<%BANNER%>

Seedling Recruitment of Large-Seeded Tropical Trees Planted As Seeds in the Ecuadorian Amazon

Permanent Link: http://ufdc.ufl.edu/UFE0025046/00001

Material Information

Title: Seedling Recruitment of Large-Seeded Tropical Trees Planted As Seeds in the Ecuadorian Amazon
Physical Description: 1 online resource (99 p.)
Language: english
Creator: Van Etten, Erica
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: amazon, caryodendron, cordata, densiflora, direct, dispersal, ecuador, forestry, gustavia, inga, large, macarenensis, orinocense, plinia, pouteria, predation, quararibea, recruitment, reforestation, seed, seeded, seeding, seedling, shuar, tropical
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Natural regeneration of large-seeded tropical forest trees in forests and pastures in landscapes fragmented by agriculture can be severely limited by lack of animal-mediated seed dispersal. Seeds that arrive by rare cases of long-distance dispersal may be critical to species persistence, provided they survive in the new sites. Planting seeds of these species may be an inexpensive method to add economic value to the landscape. The purpose of this study was to assess some of the factors affecting seedling recruitment of large-seeded tropical forest trees in human-modified landscapes in the Ecuadorian Amazon. In the first experiment (Chapter 2), seeds of Pouteria caimito (Sapotaceae), Quararibea cordata (Malvaceae), and Inga densiflora (Fabaceae) were planted in pastures and forests, and seed survival and seedling development was recorded at 2 and 14 weeks. In the second experiment (Chapter 3), I planted seeds of Inga densiflora, Gustavia macarenensis (Lecythidaceae), Caryodendron orinocense (Euphorbiaceae), and Plinia sp. (Myrtaceae) in secondary forests and survival and seedling development were monitored biweekly for 18 weeks. In both experiments, seeds were planted on the soil surface or were buried. A caging treatment was included in the second experiment to assess seed predation levels. Over 55% of Inga, Caryodendron and Gustavia (56% and 76% in the two experiments, 76%, and 64%, respectively), 17% of Plinia , 10% of Quararibea, and 2% of Pouteria seeds developed into seedlings and survived during the 14?18 week periods of the two experiments. Seed burial did not increase survival or seedling recruitment of any species in either experiment. In the first experiment, the location of the plots in the landscape (i.e. the indigenous community in which they were planted) affected seed removal and seedling recruitment more than did the habitat (grazed pasture, mature pasture, forest) in which the seeds were planted. Seed removal was extremely low ( < 1%) in secondary forests located < 1 km from a main road. The results of this study suggest seedlings recruit from seeds of large-seeded tropical forest species if dispersal limitations are overcome. Species differences in initial seedling recruitment can be expected, and seed predation and seedling recruitment vary among sites.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Erica Van Etten.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Kitajima, Kaoru.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025046:00001

Permanent Link: http://ufdc.ufl.edu/UFE0025046/00001

Material Information

Title: Seedling Recruitment of Large-Seeded Tropical Trees Planted As Seeds in the Ecuadorian Amazon
Physical Description: 1 online resource (99 p.)
Language: english
Creator: Van Etten, Erica
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: amazon, caryodendron, cordata, densiflora, direct, dispersal, ecuador, forestry, gustavia, inga, large, macarenensis, orinocense, plinia, pouteria, predation, quararibea, recruitment, reforestation, seed, seeded, seeding, seedling, shuar, tropical
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Natural regeneration of large-seeded tropical forest trees in forests and pastures in landscapes fragmented by agriculture can be severely limited by lack of animal-mediated seed dispersal. Seeds that arrive by rare cases of long-distance dispersal may be critical to species persistence, provided they survive in the new sites. Planting seeds of these species may be an inexpensive method to add economic value to the landscape. The purpose of this study was to assess some of the factors affecting seedling recruitment of large-seeded tropical forest trees in human-modified landscapes in the Ecuadorian Amazon. In the first experiment (Chapter 2), seeds of Pouteria caimito (Sapotaceae), Quararibea cordata (Malvaceae), and Inga densiflora (Fabaceae) were planted in pastures and forests, and seed survival and seedling development was recorded at 2 and 14 weeks. In the second experiment (Chapter 3), I planted seeds of Inga densiflora, Gustavia macarenensis (Lecythidaceae), Caryodendron orinocense (Euphorbiaceae), and Plinia sp. (Myrtaceae) in secondary forests and survival and seedling development were monitored biweekly for 18 weeks. In both experiments, seeds were planted on the soil surface or were buried. A caging treatment was included in the second experiment to assess seed predation levels. Over 55% of Inga, Caryodendron and Gustavia (56% and 76% in the two experiments, 76%, and 64%, respectively), 17% of Plinia , 10% of Quararibea, and 2% of Pouteria seeds developed into seedlings and survived during the 14?18 week periods of the two experiments. Seed burial did not increase survival or seedling recruitment of any species in either experiment. In the first experiment, the location of the plots in the landscape (i.e. the indigenous community in which they were planted) affected seed removal and seedling recruitment more than did the habitat (grazed pasture, mature pasture, forest) in which the seeds were planted. Seed removal was extremely low ( < 1%) in secondary forests located < 1 km from a main road. The results of this study suggest seedlings recruit from seeds of large-seeded tropical forest species if dispersal limitations are overcome. Species differences in initial seedling recruitment can be expected, and seed predation and seedling recruitment vary among sites.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Erica Van Etten.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Kitajima, Kaoru.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025046:00001


This item has the following downloads:


Full Text

PAGE 1

1 SEEDLING RECRUITMENT OF LARGE SEEDED TROPICAL TREES PLANTED AS SEEDS IN THE ECUADORIAN AMAZON By ERICA VAN ETTEN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIRE MENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Erica Van Etten

PAGE 3

3 To my family, and all those that guide me along the way

PAGE 4

4 ACKNOWLEDGMENTS This thesis was improved by input from my major a dvisor, Kaoru Kitajima, and committee members Jack Putz and Doug Levey. Financial support was from a Graduate Research Assistant ship from the School of Natural Resources and the Environment and a Summer Field Grant by the Tropical Conservation and Development Program at the University of Florida. Alfonso Wajuyat a Sharina Cruz worked closely with me for three years in Amazonian Ecuador, hou sed me with his family, taught m e about the forest, and secured connections for me with local landowners so that I could conduct this research. Rovin Yasaca, a student at the Universidad Central in Quito, assisted me with fieldwork, and Ana Mariscal provided institutional support at the Herbario Nacional de Ecuador, Quito. This work would not have been possible but for numer ous friends and my family who provided essential encouragement and suppor t both while I was in the field and during the process of writing.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 7 LIST OF FIGURES .............................................................................................................................. 8 ABSTRACT .......................................................................................................................................... 9 CHAPTER 1 INTRODUCTION ....................................................................................................................... 11 2 SEED AND EARLY SEEDLING SURVIVAL OF THREE LARGE -SEEDED TROPICAL FOREST TREES PLANTED AS SEEDS IN PASTURES AND FORESTS IN THE ECUADORIAN AMAZON ......................................................................................... 17 Introduction ................................................................................................................................. 17 Materials and Methods ................................................................................................................ 20 Study Site ............................................................................................................................. 20 Tree Species Selection ......................................................................................................... 22 Experimental Design ........................................................................................................... 23 Data Analyses ...................................................................................................................... 25 Results .......................................................................................................................................... 26 General Patterns of Seeding Recruitment and Survival .................................................... 26 Species Differ ence in Seedling Recruitment and Mortality .............................................. 26 Effects of Seed Burial .......................................................................................................... 28 Community Differences in Seed and Seedling Survival ................................................... 28 Seedling Recruitment in Pastures and Forests ................................................................... 29 Discussion .................................................................................................................................... 29 Lan dscape -Level Variation ................................................................................................. 30 Microhabitat Effects on Seed Predation ............................................................................. 32 Microhabitat Effects on Seedling Recruitment and Survival ............................................ 34 Differences Among the Three Large -seeded Species ....................................................... 35 Recommendations for Propagation of the Study Species ................................................. 37 Recommendations for Future Studies ................................................................................ 37 Conclusion ................................................................................................................................... 38 3 SEEDLING RECRUITMENT OF FOUR LARGE SEEDED TROPICAL FOREST TREES PLANTED AS SEEDS IN SECONDARY FORESTS IN THE ECUADORIAN AMAZON .................................................................................................................................... 46 Introduction ................................................................................................................................. 46 Materials a nd Methods ................................................................................................................ 48 Study Site ............................................................................................................................. 48

PAGE 6

6 Tree Species ......................................................................................................................... 48 Secondary Forest Plots ........................................................................................................ 50 Experimental Design ........................................................................................................... 51 Data Analyses ...................................................................................................................... 52 Results .......................................................................................................................................... 52 General Patterns of Seeding Recruitment and Survival .................................................... 52 Species Differences ............................................................................................................. 53 Seed Burial ........................................................................................................................... 53 Plots ...................................................................................................................................... 54 Caging ................................................................................................................................... 54 Species Inventory ................................................................................................................. 54 Discussion .................................................................................................................................... 55 Low Seed Removal .............................................................................................................. 55 High Germination ................................................................................................................ 57 Seed Burial ........................................................................................................................... 57 Secondary Forests ................................................................................................................ 58 Recommendations for Propagation of the Study Species ................................................. 59 Recommendations for Future Studies ................................................................................ 59 Conclusion ................................................................................................................................... 60 4 CONCLUSION ........................................................................................................................... 64 APPENDIX CHARACTERIZATION OF RESEARCH SITES AND SPECIES USED IN THE EXPERIMENTS ................................................................................................................ 67 LIST OF REFERENCES ................................................................................................................... 84 BIOGRAPHICAL SKETCH ............................................................................................................. 98

PAGE 7

7 LIST OF TABLES Table page 2 1 Results from mixed model logisti c regress ion of missing seeds, mortality and seedling recruitment at 2 weeks of three large -seeded tropical trees planted as seeds in three field types (grazed pasture, mature pasture, forest) in three indigenous communities .......................................................................................................................... 39 2 2 Results from mixed model logistic regression of mortality, seedling recruitment of three large -seed ed tropical trees at 14 weeks and 12 -week survival of Inga densiflora seedlings. Seeds were planted on the soil surface and buried in three field types (grazed pasture, mature pasture, forest) ............................................................................... 40 3 1 Results from mixed model logistic regression of seedling recruitment, total mortality, and mortality of r emaining seeds at 18 wks of four large -seeded tropical tree species planted as seeds in three secondary forest p lots. .................................................................. 61 A 1 Av erage monthly precipitation, air temperature a nd number of days with rainfall at the Puyo weather station M008 of the Instituto Nacional de Meteorologa en Hidrolgica (INAH MI), Pastaza Province, Ecuador. .......................................................... 67 A 2 Common names, seedling types, cotyledon characteristics, seed sizes, and seedling recruitment of the six large -seeded tropical forest species from the Ecuadorian A mazon used in th is experiment. ......................................................................................... 68 A 3 Nu mber of seeds that were missi ng, dead as seeds and seedlings, and alive as seeds and seedlings at 2 and 14 weeks from planting of three large -seeded tropical forest trees that were planted on the soil surface and buried in three field types (grazed pasture, mature pasture and forest) in three indigenous communities (Kenkuimi, Kunkuki, San Ramon) in the Ecuadorian Amazon. ............................................................ 69 A 4 Vegetation characteristics of three secondary forest plots in the Ecuadorian Amazon calculated from modified Gentry transects. .......................................................................... 72 A 5 Numbers of woody species with dbh encount ered in four 50 x 2 m transects (400 m2) established in each of three secondary forest plots in Pastaza Prov ince in the Ecuadorian Amazon. ........................................................................................................ 73 A 6 Numbers of seeds of four t ropical tree species that were planted, died as seeds, were alive and live seedlings at 18 weeks planted as seeds in secondary forest in the Ecuadorian Amazon. .............................................................................................................. 77 A 7 Comparis on between the tw o experiment s presented in Chapter 2 and Chapter 3 in which species of large -seeded tropical forest trees were planted in a variety of human modified landscapes in the Ecuadorian Amazon. ................................................................ 78

PAGE 8

8 LIST OF FIGURES Figure page 2 1 Map of the study area in Morona Santiago and Pastaza Provinces, Ecuador. .................... 41 2 2 Schematic diagram of the experimental design showing nesting of field plots (grazed pasture, mature pasture, forest) in three indigenous communities ( Kenkuimi, Kunkuki, San Ramon) in the Ecuadorian Amazon. ............................................................. 42 2 3 Proportion of se eds of three large -seeded tropical tree species that were missing at 2 weeks from three field types (grazed pasture, mature pasture, forest) in three indigenous communities (Kenkuimi, Kunkuki, San Ramon) 14 18 km a part in the Ecuadorian Amazon. ............................................................................................................. 43 2 4 Seed fates at 2 weeks of three large -seeded tropical tree species that were planted as seeds in forests and pastures in three indigenous communities (Kenkuimi, Kunkuki, San Ramon) 14 18 km apart in the E cuadorian Amazon (Figure 2 1). ...... 44 2 5 Seed fates at 14 weeks of three large -seeded tropical tree species that were planted as seeds in forests and pastures in three indigenou s communities (Kenkuimi, Kunkuki, San Ramon) 14 18 km apart in the Ecuadorian Amazon ........................... 45 3 1 Seed fates at 18 weeks of four species of large -seeded tropical trees that were planted on the soil surface and buried in three secondary forests plots in the Ecuadorian Amazon. .................................................................................................................................. 62 3 2 Seedling development of four species of large -seeded tropical trees that were planted as seeds on the soil surface and buried in secondary for ests in the Ecuadorian Amazon. .................................................................................................................................. 63 A 1 Average monthly rainfall an d precipitation (19742007) at the Puyo weather station M008 of the Instituto Naci onal de Meteorologa en Hidrolgica (INAH MI), Pastaza Province, Ecuador. ................................................................................................................. 79 A 2 Photos of vegetation in grazed pasture mature pasture and forest plots at the time the experiment descri bed in Chapter 2 was established (June 2007). ........................ 80 A 3 Average percent canopy cover of three types of field plots (grazed p asture, mature pasture, forest) in three indigenous communities measure d with a handheld spherical densitometer held at ground level at 20 points per plot ...................................................... 81 A 4 Photos of the Plinia species ( Myrtaceae ) planted in the expe riment described in Chapter 3. ................................................................................................................................ 82 A 5 Average percent canopy cover of three secondary forests plots measured with a handheld spherical densitometer held at 1 m above groun d level at 20 points per plot. ... 83

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science SEEDLING RECRUITMENT OF LARGE SEEDED TROPICAL TREES PLANTED AS SEEDS IN THE ECUADORIAN AMAZON By Erica Van Etten August 2009 Chair: Kaoru Kitajima Major: Interdisciplinary Ecology N atural regeneration of large -seeded tropical forest trees in forests and pastures in landscapes fragmented by agriculture can be severely limit ed by lack of animal -mediated seed dispersal. Seeds that arrive by rare cases of long distance dispersal may be critical to species persistence, provided they survive in the new sites. In addition to their value for wildlife, many tropical trees with lar ge -seeds have fruits co nsumed by people. Thus, p lant ing seeds of these species may be an inexpensive method to add economic value to the landscape. The purpose of this study was to assess some of the factors affecting seedling recruitment of large -seeded tropical forest trees in human -modified landscapes in the Ecuadorian Amazon. In the first experiment (Chapter 2), seeds of Pouteria caimito (Sapotaceae), Quararibea cordata (Malvaceae) and Inga densiflora (Fabaceae) were planted in pastures and forests, and seed survival and seedling dev elopment was recorded at 2 and 14 weeks In the second experiment (Chapter 3), I planted seeds of Inga densiflora Gustavia macarenensis (Lecythidaceae), Caryodendron orinocense (Euphorbiaceae), and Plinia sp. (M yrtaceae) in secondary forests and survival and seedling development were monitored biweekly for 18 weeks In both experiments,

PAGE 10

10 seeds were planted on the soil surface or were buried. A caging treatment was included in the second experiment to assess seed predatio n levels. Over 55% of Inga, Caryodendron and Gustavia (56% and 76% in the two experiments, 76%, and 64%, respectively) 17% of Plinia 10% of Quararibea, and 2% of Pouteria seeds developed into seedlings and survived during the 14 18 week periods of th e two experiments Seed burial did not increase survival or seedling recruitment of any species in either experiment. In the first experiment, the location of the plots in the landscape ( i e. the indigenous community in which they were planted) affected seed removal and seedling recruitment more than did the habitat (grazed pasture, mature pastur e, forest) in which the seeds were planted. Seed removal was extremely low (< 1 %) in secondary forests located < 1 km from a main road. The results of this stud y suggest seedlings recruit from seeds of large -seeded tropical forest species if dispersal limitations are overcome. Species differences in initial seedling recruitment can be expected, and seed predatio n and seedling recruitment vary among sites.

PAGE 11

11 CHAPTER 1 INTRODUCTION Tropical forests are home to over half of the worlds described species (Dirzo and Raven 2003) contain 46% of the living terrestrial carbon pool (Soepadmo 1993) and are located in countries with tw o thirds of the worlds human population (Wright 2005) By 2030, an additional two billion people will live in tropical countries (United Nations 2004) Economic development in these countries, combined with expanding global economies, increase the demand for forest products (FAO 2005; Fox 2000) resulting in further fragmenta tion and degradation of existing forests (Wade et al. 2003) The future tropical forest s depends in part on the availability of restoration techniques that sustain economic values associated with forests, su pport local livelihoods, improve landsca pe level ecological health and catalyze natural regeneration processes (Chazdon 2008) In this thesis, I report results from experiments in which seeds of eco nomically valuable tropical forest trees are planted in various human -modified landscapes. Survival through the early stages of seedling establishment is critical to the success of this reforestation method. In two field experiments, I tested the influen ce of seed burial and habitat type on the probability of seed predation, seedling recruitment and early seedling survival. Seed dispersal of large -seeded tropical forest trees is often severely interrupted by tropical deforestation and fragmentation. Ma ny trees with large seeds depend on relatively large animals such as tapirs (Fragoso and Huffman 2000) primates (Nunez -Iturri and Howe 2007) and large birds (Howe and Schupp 1985; Meehan et al. 2002) for seed dispersal away from parent trees. Large bodied frugivores are highly susceptible to local extinction in fragmented landscapes because they are generally preferred by hunters (Jerozolimski and P eres 2003; Peres and Palacios 2007) and often require large, continuous tracts of forested habitat (Chiarello 1999) In

PAGE 12

12 agricultural landscapes containing only remnant forest patches, the rarity of seed dispersal by large animal s (White et al. 2004) frequently confines regeneration of la rge -seeded species to < 100 m from forest edges (Gunter et al. 2007; Wunderle 1997) L arge seeds rarely persist in the soil seed bank for more than one season (Garwood 1989; Hopkins and Graham 1984) and thus are dependent on animal -mediated seed dispersal for regeneration. Without hum an intervention, large -seeded tropical species can be vulnerable to local extinction in fragmented and degraded landscapes (Galetti et al. 2006; Terb orgh et al. 2008) Lack of large -seeded species in forests alters successional trajectories (Chazdon 2003), reducing biodiversity recovery during secondary forest succession (Turner et al. 1997) In addition to supporting wildlife populations, fleshy fruits of many large -seeded species provide food and income for millions of people living in or near tropical forests (Chomitz 2007; FAO 2005) In the Amazon, dozens of tropical forest tree species were in the process of domestication since well before the time of European contact (Clement 1999a) Currently, n umerous species previously cultivated only by small -scale farmers are currently being investigated for their economic potential as commercial crops (Akinnifesi et al. 2004; Simons and Leakey 2004) Enrichi ng the landscape with fleshy-fruited species can generate sources of income for local people (Ricker et al. 1999) improve wildlife habitat (Bowen et al. 2007) and conserve genetic diversity of future food sources (Dawson et al. 2009) Multiple methods of catalyzing the colonization of large -seeded species in fragmented tropical la ndscapes have been proposed. For example, c reating forested corridors and protecting waterways prom otes the movement of animals among existing forest patches (Da Silva et al. 2005; Keuroghlian and Eaton 2008) Agricultural practices that increase forest cover such as shade -grown coffee farms (Williams Guillen et al. 2006) and planting trees to create living

PAGE 13

13 fences (Zahawi 2005) can also improve the habitat quality of agricultural landscapes for seed dispersing animals (Perfecto and Vandermeer 2008) The colonization of trees into pastures may be facil itated by nurse trees (Toh et al. 1999; Vieira et al. 1994) or the creation of bird and bat perches (Kelm et al. 2008; Shiels and Walker 2003) Dense planting of native trees in patches, alternatively called woodland islets (Benayas et al. 2008) or tree islands (Zahwai and Augspurger 2006) has also been promoted as a method to increase the colonization and survival of large -seeded species. The success of these strategies depen ds on the distance to seed sources, the existing seed disperser community, and the degree of site degradation (Hooper et al. 2005; Lamb et al. 2005) When the degree of isolation from seed sources is too great or animal dispers ers are too scarce to ensure the a rrival of large -seeded species, forest restoration require s planting of the missing species (Wunderle 1997) I n addition, people may want more control over future species com position rather than to only promote the natural colonization of large seeded species. The propagation of locally rare tree species within their native habitat is commonly called enrichment planting, and includes planting vegetative stakes (Zahawi 2008) seeds (Ochsner 2001) or seedlings (Raman et al. 2009) Compared to transplanting nursery-grown seedlings, planting seeds (direct seeding) into restor ation sites eliminates the costs of nurser ies a nd decreases the l abor and transportation expenses (Schmidt 2008) Furthermore, s urvival of direct seeded trees can be greater than transplants due to reduced root damage during planting and better acclimation to the planting site (Kitao et al. 2006) Direct seeding also allows for the reforestation of larger and less accessible regions because seeds are smaller and easier to transport than seedlings, and seeds often cost subs tantially less than seedlings (Ochsner 2001)

PAGE 14

14 The success of d irect seed ing depends on survival through early st ages of seedling develo pment in the field. If seed losses are heavy or too many seedl ings die planting of nurserygrown seedlings may be a better option. Challenges to planted seeds are similar to those of naturally dispersed seeds. Causes of seed mortality include vertebrat e predation (Notman and Gorchov 2001) pathogen infection (Pringle et al. 2007) insect infestation (Camargo et al. 2002) and desiccation (Hammond 1995) Young seedlings are susceptible to pathogens (Augspurger and Kelly 1984) as well as competition with existing vegetation for light and soil resources (Dantonio and Vitousek 1992; Hooper et al. 2005) These biotic and abiotic constraints differ among ha bita ts and th eir relative importance varies among species (Dupuy and Chazdon 2008; Myster 2003) If seeds are dispersed, large seeds of t ropical forest trees may develop seedlings and survive in human-modified environments. Substantial ene rgy and nutrient res erves allow seed s to survive partial predation (Vallejo Marin et al. 2006) and seedlings to resprout repeatedly if the stem is dam aged (Harms and Dalling 1997) Furthermore, large seeds make large seedlings (Zhang and Maun 1993) that can better emerge through thick leaf litter (Scarpa and Valio 2008) survive herbivory (Armstrong and Westoby 1993) develop large root systems (Leishman and Westoby 1994) and compete effectively with existing veget ation Seedlings from large -seeded species also tend to survive well in the shade (Myers and Kitajima 2007) and have greater drought tolerance than seedlings from smaller seeded species (Moles and Westoby 2002) In this thesis, I study seed and seedling surviva l of six species of tropical forest trees planted as seeds in pastures and forests. I investi gate the relative importance of seed placement (s urface planting and seed burial) and habitat (pastures and forests) on seed predation, seed survival, seedling recruitment, and early seedling survival. All researc h was conducted on

PAGE 15

15 communally owned land in indigenous Shuar communities in the Ecuadorian Amazon In this region, previous reforestation programs that required even moderate financial and labor investments in nursery propagation failed (Rudel 2006) Local residents were interested in low cost methods of propagating native fruit trees that would bear fruits for local consumption, market sale, and increase wildlife populations to improve subsistence hunting In Chapter 2, I report results from a field study comparing seedling recruitment of three species of tropical forest trees with large, recalcitrant seeds that were planted as seeds in recently grazed pastures (2 6 weeks of vegetation regrowth), mature pastures (3 6 months of vegetation regro wth), and selectively logged forests. Seed burial was hypothesized to improve seedling establishment by reducing seed predation and by mitigating harsh microclimate conditions. The species planted were Pouteria caimito (Sapotaceae), Quararibea cordata (M alvaceae) and Inga densiflora (Fabaceae). Seeds were planted at low dens ities to maximize area coverage and to mimic densities generated by rare cases of natural seed dispersal by animals. The experiment was conducted in three indigenous communities tha t differed in degree of forest cover. Seed removal, germination, seedling recruitment and seedling survival were recorded at 2 and 14 weeks In Chapter 3, I report results from a second experiment c onducted the following year with a modified experimental design, using one species also planted in the first experiment (Inga densiflora ), and three other species with large recalcitrant seeds: Gustavia macarenensis (Lecythidaceae), Caryodendron orinocense (Euphorbiaceae), and a species of Plinia (Myrtaceae). I added a caging treatment to isolate the effect of seed pre dation on seedling recruitment and checked seeds biweekly for 18 weeks to assess trends of seedling development and mortality. In this experiment I planted seeds only in secondary forests because the first

PAGE 16

16 study showed no differences in seedling recruitment among field types (pastures vs. forests), and because naturally regenerating forests are extensive in the tropics and likely to be of lower value to landowners than pastures under active use. In Chapter 4 of this thesis I summarize my main findings and discuss broader implications of this research for conservation and restoration programs.

PAGE 17

17 CHAPTER 2 SEED AND EARLY SEEDLING S URVIVAL OF THREE LARGE -SEEDED TROPICAL FOREST TREES PLANTED AS SEE DS IN PASTURES AND F ORESTS IN THE ECUADORIAN AMAZON Introduction Much of the land surface in the tropics is composed of a mosaic of large cattle ranches, industrial plantations, smaller -scale agricultural fields and degraded and secondary forests (McNeely and Scherr 2003) Within these landscape s m a ny rural people depend on relatively small landholdings for subsistence farming and sale of agricultural and non-timber forest products. Propagation of locally valuable native fruit trees can provide people with food and potential market crops (Ricker et al. 1999) as well as enhance the value of the landscape for wildlife (Piotto 2007) A simple and inexpensi ve method of enriching degraded areas with desired species is to plant tree seeds directly in the landscape. This technique, called direct seeding, may be used in place of the more resource intensive process of nurse ry propagation and transplantation of seedlings (Shen and Hess 1983) if chance of seedling recruitment is sufficiently high in the field. Seedlin g recruitment can b e constrained by seed predation, seed desiccation, and competi tion with existing vegetation. L arge-seeded tropical species may be suitable for direct seeding (Camargo et al. 2002; Doust et al. 2006) because they may be better able to survive early chall enges to seedling establishment than smaller seeds in both pastures (Hooper et al. 2002) and degraded or secondary forests (Moles and Westoby 2004) The purpose of this study was to assess the likelihood of seedling recruitment of three species of large -seeded tropical trees that were planted as seeds in pastures and forests. I tested seed burial as a treatment for improving seedling recruitment by reducing seed predation and desiccation.

PAGE 18

18 Large energy and nutrient reserves in large seeds may be advantageous for seed l ing recruitment. Storage carbohydrates allow for survival through partial seed predation (Vallejo Marin et al. 2006) and longer survival of seedlings in the shade (Myers and Kitajima 2007) Rapid germination of many large -seeded species can all ow for quick escape from seed predators (Daws et al. 2005) After germination, large seeds make large seedlings (Zhang and Maun 1 993) which can emerge through thick leaf litter (Scarpa and Valio 2008) survive herbivory (Armstrong and Westoby 1993) resprout (Harms and Dalling 1997) and produce large root systems (Leishman and W estoby 1994) Once seeds are planted, seed predation by vertebrates can be one of the first major barriers to seed survival. Seed removal levels of 75 100% for large -seeded species have been documented in tropical forests (Blate et al. 1998; Pea Claros and De Boo 2002) and pastures (Holl and Lulow 1997; Vieira and Scariot 2006) though seed predation levels of ten vary widely among species planted in the same environment (Janzen 1969; Jones et al. 2003b) P opulat ions of seed eating animals also vary, depending on habitat availability (Lambert et al. 2006) and hunting pressures (Carlos 2001) In landscapes devoid of large seed predators, small rodents can become important consumers and secondary disperse r s of large seeds (Brewer and Rejmanek 1999) Burial of seeds can greatly reduce seed predation (Crawley 2000; Vander Wall 1990) though this effect may be less for species with large seeds that are easier for seed predators to detect by odor than smaller seeds (Thompson 1987) It should be noted that not all seeds that are removed by animals die, as some may be cache d or secondarily dispersed to sites where germination is possible (Vander Wall et al. 2005) In addition to seed predation by v ertebrates, mortality of seeds can also be caused by seed desiccation (Hammond 1995) fungal infection, physical damage, insect predation (Woods and

PAGE 19

19 Elliott 2004) anoxia in waterlogged soils (Fenner and Thompson 2005) and fire (Hooper et al. 2005) The thin seed coat of many tropical forest tree species makes them especially sensitive to desiccation (Daws et al. 2006) Mortality of large tropical seeds is often higher in tropical pastures (Vieira and Scariot 2006) and agricultural fallows (Notman and Gorchov 2001) than forests, primarily due to increased air temperature, air vapor pressure deficit and soil moisture stress in exposed environments (Benitez -Malvido et al. 2005; Holl 1999; Nepstad et al. 1996) Moisture stress can be less extreme if the seed is covered by pasture vegetation (Holl 1999) While seeds may die of desiccation in pastures, higher levels of pathogen attack of ungermi nated seeds are associated with shade and forested environments (Pringle et al. 2007) If seeds survive to germinate mortality can still be high during germination and early seedling development in both pastures and forests In pastures, mats of dead grass can prevent the radicle from reaching the soil, and below -ground competit ion with roots of pasture grasses for available soil water can limit seedling survival and growth (Dantonio and Vitousek 1992) Leaf herbivory, particularly by leaf -cutter ants, can be extensive in tropical pastures (Nepstad et al. 1991) In forests, dea th of seedlings can be caused by mechanical damage from falling vegetation (Scariot 2000) trampling (Beck 2006) herbivory (Lopez and Terborgh 2007) and fungal infection (Augspurger 1984b; Benitez Malvido e t al. 1999) As seed reserves are depleted, the young seed l ing must photosynthesize to maintain a positive carbon balance (Myers and Kitajima 2007) and al though abiotic stresses may be higher in pastures, low light levels in forests results in reduced seedling growth (Augspurger 1984a) Se ed burial may improve seedling recruitment by mitigating both abiotic and biotic st resses on seeds and young seedlings. Burial of seeds can reduce moisture loss, moderate temperature extremes and provide protection from insect infestation (Forget 1990; Thompson

PAGE 20

20 1987) P redation of storage cotyledons can kill young seedlings (Alvarez Clare and Kitajima 2009) and seed burial may help protec t seed reserves from post germination predation. The purpose of this study was to assess the likelihood of seedling recruitment of three large -seeded tropical forest tree species planted as seeds in pastures and forests. Seeds were planted at low densitie s in recently grazed pasture, pasture with 3 6 months of vegetation re growth, and forests in three indigenous Shuar communities in the Ecuadorian Amazon. Seeds were planted on the soil surface and shallowly buried. At 2 and 14 weeks, seed s were checked for removal germination, seedling recruitm ent and mortality. I selected the three species based on seed availability of native s pecies with recalcitrant seeds and interests of the local indigenous landowners in propagation of trees with edible fruits Re search Questions How likely is seedling recruitment from seeds of large -seeded tropical forest trees planted in forests and pastures? How does habitat (grazed pasture, mature pasture, and forest) affect the probability of seedling recruitment? Does shallow seed burial increase seeding recruitment? Are results consistent among three indigenous communities with differences in landuse patterns of the surrounding landscape? Materials and Methods Study Site The study was conducted in three indige nous Shuar com munities in the Morona Santiago and Pastaza provinces of the Ecuadorian Amazon. All research plots were located within 15 km of the village of Tsurak (1 4831S, 774950W), 51 km south of the provincial capital of Puyo on the Puyo-Macas Rd (Figure 2 1). The elevation is 850 950 m amsl and average monthly temperature is 20.1 1 C. The mean annual rainfall is ca 4600 mm/yr, with 265

PAGE 21

21 days/yr of precipitation of at least 1mm. Ave rage monthly rainfall is slightly lower from December February (327 23 mm/month ) and July September ( 342 30 mm/month), but there is no clearly defined dry season (Table A 1 and Figure A 1). A recent analysis revealed a trend of increasing temperature and evaporation/precipitation ratio over the last 30 years (Mill n et al. 2008) but the climate is still wet. Soils in the region are Humic Andosols ( Hydrandepts ) of volcanic o rigin with low fertility, high susceptibility to leac hing and erosion, and often low levels of aluminum toxicity (Custode 1983) Natural vegetation in the study area is very humid premontane forest of high biodiversity (Caadas and Estrada 1978) Prior to the 1900s, the land was inhabited by low density mobile settlements of Shuar Amerindians (Rudel et al. 2002) Road construction and agricultural sub sidies in the 1960s and 1970s led to timber extraction and larger -scale forest clearing for permanent settlements, pastures and small -scale commercial agriculture (Rudel and Horowitz 1993) Suspension of agr icultural subsidies in the 1980s led to reduction in cattle and increased cultivation of cash crops (primarily of naranjilla Solanum quitoense ). Subsequent pest outbreaks on agricultural crops led to an increased reliance on revenue from selective logging in the 1980s, and migr ation to cities in the 1990s (Rudel and Horowitz 1993) Abandonment of agricultura l lands has resulted in development of large areas of second ary forests over the last 20 years (Rudel 2006) The current landscape is a patch work of secondary and selectively logged forests, with clearings for road s, houses, horse pastures, and home gardens. To assess the likelihood of seedling recruitment of three large -seeded tropical forest tree s, I placed seeds in pastures and forests in three indigenous communities located approximat ely 15 km apart (Figure 2 1) Based on my observations, the three communities (Kenkuimi, Kunkuki and San Ramon) have approximately the same populat ion size (~100 adults ), but differ in land

PAGE 22

22 uses and livelihoods of the resident s. The community of Kenkuimi is surrounded by secondary and selectively logged forest, and at the time of the study the village was only accessible by foot Most houses are located in a central clearing surrounded by secondary and selectively logged forest, and the majority of people practice a subsistence li velihood based on small home gardens (0.5 1 ha) located in forest clearings. In contrast, the community of Kunkuki is located on t he interprovincial Puyo -Macas Road with many single -family homes in clearings along the main road. The surrounding landsca pe is a mix of horse pastures and secondary forests. The nearest mature forest is approximately 10 km to the east, accessible by foot and mule trails. Many residents of Kunkuki work during the week in the provincial capital of Puyo (55 km north) or the s mall city of Macas (60 km south) Landowners maintain pastures larger than needed for use by their own horses, and often rent pastures to logging companies and individuals using horses for transport of supplies into roadless areas to the east. The commun ity of San Ramon is accessible by dirt road, and the village is mostly surrounded by cattle pasture, with secondary forest and selectivel y logged mature forest within 0.5 km of the village center. It has more s urrounding forest than Kunkuki and less than Kenkuimi. Tree Species Selection I selected native species that were in fruit at the onset of the experiment (June 2007), that were known to have recalcitrant seeds and that had edible fruits consumed by local people (Bennett et al. 2002) The three species chose n were Pouteria caimito (Ruiz and Pavon) Radlk. (Sapotaceae), Quararibea cordata Vischer ( syn. Matisia cordata Humb. & Bonpl .; Malvaceae) and Inga densiflora Benth (Fabaceae). Pouteria caimito is a shadetolerant, long -lived canopy tree (Benitez -Malvido and KossmannFerraz 1999) cultivated for its fruits throughout its native range in the tropical lowlands of Ecuador, Colombia, Peru, Venezuela and Brazil (Morton 1987a) Breeding programs for commercial fruit product ion began in Brazil in the 1960s

PAGE 23

23 (Clement et al. 2008) Pouteria caimito is considered a pre -Colombian domesticate from NW Amazonia by Clement (1999). Quararibea cordata is a fast growing tree also from NW Amazonia and widely cultivated for its edible fruits (Morton 1987b) It may also have been partially domesticated by pre Colombian people in NW Amazonia (Clement 1999b) Inga densiflora is a medium sized tree ( America, and western South America for the sweet sarcotesta surrounding its seeds (Pennington and Fernandes 1998) Inga is often planted as a nitrogen -fixing shade tree in coffee plantations, and it is valued for firewood production (Pennington and Revelo 1997) Due to extensive cultivation throughou t Latin America, the natural distribution of this species is unknown (Pennington 1997) Hereafter, the species are referred to by generic epithets. Experimental Design In each of the three indigenous communities (Kunkuki, Kenkui mi, San Ramon), one 0.5 ha plot was established in each of three field types: grazed pasture (2 6 weeks of vegetation regrowth); mature pasture (3 6 mo of vegetation regrowth) ; and forest (mature forest selectively logged w ithin the last 10 years ). P lots within communitie s were separated by 100 500 m and communities were separated by 15 20 km. Wi thin each plot, 50 100 m parallel transects were established at 3 8 m intervals, and individu al seeds were sown at 1 m intervals, 1 m to the left or righ t of the transect (Figure 2 2 ). The pasture plots consisted primarily of the American perennial pas ture grass Axonopus scoparius (Flgge) Kuhlm, called gramalote rojo in Spanish (Zuloaga et al. 2003) an d sak in Shuar. Due to the wet climate, fire is used to burn woody material only when the land is first cleared, and subsequent control of woody vegetation in pastures is by machete. Pasture plots had been maintained for at least 10 years and the land was initially cleared in the 1970s 1990s. All pasture plots contained some colonizing trees, shrubs and large herbaceous plants ( e. g.

PAGE 24

24 Miconia Ochroma Lonchocarpus Piptocoma, and Heliconia). The presence of pasture grass, woody colonizing plants a nd herbaceous vegetation created a heterogeneous vegetation structure in pasture plots (Figure A 2). The forest plots had been selectively logged within the last 10 years and a relatively continuous canopy cover was present at the time of the experiment ( Figure A 2) Percent vegetation cover was measured with a hand -held spherical densitometer held at ground level at 20 points in each plot. Canopy cover was lowest and most variable in the grazed pasture plots ( 43 20% ). Percent vegetation cover ranged from 62 88% in the mature pasture plots and 8688% in the forest plots (Figure A 3 ). In June of 2007, seeds were collected from fresh fruits of trees growing along roadsides and in home gardens i n the Sucua -Macas area approximately 50 75 km south of the ex perimental plots (elevation 1050 m). Seeds were harvested from ripe fruits picked from trees and from freshly fallen fruits. Fruits were collected from at least three trees of each species and mixed together before planting Fruit pulp was removed, see ds were inspected for damage, washed in freshwater, float -tested for viability, and planted within one week of harvest. Average seed sizes were 2.9 cm x 1.5 cm for Pouteria 3.9 x 2.2 cm for Quararibea, and 4.5 cm x 2.2 cm for Inga. Seed and seed ling cha racteristics, as well as common names are presented in Table A 2. See ds were planted in surface and buried treatments. In the surface treatment, seeds were placed on the ground surface, without disturbing the vegetation, litter or underlying soil Buried seeds were placed in a 1 2 cm deep hole and covered with ~1 cm of soil. Seed location was marked by a 20 cm 14-gauge wire inserted into the ground beside each seed (within 1 2 cm of the planting location). Species, treatment, location along the transect, and left/right planting were completely randomized in each plot. An ave rage of 25 replicates (range 12 40) of each

PAGE 25

25 species x treatment combination were planted in each of the 9 plots ( 3 communities x 3 field types). At each planting location, seed disappearance (only fragments of the seed coat remaining, or no seed w ithin 30 cm of the planting site), germination (radicle emergence), seedling recruitment (survival to development of at least one true leaf ) and mortality were recorded by me in early July (2 weeks ) and by a trained a ssistant in late September (14 weeks ). At 2 weeks seeds were carefully removed to visually inspect for germination (appearance of the radicle) and returne d to their original location (placed on the ground substrate or buried). At 14 weeks all ungerminated seeds (buried and surface-planted ) were cut open and inspected for damage. Seeds that were intact with solid, undamaged tissue were considered alive. Data Analyses Treatment effects on the proportions of seeds that had disappeared by 2 weeks died or were missing (mortality), and were recruited to seedling stage at 2 and 14 weeks were assessed by logistic regression using PROC GLIMMIX in SAS 9.1.3 ( 2005, SAS Institute, Cary, NC). Species, burial treatment (surface, buried), field (grazed pasture, mature pasture, forest), and community (Kenkuimi, Kunkuki, San Ramon) were treated as fixed effects, and fields were nested within each communities. Althou gh environmental differences among multiple communities may vary randomly, the small sample size (n = 3 communities) and variation in the factors tested precluded the treatment of community as a random effect, as the models would not converge. I first te sted full factorial models including main effects and all interactions, and subsequently reduced models by removing non -significant intera ctions (p > 0.05). Reported analyses are of main effects and significant interactions. Treatment effects were furthe r eva luated with post -hoc Bonferroni adjusted pairwise comparisons. Strong differences among

PAGE 26

26 species made models imbalanced for most analyses, and thus when species was a highly significant main effect (p < 0.0001), individual species were analyzed separ ately as subsets for treatment effects. In cases of quasi -separation of the data (lack of variability in the data due to zeros in many cells), data were analyzed in subsets (Lamotte 2005) or the results reported with descriptive statistics when analysis with logistic regression was not possible. Seed disappearan ce, mortality (the total of missing seeds, dead seeds and dead seedlings combined), and seedling recruitment were analyzed at 2 weeks I analyzed seed disappearance only at 2 weeks because it was not possible to distinguish between remov al of viable seeds, decay of inviable seeds, or decay of seedlings that were recruited and died between the 2 and 14 week censuses. Seedling recruitment and mortality of all species were measured at 14 weeks. In addition, Inga had enough seedlings alive at 2 weeks (239) to analyze survival of the cohort of early recruiting seedlings between weeks 2 and 14. Results General Patterns of Seeding Recruitment and Survival In the first 2 weeks only 49 of 736 seeds disa ppeared from planting sites, but communi ties differed in the proportions of seeds missing (Figure 2 3). The majority of seedling recruitment of Inga occurred in the first 2 weeks whereas seedling recruitment was slower for Quararibea and Pouteria Seedling recruitment did not differ among fi eld types (pastures and forests) at 14 weeks or between surface-planted and buried seeds at 2 or 14 weeks (Tables 2 1 & 2 2). Species Difference in Seedling Recruitment and Mortality Species differed in seedling recruitment at 2 weeks (p < 0.0001; Table 2 1) and at 14 weeks (p < 0.0001; Table 2 2). At each census, there were more seedlings of Inga than Quararibea, and more seedlings of Quararibea than Pouteria (p < 0.01 in all post -hoc pairwise comparisons between species). At 14 weeks 56% (171 of 307) of Inga, 10% (18 of 187) of

PAGE 27

27 Quararibea, and 2% (5 of 242) of Pouteria were live seedlings (Figure 2 5). When observations from the 2 and 14 week censuses were combined, 91% of Inga, 25% of Quararibea and 3% of Pouteria seeds had germinated ( found as seedlings or seeds with a radicle emergence ). Additional seeds may have germinated and decayed or been eaten between the 2 and 14 week censuses. Species also di ffered in the time to seedling recruitment Of the 171 live seedlings of Inga at 14 weeks 89% wer e seedlings by 2 weeks. Only 18 additional Inga seedlings were recruited after 2 weeks In contrast, seedling recruitment was slower for Quararibea and Pouteria Of the 18 live Quararibea seedlings at 14 weeks 11 were recruited after 2 weeks and 4 of the 5 Pouteria seedlings alive at 14 weeks were recruited between weeks 2 and 14 M ortality was calculated as the sum of all missing seeds, dead seeds and dead seedlings divided by all seeds planted. The mortality differed among species at 2 and 14 weeks At 2 weeks mortality was low for all species (7 9% missing or dead in any species; Table A 3) and species differed in mortality (p = 0.0304; Table 2 1). At 14 weeks species differences in mortality were clear (p < 0.0001; Table 2 2). The vast majo rity of Quararibea and Pouteria seeds planted were either dead or missing (93 dead and 49 missing of 187 Quararibea seeds, and 107 dead and 87 missing of 242 Pouteria seeds ) whereas the majority of Inga seeds were alive as seedlings at 14 weeks (239 of 307; Table A 3). Of the 239 Inga seedlings that were alive at 2 weeks only 28 died by week 14. S pecies differences in mortality at 14 weeks were nearly equivalent to species differences in seedling recruitment, because nearly all seeds were either recruit ed to seedlings or died, and less than 20 seeds were alive and ungerminated at 14 weeks for any species (Table A 3).

PAGE 28

28 Effects of Seed Burial Surface -planted and buried seeds did not differ in seedling recruitment at 2 or 14 weeks in mortality at 2 or 14 we eks or in the proportions of seeds that were missing at 2 weeks (Tables 2 1 and 2 2). Community Differences in Seed and Seedling Survival At 2 weeks more seeds were missin g and seedling recruitment was lowest in Kenkuimi than in the other two communities (Figure s 2 3 & 2 4) Almost all seedlings that were recruited by 2 weeks were Inga (239 of 267; Table A 3), thus differences in seedling recruitment among communities at 2 weeks (p = 0.0115; Table 2 1) are largely due to differences in recruitment of Ing a seedlings (Figure 2 4). Lower seedling recruitment in Kenkuimi than San Ramon (p = 0.0085) appears to be partially explained by higher levels of seed disappearance in Kenkuimi than San Ramon (Figure 2 3). At 14 weeks differences among communities in seedling recruitment were clear er (p < 0.0001; Table 2 2; see Figure 2 5). More seedlings were recruited in Kunkuki than in Kenkuimi (p < 0.0001) or San Ramon (p < 0.0001), and more seedlings were recruited in San Ramon than Kenkuimi (p = 0.0060; Table 2 2). Mortality of Inga seedlings between weeks 2 and 14 differed among the three communities (p < 0 .0001; Table 22). Of 239 Inga seedlings alive at 2 weeks 86 were dead at 14 weeks More died in Kenkuimi than Kunkuki ( p < 0 .0001) or San Ramon (p = 0.0043), and more died in San Ramon than Kunkuki (p = 0.037; Table 22). Differences in mortality at 14 weeks among the three communities (p = 0.0174) are positively correlated with differences in 14 week seedling recruitment, because the majority of live individuals were seedlings at 14 weeks and few live seeds remained (Figure 2 5). In post hoc pairwise comparisons between communities, mortality at 14 weeks was higher in Kenkuimi

PAGE 29

29 than Kunkuki (p = 0.0233; Table 2 2). Higher mortality in Kenkuimi reflect s both higher levels of seed disappearance at 2 weeks and lower survival of Inga seedlings between weeks 2 and 14. Seedling Recruitment in Pastures and Forests Some differences in seed disappearance, seed survival and seedling recruitment between pastures and forests were observed at 2 weeks but by 14 weeks seedling recruitment and mortality did not differ among field types in any community (Tables 2 1 & 2 2). At 2 weeks more seeds were missing from the forests than the grazed plots in Kenkuimi (p = 0.0049) and Kunkuki In San Ramon, more seeds were missing from the pastures than the forest plot ( Figure 2 5). At 2 weeks mortality differed among fields within communities (p = 0.007; Table 2 1). The only significant difference in post -hoc pairwise compa risons between field types within each community was in the community of Kenkuimi, where higher mortality occurred in the forest than in the grazed pasture (p = 0.0057; Table 2 1). This result is partially explained by higher levels of seed disappearance in the forest than pasture plots in Kenkuimi (Figure 2 3) and some differences in seedling recruitment among field types (p = 0.0454; Table 2 1). Survival of Inga seedlings between weeks 2 and 14 did not differ among fields (Table 2 2). Discussion Indige nous Amazonian cultures have a long history of enriching forested landscapes with useful species by managing naturally regenerating species (Gadgil et al. 199 3; Posey 1985; Redford and Padoch 2000; Toledo and Salick 2006) Natural seed dispersal of large-seeded species in landscapes heavily used by people is often reduced due to distance from seed sources (White et al. 2004) and hunting of seed-dispersing animals (Chapman and Onderdonk 1998) When natural seed dispersal is rare, occasional natural seed dispersal or seed planting can make large differences in the regeneration of the species. If seed s grow into tree s that bear fruit, they

PAGE 30

30 can serve as seed sources for further population expansion by natural seed dispersal. Cultivated and naturally regenerated fruit trees also provide food and potential sources of income to rural landowners. H owever, a seed is not a tree, and many obstacles must be overcome, the first of wh ich is survival of the seed to the seedling stage The purpose of this experiment was to assess the probability of seedling recruitment among three tropical tree species wit h large seeds that were planted in a variety of potential dispersal sites in three indigenous communities in a human -modified landscape. Seedling recruitment differed among the three species, as well as among the three communities. Inte restingly, the com munity in which seeds were planted affected seed disappearance and seedling recruitment more than whether seed s were planted in pastures or forests. Contrary to my initial expectations, seed burial did not affect seed disappearance at 2 weeks or seedling recruitment at 2 or 14 weeks Landscape -Level Variation The indigenous community in which the seeds were planted affected the proportion of seeds that were missing at 2 weeks more tha n any other treatment factor. Density and species composition of seed predator communities are known to vary with changes in habitat and hunting pressures (Asquith et al. 1997) and the three communities were separated by 10 15 km and differed in surrounding forest cover and land use. S eed disappearance was highest in the commun ity of Kenkuimi (Figure 2 3) which, a t th e time of the study, was not accessible by road and was surrounded by relatively continuous secondary and logged forests. Local residents informed me that there was better hunting of guatusas ( Dasyprocta fuliginosa ), and guantas (Agouti sp ) in Kenkuimi than in the other two communities. The guatn (likely Myoprocta pratti) was only noted as being present in Kenkuimi (Eisenber g and Redford 1999). All these animals eat seeds (Asquith et al. 1997; Beck and Terborgh 2002; Jansen et al. 2004; Silvius and

PAGE 31

31 Fragoso 2003) and higher populations of these rodents may have caused higher levels of seed disappearance in Kenkuimi. The oth er two communities have road access and higher densities of people, and forest pat ches that are interspersed among larger h orse and cattle pastures. D ifferences in seed disappearance amo ng the three communities highlight the importance of considering small -scale, regional differences in surrounding land use patterns on ecological processes within a landscape. Similar to seed disappearance, seedling recruitment at 14 weeks varied among the three communities but not between pastures and forests within any community (Table 2 2). M any m ore seedlings of Inga were recruited by 14 weeks compared to the other species, thus patterns of seedling recruitment discussed henceforth are largely driven by this species (Figures 2 4 & 2 5). Initial seedling recruitment was highest in San Ramon and lowest in Kenkuimi. Survival of Inga seedlings that were recruited at 2 weeks differed among the three communities, with the greatest mortality in Kenkuimi and lowest mortality in Kunkuki (Table 2 2). Low mortality between the 2 and 14 week censuses resulted in greater seedling recruitment at 14 weeks in Kunkuki than the other two communities ( p < 0 .0001). Pasture and forest plots did not differ in seedling rec ruitment within any community. These results suggest microhabitats affect early seeding survival less than larger scale, regional differences among the communities. For e xample, although the communities were only 10 20 km apart, rainfall p atterns could h ave differed Lack of rainfall would be expected to decrease seed and early seedling survival of desiccation -sensitive seeds. Other abiotic factors such as soil type and temperature extremes may also have varied among the communities, but these data were not collected. Finally, forest cover and surrounding land use s differed among the three communities and this variation could be correlated with biotic causes o f early seed and seedling death such as higher population

PAGE 32

32 densities of pathogens and seed preda tors in some communities. The order of planting (Kenkuimi, Kunkuki, and San Ramon) did not correlate with patterns of mortality or seedling recruitment, as the lowest mortality and highest seedling recruitment were both in Kunkuki. Microhabitat Effects o n Seed Predation The proportion of seeds that disappeared varied among communities, and also between field types within communities (Figure 23). In Kenkuimi and Kunkuki, more seeds were mi ssing from forests than pastures, and in San Ramon, m ore seeds wer e missing from pastures than forest (Figure 2 5). Previous studies in the tropics frequently report higher levels of seed predation in earlier successional habitats than mature forests (Nepstad et al. 1996; Pea Claros and De Boo 2002) although some report higher seed predation in forest fragments than open fields (Cole 2009) or no difference between habitat types (Holl and Lulow 1997) Potential seed predators must be able to access seeds, while minimizing risk of exposure to their own predators (Taylor 1984) The heterogeneity of vegetation within pasture plots, as well as the location of the plots in the landscape, likely influenced patterns of seed disappearance. Forest vegetation may have provided greater protection for rodents from thei r p redators, even if seeds were easier to find in open field s Pasture plots with more woody vegetation or in closer proximity to forests could b e expected to experience greater seed predation than open plots removed from protective cover. More seeds disappeared from pastures in Kenkuimi and San Ramon, where the pasture plots were adjacent to selectively logged forests, and almost no seeds disappeared from pastures in Kunkuki, where the two pasture plots were within larger pastures Higher levels o f seed predation by mice have been observed in temperate fields with greater woody cover (Ostfeld et al. 1997) and higher levels of seed removal were found under woody vegetation within tropical pastures in Costa Rica (Holl 2002)

PAGE 33

33 Seed removal by animal s does not necessarily indicate seed death (see review in Vander Wall et. al. 2005), since removed seeds may be cached and later germinate (Brewer and Rejmanek 1999; Jansen and Forget 2001) This s e condary seed dispersal has been reported to be relatively frequent in fore st fragments and rare in pastures and secondary forests (Cole 2009) Some of the disappeared seeds may have been scatterhoarded, particularly in forest plots and in the community of Kenkuimi, which was surrounded by more forest and believed to have higher populations of scatterhoarding animals tha n the other two communities. Seed predation patterns are often variable and species specific (Guariguata et al. 2000) even within apparently homogeneous environm ents. Similar studies in both Costa Rica (Jones et al. 2003b) and Mexico (Garcia Orth and Martinez -Ramos 2008) observed different seed predation leve ls between four pasture plots located within 3 4 km of each other. Fluctuating populations of seed predators and dispersers during the year also influence seed predation levels (DeMattia et al. 2006) and results of my 2 -week experiment may not apply to other times of the year. In 2 weeks only 49 of 739 seeds were remove d in this study (Table A 3), thus limited interpretation is possible regarding treatment effects on missing seeds. Although s eed predators generally prefer fresh seeds (Vander Wall 1990) additional seeds may have been eaten after 2 weeks Finally, although seed disappearance did not vary significantly among the three species tested, seed predators are known to vary in preferred foods (Brewer and Rejmanek 1999; Vieira et al. 2006) and a longer study may have revealed species differe nces in seed predation. Contrary to my original expectations, seed burial did not appear to reduce the proportion of seed s removed by seed predators. Burial reduces the odor of fresh se eds and more deeply buried seeds are more difficult to detect (Vander Wall 1990) but rodents can still find buried seeds

PAGE 34

34 using only olfaction (Howard et al. 1968) In a study at Barro Colorado Island, Panama, agoutis (Dasyprocta punctata) detected caches of 6 12 maize kernels buried 3 7 cm deep (Murie 1977) Rodents also detect seeds in moist soil more readily (Wall 2003) because as seeds imbibe water, volatil e compounds are released (Duke et al. 1983) The moist conditions of fresh seeds, moist soil, and shallow burial of the seeds may have been insufficient to mask seed odors from foraging animals. Although only a small area of soil was disturbed at the planting sites, visual (but not olfactory) cues have been shown to attract agoutis to experimentally buried food caches (Murie 1977) Finally, it is possible some buried seeds were not counted simply because they are more difficult to find than surface planted see ds, resulting in an inflated numbers of removed buried seeds. Microhabitat Effects on Seedling Recruitment and Survival Seedling recruitment and mortality did not differ among the grazed pasture, mature pasture or forest plots at 14 weeks (Table 2 2). Variation of vegetation cover within plots, numerous potential causes of seed and seedling death, as well as possible similarities in m icrosite conditions in forests and pastures may all h ave obscured detection of field -specific patterns of seedling recruitment (if patterns existed ) Veget ation in pasture plots was a mix of woody and herbaceous species, resulting in large variation in cover at pla nting locations (Figure A 2). Microclimate differences at planting sites within pasture plots could have c aused variability in seed and seedling survival due to fungal infection and light competition at heavily shaded planting sites and desiccation where seeds were more exposed. In addition, quick regrowth of pasture grasses at the grazed sites may have resulted in similar moisture conditions at g round level in grazed and mature pasture plots At the scale of a seed, microsite conditions in pastures and forests may have been more similar than suggested by obvious larger -scale differences between the habitat types. Higher temperatures and great er moisture stress in

PAGE 35

35 pasture s than forests are expected (Nepstad et al. 1996) but pasture grasses can modify microclimates such that forest and pasture sites experience similar temperature and moisture conditions at ground level (Holl 1999) Surface planted seeds in both forests and pastures were placed without disturbance of the vegetation, which resulted in som e seeds lying on litter in the forest, and on mats of dead grass in the pastures. In both habitats, these seeds were at higher risk of death by desiccation, increasing the variability of survival within both pasture s and forest s Overall, variability in microclimate within both forests and pastures and possible similarities between pasture and forest environments in factors affecting seed and seedling survival may have resulted in similar seedling recruitment between field types. Seedlings were recruited equally well from surface -planted and buried seeds (Table 2 2), indicating that any potential benefit of seed burial (moisture retention, soil contact, etc.) was either unnecessary (for Inga) or insufficient to compensate for other mortality agents (for Po uteria and Quararibea). It is also possible that seed burial affected germination or early seedling survival of Pouteria or Quararibea, but this effect was not able to be detected due to mortality and subsequent decay of seeds between the 2 and 14 week ce nsuses. Differences Among the Three Large -seeded Species The three large seeded species in this study differed in time to germination and seedling recruitment as well as overall patterns of seed and seedling mortality. Inga germinated rapidly and seedlings developed earlier and in greater numbers than the other species. At 2 weeks, 93 % (262 of 281) of Inga seeds present were germinated, and 91% (239) were already plants with at least one leaf (Figure 2 4). In contrast, only 24% (41 of 173) of live seeds of Quararibea and 3 of 216 live seeds of Pouteria germinated by 2 weeks Rapid germination and early seedling recruitment of Inga apparently led to greater success in total seedlings alive at 14 weeks compared to the other species. Quick seedling development reduces the duration of seed

PAGE 36

36 exposure to seed predators and pathogens, and allows for quick onset of photosynthesis. For example, over 75% of seeds of Quararibea cordata were attacked by fung i in a three -month greenhouse trial with forest soil (Pringle et al. 2007) Levels of germination and seedling recruitment were similar to other studies for Inga, but were lower than usual for Quararibea and Pouteria Germination levels of 95 100% are common in many species of Inga (Pennington and Revelo 1997) and the 93% germination observed in this experiment is similar to the 99% germ ination level for Inga in the 2008 experiment (Chapter 3). Seedling recruitment of the majority of Inga seeds within 4 weeks was also observed in both experiments (Table A 3 and Figure 3 2). In contrast, at 14 weeks 83% of Quararibea and 91% of Pouteria seeds had died or disappeared (Figure 2 5). In outdoor trials with forest soil, 85% of Pouteria caimito seeds germinated with in 15 days (Snchez et al. 2003) and most Q. cordata seeds germinated in 16 18 days (Pringle et al. 2007) suggesting germination of these species would have been occurring near the time or just after the 2 week census. The 12 week gap between the 2 and 14 week censuses in my study did not allow for accurate assessment of final germination percentages. Nevertheless, very low seedling recruitment at 14 weeks suggests lower survival of Pouteria a nd Quararibea than observed in other studies. Although seeds were planted within 1 week of harvest from fresh fruits it is possible that seeds of these species had low initial viability at the time of planting. Survival in the field can also be expected to be less than under more controlled conditions. Differences in seed mortality and seedling recruitment among these three species highlight the limitation of generalizing survival patterns for the group of large -seeded species, though this issue may be resolved by including more species in future studies.

PAGE 37

37 Recommendations for Propagation of the Study Species In this study, Inga densiflora had high survival of seeds to seedlings under all experimental conditions, and planting of this species as seeds in p astures and forests is likely to be successful. Volunteer seedlings of Inga densiflora were fairly common in pastures and along roadsides in the study area, suggesting seedlings can survive later stages of development in pastures with minimal management In contrast, the vast majority of Pouteria caimito seeds either disappeared or died during the experiment, and very few seedlings established in the field. For Pouteria transplantation of nursery-grown seedlings may be necessary to ensure seedling estab lishment. Quararibea cordata had relatively low seedling recruitment at 14 weeks (10%), but this may be acceptable for direct seeding if seed supplies are plentiful. Given that seedling recruitment did not differ between pastures and forests, additional f actors may be more important when considering planting sites. For example, faster growth of seedlings would be expected in pastures than forests, and planting in open sites may be preferred to minimize time to onset of fruit production. Also to be considered is the relative value of the current land use. Within 6 months of the end of this study, landowners re -introduced horses into several of the pastures. U nless the seedling s were protected from trampling and browsing most recruited seedlings probably died. Depending on local land use priorities, enriching forests may add greater value to the landscape than adding trees to pastures. Recommendations for Future Studies This study woul d be improved by including more species and more frequent censuses of seedling development and survival. Caging of seeds w ould prevent the removal of viable seeds and could be used to isolate seed predation from other causes of mortality Simultaneous greenhouse trials should be included to establish baseline seed viabilit y data on seeds planted in the field. Including seedling transplants and a financial comparison between direct seed ing and

PAGE 38

38 transplant planting would give more information about planting options to land managers. Repeating the study at more sites would al low community location to be tested as a random factor to evaluate variation at the landscape level on seed mortality and seedling recruitment. Longer term studies would obviously be useful to evaluate survival of recruited seedlings in the different envi ronments. Conclusion Survival to the seedling stage was most strongly affected by the species planted and secondly by the indigenous community into which the seed was planted. For these three large seeded species, microhabitat variations such as seed bu rial and planting in specific habitats (pastures or forests) did not affect the success of seedling development over the first 14 weeks Seedling recruitment at 14 weeks from planted seeds was high for Inga (56%), low for Quararibea (10%) and extremely l ow for Pouteria (2%). Species are known to differ in seedling recruitment in human -impacted landscapes (Nepstad et al. 1991; Ramos and Delamo 1992) but research on the influence of the agricultural matrix surrounding tropical forest fragments has only recently gained attention in the fields of restoration ecology and conservation biology (Chazdon et al. 2009; Perfecto and Vandermeer 2008) After species selection, consideration of the landscape surrounding planting sites may be more important for early seedling recruitment than management of microsite conditions at the planting site.

PAGE 39

39 Table 2 1. Results from mixed model logistic regression of seeds missing, mortality (missing or dead seeds and dead seedlings ) and seedling recruitment at 2 weeks of th ree large seeded tropical trees planted as seeds ( surface planted and buried) in three field types (grazed pasture, mature pasture, forest) in three indigenous communities in the Ecuadorian Amazon. Results are for reduced models that include only main ef fects because no interactions were significant (p ). Source Num df Den df F value P > F Num df Den df F value P > F Num df Den df F value P > F Species 2 233 2.11 0.1239 2 707 3.51 0.0304C2 724 88.06 <0.0001E Burial 1 233 0.01 0.9371 1 707 0.91 0.3397 1 724 0.96 0.3265 Field (Community) 2 233 6.31 0.0021B6 707 2.98 0.0070D6 724 2.16 0.0454F Community 2 707 0 0.9994 2 724 4.49 0.0115G Seed disappearance at 2 wks in Kenkuimi A Mortality at 2 wks Seedling recruitment at 2 wks A Only Kenkuimi was a nalyzed because so few seeds were missing in some fields in Kunkuki and San Ramon, thus analysis by logistic regression was not possible for these communities (Figure 2 3) B In Kenkuimi, more seeds were missing in forest than grazed plot (p = 0.0049) in post hoc pairwise comparisons. In Kunkuki, more seeds were missing in the forest than grazed plots, and in San Ramon, more seeds were missing from the pastures than the forest plot (Figure 2 5) C No pairwise comparisons significant in post hoc tests. D In Kenkuimi, higher mortality in the forest plot than the grazed pasture (p = 0.0057) in post hoc pairwise comparisons within communities. E More recruitment of Inga s eedlings than Quararibea (p < 0.0001) or Pouteria ( p < 0.0001), and more Quararibea seedlings than Pouteria (p = 0.0004) in post hoc pairwise comparisons. F In Kunkuki, more seedlings were recruited in the forest than the mature pasture plot (p = 0.0171) in post hoc pairwise comparisons within communities. G More seedling recruitment in San Ramon than in Kenkuimi (p = 0.0085) in post hoc pairwise comparisons.

PAGE 40

40 Table 2 2. Results from mixed model logistic regression of mortality and seedling recruitment of three large -seeded tropical trees at 14 weeks Also shown are results of logistic regression on 14 week survival of Inga densiflora seedlings that were recruited at 2 weeks Seeds were planted on the soil surface and buried in three field types (grazed pasture, mature pasture, forest) in three indigenous communities in the Ecuadorian Amazon. R esults are for reduced models including only main effects and significant interactions (p 0.05). Source Num df Den df F value P > F Num df Den df F value P > F Num df Den df F value P > F Species 2 703 63.53 0.0001A2 724 73.05 <0.0001D Burial 1 703 0.69 0.4058 1 724 0.02 0.8844 1 229 0.06 0.0869 Field (Community) 6 703 0.88 0.5118 6 724 0.99 0.4316 6 229 0.91 0.4889 Community 2 703 4.08 0.0174B2 724 22.13 <0.0001E2 229 12.51 <0.0001G Species*Community 4 703 4.77 0.0008C Mortality at 14 wks 14 wk seedling recruitment of all seeds planted 14 wk survival of Inga seedlings recruited at 2 wksF A Higher mortality of Pouteria than Inga (p < 0.0001) or Quararibea (p = 0.0411), and higher mortality of Quararibea than Inga ( p < 0.0001) in post hoc pairwise comparisons between species. In subsets of individual species, mortality of Inga differed among communities (p < 0.0001), with higher mortality in Kenkuimi than San Ramon (p = 0.0029) or Kunkuki (p < 0.0001), and higher mortality in San Ramon than Kunkuki (p = 0.0139) in post hoc pairwise comparisons between communities. B Higher mort ality in Kenkuimi than Kunkuki (p = 0.0233) in post hoc pairwise comparisons between communities. C In all communities, higher survival of Inga than Quararibea and Pouteria (p < 0.05 in all pairwise comparison between species within communities in post hoc tests). Within Kunkuki higher survival of Quararibea than Pouteria (p = 0.0373) in post hoc comparisons within communities. D More Inga than Quararibea (p < 0.0001) or Pouteria (p < 0.0001) seedlings were recruited, and more Quararibea than Pouteria (p = 0.0050) seedlings were recruited in post hoc pairwise comparisons between species. In subsets of individual species, seedling recruitment differed among communities for Inga (p < 0.0001), with more Inga seedlings were recruited in Kunkuki than Kenkuimi (p < 0.0001) or San Ramon (p < 0.0095), and more Inga seedlings were recruited in San Ramon than Kenkuimi (p < 0.0001) in post hoc pairwise comparisons between communities. E More seedlings were recruited in Kunkuki than in Kenkuimi ( p < 0.0001) or San R amon ( p < 0.0001), and more seedlings were recruited in San Ramon than Kenkuimi (p = 0.0060) in pairwise post hoc comparisons between communities. F Only mortality of Inga seedlings was analyzed due to low number of seedlings of Quararibea (27) and Pouter ia (1) at 2 weeks G Higher survival of Inga seedlings in Kunkuki than Kenkuimi (p < 0.0001) or San Ramon (p = 0.0370), and higher survival in San Ramon than Kenkuimi (p = 0.0043) in post hoc pairwise comparisons between communities.

PAGE 41

41 Figure 2 1. Map of the study area in Morona Santiago and Pastaza Provinces, Ecuador. Stars indicate the locations of i ndigenous communities and towns The community of Tsura k (1 4831S, 774950W) had approximately 200 residents in 2008, and is located on the Puyo-Macas Rd, 51 km south of the provincial capital of Puyo, and 78 km north of the town of Macas. The division between the provinces in this area is the Palora River : Pastaza P rovince is east of the river. Palor Rive Pu yo M ac Ro ad Ri ve Kunkuki Kenkuim Nort Palora River San Ramon Kenkuimi North Research site Quito Pastaza River Puyo Macas Rd Puyo Macas Tsur ak Indigenous community San Antonio Kunkuki 0 5 10 15 Scale in km Town Palora

PAGE 42

42 Figure 2 2. S chematic dia gram of the experimental design s howing nesting of field plots (grazed pasture, mature pasture, forest) in each indigenous community (Kenkuimi, Kunkuki, San Ramon). Each field plot contained si x 50 m transects separated by 4 8 m. Individual s eeds were planted at 1 m intervals at locations marked with a metal wire and 1 m from the transect ( L/R side of transect, specie s, and treatment were randomized )

PAGE 43

43 Figure 2 3. Proportion of seeds of three large -seeded tropical tree species that were m issing at 2 weeks from three field types (grazed pasture, mature pasture, forest) nested in three indigenous communities (Kenkui mi, Kunkuki, San Ramon) 1418 km apart in the Ecuadorian Amazon. More seeds disappeared from the forest than fro m grazed pastur e in Kenkuimi (p = 0.0049; Table 2 1). Seed disappearance in the other two communities could not be statist ically analyzed due to lack of seed removal from several field types. Surface-planted and buried seeds did not differ in seed disappearance (Table 2 1) and are combined in the figure.

PAGE 44

44 Figure 2 4. Seed fates at 2 weeks of three large -seeded tropical tree species that were planted as seeds in forests and pastures in three indigenous communities (Kenkuimi Kunkuki, San Ramon) 14 18 km apart in the Ecuadorian Amazon (Figure 2 1). Seedling recruitment differed among the species ( p < 0 .0001) and communities (p = 0.0115; Table 2 1). Surface -planted and buried seeds are combined in the figure because seed burial did not affect mortality or seedling recruitment (Table 2 1). Seed disappearance differed among field t ypes (forest and pasture plots) and is illustrated in Figure 2 5, but field types are combined in this figure.

PAGE 45

45 Figure 2 5 Seed fates at 14 weeks of three large -seeded tropical tree species that were planted as seeds in forests and pastures in three indigenous communities (Kenkuimi, Kunkuki, San Ramon) 1 4 18 km apart in the Ecuadorian Amazon (Figure 2 1). Seedling r ecruitment differed among species ( p < 0 .0001) and communities ( p < 0 .0001; Table 2 2). Seedling recruitment did not differ among field types or surface and buri ed seeds (Table 2 2), and treatments are comb ined in the figure.

PAGE 46

46 CHAPTER 3 SEEDLING RECRUITMENT OF FOUR LARGE SEED ED TROPICAL FOREST TREES PLANTED AS SEEDS IN SECONDARY FORESTS IN THE ECUADORIAN AMAZON Introduction Tropical forest mana gement requires an ever greater understanding of secondary forests, which are expanding in area and increasingly used by people for forest products. Tropical countries are becoming more urbanized (United Nations 2004) a trend associated with abandonment of less productive agricultural land (Mather and Needle 1998) and expansion of secondary forest (Aide and Grau 2004) L ocal people are often dependent on these forests for timber, firewood, a nd nontimber forest products ( e.g. fruits and medicine s ) for both market and subsistence uses (FAO 2005). As mature forest cover is reduced (Achard et al. 2002; Hansen et al. 2008) and increasingly fragmented by human activities (Wade et al. 2003) more management and conservation efforts are being directed toward secondary forests (Bowen et al. 2007) The initial phases of secondary succession in tropical forests are characterized by structural recove ry, a process affected by the nature of the d isturbance (agricultural practices hurricane, fire etc.; Chazdon, 2003), soil fertility (Moran et al. 2000) seed bank (Dalling and Hu bbell 2002; Wijdeven and Kuzee 2000; Zahawi and Augspurger 1999) presence of resprouting trees (Zimmerman et al. 1994) and propagule input from the surrounding landscape (Cubina and Aide 2001; GalindoGonzalez et al. 2000) ; see review in (Dalling and John 2008) As succession continues, changes in species composition will largely depend on colonization of non-pioneer species (Gauriguata and Ostertag 2001; Norden et al. 2009) Lack of dispersal of later successional species with large seeds can severely constr ain species composition in subsequent phases of forest succession (Wijdeven and Kuz ee 2000) In landscapes that lack animal dispersers and nearby seed sources, recovery of original forest species diversity may take centuries (Finegan 1996) or not occur at all (Turner et al. 1997) Augmenting natural seed rain

PAGE 47

47 by plan ting large animal -dispersed species can accelerate natural successional processes (Camargo et al. 2002; Martinez Garza and Howe 2003) Even in landscapes where seed s arrive to successional sites seeds must pass through several post -dispersal filters on survival and development to reach the seedling stage. Seed pre dation in tropical secondary forests can be as high as 100% of dispersed seeds (Crawley 2000; Holl and Lulow 1997; Pea Claros et al. 2002) though variability among species is common. Even if seeds are not consumed, they may be killed by desiccation, pathogens (Pringle et al. 2007) or insects (Galetti et al. 2006; Wright et al. 2000) I investigate d factors that can constrain seeding recruitment of large -seeded species in secondary forests once dis persal limitations are overcome. In the study reported in Chapter 2, I found strong species differences in post -dispersal survival but little difference among habitat types in seedl ing recruitment. In the study reported here, I conducted a similar experi ment with a mo dified experimental design, using one species shared with the previous study ( Inga densiflora ), and three others with large recalcitrant seeds ( Gustavia macarenensis, Caryodendron orinocense and a Myrtaceae species ). Seeds were planted in se condary forests, due to the extent of this forest type in the tropics and lack of variability of early survival among habitat types (pastures and forests) found in the previous study. In Chapter 2, I discussed differences in seedling recruitment found among the three indigenous communities located 1015 km from each other. To control for this regional variability, this study was conducted in three secondary forest plots located within 5 km of each other within the region of the first study. A caging trea tment was included to test for vertebrate seed predation which can greatly reduce the numbers of seeds available for seedling rec ruitment in tropical forests. T he test of seed burial to reduce seed predation and increase s eedling recruitment was repeated but biweekly censuses

PAGE 48

48 were made to more accuratel y assess mortality and trends in seedling recruitment. To better understand successi onal processes I inventoried all woody species > 10 cm dbh encountered in four 50 x 2 m transects in each plot. As in the previous study, the four species planted were valued by local people for production of edible fruits. Research Questions How like ly is seedling recruitment from seeds of large -seeded trop ical forest trees planted in secondary forests? How does shall ow seed burial affect seed removal, seedling recruitment and mortality ? Are results consistent among the four species of large-seeded tropical forest trees tested? Are results consistent among three secondary forest plots located 15 km apart? Materials and Methods Study Site This study was conducted in the Shuar indigenous communities of San Antonio and Kunkuki in the Morona Santiago P rovince of the Ecuadorian Amazon, about 55 km south of the provincial capital city of Puyo (1 200S, 78010W ; Figure 2 1 ). T wo study plots were located in secondary for ests in San Antonio, and one plot was in Kunkuki. The climate, soils and vegetation type of the region are described in Chapter 2. Tree Species Species were selected from native trees with large, re calcitrant seeds that were in fruit at the onset of the experiment (January 2008). The four species were Inga densiflora Benth (Fabaceae), Gustavia macarenensis Philipson (Lecythidaceae), Caryodendron orinocense Karst (Euphorbiaceae) and a Myrtaceae ide ntified as a species of Plinia by Bruce K. Holst (Selby Botanical Gardens) All species will be referred to by their generic name after description in the next paragraph. Average fresh w eigh t of seeds ranged from 2.3 grams to 19 grams (Table A 2).

PAGE 49

49 Vouch er specimens of the four species were deposited in the National Herbarium of Ecuador, Quito (QCNE) in July of 2008, and duplicates were packed for shipment to the University of Florida Herbarium (FLAS) via the Missouri Botanical Gardens. All species have edible f ruits that are consumed locally and occ asionally sold at small markets in nearby towns (Palora and Puyo). Fruit collection is usually from ostensibly wild trees, but the species are all occasionally planted and natural ly regenerating seedlings a re sometimes transplanted to forest gardens (Byg and Balslev 2006) Inga densiflora was used in the previous experiment, and is described in detail in Chapter 2. Gustavia macarenensis is a medium sized (10 20 m heig ht ) forest tree found in Ecuador, Peru and Venezuela (Prance and Mori 1979) It has pink red indehiscent globose fruits of 5 15 cm containing 3 7 large polygonal seeds. The pulpy orange mesocarp surrounding the seeds is eaten raw by the Shuar (Bennett et al. 2002) Caryodendron orinocense is a slow -growing tree, native to the eastern foothills of the Andes in Venezuela, Ecuador and Colombia (Neito and Rodriguez 2002) The fruits are capsules that open when the fruit falls to release (typically) 3 seeds. T h e seeds are high in protein and oil (Padilla et al. 1998) and are eaten both toasted and raw by people (Schnee 1973) The ground seeds have been studied as a potential substitute for commercial soybean protein meal (Padi lla et al. 1996) The specimens of Plinia sp. used i n the experiment had ovoid berry fruits (1 seed per fruit) that were a pproximately 10 x 15 cm with thin yellow to or ange skin and a flesh, white slightly aromatic mesocarp that is consumed raw (Figure A 4 ). The species is cultivated on a small scale around Palora, Ecuador and is known locally as shawi. It is similar to a fruit called mulchi, cultivated in the Napo Province (Ecuador), about 80 km NW of Palora. Prior to this study, no herbarium spec imens of this species we re in the National Herbarium of Ecuador (QCNE).

PAGE 50

50 Seeds were collected from fresh fruits of 3 trees of Gustavia and Caryodendron trees growing in mature forest approximately 15 km west of the experimental plots. Inga seeds were col lected from five roadside trees in Kunkuki, and Plinia seeds were collected from two trees growing in pastures near the town of Palora (see Figure 2 1 for location of towns). Fruit pulp was removed, seeds were inspected for insect holes and damage, washe d in freshwater, float tested for viability, and planted within one week of harvest. Due to lack of seed availability at the start of the experiment (February 2008), Caryodendron seeds were planted 2 weeks after the other three species. All data are repo rted as weeks since planting. Thirty seeds of each species were also planted 1 cm deep in outdoor raised planting beds of forest soil in full sunlight. Secondary Forest Plots Three 0.5 ha study plots were established in 10 15 yr old secondary forests in February 2008. Two plots (Hartensia and Cascada) were located in the Shuar indigenous community of Kunkuki and one (Centro Semillas) in San Antonio (Figure 2 1). All plots were within 0.5 km of the PuyoMacas Road, and separated by 1 5 km. The sec ondary forests were located within larger tracts of secondary and selectively logged forest. The landscape surrounding Kunkuki is described in Chapter 2. The secondary forests were all natural regrowth from either horse pasture (Cascada plot in Kunkuk i) or abandoned home gardens of maize, yucca and pltano (Hartensia plot in Kunkuki and Centro Semillas plot in San Antonio). Vegetation in the secondary forest plots was charact erized by percent canopy cover measur ed with a spherical densitometer he ld at 1 m above ground level at 20 points per plot, and by height and dbh measurements of woody stems m easured in four 50 m x 2 m transects per 0.5 ha plot (modified Gentry transects; Phillips et al. 2000). Canopy cover ranged from 77 8% (Cascada) to 83 4% (Centro Semillas), and differed among the three plots (p = 0.0011; Figure A 5 ). Woody stem density of trees > 10 cm dbh was 425875 stems/ha and average height of

PAGE 51

51 trees encountered was 12 6 m. Woody tree density of stems 2.5 cm dbh was estimated to be between 3225 3625 stems/ha (Table A 4). Experimental Design A 2 x 2 factorial design was used to test seedl ing recruitment and mortality of buried vs. unburied and caged vs. uncaged seeds, resulting in four treatment co mbinations (surface & caged, surface & uncaged, buried & caged, buried & uncaged). In the surface treatment, seeds wer e placed on the ground surface after removal of woody debris. B uried seeds were placed in a 1 2 cm deep hole and covered with ~1 cm of soil. Cylindrical cages were made of 24 gauge galvanized 0.5 mm wire mesh, and measured approximately 15 cm diameter x 30 cm tall. Cages were open on the bottom and closed on the top. Cages were inserted 1 cm into the ground and anchored in place by three 20 cm hook-shaped 14 gauge wires. L ocation s of uncaged seeds were marked by 20 cm 14-gauge wire s inserted into the ground beside each seed. In each of the three secondary forest plots, 50 100 m long parallel transects were established at 3 8 m inte rvals as described in Chapter 2 (See Figure 2 2). Species, burial treatment, caging and left/right planting were randomized along transects, with an average of 20 replicates of each species x treatment combination planted in each of the three plots. Se ed disappearance (defined as no seed found within 30 cm of the wire that marked the planting location ), germination, seedling recruitment and mortality were recorded at 2, 4, 6, 8, 10, 16, and 18 weeks after planting for Caryodendron, and at weeks 2, 4, 6, 8, 10, and 18 for the other three species. Germinati on (emergence of radicle or epi cotyl from the seed) was recorded at each census without disturbing seeds. In most cases, germination of buried seeds could not be detected until the stem emerged above t he ground surface Seedling recruitment was defi ned as development of at least one true leaf long (or photosynthetic cotyledon, in the case of Caryodendron; see Table A 2 for cotyledon and seedling types). At 18 weeks remaining seeds,

PAGE 52

52 both burie d and nonburied, were cut open and vis ually inspected to determine germination and viability. Seeds with solid, intact, and undamaged endosperm were considered alive. To determine if any seed -eating animals were present at the study site, 9 kernels of maize were placed on the soil surface at 20 locations in each plot in early May, and the number s of kernels remaining were recorded 14 days later. Data Analyses E ffects of species, seed burial, and plot location on seedling recruitment and mortality were assessed by logistic regression using PROC GLIMMIX in SAS 9.1.3 (2003, SAS Institute, Cary, NC). I first tested full factorial models including main effects and all interactions, and subsequently reduced models by removing non -significant interactions (p > 0.05). I r eport analyses of main effe cts and significant interactions S ignificant (p further e valuated in post -hoc Bonferroni adjusted pairwise comparisons. When species was a highly significant main effect (p < 0.0001) treatments were tested separately for each species in mixed models. Data were discarded if the wire marking the planting location was not found in the final census. Results General Patterns of Seeding Recruitment and Survival Over the first 2 weeks only 5 of 410 uncaged seeds disappeared from their planting sites. None of the 421 caged seeds and no additional uncaged seeds disappeared by 18 weeks S urvival of seeds and seedl ings to 18 weeks differed among species (p < 0.0001; Table 3 1), and ranged from 73% ( Plinia ) to 94% ( Gustavia ; Figure 3 1). At 18 weeks, seedling recruitme nt also differed among the four species (p < 0.0001; Table 3 1), and ranged from 17% ( Plinia ) to 76% ( Inga and Caryodendron; Figure 3 2). Very few (16 of 498) seedlings died during the experiment (Figure 32).

PAGE 53

53 Species Differences Seedling recruitment at 18 weeks differed among species (p < 0.0001; Table 31) and was lowest for Plinia ( p < 0 .0001 in post -hoc pairwise comparisons with each of the other species; no other pairwise comparisons between species were significant ; Table 3 1). At 18 weeks 76% (174 of 228) of Caryodendron, 64% (128 of 198) of Gustavia 76% (144 of 189) of Inga, and 17% (36 of 216) of Plinia seeds planted were present as live seedlings (Figure 3 1). Seedling recruitment at 18 weeks was positively correlated with early germinatio n. At 2 weeks, germination was observed in 90% of surface -planted Caryodendron, Gustavia and Inga seeds, and only 19 % of surface -planted Plinia seeds (only germination of unburied seeds could be assessed at 2 wks without disturbing developing seedlings) Most Caryodendron and Inga seeds develop ed to seedlings by week 6, with few additional seeds becoming seedlings after 10 weeks In contrast, seedling development was still continuing for Gustavia and Plinia at 18 weeks (Figure 3 2). Species differed in mortality ( dead seeds and seedlings and m issing seeds ) at 18 wks (p < 0.0001, Table 31; Figure 3 1). Gustavia had higher survival than Caryodendron, Inga or Plinia (p = 0.0002, p = 0.0003, p < 0.0001, respectively) and Inga had greater survival than Plinia (p = 0.0003; Table 3 1). Mortality of seeds that had not become seedlings in 18 weeks also differed among species (p < 0.0 001; Table 3 1). Almost all (38 of 44) of Caryodendron seeds that did not become seedlings we re dead, whereas 84% (58 of 69) of Gustavia 63% (25 of 40) of Inga, and 67% (121 of 180) of Plinia seeds that had not been recruited to seedlings were alive at 18 weeks (Figure 3 1). Seed Burial At 18 weeks seedling recruitment was higher for seeds planted on the soil surface t han for buried seeds (p < 0.0001; Table 3 1 Figure 3 1). This effect was primarily due to lower seed l ing

PAGE 54

54 recruitment of Gustavia and Inga from buried seeds (p < 0.0001 and p = 0.0015, respectively, in individual species subsets; Table 3 1 ). Of seeds that had not developed to the seedling stage by 18 weeks more buried than nonburied seeds were dead (p = 0.04; Table 3 -1). This finding was almost entirely due to higher mortality of buried Plinia seeds (p = 0.0203; no other species showed effect of burial treatm ent; Table 3 1; see Figure 3 1) Plots The thr ee secondary forest plots did not differ in seedling recruitment, total mortality, or mortality of seeds that did not become seedlings (Table 3 1). Caging Caged and uncaged seeds did not differ in seedling recruitment, total mortality, or mortality of seed s that did not become seedlings (Table 3 1, Table A 6). Species Inventory A total of 129 species of trees > 2.5 cm dbh representing 69 genera and 3 4 families were recorded in twelve 50 x 2 m transects (4 transects x 3 plots). The most frequently encounte red species > 10 cm dbh in all plots was Piptocoma discolor (pig Asteraceae; Table A 5) Common understory species (2.5 10 cm dbh) were Calliandra trinervia (Fabaceae), Casearia spp. (Flacourtiaceae), Hedyosmum racemosum (Chloranthaceae), Inga spp. ( Fabaceae), Miconia spp. and Ossaea spp. (Melastomataceae), Palicourea spp. (Rubiaceae ), Piper obliquum (Piperaceae), Saurauia spp. (Actinidiaceae), Schefflera morototonia (Araliaceae) and Tournefortia sp. (Boraginaceae). Large -seeded species encountered i n transects included Caryodendron orinocense (one of the study species), multiple Inga species, Socretea Ocotea Guarea Brosim um and Virola (Table A 2).

PAGE 55

55 D iscussion The majority of forests in the tropics are degraded by human activities or in a state of secondary succession (I.T.T.O. 2002) Mature tropical forests continue to be fragmented and reduced by loggi ng new plantation s agricultural expansion (Wassenaar et al. 2007) and increased frequency and intensity of fires (Mayaux et al. 2005) At the same time, secondary forests in the tropics are expanding, reclaiming former agricultural land (Aide and Grau 2004) and as much as 10% of forest cover claimed by logging in the last two decades (Wright 2005) Future species composition of these forests will depend on natural successional trajectories based on prior land use and ecology of the surrounding landscape, combined with management interventions to increase the growth and survival of desired species. The purpose of this experi ment was to assess the likelihood of seedling recruitment of four late -successional s pec ies that were experimentally dispersed into 1015 yr old secondary forests. High levels (> 6 0%) of seedling recruitment of three of the four species resulted from extremely low levels of seed removal ( 5 of 410 uncaged seeds), high germination (> 90% of non-buried seeds) and low mort ality of newly recruited seedlings ( 16 of 482 seedlings). Seed burial did not increase seedling recruitment and was associated with increased seed mortali ty of Plinia the species with the low est seedling recruitment at 18 weeks (17%) Low Seed Removal Seed predation can dramatically reduce the number of seeds available for seedling recruitment (Notman and Gorchov 2001; Pea -Claros and De Boo 2002) However, in my experiment, very few seeds disappeared from plantin g sites and thus s eed predation presented virtually no obstacle to seedling recruitment. The near complete lack of seed removal probably reflects the lack of large mammalian seed predators (and secondary disper sers) in the study area. All plots were < 0.5 km from hou ses a long the Puyo -Macas road, and I was told by local

PAGE 56

56 residents that large animals such as peccaries and monkeys were uncommon in the area. I also observed that medium to large wild animals were nearly always hunted when encountered (agoutis, peccaries, birds etc.). P opulations of large animals that consume large seeds can be dramaticall y decreased by hunters (Jerozolimski and Peres 2003) and resulting low levels of seed predation of large -seeded species have been documented in other defaunated tropical rainforests (Dirzo et al. 2007) Some evidence for lower seed removal in this area was also found in the experime n t described in Chapter 2 where the lowest level of seed removal among three communities was in Kunkuki (Figure 2 3), the village closest to the study plots in this experiment (Figure 21) The observed infrequency of seed removal could also indicate that seeds were unpalatable to seed predators Small rodents are often common in disturbed h abitats where large mammals are extirpated (Dirzo et al. 2007) but tend to prefer small er seeds than those used in this experiment (Mendoza and Dirzo 2007; Vieira et al. 2006) G ranivores that consume small seeds were present since 68% ( 23.8% ) of maize kernels were r emoved in two weeks. These were likely small rodents, although birds could also have removed seeds. Another explanation for lack of seed predation of the planted seeds is that s econdary compounds in the seeds could also have deterred predation. For exam ple, the Plinia seeds smelled of volatile oils, and other Myrtaceae species (Eucalyptus Melaleuca etc.) are known to produce aromatic compounds that are unpalatable to most mammals (Jones et al. 2003a) S eed predation may also have been reduced by qu ick germi nation of three of the four study species, which minimized the time the seeds were exposed to potential seed consumers At least 90% of Inga, Gustavia and Caryodendron seeds placed on the soil surface germinated in the first 2 weeks Within days of germ ination all seed resources in Caryodendron seeds were

PAGE 57

57 converted into the young plant (only the seed coat remained). Although the chemical composition of seed tissue changes during germination (Bewley an d Black 1994) in ways that generally may make them less attractive to seed predators (Forget 1990) ge rmination itself does not provide complete escape from seed predators for all large seeds because storage cotyledons may still be consumed after the seedling has developed (Alvarez Clare and Kitajima 2009) Inga, Gustavia and Plinia all ha d large storage cotyledons that were still attached to recruited seedlings at 18 weeks although none were damaged by rodents. High Germination By 18 weeks over 90% of In ga, Gustavia and Caryodendron seeds and at least 79% of Plinia seeds had germinated. These levels are similar to those found in previous studies. Many Inga species regularly exhibit germination levels of 95 100%, often within days of removal from the fru it (Pennington and Revelo 1997) Likewise, 100% germination has been observed for Caryodendron seeds within 1215 days of planting (Jimenez and Bernal 1989; Neito and Rodriguez 2002) In Panama, 463 of 500 (93%) seeds of another Gustavia species, (G superba) germinat ed during an 8 -week study (Dalling and Harms 1999) At 12 months, germination of morphologically similar Brazilian species of Pli nia (also cultivated for fruit production) was 81% in one study (Danner et al. 2007) and 70 85% in another (Andrade and Martins 2003) Seed Burial As was also found in the study presented in Chapter 2, seed burial did not result in higher germination or seedling recruitment for any species Seed burial can reduce the probability of detection by seed predators, but seed removal of study species was minimal and did not present an obstacle to seedling recruitment. Seed burial prevents moisture loss (Fenner and Thompson 2005) which can be a major cause of post -disp ersal mortality for desiccation -sensitive seeds in disturbed or exposed sites (Vieira and Scariot 2006) but desiccation of seeds was rarely

PAGE 58

58 observed during this exp eriment Moisture loss can be avoided by rapid germination (Daws et al. 2005) and at 2 weeks over 90% of surface -planted Caryodendron, Gustavia and Inga seeds were germinated. The combination of rapid germination and an overall wet climate (average of ~400 mm of rain in March; Table A 1) may have minimized any moisture retention benefit of seed burial. Rather, it appears buried seeds were more susceptib le to pathogens since more seed death was observed in buried seeds. At 18 wks, all 26 dead Plinia seeds and nea rly all 58 live Plinia seeds that were removed from the soil showed signs of decay i n the form of discolored tissue. In contrast, 29 of 30 seeds buried in outdoor raised planting beds of forest soil that were exposed to full sunlight were live seedlings b y 12 weeks after planting (Table A 2). Moist and shaded conditions under forest cover are associated with increased pathogen attack (Augspurger 1984b) and this is probably why s eed surv ival and seedling recruitment were higher in planting beds than in the forest. Secondary Forests The res ults of this study indicate that young seedlings will recruit from large -seeded tropical forest trees if the seeds are dispersed into secondary fore sts. Many large -seeded species (including two of the study species) were found in the surveys of th e plots ( e. g. Caryodendron orinocense multiple Inga species, Socretea Ocotea Guarea Brosim um Virola ). This indicates that seed s of large -seeded sp ecies were naturally dispersed into these forests and survived to establish as young trees during the 10 15 years since abandonment of agricultural use. N atural regeneration of common large -seeded species in young secondary forests was also fou nd by Nor den et. al. (2009) in Costa Rica in an agricultural landscape that included fragments of mature forest and an intact animal disperser community. Although larger -bodied animals were hunted in the region of my study, the presence of large -seeded trees in 10 15 yr old secondary forests indicates at least some large -seeded species are being dispersed from nearby remnant

PAGE 59

59 trees or patches of selectively logged forest, and possibly from the mature forest located approximately 10 km to the east. In addition, some species may have regenerated by sprouting (Mwavu and Witkowski 2008) but evidence of this was not noted in the surveys. R ecommendations for Propagation of the Study Species Higher initial mortality of seedlings can be expected when species are planted in secondary forests rather than in nurseries but this may be wor th the tradeoff in avoiding extra expenses associated with nursery propagation. At 18 weeks seedling recruitment of Inga, Caryodendron and Gustavia ranged from 65 76% and seed loss es were relatively minimal (15 %, 5% and 8 %, respectively; Table A 6 ). These species may be suitable for direct -seeding. The fourth species, Plinia had the lowest (17%) and slowest recruitment to seedlings (Figure 3 -2) and seeds were subject to pathogen infection when planted in secondary forests. For this species nursery planting may be advantageous if seed sources are scarce. Fo r all species, seedling recruitment levels were high (87 100%) when seeds were planted below the soil surface in raised outdoor planting beds of forest soil in full sunlight (Table A 2 ). Seed burial was detrimental to seed and seedling survival in seconda ry forests, and is not recommended for any species if seeds are being planted in forests. After seedlings have established in the forest removal of some overstory trees is generally recommended in enrichmen t planting to promote growth of desired species (Pea Claros et al. 2002; Romell et al. 2008) Recommendations for Future Studies Results of this study suggest that regeneration of large -seeded species in secondary forests in this region is not severely constra ined by mortalit y during seedling recruitment. T o better understand constraints on establishment of large -seeded species beyond the stage of seeding dev elopment the experimental plantings would need to be monitored over a longer time period ( 1 yr) See d trap studies and more extensive vegetation surveys of the seedling and sapling

PAGE 60

60 u nderstory could be combined with the experimental plantings to better assess the relative impact of experimental seed augmentation on successional processes and forest divers ity. Conclusion Results of this research have implications for understanding of successional processes in young tropical secondary forests. The majority (> 60%) of seeds of three out of four large seeded, later succes sional tree species ( Inga, Gustavia an d Caryodendron) became seedlings by 18 weeks Seedling development was slower for Plinia but still over 15 % of seed s planted were live seedlings at 18 weeks Seed predation was extremely low (< 1 % of seeds), and seed burial did not improve survival or s eedling recruitment for any species. Large -seeded species that overcome dispersal barriers to reach secondary forests near human settlements appear to have a high cha nce of seedling establishment. H unting reduces populations of animal dispersers, but may foster succession by reducing seed predation pressure on occasionally dispersed (or planted) seeds Recovery of large -seeded species in secondary forests can be augmented by planting seeds, improving the value of the landscape for both wildlife and human s

PAGE 61

61 Table 3 1. Results from mixed model logistic regression of seedling recruitment, total mortality (missing seeds, dead seeds and seedlings), and mortality of remaining seeds at 18 wks of four large -seeded tropical tree species planted as seeds in three secondary forest plots in the Ecuadorian Amazon. Seeds that disappeared (n = 5) are counted as dead in the analysis of total mortality and are excluded from the analysis of mortality of the remaining seeds. Results are for reduced models of only main ef fects and significant interactions (p Source Num df Den df F value P > F Num df Den df F value P > F Num df Den df F value P > F Species 3 817 21.66 < 0.0001A 3 823 12.84 < 0.0001E3 325 13.72 < 0.0001G Plot 2 817 0.1 0.9076 2 823 1.92 0.3381 2 325 2.72 0.1089 Cage 1 817 0.01 0.033 1 823 0.69 0.7546 1 325 2.4 0.2029 Buried 1 817 19.57 < 0.0001B1 823 9.04 0.0126F1 325 5.97 0.0400H Species*buried 3 817 3.77 0.0105C Cage*species 3 817 2.85 0.0364D 18-week seedling recruitment of all seeds planted 18-week mortality of all seeds planted 18-week mortality of remaining seeds that had not developed into seedlings A Fewer Plinia seedlings were recruited than the other three species (p < 0.0001 in all post hoc tests). B Fewer seedlings from buried seeds. C In individual species subsets, fewer seedlings were recruited from buried than surface plant ed seeds for Gustavia (p < 0.0001) and Inga (p = 0.0015). D In individual species subsets, fewer Plinia seedlings were recruited from caged than uncaged seeds (p = 0.0166). E Higher mortality of Plinia than Gustavia (p < 0.0001) or Inga (p = 0.0003), and higher mortality of Caryodendron than Gustavia (p = 0.0002) or Inga (p = 0.038) in post hoc pairwise comparisons between species. In individual species subsets, an interaction between caging and burial was observed only for Plinia (p = 0.0348), with higher mortality of caged (p = 0.0169) and buried (p = 0.0004) Plinia seeds. F More buried t han surfaceplanted seeds died. G More Caryodendron than Gustavia Inga or Plinia seeds died (p hoc pairwise comparison ). More Plinia t han Gustavia seeds died ( p = 0.0203). In individual species subsets, more buried than surfaceplanted Plinia seeds died (p = 0.0203). H More buried than surface planted seeds died.

PAGE 62

62 Figure 3 1. Seed fates at 18 weeks of four species of large-seeded tropical trees that were planted on the soil surface and buried in three secondary forests plots in the Ecuadorian Amazon. Live seedlings and dead seedlings, live and dead seeds and seeds that disappeared are indicated (see key). Fewer seedlings were recruited from buried seeds (p < 0.0001; Table 3 1). Cages and uncaged seeds did not differ in seedling recruitment or survival (Table 3 1 ) and are combine d in the figure

PAGE 63

63 Figure 3 2. Seedling development of four species of large -s eeded tropical trees that were planted as seeds on the soil surface and buried in secondary forests in the Ecuadorian Amazon. Live seedlings are indicated in grey, dead seedlings in black, and seeds that disappeared from planting sites by the hashed shading. The white areas are seeds that did not become seedlings during the 18 week experiment (see Figure 3 1 for 18-week mortality of seeds) Data are presented for weeks 2, 4, 6, 8, 10, 16 and 18 for Caryodendron, and weeks 2, 4, 6, 8, 10, and 18 for the other three spec ies. Caged and uncaged seeds did not differ in seedling recruitment or survivorship (Table 3 1 ) and are combined in the figure.

PAGE 64

64 CHAPTER 4 CONCLUSION N atural regeneration of large -seeded tropical forest trees in degraded forests and agricultural landscap es i s often severely limited by lack of animal -mediated seed dispersal. The substantial energy and nutrient reserves in large seeds may allow for relatively high levels of seedling recrui tment when dispersal limitation is overcome. In fragmented tropical forests near human settlements, arrival of mature forest seeds by rare cases of long -distance dispersal may be critical to species persistence, providing seeds survive at the new sites. In addition to being of value to wildlife, many tropical fruit trees with large -seeds have fruits consumed by people, and if successful, planting seeds of these species into the landscape (direct seedling) is an inexpe nsive method to propagate large -seeded species. The goal of this thesis was to assess the chance of seedling recruitment of large -seeded tropical forest trees in human -modified landscapes in the Ecuadorian Amazon. Seeds of six species of native large -seeded tropical trees with recalcitrant seeds were planted in a variety of habitats, and seedling recruitment and survival were assessed for 3 5 months. Species sele ction reflected the interest of local indigenous landowners in propagating fruit -bearing trees. Seed burial was hypothesized to increase seedling recruitment by reducing seed predation and desiccat ion. Over 55% of Inga, Caryodendron and Gustavia (56 & 76% in the two experiments, 76%, and 64%, respectively), 17 % of Plinia 10% of Quararibea, and 2% of Pouteria seeds developed into seedlings and survived during the 14 18 week periods of the two e xperiments Seed burial did not increase survival or seedling recruitment of any species. The location of the plots in the landscape (the indigenous community in which seeds were planted) affected seed removal and seedling recruitment more than the habit at in which seeds were planted ( grazed

PAGE 65

65 pasture (2 6 weeks vegetative regrowth at time of planting) mature pasture ( 3 6 months of vegetative regrowth ), and selectively logged forest ; Chapter 2). In both studies, s eed removal was rare (0 10% ) and was not reduced by seed burial. Large -seeded species were present in surveys of 10 15 yr old secondary forest, confirming that large -seeded species were able to recruit from naturally dispersed seeds (Chapter 3). Direct seeding of large -seeded species may b e a useful technique for large -scale restoration projects. Planting seeds instead of transplanting seedlings allows for the reforestation of larger and less accessible regions because seeds are smaller and easier to transport than seedlings, and the cost of seeds is often substantially less than the cost of seedlings. Large -seeded tropical species often have fleshy fruits which provide food for larger -bodied animals, which can help catalyze further seed dispersal of mature forest species from remaining fo rest patches. Ironically, the chances of establishment of large -seeded species from seeds may be higher in regions most disturbed by human activities because hunti ng and habitat loss also reduce popula tion s of seed predators of large seeds. Direct -seedin g can be an advantageous strategy when the longterm investment in forest restoration is not possible due to logistical, financial or political limitations For example, if an NGO only has funding for a 6 -month project, direct seedling allows for immediat e tree planting at the project outset when outside resources as well as local interest in new projects are often highest. During the 2 years of this research, at least two forest restoration projects using seedling nurseries were started and abandoned by internationally -funded NGO groups operating in the study area. Many more seedlings died nurseries than were planted in the field. Finally, forest restoration is a longterm endeavor and predicting future land use in the tropics is often difficult. The simplicity of direct seeding reduces the costs of plant ing at many sites, thus allowing

PAGE 66

66 landowners and managers to account for the inevitable conversion of some sites to agriculture or other uses and still accomplish forest restoration at other sites The results of thi s research suggest human -modified landscapes can be managed to help recover tropical forest biodiversity. Large -seeded tree species reestablish ed within a decade of agricultural abandonment, and seed planting of five of six species was large ly successful in early seedling r ecruitment and survival. Longterm studies will be useful to evaluate later constraints on seedling survival and provide further management suggestions for the propagation of large seeded species.

PAGE 67

67 APPENDIX CHARACTERIZATION OF RESEARCH SITE S AND SPECIES USED IN THE EXPERIMENTS Table A 1 Average monthly precipitation and air temperatures (19742007) and number of days with rainfall (19611990) from the Puyo weather station M008 of the Instituto Nacional de Meteorologa en H idr olgica (INAHMI), Pastaza Province, Ecuador (S 130'27", W 7756'38, 960 m). The station is 3045 km northwest of the experimental plots. Month Air temperature (C) Precipitation (mm) Max Min Average Days with at least 1 mm rainfall Max Min Avera ge January 22.3 19.8 21.2 21.1 476.3 88.2 306.7 February 22.4 19.1 21.0 20.0 570.8 88.6 322.2 March 22.8 20.2 21.2 23.0 584.5 71.5 396.8 April 22.2 20.5 21.2 23.0 761.7 254.1 493.8 May 22.1 20.3 21.0 24.0 782.3 227.0 459.3 June 21.3 19.4 20.4 24.0 83 5.7 275.8 472.4 July 20.8 14.0 19.8 22.0 634.8 210.4 367.4 August 24.9 19.4 20.6 21.0 436.8 114.8 309.1 September 29.9 20.0 21.2 21.0 489.4 115.0 349.4 October 31.8 20.4 21.7 22.0 607.5 171.9 394.5 November 22.0 20.6 21.4 22.0 693.5 271.3 392.6 Decem ber 22.8 20.4 21.2 22.0 599.7 87.6 351.5 Average monthly 23.8 3.5 19.5 1.8 20.1 1.0 22.1 1.2 622.8 125.8 164.7 79.2 384.7 63.1 Total 265.1 4615.8

PAGE 68

68 Table A 2. Common names, seedling types, cotyledon characteristics, seed sizes and seedling recruitment of six large -seeded tropical forest species from the Ecuadorian Amazon used in the two studies. Superscripts of 1 are species used in the 2007 study (Chapter 2), and 2 in the 2008 study (Chapter 3). cotyledon enclosure by seed coat after germination Cotyledon position cotyledon morphology phanerocotylar (free of seed coat) or cryptocotylar (enveloped by seed coat) epigeal or hypogeal foliaceous or reserve planted 3/9/08 seedling 5/28/08 % recruited Length (cm) Width (cm) Caryodendron orinicense Euphorbiaceae man del arblC nampi B PEF phanerocotylar epigeal foliaceous 30 29 97 2.30.7 Gustavia macarenensis Lecythidaceae sachi avocate inik A PHR phanerocotylar hypogeal reserve 30 26 87 19.44.9 Inga densiflora Fabaceae guaba machetona sampi A PHR phanerocotylar hypogeal reserve 30 30 100 4.51.0 2.20.4 7.21.4 Plinia sp. Myrtaceae shawi shawiD CHR cryptocotylar hypogeal reserve 30 29 97 14.24.0 Pouteria caimito Sapotaceae caimito yas B CHR cryptocotylar hypogeal reserve 2.91.6 1.50.2 Quararibea cordata Malvaceae sapote saput B CHR cryptocotylar hypogeal reserve (huastorial) 3.90.7 2.20.3 Species Family Spanish name (in Ecuador) Shuar name Cotyledon functional type Seedling type (Garwood, 1996) Seed size (n=20) Seedling recruitment in planting bed (n=30) Seed fresh wt (g) (n=25) A Van den Eynden, V., E. C ueva, et al. (2003) B Bennett, B. C., M. A. Baker, et al. (2002) C Jimenez, L. C. & H. Y. Bernal (1989) D Local name near Palora, Ecuador. Species was undescribed in Ecuador at time of the study (2008).

PAGE 69

69 Table A 3. Number of seeds that were planted mis sing, dead as seeds and seedlings, and alive as seeds and seedlings at 2 and 14 weeks after planting of three large -seeded tropical forest trees that were planted as seeds (buried and surface-planted) in three field types (grazed pasture, mature pasture an d forest) nested in three indigenous communities (Kenkuimi, Kunkuki, San Ramon) in the Ecuadorian Amazon. Community Field Treatment Planted Missing at 2 wks Dead seed at 2 wks Dead seedling at 2 wks Live ungermin ated seed at 2 wks Live germinat ed seeds at 2 wks Live seedling at 2 wks Missing between wks 2 and 14 Died between wks 2 and 14 (seed or seedling) Live but not seedling at 14 wks Live seedling at 14 wks Kenkuimi Forest Surface 17 4 0 0 1 0 12 1 7 0 5 Buried 19 6 0 0 1 0 12 1 6 0 6 36 10 0 0 2 0 24 2 13 0 11 Grazed Surface 17 0 1 5 0 0 11 0 7 0 4 Buried 12 1 0 0 1 2 8 0 6 2 3 29 1 1 5 1 2 19 0 13 2 7 Mature pasture Surface 11 1 0 1 0 0 9 0 5 0 4 Buried 17 1 0 0 1 4 11 1 9 1 5 28 2 0 1 1 4 20 1 14 1 9 93 13 1 6 4 6 63 3 40 3 27 Kunkuki Forest Surface 20 1 0 0 4 2 13 0 3 0 16 Buried 17 0 0 0 3 4 10 0 5 0 12 37 1 0 0 7 6 23 0 8 0 28 Mature pasture Surface 18 0 0 0 2 1 15 2 4 0 12 Buried 15 1 0 0 0 3 11 1 4 0 9 33 1 0 0 2 4 26 3 8 0 21 Mature pasture Surface 16 0 0 0 0 1 15 0 2 0 14 Buried 13 0 0 0 0 0 13 0 0 0 13 29 0 0 0 0 1 28 0 2 0 27 99 2 0 0 9 11 77 3 18 0 76 SanRamon Forest Surface 20 0 0 0 2 0 18 1 4 0 15 Buried 16 0 0 0 0 1 15 1 8 0 7 36 0 0 0 2 1 33 2 12 0 22 Grazed Surface 20 0 0 0 0 2 18 1 9 0 10 Buried 21 0 1 0 0 2 18 2 2 0 16 41 0 1 0 0 4 36 3 11 0 26 Mature pasture Surface 20 0 0 0 3 1 16 3 5 0 12 Buried 18 2 0 1 1 0 14 1 6 0 8 38 2 0 1 4 1 30 4 11 0 20 115 2 1 1 6 6 99 9 34 0 68 159 6 1 6 12 7 127 8 46 0 92 148 11 1 1 7 16 112 7 46 3 79 307 17 2 7 19 23 239 15 92 3 171 Inga densiflora SanRamon Total Surface Total Grazed Total Forest Total Kenkuimi Total Kunkuki Total Mature pasture Total Mature pasture Total Buried Total Forest Total Grazed Total Mature pasture Total Forest Total Grazed Total Grand Total

PAGE 70

70 Table A 3 Continued. Community Field Treatment Planted Missing at 2 wks Dead seed at 2 wks Dead seedling at 2 wks Live ungermin ated seed at 2 wks Live germinat ed seeds at 2 wks Live seedling at 2 wks Missing between wks 2 and 14 Died between wks 2 and 14 (seed or seedling) Live but not seedling at 14 wks Live seedling at 14 wks Kenkuimi Forest Surface 11 1 1 0 7 2 0 6 2 0 1 Buried 9 1 0 0 5 3 0 1 5 1 1 20 2 1 0 12 5 0 7 7 1 2 Grazed Surface 9 1 0 0 5 1 2 5 3 0 0 Buried 9 0 0 0 6 0 3 1 7 1 0 18 1 0 0 11 1 5 6 10 1 0 Mature pasture Surface 9 0 2 0 4 2 1 5 1 0 1 Buried 9 0 1 0 8 0 0 1 6 1 0 18 0 3 0 12 2 1 6 7 1 1 56 3 4 0 35 8 6 19 24 3 3 Kunkuki Forest Surface 12 0 0 0 9 2 1 2 9 0 1 Buried 10 0 1 0 8 0 1 0 4 3 2 22 0 1 0 17 2 2 2 13 3 3 Mature pasture Surface 9 0 0 0 6 0 3 2 4 1 2 Buried 8 0 0 0 7 0 1 5 2 0 1 17 0 0 0 13 0 4 7 6 1 3 Mature pasture Surface 11 0 0 0 8 1 2 3 7 0 1 Buried 9 0 0 0 7 2 0 3 3 0 3 20 0 0 0 15 3 2 6 10 0 4 59 0 1 0 45 5 8 15 29 4 10 SanRamon Forest Surface 17 0 0 0 13 0 4 6 9 0 2 Buried 9 0 1 0 7 0 1 1 6 0 1 26 0 1 0 20 0 5 7 15 0 3 Grazed Surface 13 1 0 0 9 1 2 1 9 1 1 Buried 13 2 1 0 6 0 4 3 7 0 0 26 3 1 0 15 1 6 4 16 1 1 Mature pasture Surface 12 0 0 0 11 0 1 4 6 2 0 Buried 8 1 0 0 6 0 1 0 3 3 1 20 1 0 0 17 0 2 4 9 5 1 72 4 2 0 52 1 13 15 40 6 5 103 3 3 0 72 9 16 34 50 4 9 84 4 4 0 60 5 11 15 43 9 9 187 7 7 0 132 14 27 49 93 13 18 Quararibea cordata Grazed Total Mature pasture Total Kenkuimi Total Forest Total Mature pasture Total Grazed Total Forest Total Grazed Total Mature pasture Total Kunkuki Total Forest Total SanRamon Total Surface Total Buried Total Grand Total

PAGE 71

71 Table A 3 Continued. Community Field Treatment Planted Missing at 2 wks Dead seed at 2 wks Dead seedling at 2 wks Live ungermin ated seed at 2 wks Live germinat ed seeds at 2 wks Live seedling at 2 wks Missing between wks 2 and 14 Died between wks 2 and 14 (seed or seedling) Live but not seedling at 14 wks Live seedling at 14 wks Kenkuimi Forest Surface 17 7 0 0 10 0 0 5 2 3 0 Buried 12 1 0 0 10 1 0 5 6 0 0 29 8 0 0 20 1 0 10 8 3 0 Grazed Surface 17 1 0 0 16 0 0 12 3 1 0 Buried 12 0 0 0 12 0 0 3 8 1 0 29 1 0 0 28 0 0 15 11 2 0 Mature pasture Surface 17 1 0 0 15 0 1 8 4 3 1 Buried 15 5 0 0 10 0 0 1 6 3 0 32 6 0 0 25 0 1 9 10 6 1 90 15 0 0 73 1 1 34 29 11 1 Kunkuki Forest Surface 17 5 0 0 12 0 0 3 7 2 0 Buried 14 3 1 0 10 0 0 2 6 2 0 31 8 1 0 22 0 0 5 13 4 0 Mature pasture Surface 17 0 0 0 17 0 0 12 4 0 1 Buried 14 0 0 0 14 0 0 4 9 0 1 31 0 0 0 31 0 0 16 13 0 2 Mature pasture Surface 17 0 0 0 16 1 0 6 11 0 0 Buried 9 0 0 0 9 0 0 5 3 1 0 26 0 0 0 25 1 0 11 14 1 0 88 8 1 0 78 1 0 32 40 5 2 SanRamon Forest Surface 8 0 0 0 8 0 0 2 6 0 0 Buried 12 0 0 0 12 0 0 3 9 0 0 20 0 0 0 20 0 0 5 15 0 0 Grazed Surface 13 0 0 0 13 0 0 4 9 0 0 Buried 13 0 0 0 13 0 0 6 6 0 1 26 0 0 0 26 0 0 10 15 0 1 Mature pasture Surface 10 0 0 0 10 0 0 5 5 0 0 Buried 8 2 0 0 6 0 0 1 3 1 1 18 2 0 0 16 0 0 6 8 1 1 64 2 0 0 62 0 0 21 38 1 2 133 14 0 0 117 1 1 57 51 9 2 109 11 1 0 96 1 0 30 56 8 3 242 25 1 0 213 2 1 87 107 17 5 Pouteria caimito Forest Total Grazed Total Mature pasture Total Kenkuimi Total Forest Total Grazed Total Mature pasture Total Kunkuki Total Forest Total Buried Total Grand Total Grazed Total Mature pasture Total SanRamon Total Surface Total

PAGE 72

72 Table A 4 Vegetation characteristics of three secondary forest plots in the Ecuadorian Amazon from four 50 x 2 m modified Gentry transects (Phillips et al. 2002) established in each of three plots. The average dbh and heights ( 1 S.D.) of woody stems of two size classes (2.5 10 cm, >10 cm), number of dead trees, vines and palms are presented for each plot Individuals between 1 2.5 cm dbh were surveyed in one of the 50 x 2 m transects per plot. Heights were estimated in the field. # stems / 400 m2 # stems / 400 m2 Hartensia 106 4.5 1.8 4.2 1.9 17 20 6 14 4.7 1 2 11 3 Cascada 111 4.6 1.8 5.1 2.5 19 17 7.2 12 4.1 12 0 28 3 Centro Semillas 93 4.4 1.7 3.7 1.4 35 20 7.5 12 7.3 1 1 17 3 Average 103 4.5 1.8 4.4 2.1 71 19 7.1 12 6 4.7 1 18.7 3 dbh (cm) height (m) Plot name woody stems 2.5 cm 10 cm dbh woody stems >10 cm dbh # stems 1 and < 2.5 cm dbh / 100 m2 # palms 2.5 cm dbh / 400 m2 # dead stems cm dbh / 400 m2 # vines 2.5 cm dbh / 400 m2 dbh (cm) height (m)

PAGE 73

73 Table A 5 Numbers o f w oody species with 2.5 cm dbh encount ered in four 50 x 2 m transects (400 m2) established in each of three secondary forest plots in Pastaza Province in the Ecuadorian Amazon. Identifications were made at the Herbario Nacional del Ecuador in Quito (QCNE). Unidentified spe cies were classified to morphospecies. Family Genus Species Plot name Total Hartensia Cascada Centro Semillas Actinidiaceae Saurauia sp #1 1 6 1 8 Anacardiaceae Tapirira guianensis 3 2 1 6 Annonaceae Annona duckei 1 1 edulis? 1 1 Porc elia mediocris 1 1 Rollinia dolichopetala 1 1 mucosa 3 3 pittieri 2 2 Araliaceae Schefflera morototonia 2 1 6 9 Arecaceae Ireatea deltoides 1 1 Socratea exorrhiza 1 1 Wettinia sp #1 1 1 Asteraceae Mikania sp #1 1 1 Piptocoma discolor 6 14 18 38 Bignoniaceae Jacaranda copaia 2 2 Boraginaceae Cordia nodosa 1 1 ucayaliensis 1 1 Tournefortia sp #1 3 1 3 7 Caryocaraceae Caryocar glabrum 1 1 Cecropiaceae Cecropria sp #1 3 3 sp #2 1 1 s p #3 1 1 sciadophylla 2 2 Pourouma minor 1 1 Chloranthaceae Hedyosmum racemosum 9 9 Clusiaceae Chrysochlamys membranacea 1 1 Tovomita sp #1 1 1 Vismia baccifera 1 1 sp #1 1 2 3 sp #2 2 1 3 Erythroxylaceae Erythroxy lum macrophyllum 1 1

PAGE 74

74 Table A 5 Continued. Family Genus Species Plot name Total Hartensia Cascada Centro Semillas Euphorbiaceae Caryodendron orinocense 5 5 Croton matourensis 1 1 Hyeronima alchorneoides 1 1 sp #1 1 1 oblong a 1 1 Mabea standleyi 1 1 Margaritaria nobilis 1 1 Sapium sp #1 1 1 Tetrorchidium sp #1 2 2 Fabaceae Calliandra trinervia 19 1 20 Dalbergia frutescens 3 3 Dussia tessmannii 1 1 Inga sp #1 2 2 sp #2 2 2 sp #3 1 1 sp #4 1 1 sp #5 1 1 sp #6 1 1 sp #7 1 1 sp #8 1 1 nobilis 1 1 2 vismifolia 1 1 Senna bacillaris 3 5 8 Flacourtiaceae Banara sp #1 1 2 3 guianensis 1 1 guianensis? 1 1 nitida 1 1 Casearia sp #1 2 1 5 8 sp #2 1 1 sp #3 1 1 sylvestris 1 2 3 Hasseltia floribunda 1 6 7 Neosprucea sp #1 2 3 5 Tetrathylacium macrophyllum 1 1 Lauraceae Ocotea sp #1 1 1 Genus #1 sp #1 1 1 Lecythidaceae Eschweilera sp #1

PAGE 75

75 Table A 5 Continued. Family Genus Species Plot name Total Hartensia Cascada Centro Semillas Melastomataceae Clidemia sp #1 1 1 Miconia cercophora 2 2 sp #1 1 1 2 sp #2 1 2 3 sp #3 1 4 5 sp #4 2 2 sp #5 1 2 3 sp #6 1 1 sp #7 1 1 sp #8 1 1 sp #9 1 1 sp #10 1 1 punctata? 1 1 Ossaea sp #1 1 1 sp #2 1 1 sp #3 7 7 Meliaceae Cabralea canjerana 1 1 Guarea kunthiana 1 1 Mimosaceae Abarema jupunba 1 1 Monimiaceae Siparuna decipiens 2 2 schimpffi 2 2 schimpffii 2 2 sp #1 4 4 Moraceae Brosimum sp #1 2 2 Ficus c.f. tonduzii 1 1 sp #1 1 1 sp #2 1 1 Perebea guianensis 1 1 2 Pseudolmedia rigida 1 1 Trymatococcus amazonicus 2 2 Myristicaceae Virola sebifera 2 2 4 sp #1 1 1 Myrsinaceae Parathesis sp #1 1 1 Myrtaceae Myrcia sp #1 1 1 sp #2 1 1 Genus #1 sp #1 1 1 Nyctaginaceae Neaa divaricata 1 1 sp #1 1 1 sp #2 2 2 sp #3 2 2 Piperaceae Piper obli quum 5 7 2 14 sp #1 2 2

PAGE 76

76 Table A 5 Continued. Family Genus Species Plot name Total Hartensia Cascada Centro Semillas Rubiaceae Agouticarpa sp #1 2 2 Chimarrhis glabriflora or hookeri 2 2 Palicourea sp #1 3 1 4 subalatoides 1 1 4 6 Pentagonia amazonica 2 2 1 5 Psychotria sp #1 1 1 Genus #1 sp #1 1 1 Rutaceae Zanthoxylum sp #1 1 1 sp #2 1 1 Sapindaceae Allophylus sp #1 1 1 sp #2 1 1 Solanaceae Cestrum megalophyllum 1 2 3 silvaticum 3 3 Solanum grandiflorum 1 1 sp #1 1 1 sp #2 1 1 Genus #1 sp #1 1 1 2 Sterculiaceae Herrania sp #1 1 1 Verbenaceae Vitex sp #1 Total Families 24 22 26 34 Total Genera 33 35 34 69 Total spe cies 40 84 70 129

PAGE 77

77 Table A 6 Numbers of seeds of four tropical tree specie s that were planted, died as seeds were alive (sum of live seeds and seedlings), and were live seedlings at 18 weeks under two treatments (surface-planted and buried) in thr ee secondary forest plots in the Ecuadorian Amazon. The f ive seeds disappeared from planting sites (4 surface-planted Caryodendron seeds and one surface-planted Gustavia seed), and are included in the total Planted Plot name Planted Died as seed Live Live seedling Planted Died as seed Live Live seedling Planted Died as seed Live Live seedling Planted Died as seed Live Live seedling Surface, uncaged16 3 10 10 16 1 14 10 18 1 15 12 20 1 19 6Surface, caged21 2 18 18 17 0 17 13 18 2 16 15 21 4 17 4 37 5 28 28 33 1 31 23 36 3 31 27 41 5 36 10 18 5 13 12 14 1 12 7 13 1 12 9 19 5 14 3 19 1 18 15 17 0 17 6 15 1 14 14 15 4 11 3 37 6 31 27 31 1 29 13 28 2 26 23 34 9 25 6Hartensia total74 11 59 55 64 2 60 36 64 5 57 50 75 14 61 16Cascada Surface, uncaged18 0 16 15 22 1 21 16 22 1 20 19 24 4 20 8Surface, caged24 3 20 20 21 1 20 20 18 0 17 16 21 6 15 1 42 3 36 35 43 2 41 36 40 1 37 35 45 10 35 9Buried, uncaged15 0 15 14 151 14 7173 14 9169 7 1 Buried, caged19 3 14 14 183 15 7162 14 9136 7 0 34 3 29 28 33 4 29 14 33 5 28 18 29 15 14 1Cascada total76 6 65 63 76 6 70 50 73 6 65 53 74 25 49 10 Centro Semillas Surface, uncaged20 6 11 11 16 0 16 14 16 0 16 16 18 0 18 4 Surface, caged21 2 17 17 13 0 13 9 15 0 14 13 17 7 10 2 41 8 28 28 29 0 29 23 31 0 30 29 35 7 28 6Buried, uncaged20 5 11 15 14 1 13 7 10 1 9 7 13 4 9 3Buried, caged17 4 17 13 15 1 14 12 11 3 8 5 19 9 10 1 37 9 28 28 29 2 27 19 21 4 17 12 32 13 19 4Centro Semillas Total78 17 56 56 58 2 56 42 52 4 47 41 67 20 47 10 Uncaged total 107 19 76 77 97 5 90 61 96 7 86 72 110 23 87 25 Caged total 121 15 104 97 101 5 96 67 93 8 83 72 106 36 70 11 Surface total 120 16 92 91 105 3 101 82 107 4 98 91 121 22 99 25 Buried total 108 18 88 83 93 7 85 46 82 11 71 53 95 37 58 11 Total 228 34 180 174 198 10 186 128 189 15 169 144 216 59 157 36 Buried Total Surface Total Buried Total Treatment Buried, uncaged Buried, caged Surface Total Buried Total Surface Total Caryodendron orinocense Gustavia macarenensis Inga densiflora Plinia sp. Hartensia

PAGE 78

78 Table A 7 Comparison between the two seed planting experiments presented in Chapter 2 (2007) and Chapter 3 (2008) of this thesis, in which species of large -seeded tropical forest trees were planted in a variety of human modified landscapes in the Ecuadorian Amazon. 2007 2008 Species In ga densiflora Quararibea cordata, Pouteria caimito Inga densiflora Gustavia macarenensis Caryodendron orinocense Plinia sp. Field type grazed pasture (2 6 wks regrowth) mature pasture (3 6 mo regrowth) selectively logged forest 10 15 yr old secondary forest Treatments Surface -planted vs. buried 2x2 factorial, Surface planted vs. buried, caged vs. uncaged Indigenous c ommunity 3 communities, 14 18 km apart (Kenkuimi, Kunkuki, San Ramon) 2 communities, 5 km apart on Puyo -Macas Rd. (Kunkuki, S an Antonio) Number of plots 9 (3 field types x 3 communities) 3 plots Distance between plots 14 18 km between communities, 50 500 m between plots 1 5 km between plots Planting site pre treatment Seed planted without removal of surface debris Surface deb ris cleared Planting dates mid June 2007 late Feb/early March 2008 Data collected Seed disappearance germination, seed/seedling death Seed disappearance, germination seed/seedling death Data collection schedule At 2 & 14 wks (14 wk check by trained ass istant) Biweekly to 18 weeks Duration of experiment 14 weeks 18 weeks

PAGE 79

79 0 100 200 300 400 500 600 700 January February March April May June July August September October November December Month Precipitation (mm) Figure A 1 Average monthly rainfall and precipitation (1974 2007) from the Puyo weather station M008 of the Instituto Nacional de Meteorologa en Hi drolgica (INAHMI), Pastaza Pr ovince, Ecuador (S 130'27", W 7756'38, 960 m). The station is 30 45 km northwest of the experimental plots. Error bars represent 1 S.D.

PAGE 80

80 Figure A 2 Photos of the vegetation in the field plots in three indigenous communities described in Chapter 2 ( Figure s 2 1 & 2 2) Pictured are grazed pasture plots in Kenkuimi (A), San Ramon (B) and Kunkuki (C), mature pasture plots in Kunkuki (D) and Kenkuimi (E), a pasture abutting forest in Kunkuimi (F), and forest plots in Kunkuimi (G) and Kunkuki (H). The grazed pasture plots had 2 6 weeks of vegetation regrowth and the mature pasture plots had 3 6 months of regrowth at the time the experiment was established (June 2007).

PAGE 81

81 0 10 20 30 40 50 60 70 80 90 100 Kenkuimi Kunkuki San Ramon Communities Percent canopy coverage Grazed pasture Mature pasture Forest Figure A 3. Average percent canopy cover of three types of field plots ( grazed pasture, mature pasture, and forest ) in three indigenous communities (Kenkuimi, Kunkuki, San Ramon) in the Ecuadorian Amazon. Percent canopy cover was measured with a handheld spherical densitometer (Robert E. Lemmon Forest Densitometer, Model C ) h eld at ground level at 20 points per plot. Error bars represent 1 S.D. Fields were nested within communities, and letters represent differences between fields (n = 3) within each community in pair -wise Bonferroni adjusted t test comparisons assuming unequal variances on untransformed data. a b b b b a c b a

PAGE 82

82 Figure A 4. Photos of the Myrtaceae species planted in the experiment described in Chapter 3. Pictures are photos of the tree trunk (A), branch with fruit (B), leaves and mature fruits (C), longitudinal cut through fruit (D), fresh seeds (E), and consumption of fresh fruits (F). The species was identified (from photos and description) to be in the genus Plin i a by Bruce K. Holst, Collections Manager of the Marie Selby Botanical Gardens in Sarasota, Florida Herbari ums specimens were deposited in the Herbario Nacional del Ecuador in Quito.

PAGE 83

83 Figure A 5 Average percent canopy cover of three secondary forests measured with a handheld spherical densitometer (Robert E. Lemmon Forest Densiometer, Model C) held at 1 m above ground level at 20 points per plot. Error bars represent one S.D. Percent canopy cover differed among the three plots (p = 0.0011, ANOVA on arcsine transformed data), and le tters in graph show results from Tukeys test. 0 10 20 30 40 50 60 70 80 90 100 Hartensia Cascada Centro Semillas Plot name Percent canopy cover a b ab

PAGE 84

84 LIST OF REFERENCES Achard, F., H.D. Eva, H.J. Stibig, P. Mayaux, J. Gallego, T. Richards, and J.P. Malingreau. 2002. Determination of deforestation rates of the world's humid tropical forests. Science 297(5583):9991002. Aide, T.M., a nd H.R. Grau. 2004. Ecology, globalization, migration, and Latin American ecosystems. Science 305(5692):19151916. Akinnifesi, F.K., F.R. Kwesiga, J. Mhango, A. Mkonda, T. Chilanga, and R. Swai. 2004. Domesticating priority miombo indigenous fruit trees as a promising livelihood option for small -holder farmers in Southern Africa. p 15 30 in XXVI International Horticultural Congress: Citrus and Other Subtropical and Tropical Fruit Crops: Issues, Advances and Opportunities, Albrigo, L.G., and V. Galn Saco (eds.). ISHS, Toronto, Canada Alvarez -Clare, S., and K. Kitajima. 2009. Susceptibility of tree seedlings to biotic and abiotic hazards in the understory of a moist tropical forest in Panama. Biotropica 41(1):47 56. Andrade, R.A., and A.B.G. Martins. 2003. Influence of the temperature in germination of seeds of jabuticaba tree. Revista Brasileira de Fruticultura 25(1):197198. Armstrong, D.P., and M. Westoby. 1993. Seedlings from large seeds tolerate defoliation better: a test using phylogenetically independent contrasts. Ecology 74(4):10921100. Asquith, N.M., S.J. Wright, and M.J. Clauss. 1997. Does mammal community composition control recruitment in neotropical forests? Evidence from Panama. Ecology 78(3):941 946. Augspurger, C.K. 1984a. Light requirement s of neotropical tree seedlings: a comparative study of growth and survival. Journal of Ecology 72(3):777795. Augspurger, C.K. 1984b. Seedling survival of tropical tree species: interactions of dispersal distance, light -gaps, and pathogens. Ecology 65(6): 17051712. Augspurger, C.K., and C.K. Kelly. 1984. Pathogen mortality of tropical tree seedlings: experimental studies of the effects of dispersal distance, seedling density, and light conditions. Oecologia 61(2):211217. Beck, H. 2006. A review of peccary-palm interactions and their ecological ramifications across the Neotropics. Journal of Tropical Forest Science 87(3):519530. Beck, H., and J. Terborgh. 2002. Groves versus isolates: how spatial aggregation of Astrocaryum murumuru palms affects seed remov al. Journal of Tropical Ecology 18:275288. Benayas, J.M.R., J.M. Bullock, and A.C. Newton. 2008. Creating woodland islets to reconcile ecological restoration, conservation, and agricultural land use. Frontiers in Ecology and the Environment 6(6):329336.

PAGE 85

85 Benitez-Malvido, J., G. Garca Guzmn, and I.D. Kossmann -Ferraz. 1999. Leaf -fungal incidence and herbivory on tree seedlings in tropical rainforest fragments: an experimental study. Biological Conservation 91(23):143 150. Benitez-Malvido, J., and I.D. Kos smann Ferraz. 1999. Litter cover variability affects seedling performance and herbivory. Biotropica 31(4):598606. Benitez-Malvido, J., M.M. Martinez Ramos, J.L.C. Camargo, and I.D.K. Ferraz. 2005. Responses of seedling transplants to environmental variati ons in contrasting habitats of central Amazonia. Journal of Tropical Ecology 21:397406. Bennett, B.C., M.A. Baker, and P.G. Andrade. 2002. Ethnobotany of the Shuar of eastern Ecuador. New York Botanical Garden Press. 299 p. Bewley, J.D., and M. Black. 199 4. Seeds: physiology of development and germination. Plenum Press, New York and London. 445 p. Blate, G.M., D.R. Peart, and M. Leighton. 1998. Post -dispersal predation on isolated seeds: a comparative study of 40 tree species in a southeast Asian rainfores t. Oikos 82(3):522538. Bowen, M.E., C.A. McAlpine, A.P.N. House, and G.C. Smith. 2007. Regrowth forests on abandoned agricultural land: A review of their habitat values for recovering forest fauna. Biological Conservation 140(34):273296. Brewer, S.W., a nd M. Rejmanek. 1999. Small rodents as significant dispersers of tree seeds in a Neotropical forest. Journal of Vegetation Science 10(2):165174. Byg, A., and H. Balslev. 2006. Palms in indigenous and settler communities in southeastern Ecuador: farmers p erceptions and cultivation practices. Agroforestry Systems 67(2): 147158. Camargo, J.L.C., I.D.K. Ferraz, and A.M. Imakawa. 2002. Rehabilitation of degraded areas of central Amazonia using direct sowing of forest tree seeds. Restoration Ecology 10(4):636644. Caadas, L.C., and W.A. Estrada. 1978. Mapa ecolgica. 1:1,000,000., Programa Nacional de Regionalizaci on Agraria PRONAREG, Quito. OSTROM, France. Carlos, A.P. 2001. Synergistic effects of subsistence hunting and habitat fragmentation on Amazonian fo rest vertebrates. Biological Conservation 15(6):14901505. Chapman, C.A., and D.A. Onderdonk. 1998. Forests without primates: primate/plant codependency. American Journal of Primatology 45(1):127141. Chazdon, R.L. 2003. Tropical forest recovery: legacies of human impact and natural disturbances. Perspectives in Plant Ecology Evolution and Systematics 6(1 -2):5171.

PAGE 86

86 Chazdon, R.L. 2008. Beyond deforestation: restoring forests and ecosystem services on degraded lands. Science 320(5882):14581460. Chazdon, R.L. C.A. Harvey, O. Komar, D.M. Griffith, B.G. Ferguson, M. Martinez -Ramos, H. Morales, R. Nigh, L. Soto -Pinto, M. van Breugel, and S.M. Philpott. 2009. Beyond reserves: a research agenda for conserving biodiversity in human -modified tropical landscapes. Biotropica 41(2):142153. Chiarello, A.G. 1999. Effects of fragmentation of the Atlantic forest on mammal communities in south -eastern Brazil. Biological Conservation 89(1):7182. Chomit z, K.M. 2007. At loggerheads?: a gricultural expansion, poverty reduction, and environment in the tropical forests. World Bank Publications, Washington, D.C. 284 p. Clement, C.R. 1999a. 1492 and the loss of Amazonian crop genetic resources. I. The relation between domestication and human population decline. Economic Botany 53(2) :188202. Clement, C.R. 1999b. 1492 and the loss of Amazonian crop genetic resources. II. Crop biogeography at contact. Economic Botany 53(2):203216. Clement, C.R., J.P. Cornelius, M.H. Piedo -Panduro, and K. Yuyama. 2008. Native fruit tree improvement in Amazonia: an overview. p 100120 in Indigenous fruit trees in the tropics: Domestication, utilization and commercialization, Akinnifesi, F.K., R.B. Leakey, O. Ajayi, G. Sileshi, Z. Tchoundjeu, P. Matakala, and F.R. Kwesiga (eds.). Oxford University Press Oxford. Cole, R.C. 2009. Post -dispersal seed fate of tropical montane trees in an agricultural landscape, southern Costa Rica. Biotropica 41(3):319327. Crawley, M. 2000. Seed predators and plant populations dynamics. p 167182 in Seeds: The ecology and regeneration in plant communities, Fenner, M. (ed.). CABI Publishing, Wallingford, UK. Cubina, A., and T.M. Aide. 2001. The effect of distance from forest edge on seed rain and soil seed bank in a tropical pasture. Biotropica 33(2):260267. Custode, E. 1 983. Mapa morfo-edafologico. Provincia de Morona Santiago. (Zona Norte). 1:500,000. Ministerio de Agricultura y Ganaderia. Programa Nacional de Regionalizacion Agraria PRONAREG, Q. (ed.). ORSTOM, France. Da Silva, J.M.C., A.B. Rylands, and G.A.B. Da Fon seca. 2005. The fate of the Amazonian areas of endemism. Conservation Biology 19(3):689694. Dalling, J.W., and K.E. Harms. 1999. Damage tolerance and cotyledonary resource use in the tropical tree Gustavia superba. Oikos 85(2):257264. Dalling, J.W., and S.P. Hubbell. 2002. Seed size, growth rate and gap microsite conditions as determinants of recruitment success for pioneer species. Journal of Ecology 90(3):557 568.

PAGE 87

87 Dalling, W.D., and R. John. 2008. Seed limitation and the coexistence of pioneer tree spec ies. p 242253 in Tropical forest community ecology, Carson, W.P., and S.A. Schnitzer (eds.). Wiley -Blackwell, Oxford. Danner, M.A., I. Citadin, A.F. Junior, A.P. Assmann, S.M. Mazaro, and S.A. Zolet -Sasso. 2007. Formao de mudas de jabuticabeira ( Plinia sp.) em diferentes substratos e tamanhos de recipientes. Revista Brasileira de Fruticultura 29(1):179182. Dantonio, C.M., and P.M. Vitousek. 1992. Biological invasions by exotic grasses, the grass fire cycle, and global change. Annual Review of Ecology a nd Systematics 23:6387. Daws, M.I., N.C. Garwood, and H.W. Pritchard. 2005. Traits of recalcitrant seeds in a semi deciduous tropical forest in Panama: some ecological implications. Functional Ecology 19(5):874885. Daws, M.I., N.C. Garwood, and H.W. Prit chard. 2006. Prediction of desiccation sensitivity in seeds of woody species: a probabilistic model based on two seed traits and 104 species. Annals of Botany 97(4):667674. Dawson, I.K., A. Lengkeek, J.C. Weber, and R. Jamnadass. 2009. Managing genetic va riation in tropical trees: linking knowledge with action in agroforestry ecosystems for improved conservation and enhanced livelihoods. Biodiversity and Conservation 18(4):969986. DeMattia, E.A., B.J. Rathcke, L.M. Curran, R. Aguilar, and O. Vargas. 2006. Effects of small rodent and large mammal exclusion on seedling recruitment in Costa Rica. Biotropica 38(2):196202. Dirzo, R., E. Mendoza, and P. Ortiz. 2007. Size related differential seed predation in a heavily defaunated neotropical rain forest. Biotropica 39(3):355362. Dirzo, R., and P.H. Raven. 2003. Global state of biodiversity and loss. Annual Review of Environment and Resources 28(1):137167. Doust, S.J., P.D. Erskine, and D. Lamb. 2006. Direct seeding to restore rainforest species: microsite eff ects on the early establishment and growth of rainforest tree seedlings on degraded land in the wet tropics of Australia. Forest Ecology and Management 234(13):333343. Duke, S.H., G. Kakefuda, and T.M. Harvey. 1983. Differential leakage of intracellular substances from imbibing soybean seeds. Plant Physiology 72(4):919924. Dupuy, J.M., and R.L. Chazdon. 2008. Interacting effects of canopy gap, understory vegetation and leaf litter on tree seedling recruitment and composition in tropical secondary forests Forest Ecology and Management 255(11):37163725. Eisenber g, J. F., & Redford, K. H. (1999). Mammals of the Neotropics (Volume 3 ): the central Neotropics: Ecuador, Peru, Bolivia, Brazil. University of Chicago Press. Chicago. 599 p.

PAGE 88

88 F.A.O. 2005. Global fo rest resources assessment 2005: progress towards sustainable forest management. p 320, United Nations Rome. Fenner, M., and K. Thompson. 2005. Soil seed banks. p 76 96 in The ecology of seeds, Fenner, M., and K. Thompson (eds.). Cambridge University Pre ss, New York. Finegan, B. 1996. Pattern and process in neotropical secondary rain forests: the first 100 years of succession. Trends in Ecology & Evolution 11(3):119124. Forget, P.M. 1990. Seed dispersal of Vouacapoua americana (Caesalpiniaceae) by caviom orph rodents in French Guiana. Journal of Tropical Ecology 6:459468. Fox, T.R. 2000. Sustained productivity in intensively managed forest plantations. Forest Ecology and Management 138(1 3):187202. Fragoso, J.M.V., and J.M. Huffman. 2000. Seeddispersal and seedling recruitment patterns by the last Neotropical megafaunal element in Amazonia, the tapir. Journal of Tropical Ecology 16:369385. Gadgil, M., F. Berkes, and C. Folke. 1993. Indigenous knowledge for biodiversity conservation. Ambio 22(2 3):151156. Galetti, M., C.I. Donatti, A.S. Pires, P.R. Guimaraes, and P. Jordano. 2006. Seed survival and dispersal of an endemic Atlantic forest palm: the combined effects of defaunation and forest fragmentation. Botanical Journal of the Linnean Society 151(1):141 149. Galindo -Gonzalez, J., S. Guevara, and V.J. Sosa. 2000. Bat and bird -generated seed rains at isolated trees in pastures in a tropical rainforest. Conservation Biology 14(6):16931703. Garcia Orth, X., and M. Martinez Ramos. 2008. Seed dynamics of ea rly and late successional tree species in tropical abandoned pastures: seed burial as a way of evading predation. Restoration Ecology 16(3):435443. Garwood, N.C. 1989. Tropical soil seed banks: a review. p 149209 in Ecology of soil seed banks, Leck, M.A ., Parker, V.T., and Simpson, R.L. (eds.). Academic Press, New York. Gauriguata, M.R., and R. Ostertag. 2001. Neotropical secondary forest succession: changes in structural and functional characteristics. Forest Ecology and Management 148(1 3):185206. Gua riguata, M.R., J.J.R. Adame, and B. Finegan. 2000. Seed removal and fate in two selectively logged lowland forests with constrasting protection levels. Conservation Biology 14(4):10461054. Gunter, S., M. Weber, R. Erreis, and N. Aguirre. 2007. Influence of distance to forest edges on natural regeneration of abandoned pastures: a case study in the tropical mountain rain forest of southern Ecuador. European Journal of Forest Research 126(1):6775.

PAGE 89

89 Hammond, D.S. 1995. Postdispersal seed and seedling mortality of tropical dry forest trees after shifting agriculture, Chiapas, Mexico. Journal of Tropical Ecology 11:295313. Hansen, M.C., S.V. Stehman, P.V. Potapov, T.R. Loveland, J.R.G. Townshend, R.S. DeFries, K.W. Pittman, B. Arunarwati, F. Stolle, M.K. Steinin ger, M. Carroll, and C. DiMiceli. 2008. Humid tropical forest clearing from 2000 to 2005 quantified by using multitemporal and multiresolution remotely sensed data. Proceedings of the National Academy of Sciences of the United States of America 105(27):9439 9444. Harms, K.E., and J.W. Dalling. 1997. Damage and herbivory tolerance through resprouting as an advantage of large seed size in tropical trees and lianas. Journal of Tropical Ecology 13(4):617621. Holl, K.D. 1999. Factors limiting tropical rain fore st regeneration in abandoned pasture: seed rain, seed germination, microclimate, and soil. Biotropica 31(2):229242. Holl, K.D. 2002. Effect of shrubs on tree seedling establishment in an abandoned tropical pasture. Journal of Ecology 90(1):179187. Holl, K.D., and M.E. Lulow. 1997. Effects of species, habitat, and distance from edge on post dispersal seed predation in a tropical rainforest. Biotropica 29(4):459 468. Hooper, E., R. Condit, and P. Legendre. 2002. Responses of 20 native tree species to refore station strategies for abandoned farmland in Panama. Ecological Applications 12(6):16261641. Hooper, E., P. Legendre, and R. Condit. 2005. Barriers to forest regeneration of deforested and abandoned land in Panama. Journal of Applied Ecology 42(6):11651174. Hopkins, M., and A.W. Graham. 1984. Viable soil seed banks in disturbed lowland tropical rainforest sites in North Queensland. Australian Journal of Ecology 9:7179. Howard, W.E., R.E. Marsh, and R.E. Cole. 1968. Food detection by deer mice using olfac tory rather than visual cues. Animal Behaviour 16(1):1317. Howe, H.F., and E.W. Schupp. 1985. Early consequences of seed dispersal for a Neotropical tree (Virola surinamensis ). Ecology 66(3):781791. I.T.T.O. 2002. Guidelines for the restoration, manageme nt and rehabilitation of degraded and secondary tropical forests. International Tropical Timber Organization Policy Development Series 13:86. Jansen, P.A., F. Bongers, and L. Hemerik. 2004. Seed mass and mast seeding enhance dispersal by a neotropical scat ter -hoarding rodent. Ecological Monographs 74(4):569-589.

PAGE 90

90 Jansen, P.A., and P.M. Forget. 2001. Scatter hoarding rodents and tree regeneration. p 275 288 in Dynamics and plant animal interactions in a neotropical rainforest, Bongers, F., Charles Dominique P., Forget, P and Thry, M. (eds ). Kluwer Academic Publishers, Dordrecht, Netherlands. Janzen, D.H. 1969. Seed eaters versus seed size, number, toxicity and dispersal. Evolution 23(1):127. Jerozolimski, A., and C.A. Peres. 2003. Bringing home the biggest bacon: a cross -site analysis of the structure of hunter kill profiles in Neotropical forests. Biological Conservation 111(3):415425. Jimenez, L.C., and H.Y. Bernal. 1989. El inichi Caryodendron orinocense Karsten (Euphorbiaceae). Talleres de Editora G uadalupe Ltda., Bogota. Jones, A.S., B.B. Lamont, M.M. Fairbanks, and C.M. Rafferty. 2003a. Kangaroos avoid eating seedlings with or near others with volatile essential oils. Journal of Chemical Ecology 29(12):26212635. Jones, F.A., C.J. Peterson, and B.L Haines. 2003b. Seed predation in neotropical pre -montane pastures: site, distance, and species effects. Biotropica 35(2):219 225. Kelm, D.H., K.R. Wiesner, and O. von Helversen. 2008. Effects of artificial roosts for frugivorous bats on seed dispersal in a Neotropical forest pasture mosaic. Conservation Biology 22(3):733741. Keuroghlian, A., and D.P. Eaton. 2008. Importance of rare habitats and riparian zones in a tropical forest fragment: preferential use by Tayassu pecari, a wide ranging frugivore. Jou rnal of Zoology 275(3):283293. Kitao, M., R. Yoneda, H. Tobita, Y. Matsumoto, Y. Maruyama, A. Arifin, A.M. Azani, and M.N. Muhamad. 2006. Susceptibility to photoinhibition in seedlings of six tropical fruit tree species native to Malaysia following transp lantation to a degraded land. Trees Structure and Function 20(5):601610. Lamb, D., P.D. Erskine, and J.A. Parrotta. 2005. Restoration of degraded tropical forest landscapes. Science 310(5754):16281632. Lambert, T.D., J.R. Malcolm, and B.L. Zimmerman. 200 6. Amazonian small mammal abundances in relation to habitat structure and resource abundance. Journal of Mammalogy 87(4):766776. Lamotte, L. 2005. A note on separated data and exact likelihood ratio p -values in logistic regression. Journal of Statistical Computation and Simulation 75(8):667672. Leishman, M.R., and M. Westoby. 1994. The role of seed size in seedling establishment in dry soil conditions: experimental evidence from semi arid species. Journal of Ecology 82(2):249258.

PAGE 91

91 Lopez, L., and J. Terbor gh. 2007. Seed predation and seedling herbivory as factors in tree recruitment failure on predator -free forested islands. Journal of Tropical Ecology 23(2):129137. Martinez Garza, C., and H.F. Howe. 2003. Restoring tropical diversity: beating the time tax on species loss. Journal of Applied Ecology 40(3):423429. Mather, A.S., and C.L. Needle. 1998. The forest transition: a theoretical basis. Area 30(2):117124. Mayaux, P., P. Holmgren, F. Achard, H. Eva, H. Stibig, and A. Branthomme. 2005. Tropical forest cover change in the 1990s and options for future monitoring. Philosophical Transactions of the Royal Society 360:373384. McNeely, J.A., and S.J. Scherr. 2003. Ecoagriculture: strategies to feed the world and save wild biodiversity. Island Press, Washingt on D.C. and London. 323 p. Meehan, H.J., K.R. McConkey, and D.R. Drake. 2002. Potential disruptions to seed dispersal mutualisms in Tonga, Western Polynesia. Journal of Biogeography 29(5 6):695712. Mendoza, E., and R. Dirzo. 2007. Seed-size variation dete rmines interspecific differential predation by mammals in a neotropical rain forest. Oikos 116(11):18411852. Milln, H., A. Kalauzi, G. Llerena, J. Sucoshaay, and D. Piedra. 2008. Climatic trends in the Amazonian area of Ecuador: classical and multi -frac tal analyses. Atmospheric Research 88(34):355366. Moles, A.T., and M. Westoby. 2002. Seed addition experiments are more likely to increase recruitment in larger -seeded species. Oikos 99(2):241248. Moles, A.T., and M. Westoby. 2004. Seedling survival and seed size: a synthesis of the literature. Journal of Ecology 92(3):372383. Moran, E.F., E.S. Brondizio, J.M. Tucker, M.C. da Silva -Forsberg, S. McCracken, and I. Falesi. 2000. Effects of soil fertility and landuse on forest succession in Amaznia. Fores t Ecology and Management 139(1 3):93108. Morton, J. 1987a. Abiu ( Pouteria caimito ). p 406 408 in Fruits of warm climates, Dowling, C.F.J. (ed.). J.F. Morton, Miami, FL. Morton, J. 1987b. Chupa -chupa ( Quararibea cordata). p 291292 in Fruits of warm clim ates, Dowling, C.F.J. (ed.). J.F. Morton, Miami. Murie, J.O. 1977. Cues used for cache -finding by agoutis ( Dasyprocta punctata). Journal of Mammalogy 58(1):9596.

PAGE 92

92 Mwavu, E.N., and E.T.F. Witkowski. 2008. Sprouting of woody species following cutting and tre e -fall in a lowland semi -deciduous tropical rainforest, North Western Uganda. Forest Ecology and Management 255(3 4):982992. Myers, J.A., and K. Kitajima. 2007. Carbohydrate storage enhances seedling shade and stress tolerance in a neotropical forest. Jou rnal of Ecology 95(2):383395. Myster, R.W. 2003. Effects of species, density, patchtype, and season on post -dispersal seed predation in a Puerto Rican pasture. Biotropica 35(4):542 546. Neito, V.M., and J. Rodriguez. 2002. Caryodendron orinocense H. Karst. p 362365 in Tropical tree seed manual, Vozzo, J.A. (ed.). USDA, Forest Service. Nepstad, D.C., C. Uhl, C.A. Pereira, and J.M.C. daSilva. 1996. A comparative study of tree establishment in abandoned pasture and mature forest of eastern Amazonia. Oikos 76(1):2539. Nepstad, D.C., C. Uhl, and E.A.S. Serrao. 1991. Recuperation of a degraded Amazonian landscape: forest recovery and agricultural restoration. Ambio 20(6):248255. Norden, N., R.L. Chazdon, A. Chao, Y.H. Jiang, and B. Vilchez -Alvarado. 2009. Re silience of tropical rainforests: tree community reassembly in secondary forests. Ecology Letters 12(5):385394. Notman, E., and D.L. Gorchov. 2001. Variation in post -dispersal seed predation in mature Peruvian lowland tropical forest and fallow agricultur al sites. Biotropica 33(4):621636. Nunez Iturri, G., and H.F. Howe. 2007. Bushmeat and the fate of trees with seeds dispersed by large primates in a lowland rain forest in western Amazonia. Biotropica 39(3):348 354. Ochsner, P. 2001. Direct seeding in the tropics. in IUFRO joint symposium on tree seed technology, physiology and tropical silviculture. Danida Forest Seed Centre, University of the Philippines, Los Banos. 7 p. Ostfeld, R.S., R.H. Manson, and C.D. Canham. 1997. Effects of rodents on survival of tree seeds and seedlings invading old fields. Ecology 78(5):15311542. Padilla, F.C., M.J. Alfaro, and J.F. Chavez. 1998. Chemical composition of the nogal de Barquisimeto ( Caryodendron orinocense Euphorbiaceae) seeds. Food Science Technology Internati onal 4(4):285289. Padilla, F.C., M.T. Alvarez, and M.J. Alfaro. 1996. Functional properties of barinas nut flour (Caryodendron orinocense Karst, Euphorbiaceae) compared to those of soybean. Food Chemistry 57(2):191196.

PAGE 93

93 Pea Claros, M., R.G.A. Boot, J. Do rado Lora, and A. Zonta. 2002. Enrichment planting of Bertholletia excelsa in secondary forest in the Bolivian Amazon: effect of cutting line width on survival, growth and crown traits. Forest Ecology and Management 161(13):159168. Pea Claros, M., and H De Boo. 2002. The effect of forest successional stage on seed removal of tropical rain forest tree species. Journal of Tropical Ecology 18:261274. Pennington, T.D. 1997. The genus Inga: botany. The Royal Botanical Gardens, Kew, England 844 p. Pennington T.D., and E.C.M. Fernandes. 1998. The genus Inga: utilization. The Royal Botanic Gardens, Kew, London. 177 p. Pennington, T.D., and N. Revelo. 1997. El genero Inga en el Ecuador: Morfologia, distribucion y usos. The Royal Botanic Gardens, Kew. 193 p. Per es, C.A., and E. Palacios. 2007. Basin-wide effects of game harvest on vertebrate population densities in Amazonian forests: implications for animal-mediated seed dispersal. Biotropica 39(3):304315. Perfecto, I., and J. Vandermeer. 2008. Biodiversity cons ervation in tropical agroecosystems: a new conservation paradigm. Annals of the New York Academy of Sciences 1134:173200. Phillips, O., J.S. Miller, and V.C. Hollowell. 2002. Global patterns of plant diversity: Alwyn H. Gentry's forest transect data set. Missouri Botanical Garden Press, St. Louis. 319 p. Piotto, D. 2007. Growth of native tree species planted in open pasture, young secondary forest and mature forest in humid tropical Costa Rica. Journal of Tropical Forest Science 19(2):92102. Posey, D.A. 1 985. Indigenous management of tropical forest ecosystems: the case of the Kayapo Indians of the Brazilian Amazon. Agroforestry Systems 3(2):139158. Prance, G.T., and S.A. Mori. 1979. Lecythidaceae. Part I. The actinomorphic -flowered New World Lecythidacea e ( Asteranthos Gustavia Grias, Allantoma & Cariniana). New York Botanical Garden, New York. 272 p. Pringle, E.G., f AlvarezLoayza, and J. Terborgh. 2007. Seed characteristics and susceptibility to pathogen attack in tree seeds of the Peruvian Amazon. Plant Ecology 193(2):211222. Raman, T.R.S., D. Mudappa, and V. Kapoor. 2009. Restoring rainforest fragments: survival of mixed -native species seedlings under contrasting site conditions in the Western Ghats, India. Restoration Ecology 17(1):137147. Ramos J.M., and S. Delamo. 1992. Enrichment planting in a tropical secondary forest in Veracruz, Mexico. Forest Ecology and Management 54(14):289304.

PAGE 94

94 Redford, K.H., and C. Padoch. 2000. Interpreting and applying the "reality" of indigenous concepts: what is necessary to learn from the n atives? p 21 34 in Conservation of Neotropical forests: working from traditional resource u se Redford, K.H., and C. Padoch (eds.). Columbia University Press, NY. Ricker, M., R.O. Mendelsohn, D.C. Daly, and G. Angeles. 1999. Enriching the rainforest with native fruit trees: an ecological and economic analysis in Los Tuxtlas (Veracruz, Mexico). Ecological Economics 31(3):439448. Romell, E., G. Hallsby, A. Karlsson, and C. Garcia. 2008. Artificial canopy gaps in a Macaranga spp dominated secondary tropical rain forest: effects on survival and above ground increment of four under -planted dipterocarp species. Forest Ecology and Management 255(56):14521460. Rudel, T.K. 2006. After the labor migrants leave: the search for sustain able development in a sending region of the Ecuadorian Amazon. World Development 34(5):838851. Rudel, T.K., D. Bates, and R. Machinguiasli. 2002. Ecologically noble Amerindians? Cattle, ranching and cash cropping among Shuar and colonists in Ecuador. Lati n American Research Review 37(1):144 159. Rudel, T.K., and B. Horowitz. 1993. Tropical deforestation: small farmers and land clearing in the Ecuadorian Amazon. Columbia University Press. 234 p. Snchez, D., E. Arends, V. Garay, and U.L.A. Saber. 2003. Cara cterizacin de las semillas de seis especies frutales arbreas, usadas por la etnia Piaroa en la Reserva Forestal Sipapo, Estado Amazonas, Venezuela. Revista Forestal Venezolana 77(2):31 36. Scariot, A. 2000. Seedling mortality by litterfall in Amazonian forest fragments. Biotropica 32(4):662669. Scarpa, F.M., and I.F.M. Valio. 2008. Relationship between seed size and litter effects on early seedling establishment of 15 tropical tree species. Journal of Tropical Ecology 24:569573. Schmidt, L. 2008. A rev iew of direct sowing versus planting in tropical afforestation and land rehabilitation. Development and Environment Series 102008. University of Copenhagen, Hrsholm, Denmark. 38 p. Schnee, L. 1973. Plantas comunes de Venezuela. Universidad Central de Ven ezuela, Instituto de Botanica Agricola, Maracay. 543 p. Shen, S., and A. Hess. 1983. Sustaining tropical forest resources: reforestation of degraded lands. p 56, Assessment, O.o.T. (ed.). Congress of the United States. Shiels, A.B., and L.R. Walker. 2003. Bird perches increase forest seeds on Puerto Rican landslides. Restoration Ecology 11(4):457465.

PAGE 95

95 Silvius, K.M., and J.M.V. Fragoso. 2003. Redrumped agouti ( Dasyprocta leporina) home range use in an Amazonian forest: implications for the aggregated distr ibution of forest trees. Biotropica 35(1):7483. Simons, A.J., and R.R.B. Leakey. 2004. Tree domestication in tropical agroforestry. Agroforestry systems 61 62:167181. Soepadmo, E. 1993. Tropical rainforests as carbon sinks. Chemosphere 27(6):10251039. T aylor, R.J. 1984. Predation. Chapman and Hall, New York. 166 p. Terborgh, J., G. Nunez -Iturri, N.C.A. Pitman, F.H.C. Valverde, P. Alvarez, V. Swamy, E.G. Pringle, and C.E.T. Paine. 2008. Tree recruitment in an empty forest. Ecology 89(6):17571768. Thompson, K. 1987. Seeds and seed banks. New Phytologist 106(1):2334. Toh, I., M. Gillespie, and D. Lamb. 1999. The role of isolated trees in facilitating tree seedling recruitment at a degraded sub -tropical rainforest site. Restoration Ecology 7(3):288297. Tol edo, M., and J. Salick. 2006. Secondary succession and indigenous management in semideciduous forest fallows of the Amazon basin. Biotropica 38(2):161170. Turner, I.M., Y.K. Wong, P.T. Chew, and A. BinIbrahim. 1997. Tree species richness in primary and ol d secondary tropical forest in Singapore. Biodiversity and Conservation 6(4):537 543. United Na tions. 2004. World urbanization prospects: The 2003 revision. Dept. of Economic and Social Affairs, Population Division. United Nations Publications. 323 p. Val lejo -Marin, M., C.A. Dominguez, and R. Dirzo. 2006. Simulated seed predation reveals a variety of germination responses of neotropical rain forest species. American Journal of Botany 93(3):369376. Van den Eynden, V., E. Cueva, and O. Cabrera. 2003. Wild f oods from southern Ecuador. Economic Botany 57(4):576603. Vander Wall, S.B. 1990. Food hoarding in animals. University of Chicago Press, Chicago. 445 p. Vander Wall, S.B., K.M. Kuhn, and M.J. Beck. 2005. Seed removal, seed predation, and secondary dispers al. Ecology 86(3):801806. Vieira, D.L.M., and A. Scariot. 2006. Effects of logging, liana tangles and pasture on seed fate of dry forest tree species in Central Brazil. Forest Ecology and Management 230(1 3):197205.

PAGE 96

96 Vieira, E.M., G. Paise, and P.H.D. Mac hado. 2006. Feeding of small rodents on seeds and fruits: a comparative analysis of three species of rodents of the Araucaria forest, southern Brazil. Acta Theriologica 51(3):311318. Vieira, I.C.G., C. Uhl, and D. Nepstad. 1994. The role of the shrub Cord ia multispicata Cham as a succession facilitator in an abandoned pasture, Paragominas, Amazonia. Plant Ecology 115(2):9199. Wade, T.G., K.H. Riitters, J.D. Wickham, and K.B. Jones. 2003. Distribution and causes of global forest fragmentation. Conservation Ecology 7(2):16. Wall, S.B.V. 2003. How rodents smell buried seeds: a model based on the behavior of pesticides in soil. Journal of Mammalogy 84(3):10891099. Wassenaar, T., P. Gerber, P.H. Verburg, M. Rosales, M. Ibrahim, and H. Steinfeld. 2007. Projecti ng land use changes in the Neotropics: the geography of pasture expansion into forest. Global Environmental Change 17(1):86104. White, E., N. Tucker, N. Meyers, and J. Wilson. 2004. Seed dispersal to revegetated isolated rainforest patches in North Queens land. Forest Ecology and Management 192(2 3):409426. Wijdeven, S.M.J., and M.E. Kuzee. 2000. Seed availability as a limiting factor in forest recovery processes in Costa Rica. Restoration Ecology 8(4):414424. Williams Guillen, K., C. McCann, J.C.M. Sanch ez, and F. Koontz. 2006. Resource availability and habitat use by mantled howling monkeys in a Nicaraguan coffee plantation: can agroforests serve as core habitat for a forest mammal? Animal Conservation 9(3):331 338. Woods, K., and S. Elliott. 2004. Direc t seeding for forest restoration on abandoned agricultural land in northern Thailand. Journal of Tropical Forest Science 16(2):248259. Wright, S.J. 2005. Tropical forests in a changing environment. Trends in Ecology & Evolution 20(10):553560. Wright, S.J ., H. Zeballos, I. Dominguez, M.M. Gallardo, M.C. Moreno, and R. Ibanez. 2000. Poachers alter mammal abundance, seed dispersal, and seed predation in a Neotropical forest. Conservation Biology 14(1):227. Wunderle, J.M. 1997. The role of animal seed dispers al in accelerating native forest regeneration on degraded tropical lands. Forest Ecology and Management 99(12):223235. Zahawi, R.A. 20 05. Establishment and growth of living fence species: an overlooked tool for the restoration of degraded areas in the tropics. Restoration Ecology 13(1):92 102. Zahawi, R.A. 2008. Instant trees: using giant vegetative stakes in tropical forest restoration. Forest Ecology and Management 255(7):30133016.

PAGE 97

97 Zahawi, R.A., and C.K. Augspurger. 1999. Early plant succession in abandoned pastures in Ecuador. Biotropica 31(4):540552. Zahwai, R.A., and C.K. Augspurger. 2006. Tropical forest restoration: tree islands as recruitment foci in degraded lands of Honduras. Ecological Applications 16(2):464478. Zhang, J.H., and M.A. Maun. 1993. Components of seed mass and their relationships to seedling size in Calamovilfa longifolia Canadian Journal of Botany 71(4):551557. Zimmerman, J.K., E.M.E. Iii, R.B. Waide, D.J. Lodge, C.M. Taylor, and N.V.L. Brokaw. 1994. Responses of tree species to hurricane winds in subtropical wet forest in Puerto Rico: implications for tropical tree life histories. Journal of Ecology 82(4):911922. Zuloaga, F.O., R.J. Soreng, and S.J. Pennington. 2003. Catalogue of New World grasses (Poaceae): III Subfamilies Panicoideae, Aristidoideae, Aruninoideae, and Danthonioideae. Dept. of Botany, National Museum of Natural History, Washington, DC. 662 p.

PAGE 98

98 BIOGRAPHICAL SKETCH Erica Van Etten was born in Ithaca New York in 1971. She attended Newfield public schools until she left her small hometown to begin college at the University of California at Santa Cruz in 1989. She transferred to Cornell Univers ity in 1990 and also attended the S.U.N.Y. Environmental Science and Forestry Ranger School before graduating in 1993 from Cornell with a B.S. in Natural Resource Management. After graduation, she pursued her interest s in for estry and land management in a series of jobs with the National Park Service and National Forest Service in California and Oregon. A visi t to Ithaca in the fall and a purchase of lumber at a farm auction led to a move back to upstate NY, wher e she built a house and continued her work in natural resource management with the County Soil and Water Conserva tion District. Passionate about teaching and hands -on learning, Erica then worked for five years at schools and outdoor education programs located on farms in Massachusetts and Vermont. In this work she combined her training in biology and land management with carpentry and farming skills to help students understand how human activities interact with natural ecosystems. A desire for an international perspective on environmental issu es led her to begin travelling and working in Central and South America in 2005. After working 6 month s on a restoration project at a biological station in the rainforests of coastal Ecuador, she decided to invest her time in restoration project that work ed directly with local landowners. In 2006, she began a reforestation project with indigenous Shuar communities in the Ecuadorian Amazon. After constructing several seedling nurseries, she decided to focus her Masters research on direct seeding, a potent ially cheaper an d easier method of planting trees than using nur sery -grown seedlings. E rica turned over the management of t he reforestation project to her Ecuadorian collaborators in 2008.

PAGE 99

99 After completin g her m asters degree, Erica plans to c ontinue her work in the field of intern ational conservation and management of projects that combine her interests in teaching, research, ecological restoration, and mentorship