 100 cm treatments, herbage continued to have an impact belowground. It can therefore be inferred that herbage removal is only effective when presence of belowgr ound plant components are removed through measures such as systemic herbicide app lication and rototilling. However, a combined 129

PAGE 130

analysis of the data across herbage removal treatments showed the lowest seedling biomass yields occurred in the Control treatment whic h included no herbage rem oval, suggesting that herbage removal treatments were in general e ffective in increasing seedling growth. This indicates that herbage removal treatments indeed had an effect on seedling biomass accumulation. The three species tested in the study perfor med differently throughout the experiment. A. occidentale is quite abundant locally in the study site. Its ability to acclimat e and thrive within the conditions of the study during th e trial may have been to a certain extent an a ttribute of its inherent adaptation to the area. However, A. occidentale experienced difficulties between month 6 and month 12 of the experiment when there was only a small increase in its total biomass production while between months 12 and 24 its growth accelerated. Trenching, coupled with plastic lining, may have hindered A. occidentale growth during months 6 and 12. Only when the species was able to penetrate and overcome the pl astic barrier, perhaps at or after month 12, was it able to reach its full growth potential. A very similar response occurred in the experiment of Harrington et al. (2003) with longleaf pine. Across the species, the 6 to 12 month period saw a general decrease in biomass accumulation an unexpected result which may have occurred due to seasonal variation. The 6 to 12 month period coincided with the local dry season where pr ecipitation can fall below 13 mm monthly (Murphy and Lugo, 1995). The region where the study was conducted, called the arco seco is known to have the driest and most prolon ged dry season in the co untry extending up to 5 months. Hence, the decrease in biomass weight s observed may have been a consequence of the severe drought experienced during these mont hs and/or the result of herbivory by local herbivores. 130

PAGE 131

Stem and Root Biomass Leaf data are not reported in th is study due to the lack of ava ilability of leaf biomass for all of the species on all of the harvest dates. B. quinata and T. grandis are deciduous species; their leaves had often fallen before the harvest dates. In fact, B. quinata is known to be devoid of leaves during six months of the year. Effects of herbage removal differed for stem and root growth. It was expected that herbage removal would have a relevant im pact on root growth, as has been observed in diverse studies (Harmer and Robertson, 2003; Coll et al., 2004; Plat t et al., 2004). However, herbage removal did not have a significant impact on root growth. On the other hand, stem growth was, in fact, adversely affected by herbage removal. Th e reasons for this observation are unclear. Ratios of roots to stem varied distinctly among the species. For example, root:shoot ratio of A. occidentale differed from that of the other speci es and the ratio for the species itself differed over time. The change s in the root:shoot ratio of A. occidentale may have been a consequence of the seedlings inability to acce ss soil resources in which the inlaid plastic impeded growth of A. occidentale roots and their ability to access growth resources. The change in the root:shoot ratio of A. occidentale coincided with its marked, accelerated growth between months 12 and 24. This could be attributed to A. occidentale roots reaching a region of the soil profile with greater soil resources thereby allowing A. occidentale to distribute greater biomass to aboveground growth and forsake increase s in belowground growth (Schenk, 2006). In contrast, B. quinata maintained a constant root:shoot ratio throughout the experiment regardless of seasonal fluctuations and treatment effects. A unique trait of B. quinata is its ability to thrive under drought conditions for ex tended time periods. Cons istent allocation of more biomass to roots than stems may be one of the adaptation and survival mechanisms of this species. Conversely, T. grandis had a difficult time in this study, demonstrated by its complete 131

PAGE 132

mortality by month 24. However, during mont hs 6 and 12 the species maintained a high root:shoot ratio, roots accounting for 67 to 77% of total observed biomass; this may also have been an after-effect of being grown from bareroot stalks. Conclusion This study examined the effects of herbag e removal treatments on three tree seedling species over two years. At the initiation of the study, I hypothe sized that increasing herbage removal would lead to increases in seedling growth. The experiment results did not provide evidence to validate this hypothesis. However, with in this hypothesis, I st ated that it was likely that the Trench treatments would have the great est effect on increasing seedling growth and this hypothesis was confirmed by the resu lts. I also hypothesized that A. occidentale would be the hardiest of the three tree species in the experiment. However, this was not demonstrated in the results. In fact, B. quinata had the largest overall mean weight. While A. occidentale was a close second to B. quinata, its performance was less c onsistent than that of B. quinata Finally, the applied objective of this study was to garner information in order to make recommendations to land managers regarding appropriate herbag e removal for establishing seedlings. The study results indicate that herbage removal in general will favor seedling performance; however, the results do not provide a clear result for the approp riate, specific amount of herbage removal to optimize seedling establishment and growth. 132

PAGE 133

Table 5-1 Analysis of the effects of the repeat ed measures herbage removal (at distances of 50 cm, 100cm, and 100cm with trenching, from seedling stem), tree species ( Anacardium occidentale Bombacopsis quinata and Tectona grandis), and time (6, 12, and 24 months after planting) and their interactions on biomass accumulation of tree seedlings planted on-farm in a non-grazed pasture in Rio Grande, Cocl, Panama. Effect df F Pr > F Herbage removal 3 2.6 0.0562 Species 2 118.59 < 0.0001 Species x Herbage removal 6 0.1 0.9962 Time 1 2.51 0.1152 Season x Herbage removal 3 1.87 0.1371 Time x Species 2 110.22 < 0.0001 Time x Species x Herbage removal 6 0.61 0.7212 133

PAGE 134

Table 5-2 Comparisons of the with in-subject effects of the repeat ed measure herbage removal (0 = control (no herbage re moval), 50 = herbage rem oval 50 cm diameter around seedling stem, 100 cm = herbage remova l 100 cm diameter around seedling stem, Ditch = herbage removal 100 cm diameter around seedling stem coupled with plasticlined, back-filled trench 100 cm diam eter around seedling stem) on biomass accumulation of tree seedlings planted on-farm in a non-grazed pasture and observed over two years in Rio Grande, Cocl, Panama. Herbage removal Herbage removal t Pr > t 0 100 -0.91 0.3644 0 50 -0.98 0.3284 0 Ditch -2.51 0.0137 100 50 -0.07 0.942 100 Ditch -1.6 0.1124 50 Ditch -1.52 0.1303 134

PAGE 135

Table 5-3 Effects of the interactions of three seedling species (Cashew ( Anacardium occidentale ), Tropical cedar ( Bombacopsis quinata ), and Teak ( Tectona grandis)) with harvest time (6, 12, and 24 months af ter planting) on biomass accumulation of tree seedlings planted on-farm in a non-grazed pasture in Rio Grande, Cocl, Panama. Species Species Time t Pr > t Cashew Tropical cedar 6 -15.26 < 0.0001 Cashew Teak 6 -4.39 < 0.0001 Tropical cedar Teak 6 9.57 < 0.0001 Cashew Tropical cedar 12 -12.5 < 0.0001 Cashew Teak 12 11.08 < 0.0001 Tropical cedar Teak 12 20.29 < 0.0001 Cashew Tropical cedar 24 2.94 0.004 135

PAGE 136

0 200 400 600 800 1000 1200 Control 50 cm 100 cm 100 cm+ditch Herbage removal regimes Mean biomass per seedling (g) A. occidentale B. quinata T. grandis Figure 5-1 Responses of three species of tree seedlings to three understory-herbageremoval treatments during the first two years after tr ee planting in a field site in Rio Grande, Cocl, Panama. 136

PAGE 137

0 100 200 300 400 500 600 700 800 900 1000 6 12 24 Months Mean weight per seedling (g) A occidentale stem A occidentale roo t B quinata stem B quinata roo t T. grandis stem T. grandis roo t Figure 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, Cocl, Panama. 137

PAGE 138

100 200 300 400 500 600 700 800 900 100 200 300 400 Mean weight per seedling part (g) Shoot Root 6 months 12 months 24 months 0.405 1.185 3.31 0.426 1.465 2.0 0.29 1.21 A. occidentale B. quinata T. grandis Time after planting tree seedlings Figure 5-3 Changes in seedling biomass accumulati on in stems and roots, and root:shoot ratio (numbers above bars) changes during the tw o-year establishment of silvopastoral systems in pastures in Cocl, Panama. 138

PAGE 139

0 0.5 1 1.5 2 2.5 3 3.5 61 22 4 Time (months after planting) Root:shoot ratio A. occidentale B. quinata T. grandis Figure 5-4 Root:shoot ratios of three species of seedlings acro ss grass removal treatments during the two-year establishment pha se of silvopastoral systems planted in pastures in Rio Grande, Cocl, Panama. 139

PAGE 140

CHAPTER 6 SUMMARY AND CONCLUSIONS The focus of this effort was to addre ss the problems and challenges of improving production efficiency and environmental health that small-scale cattle farmers in Panama face today. The goal was to produce research results that could be readily adopted by farmers and adapted to their production prac tices on-farm. With this purpose and based on the premise that trees in pastures can augment production and provide beneficial environmental services, this research examined the survival of planted tr ee seedlings in active pastures, evaluated the interactions between establis hing seedlings and surrounding vege tation, and assessed the effects of large trees on forage characteristics in past ure. The principal questions that guided the research were: 1. What are the best means, in terms of tree species and planting c onfiguration design, to establish young tree seedlings into actively grazed pastures? 2. In terms of management strategies, what is the vegetation removal re gime that optimizes seedling survival? 3. What is the effect of dispersed trees in pa sture on forage characteristics and pasture production? Experimental Findings To explore possible responses to the principal research questions, research was carried out on tree seedling survival and herbivory, conse quences of large trees on forage, and the interactions between seedlings and grasses. Seedling Survival and Herbivory We found that species characterist ics played a major role in seedling survival. This was to be expected considering that rooting ability, ability to acquire resource s, carbohydrate reserves, growth type, and light needs are characteristics that ar e critical to the surv ival of a species, particularly in competitive environments. In the study, planting configuration and tree species 140

PAGE 141

played important roles in seedling survival. S eedlings in open pasture (planted in lines and diagonals) survived better and longer than those planted along fences. On the other hand, incidence of herbivory was overwhelmingly depende nt on species type. Species with palatable leaves were browsed far more often than the le ss palatable ones. Of the three species studied, Anacardium occidentale, Tectona grandis, and Bombacopsis quinata, A. occidentale, which is locally abundant, showed greater survival and had least herbivory, and it pe rformed better than the other two species at the e nd of the two-year study period. Seedlings were quite sensitive and they responde d differently to planting configurations. Seedling mortality was highest in the fence treatment (66%), followed by diagonal (51%) and line (47%). It was also clear that presence of cattle was not conducive to seedling survival. Grazing cattle present a challenge to both increas ing seedling survival and diminishing seedling herbivory in grazed pastures. When cattle were present, A. occidentale performed markedly better than T. grandis and B. quinata Effects of Large Trees on Understory Forage Three forage characteristics were examined : forage mass, digestibility, and botanical composition. Season and distance had different effects on the two tr ee species tested, Anacardium occidentale and Tectona grandis. Season was important to forage growth below and around A. occidentale in that across distances forage abundance was greater in the wet season than in the dry season. However, this was not the case for T. grandis Surprisingly, forage mass values were greater in the dry season than in the wet season below and around T. grandis crowns. The fact that forage mass values for T. grandis were generally low in comparison to those of A. occidentale forage dry season abundance was greater than in the wet season, and forage mass in open pasture was greater than it was at the drip line and close to the stem (unlike A. occidentale) indicates that there are potentially relevant interactions occurring 141

PAGE 142

between the T. grandis trees and forage which may be impinging upon forage growth. Hence, season was relevant to forage gr owth throughout. It was evident, however, that, similar to the results of seedlings survival a nd herbivory, tree species was key to differences in forage growth. Unlike forage mass, forage digestibility was impacted by distance and season under A. occidentale and only season under T. grandis Tree species had a particularly noteworthy effect on digestibility in that forage under T. grandis had consistently better digestibility than that under A. occidentale Even when close to the tree stem, species had important impacts on digestibility, unlike forage mass. However, when moving away from trees and into open pasture, only season became relevant to changes in digestibility. Forage composition was also highly affected by tree species. Quantity of grass, legume, and weed biomass was sensitive to tree species as their abundance was generally static in the T. grandis understory yet varied under A. occidentale. Changes in effects did not occur across distances, implying that distance to the tree was irrelevant while the tree species itself was the relevant factor affecting forage composition. Ho wever, the amount of necromass (dead material in the forage) was considerably sensitive to distance to T. grandis stem in that it increased at the drip line; this observation provide s a relevant insight into the re lationship between tree presence and forage botanical composition. Interactions between Seedlings and Vegetation Vegetation removal regimes had varying e ffects on tree seedling growth. Seedling biomass was affected positively by vegetation removal aboveground. However, a significant increase in seedling growth occurred only wh en belowground vegetation biomass growth was impeded, indicating the importance of belowg round competition on seedlings. The different vegetation removal regimes affected seedling stem growth but did not have significant impacts on root growth. However, interestingly, the different vegetation removal regimes affected 142

PAGE 143

seedlings differently. A. occidentale root:shoot ratio fluctuated with vegetation removal while B. quinata root:shoot ratio was consiste nt regardless of season or herbage removal. Overall, B. quinata and T. grandis allocated more biomass to roots than stems. Implications for Implementation Options for Grazing In the experiment, cattle were left on pastures to graze in an attempt to imitate the real situations on producers farms as producers are reluctant to remove their ca ttle from pastures to allow for seedling establishment and growth. Ho wever, research results revealed that cattle grazing produced deleterious effect s on seedlings. Therefore, a quanda ry exists as to how best to establish seedlings while meeting the needs and desires of producer s to allow cattle to graze. One option may include recommending that in th e wet season producers exclude cattle from pastures that have been planted with seedlings an d that cattle are allowed to graze these pastures only in the dry season. By eliminating grazing in the wet season, seedlings will be allowed seven to eight months to become established and develop their root systems that are important in preparation for potential grazing or herbivory in the dry season, free of the negative effects of cattle. At the same time, generally the wet season is the period when available forage is highly abundant. It is assumed then that a producer coul d satisfy cattle needs in other pastures leaving the seedling-planted pasture free to grow and develop during the period. Conversely, in the dry season cattle would graze the seedling-planted past ure. In the dry season, when forage availability is generally deficient, the produ cer is able to access the forage on the seedlingplanted pasture. As found in the experiment, the use of deciduous species may benefit seedling survival in active pastures during the dry season as seedlings would be devoid of leaves when cattle are present thereby reducing the pot ential for herbivory and damage. 143

PAGE 144

Manipulating Forage with Trees According to the research results, tree presence diminished necromass around trees. Necromass is undesirable in terms of productive capacity and performance of pasture. Therefore, through appropriate use of trees, producers may be able to reduce necromass abundance. Tree spacing and tree cr own architecture would be critical to generate th is benefit of reduced necromass yet, at the same time it is necessary to balance light av ailability to forage when considering tree spacing and total tree stem density. Tree Establishment The research results show that seedling establis hment, the first step in the integration of trees into pastures, is sensitive to presence of neighboring vegetation. For optimal seedling establishment, competition both aboveand belowground should be minimized. However, removal of belowground competition is not always feasible for producers due to cost and labor requirements. Response to vegetation removal within 1 m diameter around the seedling stem was beneficial to seedlings in terms of biomass accumulation although results differed by species. Based on this study, weeding within a 50 to 100 cm diameter around seedlings is the recommended regime. Future Research The overriding message from this research as it bears upon impacting directions for future research is that: 1. a dispersed tree silvopastoral system can ha ve positive impacts on extensive pasture productivity, and 2. overall, the species used in the system determines whether the system will benefit or negatively affect pasture characteristics. It is imperative that silvopastoral research be conceived within the scope of improving agricultural productivity. Given the generally lo w adoption success of silvopastoral systems in Central America, there needs to be a shift in the conception of silvopastoral research. A 144

PAGE 145

systematic framework geared toward addr essing producers botto m line, agricultural productivity, must be created that will mobili ze the transfer of this evidently beneficial technology that is still in the pr ocess of being researched, addre ssed, and adapted to the needs of producers. It is too vital that research address multiple scal es of silvopastoral research application conceiving of systems for small, medium-, and large-scale cattle producers, as each of these is abundant across the landscape. Research should focus on tree species, addressing how these affect forage characteristics. Likewise, forage species within the specific co ntext of silvopastoral systems need to be investigated. There should also be a focus on the livestock component of the system, something that was not included in this study. New studies should also examine the environmental impacts of silvopastoral systems including general biodive rsity, birds, insects, and carbon sequestration. These studies will be of part icular importance as producers begin to enter the global environmental services market. Finally, it is imperative that research broaches a systems approach to silvopastoral systems. Each of the components, tree, forage, and livestock, must be assessed in terms of their interact ions with one another and the ultimate effect these interactions have on production. 145

PAGE 146

APPENDIX A PLANTING CONFIGURATIONS OF THE TH REE TREE-SPECIES SEEDLINGS FOR ESTABLISHMENT OF A SILVOPASTORAL SYSTEM IN RIO, GRANDE, COCL, PANAMA x x x x x x x x x x x x x x x Fence x x x x x x x x x x x x x x x Fence x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x Diagonal x x x x x x x x x x x x x x x Line x x x x x x x x x x x x x x x Line x x x x x x x x x x x x x x x Line x x x x x x x x x x x x x x x x x x Diagonal x x x x x x x x x x x x x x x x x x Diagonal Fence 146

PAGE 147

APPENDIX B COMPARISONS OF MEANS OF INCIDE NCE OF TREE SEEDLING HERBIVORY ACROSS TREE SPECIES AND PLANTING CONFIGURATION Confi g urationSpecies (i)Species ( j )Mean difference (ij )SESi g DiagonalCashewTropical cedar -3.847*1.2130.005 Teak -2.4521.3140.177 Tropical cedarCashew 3.847*1.2130.005 Teak 1.3960.9050.328 TeakCashew 2.4521.3140.177 Tropical cedar-1.3960.9050.328 FenceCashewTropical cedar-1.1040.9110.537 Teak -3.0280.9480.005 Tropical cedarCashew 1.1040.9110.537 Teak -1.9240.8090.053 TeakCashew 3.0280.9480.005 Tropical cedar 1.9240.8090.053 Line CashewTropical cedar -5.346*1.0470.000 Teak -1.7361.0350.257 Tropical cedarCashew 5.346*1.0470.000 Teak 3.611*0.7620.000 TeakCashew 1.7361.0350.257 Tropical cedar -3.611*0.7620.000 147

PAGE 148

APPENDIX C FORAGE SAMPLING SCHEMATIC OF HERBAGE MASS HARVESTED AT THREE DISTANCES FROM TREE STEM IN THE FO UR CARDINAL DIRECTIONS CARRIED OUT UNDER SCATTERED TREES IN PASTURES IN RIO GRANDE, COCL, PANAMA 200% 100% 50% 148

PAGE 149

LIST OF REFERENCES Allcock, K., and D. Hik. 2004. Survival, growth, and escape from herbivory are determined by habitat and herbivore species for three Australian woodla nd plants. Oecologia 138:231241. ANAM. 1999. Memoria Anual. Autoridad Nacional del Ambiente, Panama. ANAM. 2000. Resumen Anual. Autoridad Nacional del Ambiente, Panama. Araujo Filho J.A. 1990. Manipulao da vegeta o lenhosa da caatinga para fins pastoris. Sobral, EMBRAPA-CNPC (EMBRA-PACNPC. Circular Tcnica 11). Ares, A., D. St. Louis, and D. Brauer. 2003. Tr ends in tree growth and understory yield in silvopastoral practices with southern pines. Agroforestry Systems 59:27-33. Ares, A., W. Reid, and D. Brauer. 2006. Producti on and economics of native pecan silvopastures in central United States. Agroforestry Systems 66:205-215. Bailey, J., and N. Harjanto. 2005. Teak (Tectona gr andis L.) tree growth, stem quality and health in coppiced plantations in Java, Indonesia. New Forests 30:55-65. Balderrama, S., and R. Chazdon. 2005. Light-depende nt seedling survival and growth of four tree species in Costa Rican second-growth ra in forests. Journal of Tropical Ecology 21: 383-395. Ball, M., J. Egerton, R. Leuni ng, R. Cunningham, and P. Dunne. 1997. Microclimate above grass adversely affects spring growth of seedling s now gum (Eucalyptus pauciflora). Plant, Cell and Environment 20:155-166. Ball, M., J. Egerton, J. Lutze, V. Gutsch ick, and R. Cunningham. 2002. Mechanisms of competition: thermal inhibition of tree s eedling growth by grass. Ecophysiology 133:120130. Banco Nacional de Panama. 2003. Interview with official from National Bank of Panama. Panama. Barberis, I., and E. Tanner. 2005. Gaps and root trenching increase tree seedling growth in Panamanian semi-evergreen forest. Ecology 86:667-674. Beckage, B., and J. Clark. 2003. Seedling survival and growth of three fore st tree species: the role of spatial heter ogeneity. Ecology 84:1849-1861. Behrens, R. 1996. Cashew as an Agroforestry Cr op: Prospects and Potentials. Margraf Verlag, Weikershiem, Germany. 149

PAGE 150

Belesky, D. 2005. Growth of Dactylis glomerata al ong a light gradient in the central Appalachian region of the eastern USA: I. Dry matter production and partitioning. Agroforestry Systems 65:81-90. Belsky, A. 1994. Influences of trees on savanna produc tivity Tests of shade, nutrients, and treegrass competition. Ecology 75:922-932. Benayas, J., T. Espigares, and P. Castro-Diaz. 2003. Simulated effects of herb competition on planted Quercus faginea seedlings in Mediterranean abandoned cropland. Applied Vegetation Science 6:213-222. Benitez-Malvido, J., M. Martinez-Ramos, J. Camargo, and I. Ferraz. 2005. Responses of seedling transplants to environmental variat ion in contrasting habitats of Central Amazonia. Journal of Tropical Ecology 21:397-406. Benjamin, A., and J. Quintero. 2005. Indigenas se sentarn a negociar. In La Prensa, 13 October 2005. Blair, B. 2001. Effect of soil nutrient he terogeneity on the symmetry of belowground competition. Plant Ecology 156:199-203. Burgess, S., M. Adams, N. Turner, and C. Ong. 1998. The redistribution of soil water by tree root systems. Oecologia 115(3):306-311. Cajas-Giron, Y., and F. Sinclair. 2001. Character ization of multistrata silvopastoral systems on seasonally dry pastures in the Caribbean region of Colombia. Agroforestry Systems 53:215-225. Cahill, J. 2002. Interactions between root and shoot competition vary among species. Oikos 99:101-112. Cahill, J. 2003. Lack of relationship between below-ground competition and allocation to roots in 10 grassland species. Journal of Ecology 91:532-540. Canham, C., R. Kobe, E. Latty, and R. Chazdon. 1999. Interspecific and intraspecific variation in tree seedling survival: eff ects of allocation to roots versus carbohydrate reserves. Oecologia 121:1-11. Casper, B., H. Schenk, and R. Jackson. 2003. Defining a plant's belowground zone of influence. Ecology 84:2313-2321. Censo Agropecuario. 2001. Censo Nacional Agropecuario 2000. In D. d. Estadisitica, ed. Contralora de la Republica de Panam. Chang, S., and D. Mead. 2003. Growth of radiata pi ne (Pinus radiata D. Don) as influenced by understory species in a silvopast oral system in New Zealand. Agroforestry Systems 59:4351. 150

PAGE 151

Clason, T. 1999. Silvopastoral practices sustain timber and forage production in commercial loblolly pine plantations of northwest Louisiana, USA. Agroforestry Systems 44:293-303. Coates, A. 1997. The Forging of Central America. Pp. 1-37. In A. Coates, ed. Central America: A Natural and Cultural History. Yale Univ ersity Press, New Haven, Connecticut. Coll, L., P. Balandier, and C. Picon-Cochar d. 2004. Morphological and physiological responses of beech (Fagus sylvatica) seedlings to grass-induced belowground competition. Tree Physiology 24:45-54. Collet, C., M. Lof, and L. Pags. 2006. Root sy stem development of oak seedlings analyzed using an architectural model. Effects of competition with grass. Plant and Soil 279:367383. Cooke, R. 1997. The Native Peoples of Centra l America during Precolumbian and Colonial Times. Pp. 137-176. In A. Coates, ed. Centra l America: A Natural and Cultural History. Yale University Press, New Haven, Connecticut. Cordero, L., and M. Kanninen. 2002. Wood speci fic gravity and aboveground biomass of Bombacopsis quinata plantations in Costa Ri ca. Forest Ecology and Management 165:1-9. Cordero, L., M. Kanninen, and L. Arias. 2003. Sta nd growth scenarios for Bombacopsis quinata plantations in Costa Rica. Forest Ecology and Management 59:174:345-352. Dagang, A. 2004. A Summary of Agricultural Land Use and Land Tenure in the Republic of Panama. PRORENA, Smithsonian Tropi cal Research Institute, Panama. Dagang, A., F. Herrera, and R. Gonzalez. 2003. Unbr idled expansion and the continuation of the conquest: the agricultura l frontier and consequences for the environment and indigenous populations in eastern Panama. 2003 Meeting of the Latin American Studies Association. Dallas, Texas. Dagang, A., and P. Nair. 2003. Silvopastoral resear ch and adoption in Central America: recent findings and recommendations for future directions. Agroforestry Systems:149-155. Davis, M., L. Bier, E. Bushelle, C. Diegel, A. Johnson, and B. Kujala. 2005. Non-indigenous grasses impede woody succession. Plant Ecology 178:249-264. Denslow, J., A. Uowolo, and R. Hughes. 2006. Limitations to seedling establishment in a mesic Hawaiian forest. Oecologia 148:118-128. de Souza, R., R. Ribeiro, E. Machado, R. de Oliveira, and J. da Silveira. 2005. Photosynthetic responses of young cashew plants to varying environmental conditions. Pesq. agropec. Bras. 40(8):735-744. de Villalobos, A., D. Pelaez, and O. Elia. 2005. Gr owth of Prosopis caldenia Burk: seedlings in central semi-arid rangelands of Argentina. Journal of Arid Environments 61:345-356. 151

PAGE 152

Divakara, B., B. Kumar, P. Balachandran, a nd N. Kamalam. 2001. Bamboo hedgerow systems in Kerala, India: root distribution and competition with trees for phosphorous. Agroforestry Systems (51):189-200. Douglas, G., A. Walcroft, S. Hurst, J. Potter, A. Foote, L. Fung, W. Edwards, and C. van de Dijssel. 2006. Interactions between widely spaced young poplars (Populus spp.) and introduced pasture mixes. Agroforestry Systems 66:165-178. Dulormne, M., J. Sierra, R. Bonhomme, and Y. Cabidoche. 2004. Seasonal ch anges in tree-grass complementarity and competition for water in a subhumid tropical silvopastoral system. European Journal of Agronomy 21(3):311-322. Durr, P., and J. Rangel. 2002. Enhanced forage production under Samanea saman in subhumid tropical grassland. Agroforestry Systems (54):99-102. Emerman, S., and T. Dawson. 1996. Hydraulic lift a nd its influence on the water content of the rhizosphere: an example from sugar maple, Acer saccharum. Oecologia 108(2):273-278. Espigares, T., A. Lopez-Pintor, and J. Bena yas. 2004. Is the interaction between Retama sphaerocarpa and its understorey herbaceous vegetation always reciprocally positive? Competition-facilitation shift during Retama establishment. Acta Oecologica 26:121-128. Evans, S., A. Pelster, W. Leininger, and M. Trlica. 2004. Seasonal diet sel ection of cattle grazing a montane riparian community. Journal of Range Management 57:539-545. FAO UNESCO. 1972. Soil Map of the World. Prepared by the Food and Agriculture Organization of the United Nations. Rome, Italy. FAOSTAT. 2006. Agricultural Data. Land Use. Food and Agriculture Organization. http://faostat.fao.org/ 6/16/2006. Feldhake, C. 2001. Microclimate of a natural pa sture under planted Robinia pseudoacacia in central Appalachia, West Virginia Agroforestry Systems (53):297-303. Feldhake, C. 2002. Forage frost protection potential of conifer silvopastures. Agricultural and Forest Meteorol ogy 112(2):123-130. Fernandez, M., J. Gyenge, G. Dalla Salda, a nd T. Schlichter. 2002. Silvopastoral systems in northwestern Patagonia I: growth and photosynt hesis of Stipa speci osa under different levels of Pinus ponderosa cover. Agroforestry Systems 55:27-35. Fernandez, M., J. Gyenge, and T. Schlichter. 20 04. Shade acclimation in the forage grass Festuca pallescens: biomass allocation and foliage orientation. Agroforestry Systems 60:159-166. Fernandez, M., J. Gyenge, and T. Schlichter. 20 06. Growth of Festuca palle scens in silvopastoral systems in Patagonia, Part 1: positive balance between competition and facilitation. Agroforestry Systems 66:259-269. 152

PAGE 153

Ferreira, A. 2004. Nutritive value of elephant grass silage added byproducts from the cashew juice industry. Revista Brasileira de Zoot ecnia-Brazilian Journal of Animal Science 33(6):1380-1385. Florentine, S., and J. Fox. 2003. Competition betw een Eucalyptus victrix seedlings and grass species. Ecological Research 18:25-39. Gakis, S., K. Mantzanas, D. A lifragis, V. Papanastasis, A. Papaioannou, D. Seilopoulos, and P. Platis. 2004. Effects of unde rstorey vegetation on tree establishment and growth in a silvopastoral system in northern Gr eece. Agroforestry Systems 60:149-157. Ganskopp, D., and D. Bohnert. 2006. Do pasturescale nutritional patterns affect cattle distribution on rangelands? Rangela nd Ecology and Management 58:189-196. Garrett, H., M. Kerley, K. Ladyman, W. Walter, L. Godsey, J. Van Sambeck, and D. Brauer. 2004. Hardwood silvopasture management in North America. Agroforestry Systems 61:2133. Gordon, P. 2001. Personal communication. RAIA Meeting. IDIAP. Panama. Gordon, D., and K. Rice. 2000. Competitive su ppression of Quercus douglasii (Fagaceae) seedling emergence and growth. American Journal of Botany 87:986-994. Griscom, H., P. Ashton, and G. Berlyn. 2005. Seedling survival and growth of native tree species in pastures: Implications for dry tropical fo rest rehabilitation in central Panama. Forest Ecology and Management 218:306-318. Hangs, R., J. Knight, and K. Van Rees. 2002. Interspecific competition for nitrogen between early successional species and planted white spruce and jack pine seedlings. Can. J. For. Res. 32:1813-1821. Harmer, R., and M. Robertson. 2003 Seedling root growth of six broadleaved tree species grown in competition with grass under irrigated nursery conditions. Annals of Forest Science 60:601-608. Harrington, T., C. Dagley, and M. Edwards. 2003. Aboveand belowground competition from longleaf pine plantations limits performance of reintroduced herbaceous species. Forest Science 49:681-695. Haukioja, E., and J. Koricheva. 2000. Tolerance to herbivory in woody vs. herbaceous plants. Evolutionary Ecology 14:551-562. Heckadon-Moreno, S. 1997. Spanish rule, independen ce, and the modern colonization frontiers. Pp. 177-214. In A. Coates, ed. Central Ameri ca: A Natural and Cultural History. Yale University Press, New Haven, Connecticut. 153

PAGE 154

Herrera, F. 1986. Analisis socioeconmico de la Cuenca Alta de Bayano. Pp. 121-167. In Centro Agronomico Tropical de Invest igacion y Enseanza, ed. Memorias de Taller sobre la Problematica de la Cuenca Hidrografia de Bayano. CATIE, Costa Rica. Hester, A., P. Millard, G. Ba illie, and R. Wendler. 2004. How does timing of browsing affect aboveand below-ground growth of Betula pendula, Pinus sylvestris and Sorbus aucuparia? Oikos 105:536-550. Hinde, R., G. Corti, E. Fanning, and R. Je nkins. 2001. Large mammals in miombo woodlands, evergreen forest and a young teak (Tectona gr andis) plantation in the Kilombero Valley, Tanzania. African Journal of Ecology 39:318-321. Holdridge, L. 1967. Life Zone Ecology. Tropica l Science Center, San Jose, Costa Rica. Holl, K. 1998. Effects of above and belowg round competition of shrubs and grass on Calophyllum brasiliense (Camb.) seedling growth in abandoned tropi cal pasture. Forest Ecology and Management 109:187-195. Husen, A., and M. Pal. 2006. Variation in shoot anatomy and rooting behaviour of stem cuttings in relation to age of donor plan ts in teak (Tectona grandis Li nn f.). New Forests (31):5773. Huxley, P. 1999. Tropical Agroforestry. Blackwell Science, Oxford. Isaac, S., and M. Nair. 2005. Biodegradation of leaf litter in the warm humid tropics of Kerala, India. Soil Biology and Biochemistry 37:1656-1664. Izquierdo, I., F. Caravaca, M. Alguacil, G. Hernandez, and A. Roldan. 2005. Use of microbiological indicators for ev aluating success in soil restor ation after revegetation of a mining area under subtropical conditi ons. Applied Soil Ecology 30:3-10. Jaen, O. 1985. La Colonizacin Campesina de Bos ques Tropicales de Panama. Pp. 379-392. In W. D'Arcy, and M. Correa, eds. The Bota ny and Natural History of Panama. Missouri Botanical Garden, Missouri. Jaffe, K., and E. Vilela. 1989. On ne st densities of the leaf-cutting ant Atta cephalotes in tropical primary forest 21:234-236. Heckadon, S. 1983. Cuando Se Acaban Los Montes: los Campesinos Santeos y la Colonizacin de Tonos. Impretex, SA, Panama. Jenkins, R., K. Roettcher, and G. Corti. 2003. The influence of stand age on wildlife habitat use in exotic Teak tree Tectona grandis plan tations. Biodiversity and Conservation 12 : 975990. Joly, L. 1989. The conversion of rain Forests to pa sture in Panama. Pp. 3-18. In J. Schelhas, and R. Greenberg, eds. Forest Patches in Tropi cal Landscapes. Island Press, Washington DC. 154

PAGE 155

Jordan, C. 2004. Organic farming and agrofore stry: Alleycropping for mulch production for organic farms of southeastern United States Pp. 79-90. In P. Nair, M. Rao, and L. Buck, eds. New Vistas in Agroforestry: a Co mpendium of the 1st World Congress of Agroforestry, 2004. Kluwer Academic Publishers, Dordrecht, The Netherlands. Jose, S., A. Gillespie, and S. Pallardy. 2004. Intersp ecific interactions in temperate agroforestry. Agroforestry Systems 61:237-255. Jurena, P., and S. Archer. 2003. Woody plant es tablsihment and spatial heterogeneity in grasslands. Ecology 84:907-919. Kadambi, K. 1972. Silviculture and Management of Teak. Stephen F. Austin University, Texas. Kaewkrom, P., J. Gajaseni, C. Jordan, and N. Ga jaseni. 2005. Floristic rege neration in five types of teak plantations in Thailand. Forest Ecology and Management 210:351-361. Kallenbach, R., M. Kerley, and G. Bishop-Hurley. 2006. Cumulative fora ge production, forage quality, and livestock performa nce from an annual ryegrass and cereal rye mixture in a pine-walnut silvopasture. Ag roforestry Systems 66:43-53. Kammesheidt, L. 1998. Stand structure and spatia l pattern of commercial species in logged and unlogged Venezuelan forest. Forest Ecology and Management 109:163-174. Kraenzel, M., A. Castillo, T. Moore, and C. Po tvin. 2003. Carbon storage of harvest-age teak (Tectona grandis) plantations, Panama. Forest Ecology and Management 173:213-225. Lin, C., R. McGraw, M. George, and H. Garrett. 2001. Nutritive quality and morphological development under partial shade of some fora ge species with agroforestry potential. Agroforestry Systems 53: 269-281. Linares, O. 1980. Coastal settlements on the Atlantic s and Pacific sides. Pp. 57-66. In O. Linares, and A. Ranere, eds. Adaptive Radiations in Pr ehistoric Panama. Harvard University Press, Cambridge, Massachusetts. Lindh, B., A. Gray, and T. Spies. 2003. Responses of herbs and shrubs to reduced root competition under canopies and in gaps: a trench ing experiment in old-growth Douglas-fir forests. Can. J. For. Res. 33:2052-2057. Ludwig, F., T. Dawson, H. Prins, F. Be rendse, and H. de Kroon. 2004. Below-ground competition between trees and grasses may overwhelm the facilitative effects of hydraulic lift. Ecology Letters 7:623-631. Major, J., C. Clement, and A. DiTomasso. 2005. Influence of market orientation on food plant diversity of farms located on Amazonian dark earth in the region of Manaus, Amazonas, Brazil. Economic Botany 59(1):77-86. 155

PAGE 156

Matthews, E., and L. Guzman. 1955. The Soils and Agriculture of the Llanos de Cocle. United States of America Operations Mission to Pana ma and Ministerio de Agricultura, Comercio e Industrias, Republica de Panama. Mendona, F., K. da Silva, K. dos Santos, K. Ribeiro, and A. Sant'Ana. 2005. Activities of some Brazilian plants against larvae of the mos quito Aedes aegypti. Fito terapia (76):629-636. Menezes, R., I. Salcedo, and E. Elliott. 2002. Microclimate and nutrient dynamics in a silvopastoral system of semiarid northeaste rn Brazil. Agroforestry Systems 56(1):27-38. Midoko-Iponga, D., C. Krug, and S. Milton. 2005. Competition and herbivory influence growth and survival of shrubs on old fields: Implica tions for restoration of renosterveld shrubland. Journal of Vegetation Science 16:685-692. Miller, R., and P. Nair. 2006. Indigenous agroforest ry systems in Amazonia: from prehistory to today. Agroforestry Systems 66(2):151-164. Moore, J., and G. Mott. 1974. Recovery of residual organic matter from in vitro digestion of forages. Journal of Dairy Science 57:1258-1259. Murphy, P., and A. Lugo. 1995. Dry forests of Cent ral America and the Caribbean. Pp. 9-34. In S. Bullock, H. Mooney, and E. Medina, eds. Seasonally Dry Tropical Forests. Cambridge University Press, Cambridge. Nair, P. 1993. An Introduction to Agroforestry. Kluwer Academic Publishers, Dordrecht, The Netherlands. Nair, P., R. Buresh, D. Mugendi, and C. Latt. 1999. Nutrient Cycling in Tropical Agroforestry Systems: Myths and Science. Pp. 1-32. In L. Buck, J. Lassoie, and E. Fernandes, eds. Agroforestry in Sustainable Agricult ural Systems. CRC Press, Boca Raton. Name, B. 2003. Interview with official from the Institute for Agricultural Development of Panama (IDIAP). NASA-SERVIR. 2006. The Mesoamerican Visualization and Monitoring System. NASA. http://servir.nsst c.nasa.gov/home.html 3/16/2006. Nilsson, U., and G. Orlander. 1999. Water uptake by planted Picea abies in relation to competing field vegetation and seedling rooting depth on two grass-dom inated sites in southern Sweden. Scandinavian Journal of Forest Research 14:312-319. Ngatunga, E., S. Dondeyne, and J. Deckers. 2003. Is sulphur acidifying cashew soils of South Eastern Tanzania? Agriculture, Ecos ystems, and Environment 95:179-184. Oguntunde, P., and N. van de Giesen. 2005. Wate r flux measurements and prediction in young cashew trees using sap flow data. Hydrological Processes 19:3235-3248. 156

PAGE 157

Ong, C., C. Black, F. Marshall, and J. Corlet t. 1996. Principles of Resource Capture and Utilization of Light and Water. Pp. 73-158. In C. Ong, and P. Huxley, eds. Tree-Crop Interactions: a Physiological Approach. CABI, Oxford. Pagiola, S., P. Agostini, J. Gobbi, C. de Haa n, M. Ibrahim, E. Murgueitio, E. Ramirez, M. Rosales, and J. Ruiz. 2004. Paying for biodivers ity conservation servic es in agricultural landscapes. P. 33. In E. Department, ed. The World Bank. Peltzer, D., and M. Kochy. 2001. Competitive eff ects of grasses and woody plants in mixedgrass prairie. Journal of Ecology 89:519-527. Perez, D., and M. Kanninen. 2005. Stand growth fo r Tectona grandis planta tions in Costa Rica. Forest Ecology and Management 210:425-441. Peri, P., E. Mason, K. Pollock, A. Varella, and D. Mead. 2002. Early growth and quality of radiata pine in a silvopastoral system in New Zealand. Agroforestry Systems 55:207-219. Piotto, D., F. Montagnini, L. Ugalde, and M. Kanninen. 2003. Performance of forest plantations in small and medium-sized farms in the Atlant ic lowlands of Costa Rica. Forest Ecology and Management 175:195-204. Piotto, D., E. Vquez, F. Montagnini, and M. Kanninen. 2004. Pure and mixed forest plantations with native species of the dry tropics of Costa Rica: a comparison of growth and productivity. Forest Ecology and Management 190:359-372. Piperno, D., and D. Pearsall. 1998. The Origins of Agriculture in the Lowland Neotropics. Academic Press, San Diego. Platt, K., R. Allen, D. Coomes, and S. Wiser. 2004. Mountain beech seedling responses to removal of below-ground competition and fert iliser addition. New Zealand Journal of Ecology 28:289-293. Posada, J., T. Aide, and J. Cavelier. 2000. Cattle and weedy shrubs as restor ation tools of tropical montane rainforest. Restoration Ecology 8:370-379. Priestley, G. 1981. Military Government and Po pular Participation in Panama: the Torrijos Regime 1968-1975. Ph.D. dissertation, Columbia University, New York. Rajaniemi, T., V. Allison, and D. Goldberg. 200 3. Root competition can cause a decline in diversity with increased producti vity. Journal of Ecology 91:407-416. Ramirez-Marcial, N. 2003. Survival and growth of tree seedlings in anth ropogenically disturbed Mexican montane forests. Journal of Vegetation Science 14:881-890. Ramsey, C., S. Jose, B. Brecke, and S. Merritt. 2002. Growth response of longleaf pine (Pinus palustris Mill.) seedlings to fertilization a nd herbaceous weed control in an old field in southern USA. Forest Ecology and Management 172:281-289. 157

PAGE 158

Rao, M. 2000. Variation in leaf-cutter ant (Atta sp.) densities in forest isol ates: the potential role of predation. Journal of Tropical Ecology 16:209-225. Rao, M., J. Terborgh, and P. Nuez. 2001. Increased herbivory in forest isol ates: Implications for plant community strucutre and composition. Conservation Biology 15:624-633. Rietkerk, M., R. Blijdorp, and M. Slingerland. 1998. Cutting and resprouting of Detarium microcarpum and herbaceous forage availabil ity in a semiarid environment in Burkina Faso. Agroforestry Systems 41:201-211. Rudolf, G. 1999. Panama's Poor. University Press of Florida, Gainesville, Florida. Saha, S. 2001. Vegetation composition and stru cture of Tectona grandis (teak, Family Verbanaceae) plantations and dry deciduous fore sts in central India. Forest Ecology and Management 148:159-167. Sanchez, P. 1999. Improved fallows come of age in the tropics. Agroforestry Systems 47:3-12. Sanchez, A., and B. Peco. 2004. Interference between perennial grassland and Lavandula stoechas subsp. pedunculata seedlings: a case of spatial segregation cause by competition. Acta Oecologica 26:39-44. Schenk, H. 2006. Root competition: beyond reso urce depletion. Journal of Ecology 94:725-739 Serrao, E., and J. Toledo. 1990. The Search for Su stainability in Amazonian Pastures. Pp. 195214. In A. Anderson, ed. Alternatives to Defo restation. Columbia University Press, New York. Shankar, D., S. Lama, and K. Bawa. 1998. Ecosystem reconstruction through 'taungya' plantations following commerical logging of a dry, mixed deciduous forest in Darjeeling, Himalaya. Forest Ecology and Management 102:131-142. Sharrow, S. 1999. Silvopastoralism: Competition and facilitation between trees, livestock, and improved grass-clover pastures on temperate rainfed lands. In LE Buck, JP Lassoie, and ECM Fernandes, eds. Agroforestry in Sustai nable Agricultural Systems. CRC Press, Boca Raton. Silva-Pando, F., M. Gonzalez-Hernandez, and M. Rozados-Loren zo. 2002. Pasture production in a silvopastoral system in relation with microc limate variables in the Atlantic coast of Spain. Agroforestry Systems 56:203-211. Stammel, B., K. Kiehl, and J. Pfadenhauer. 2006. Effects of experimental and real land use on seedling recruitment of six fen speci es. Basic and Applied Ecology 7:334-346. Teklehaimanot, Z., M. Jones, and F. Sinclair. 200 2. Tree and livestock pro ductivity in relation to tree planting configuration in a silvopastora l system in North Wales, UK. Agroforestry Systems 56:47-55. 158

PAGE 159

Terborgh, J., K. Feeley, M. Silman, P. Nuez, and B. Balukjian. 2006. Vegetation dynamics of predator-free land-bridge isla nds. Journal of Ecology 94:253-263. Thevathasan, N., P. Reynolds, R. Kuessner, and W. Bell. 2000. Effects of controlled weed densities and soil types on so il nitrate accumulation, spruce growth, and weed growth. Forest Ecology and Management 133:135-144. Tournebize, R. 1994. Microclimat lumieaux et tran spiration d'une association arbuste/herbe, en milieu tropical: mesures et modelisation Ph.D dissertation, University of Paris, Orsay, France. Vasconcelos, H., and J. Cherrett. 1997. Leaf-cutting ants and early forest re generation in central Amazonia: effects of herbivory on tree seedli ng establishment. Journal of Tropical Ecology 13:357-370. Villalobos, A. 2003. Interview with official from the Ministry of Agricultural Development (MIDA). Panama. Webb, S. 1997. The Great American Faunal Interc hange. Pp. 92-122. In A. Coates, ed. Central America: a Natural and Cultural History. Yale University Press, New Haven, Connecticut. 159

PAGE 160

BIOGRAPHICAL SKETCH Aly Dagang was born in Los Angeles, Calif ornia in 1972. She graduated from the American University in Washington, D.C. in 1994 where she earned her B.A. in international development and Spanish/Latin American studi es. After graduation, she worked as an agroforestry extension volunteer in the Peace Co rps in Panama until 1998. Following her return to the United States, Aly began her graduate studies in agroforestry at the University of Florida. Currently, she is the Academic Director for the School for International Training in Panama and member of the Alianza para la Conservacion y Desarrollo. 160


Permanent Link: http://ufdc.ufl.edu/UFE0018083/00001

Material Information

Title: Establishment of Silvopastoral Systems in Degraded, Grazed Pastures: Tree Seedling Survival and Forage Production under Trees in Panama
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0018083:00001

Permanent Link: http://ufdc.ufl.edu/UFE0018083/00001

Material Information

Title: Establishment of Silvopastoral Systems in Degraded, Grazed Pastures: Tree Seedling Survival and Forage Production under Trees in Panama
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0018083:00001


This item has the following downloads:


Full Text





ESTABLISHMENT OF SILVOPASTORAL SYSTEMS IN DEGRADED, GRAZED
PASTURES: TREE SEEDLING SURVIVAL AND FORAGE PRODUCTION UNDER TREES
IN PANAMA




















By

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

2007
































Copyright 2007


by

Alyson B.K. Dagang


































To my Mother and Father, whose boundless love gives me life









ACKNOWLEDGMENTS

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


page

ACKNOWLEDGMENT S ................. ...............4....___ ......


LIST OF TABLES ............ ............ ...............10...


LIST OF FIGURES ............ ............ ...............11...


AB S TRAC T ............._. .......... ..............._ 13...


CHAPTER


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....


Introduction............... ..............6
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........... ....


Introduction............... ..............8
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....
Discussion............... ...............9

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...


Introduction............... .............11
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...

APPENDIX

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


Table page

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


Figure page

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
IN PANAMA


By

Alyson B. K. Dagang

May 2007

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.









CHAPTER 1
INTTRODUCTION

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
seedling survival.

* Isolated, large trees impact mass, digestibility, and botanical composition of the understory
forage.










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.









CHAPTER 2
AGROFORESTRY AND LAND USE INT PANAMA AND A GENERAL DESCRIPTION OF
THE STUDY SITE

Agroforestry

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

chapter.

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

degradation.

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.

Silvopastoral Systems

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

capacity .

Microclimate

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

2. 1).

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

physiological outcomes.

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

silvopastoral systems.

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.

Summary

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
Presented Research

Introduction

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

use today.

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

use today.

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

landscape.

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

(Jaen, 1985).

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,

1986).

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

these issues.

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.









Pasture Proliferation

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

management.

Conclusion

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

Location

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).

Ecology

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.

Climate

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

types, including:

* 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.

Species Descriptions

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

correspondingly unique.

Tectona grandis

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.









Uses


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).

Botany

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

species.

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-

rooted species.

Burning

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

benefit.

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).

Anacardium occidentale

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

(Behrens, 1996).

Botanical description

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).

Cultivation

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

withstand drought.

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.

Uses

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.

Planting Configuration

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.





1 ~j'T~T


i


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.












5.0


4.5


4.0


3.5


3.0


Millions 2.5


2.0


1.5


1.0


0.5


0.0


5


1961 1986 1994 2003
Years

+ 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.
























402


500,000


400,000

300,000

200,000
99 160
100,000


3,8


Total farm area (ha)
No. of farms


0.5 19.99
20 -49.9950-9.9


19.9 200 -
499.99


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.

Proportion
Total of
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.
Average
household
Earnings from monthly Farmstead
Province cattle ($ IM) income ($) poulation
Bocas del Toro 2.6328.03,2
Cocle 6.44 220.60 113,764
Col~n4.41377.60 36,830
Chirqul 6.02302.10 140,909
Darie 5.28116.50 21,016
Herrer 9.56249.80 55,743

Los Santos 26.4623.04,8
Panama13.67540.40 486,201

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 (%")
| 0-25
]26-50
i 51-75
> 76
no data

6~r2 .. I~ Corregirniento boundary


N




S











































Figure 2-6 Research study site location, Rio Grande corregimiento, Cocle province, Republic of
Panama.


S ource: www.Iib.utexas.edulmap s/cia00.html


40I 80 km
40


SO ri


Bocas del
Toro


Cooo Sole

PaANAMA
Blbo
Vaucarnonte


Da~vid


.*****
*, La Palma *
....*
Sanrtiago,
Chitr6




Research study site


COL OMBI A


CO STA
raICA









CHAPTER 3
TREE SEEDLING SURVIVAL AND IMPACT OF HERBIVORY ON SILVOPASTORAL
SYSTEM ESTABLISHMENT

Introduction

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.









Literature Review


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

nutrients.

Herbivory

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

herbivory .

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

in Panama.

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

Study Site

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).

Experimental Design

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

each pasture.

Materials

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.









Establishment

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.

Measurements

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.

Data Analysis

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

herbivory data.












Seedling Survival

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

3-1).

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,

respectively .

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,


Results









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.

Herbivory

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).

Discussion

Seedling Survival

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.

Herbivory

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

congregate.

Conclusion

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

herbivory .









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
Cocle, Panama.

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












1.0 -
0.9 -
0.8 -
0.7 -
0.6 -
0.5 -
0.4 -



0.1 -
0.0
0 100 200 300 400 500 600 700 800 900
days

Diagonal

1.0
0.9 -
0.8 -
0.7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
0.0 11111111
0 100 200 300 400 500 600 700 800 900
days

Fence


1.0
0.9 -
0.8 -



0.5 -



0.2 -
0.1 -
0.0 lillllli
0 100 200 300 400 500 600 700 800 900
days

Line
- 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

SB. quinata


50



40


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.















O Cattle
MLeaf-cutter ants
O Other


700



600



500



Incidence of herbnvory 400


B. quinata

Species


T. grandzs


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.


A. occidentale

















450


400


350


300


Incidence of herblvory 250


200


150


100


50


O Cattle

SLeaf-cutter ants

O Other


Diagonal


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
by herbivores.









CHAPTER 4
EFFECTS OF SCATTERED LARGE TREES IN PASTURES ON A Hyparrhenia rufa-
DOMINATED MIXED SWARD

Introduction

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.

Literature Review

Light

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

were similar.

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.

Biomass Allocation

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.

Belowground Factors

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

Study Site

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.

Experimental Design

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.

Measurements

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.

Data Analysis

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.

Results

Forage Mass

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

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).

Forage Composition

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).









Discussion


Forage Mass

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

investigation.

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.

Forage Digestibility