Distribution, Genetic Diversity and Eurycomanone Content of Eurycoma Longifolia Jack in the Province of Thua Thien Hue, Vietnam

Faculty of Environmental Sciences, Institute of Forest Botany and Forest Zoology

Distribution, genetic diversity and eurycomanone content of Eurycoma longifolia Jack in the province of Thua Thien Hue, Vietnam

Dissertation in fulfillment of the requirements for the Degree of Doctor of Forest Science (Dr. rer. silv.)

submitted by: M.Sc. Yen Van Thi Born on 02.06.1983 in Thua Thien Hue, Vietnam

Supervisor: Prof. Dr. Doris Krabel, Dresden University of Technology Co-supervisor 1: Prof. Andreas Roloff, Dresden University of Technology Co-supervisor 2: Prof. Dr. Nguyen Hoang Loc, Hue University of Science

Date of defense:

Declaration of conformity

I hereby confirm that this copy conforms to the original dissertation on the topic: “Distribution, genetic diversity and eurycomanone content of Eurycoma longifolia Jack in the province of Thua Thien Hue, Vietnam”

…………………………………… Location, date

…………………………………… PhD candidate’s signature

Acknowlegement ACKNOWLEDGEMENT

First, I would like to express my gratitude and deepest thanks to my first supervisor, Prof. Dr. Doris Krabel, for her guidance, teaching, useful advice, support and encouragement for the entire doctoral study at the Dresden University of Technology. Without her support, I would not have enough strength to complete my thesis. I am so grateful to be her student. I want to thank to my co-supervisors, Prof. Dr. Andreas Roloff and Prof. Dr. Nguyen Hoang Loc, for the willingness to give valuable comments and remarkable advice on the dissertation. Especially, Prof. Dr. Doris Krabel and Prof. Dr. Nguyen Hoang Loc have been strongly supported the valuable research budgets and the laboratory facilities.

I also would like to acknowledge the scholarship program from the Vietnamese Government (VIED) and the financial supports from the Graduate Academy (GA) and Association of Friends and Sponsors of TUD, e.V. (GFF) of TU Dresden, the family foundation of Nyamjav Ayur in Mongolia; my bosses, Ass. Prof. Dang Thai Duong and Dr. Hoang Huy Tuan - Faculty of Forestry, Hue University of Agriculture and Forestry (HUAF) for their support, encouragement and facilitation of my application for the VIED scholarship and WWF (World Wide Fund) for their support in the next steps of this dissertation.

I respectfully offer my sincerest thanks to my colleagues at the Chair of Forest Botany, Dresden University of Technology for their help and encouragement during my stay in Dresden. Special thanks go to Ms. Kristin Morgenstern, Mr. Jens-Ulrich Polster and the working group “Molecular Physiology of Woody Plants” for their guidance and support in lab work and data analysis. Besides, I would like to express my thank to Prof. Dr. Karsten Wesche and Dr. Veit Herklotz from Senckenberg Museum of Natural History (Görlitz) and Dr. Cao Thi Thu Hien from University of Vietnam National Forestry, for their valuable teaching, support and advise in data analysis. I also want to thank to Mr. Alexander Solger, Dr. Britt Kniesel, Dr. Matthias Mayer, Ms. Sri Astutik, Ming, Khiem, Bayartaa, Nicole, Anne and Birgit for the great support and valuable discussions.

I am thankful to all my colleagues Hue University of Agriculture and Forestry for providing useful information of my research in Vietnam, especially Dr. Tran Minh Duc, M.Sc. Nguyen Hoi, Ass. Prof. Dr. Nguyen Van Loi and to my students from HUAF who helped me between 2016 and 2018 of my field work: Dien, Mot, Huy, Nhat, Dang, Quang, Hau, Dieu, Cuong, Minh, Thuy, Tien and Ngoc Hoa.

I Acknowlegement Special thanks to the research group of Prof. Dr. Nguyen Hoang Loc, especially Dr. Hoang Tan Quang, Mr. Trinh Huu Tan, Dr. Nguyen Ngoc Luong, Dr. Le Thi Ha Thanh and all his students from Hue University of Science: Tuan, Tue and Uyen for the valuable support of my laboratory work in Vietnam.

I also respect and appreciate the kind support of the managers of the nursery garden in Huong Thuy town, Bach Ma National Park, forest rangers in Nam Dong, A Luoi Saola Nature Reserve and the People’s Committee of Phong Dien district for my field work.

Most importantly, to my husband, my little son, my parents and the family members who always support me by their love, beliefs and encouragement.

Thank you very much again. Vielen Dank! Chân thành cảm ơn!

Van Thi Yen

II Table of Contents TABLE OF CONTENTS

ACKNOWLEDGEMENT………………………………………………………………………I TABLE OF CONTENTS……………………………………………………………………..III LIST OF FIGURES…………………………………………………………………………...VI LIST OF TABLES…………………………………………………………………………..VIII LIST OF APPENDICES……………………………………………………………………….X ABBREVIATIONS.…………………………………………………………………………...XI

1 General introduction ...... 1 1.1 Status of medical plant resources in Vietnam ...... 1 1.2 The E. longifolia tree and its importance ...... 2 1.3 Specific approaches ...... 3 1.3.1 Objectives ...... 3 1.3.2 Scope of study ...... 4 1.3.3 Hypotheses ...... 4 2 Literature review ...... 6 2.1 Species description ...... 6 2.1.1 Biology ...... 6 2.1.2 Ecology and distribution ...... 7 2.2 Importance and use ...... 9 2.2.1 Traditional use ...... 9 2.2.2 Modern extracts ...... 10 2.2.3 Eurycomanone content ...... 11 2.3 Cultivation and propagation of the species ...... 12 2.3.1 Cultivation ...... 12 2.3.2 Propagation and germination of seeds ...... 14 2.4 Pests and diseases...... 16 2.5 Genetic knowledge ...... 17 2.5.1 Assessment of genetic variation ...... 17 2.5.2 Parameters for comparing the utility of markers ...... 20 2.5.3 Genetic variation – Selected Examples ...... 20 3 Material and research methods ...... 22 3.1 Study Site ...... 22 3.1.1 Overview ...... 22 3.1.2 Nam Dong district ...... 24 3.1.3 Bach Ma National Park...... 26 3.1.4 A Luoi district ...... 26 3.1.5 Phong Dien ...... 27 3.2 Field and lab design ...... 29 3.2.1 Distributive survey ...... 29 III Table of Contents

3.2.2 Plant material ...... 30 3.2.3 Propagation ...... 35 3.3 Data collection ...... 36 3.3.1 Distributional data ...... 36 3.3.2 Eurycomanone quantification ...... 38 3.3.3 DNA isolation ...... 38 3.3.4 RAPD analysis and primer screening ...... 41 3.3.5 SCoT and BPS analysis and primer screening ...... 41 3.4 Data analyses...... 43 3.4.1 Data of distribution, phenotype, propagation and eurycomanone component ... 43 3.4.2 Analysis of genetic variation and population structure ...... 44 4 Results ...... 51 4.1 Population distribution status of E. longifolia ...... 51 4.1.1 Terrain and soil ...... 51 4.1.2 Population distribution ...... 52 4.2 Phenotype ...... 60 4.2.1 Leaf morphology and anatomy of mature trees ...... 60 4.2.2 Leaf characteristics of seedlings and seedling’s growth performances ...... 69 4.3 Genetic diversity ...... 79 4.3.1 RAPD, SCoT and BPS amplification ...... 79 4.3.2 Genetic diversity in E. longifolia population ...... 82 4.3.3 Genetic differentiation ...... 83 4.3.4 Principle Coordinate Analysis (PCoA) ...... 84 4.3.5 Population structure ...... 85 4.3.6 Genetic cluster analysis ...... 87 4.3.7 Genetic variation across different generations ...... 89 4.4 Eurycomanone content ...... 94 4.4.1 Biological traits and root water content ...... 94 4.4.2 Eurycomanone content of E. longifolia root tissues ...... 96 4.4.3 Correlation between the eurycomanone content and plant traits and geographical factors ...... 99 5 Discussion ...... 101 5.1 Ecological distribution of E. longifolia population ...... 101 5.1.1 Population density ...... 101 5.1.2 Natural regeneration ...... 103 5.2 Morphological and anatomical adaptability of E. longifolia to different distribution areas ...... 105 5.2.1 Leaf morphology and anatomy of mature trees under different site conditions ...... 105 5.2.2 Leaf characteristics of seedlings and mature trees of E. longifolia ...... 110 5.2.3 Seedling growth ...... 111 IV Table of Contents

5.3 Genetic diversity of E. longifolia ...... 114 5.3.1 Comparative utility of different marker systems ...... 114 5.3.2 Genetic diversity between different distribution areas ...... 115 5.3.3 Genetic structure of different distribution areas ...... 117 5.3.4 Comparison of the genetic variation between mature trees and seedlings ...... 120 5.4 The correlation between eurycomanone content and the distribution areas .... 122 6 Conclusions and recommendations ...... 125 6.1 Conclusions ...... 125 6.2 Suggestions for further research ...... 127 7 Summary ...... 129 8 Zusammenfassung ...... 133 9 Tóm tắt ...... 137 References...... 141 Appendices…………………………………………………………………………….....… 156

V List of Figures LIST OF FIGURES

Figure 2.1 Natural occurrence of E. longifolia in a mountainous area in Vietnam (A) and typical flowers (B) and fruits (C) ...... 7 Figure 2.2 Distribution map of E. longifolia in Asia ...... 9 Figure 2.3 Geographic distribution of six E. longifolia populations selected in Peninsular, Malaysia ...... 21 Figure 3.1 Map of the province of Thua Thien Hue located in central Vietnam ...... 23 Figure 3.2 Average temperatures in Nam Dong, A Luoi, Bach Ma and Phong Dien between 2013 and 2017 ...... 28 Figure 3.3 Average precipitation in different areas between 2013 and 2017 ...... 28 Figure 3.4 Diagram of baseline and plot design ...... 30 Figure 3.5 Five different positions of leaflets from one leaf to estimate the number of stomata through microscope and microscope slides in the paper-clip boxes ...... 32 Figure 3.6 Measurement of the size of the leaves and the number of leaflets in the field ...... 32 Figure 3.7 a) Before taking the root sample and b) after digging the hole and taking the root sample ...... 34 Figure 3.8 Geographical distribution map of DNA samples, mother trees and root samples of four (wild) E. longifolia populations in the province of Thua Thien Hue ...... 40 Figure 3.9 Research Design Framework ...... 49 Figure 4.1 Topographic distribution of E. longifolia based on elevation (m) ...... 52 Figure 4.2 Distribution map of E. longifolia Jack in the province of Thua Thien Hue...... 54 Figure 4.3 E. longifolia occurencies along the observation baselines ...... 54 Figure 4.4 Distribution of trees and saplings in diameter (a) and height classes (b) ...... 56 Figure 4.5 Differences of sapling diameter and height between sandy and mountainous areas ...... 57 Figure 4.6 The increase of tree height and diameter with elevation, slope steepness and the positive correlation of trees and saplings between height and diameter ...... 57 Figure 4.7 E. longifolia in mountainous area (A) and sandy area (B) ...... 58 Figure 4.8 Comparison of seedling density and height between mountains (N = 1,752) and sandy areas (N = 99) ...... 59 Figure 4.9 Distribution of seedling number along the height classes (cm) ...... 59 Figure 4.10 Leaf area of E. longifolia mature trees ...... 61 Figure 4.11 Leaf morphology in different areas: A Luoi (1), Nam Dong (2), Bach Ma (3) and Phong Dien (4) ...... 61 Figure 4.12 Leaf area differences in moist (mountain) and dry (sand) sites ...... 62 Figure 4.13 Stomatal density differences of mature trees at the four distribution areas ...... 64 Figure 4.14 Stomatal density differences of mature trees in moist and dry sites ...... 64 Figure 4.15 Different distribution of stomatal density in dry (sandy) site (a) and moist (mountainous) site (b) ...... 65

VI List of Figures Figure 4.16 Cross section of the leaves: a) pit-type stomata in dry site and b) flat-type stomata in moist site ...... 66 Figure 4.17 Negative correlation of leaf area with stomatal density of mature trees ...... 66 Figure 4.18 Positive correlation of leaf length (cm) with leaf width (cm) ...... 69 Figure 4.19 Differences between mature trees and seedlings in aspects of a) leaf area and b) stomatal density in different sites ...... 71 Figure 4.20 Typical examples of fruit from a) sandy area (dry site) and b) mountainous area (moist site)...... 72 Figure 4.21 Germination response of the ripe seeds of E. longifolia from different provenances in the period between 15 and 55 days ...... 75 Figure 4.22 The growth pattern of E. longifolia seedlings in terms of height, collar diameter and number of leaves ...... 76 Figure 4.23 Correlation between seedling height and seedling collar diameter, from different sites ...... 77 Figure 4.24 Images of seedlings from different stages ...... 78 Figure 4.25 A gel image generated by SCoT21 primer (12 samples) and one sample investigated with five different primers ...... 80 Figure 4.26 Principle Coordinate Analysis (PCoA) based on genetic relationship of four subpopulations of 276 E. longifolia accessions ...... 85 Figure 4.27 STRUCTURE analyses ...... 86 Figure 4.28 STRUCTURE Q plots generated utilizing the maximum value of ΔK indicate E. longifolia samples ...... 86 Figure 4.29 Dendrogram agglomerative clustering using Ward’s method and Euclidean distances among individuals ...... 88 Figure 4.30 Dendrogram of 15 E. longifolia mother tree samples and 269 seedlings analysed from four different sites ...... 93 Figure 4.31 The weight of all the root samples from four sites...... 94 Figure 4.32 a) Difference of water content in moist and dry areas and b) positive correlation between root water content and elevation factor ...... 95 Figure 4.33 Standard curve of eurycomanone ...... 96 Figure 4.34 HPLC chromatogram of eurycomanone standard at 300 µg/µL ...... 96 Figure 4.35 HPLC chromatogram of eurycomanone extract from root ...... 97 Figure 4.36 a) Eurycomanone content of the root of E. longifolia in four different areas and b) eurycomanone content in moist and dry sites ...... 98 Figure 4.37 Positive correlation between eurycomanone and root water content ...... 100 Figure 4.38 Root diameter dependence on eurycomanone content ...... 100

VII List of Tables LIST OF TABLES

Table 2.1 Summary of several main bioactive compounds and pharmacological effects from E. longifolia ...... 10 Table 2.2 Summary of the previous studies of genetic diversity in E. longifolia ...... 19 Table 2.3 Main indices of genetic diversity in E. longifolia populations studied by Osman et al. (2003) ...... 21 Table 3.1 Forest status in the province of Thua Thien Hue ...... 24 Table 3.2 Population name, geographical location and sample size of DNA samples ...... 33 Table 3.3 Number of different provenances, propagated and natural seedlings and seedlings, for estimating germination, growth pattern and total DNA isolation ..... 36 Table 3.4 Variables measured of baselines, plots and tools for these measurements ...... 37 Table 3.5 PCR pipetting instructions and cycling protocol for GoTaq® Green Master Mix used for SCoT21, SCoT36, OPC02, OPB05 primers and TopTaq Master Mix used for LA2a primer ...... 42 Table 3.6 List of RAPD, SCoT and BPS primer sequences tested ...... 43 Table 3.7 Summary of the methodologies of data collection ...... 50 Table 4.1 Soil factors and forest status ...... 51 Table 4.2 Tree density (including saplings), tree size and the frequency of individuals based on elevation in four different areas ...... 55 Table 4.3 The diameter and height classes of trees and saplings ...... 58 Table 4.4 Stomatal density (stomata/mm2) and the number of stomata per leaf for moist- and dry-grown E. longifolia trees ...... 63 Table 4.5 Average ± standard deviation, range of leaf length, leaf width and number of leaflets of E. longifolia in four different sites and statistical differences ...... 67 Table 4.6 Average ± standard deviation, range of leaf length, leaf width and number of leaflets of E. longifolia in moist and dry sites ...... 68 Table 4.7 Spearman’s correlation coefficient (r2) and p of correlation among leaf characteristics (N = 505) ...... 68 Table 4.8 Average ± standard deviation of seedling characteristics from four different provenances ...... 70 Table 4.9 Average ± standard deviation, range of fruit morphology of moist and dry sites .....72 Table 4.10 Seedling’s provenances and seed maturity from the seed germination (days) within a 55-day period and survival rate >22 months ...... 74 Table 4.11 Details of banding pattern revealed through RAPD, SCoT and BPS markers ...... 81 Table 4.12 Summary of the genetic variation as revealed through RAPD, SCoT and BPS markers among four populations of E. longifolia ...... 83 Table 4.13 Analysis of molecular variance (AMOVA) using RAPD, SCoT and BPS markers ...... 84

VIII List of Tables Table 4.14 Pairwise genetic differentiation (Φst) of E. longifolia populations (above diagonal) and pairwise population matrix of Nei‘s genetic distance (below diagonal) ...... 84 Table 4.15 Analysis of Principle Coordinates (PCoA) based on RAPD, SCoT and BPS markers ...... 85 Table 4.16 a) Genetic diversity across different generations and b) Pairwise Matrix of Nei Genetic Distance (above diagonal) and genetic differentiation (Φst) of E. longifolia mother and seedling populations (below diagonal) ...... 89 Table 4.17 Genetic diversity at seedling generation in comparison with mature trees and the differences of its variation between propagated and natural seedlings from different provenances ...... 91 Table 4.18 Pairwise Matrix of Nei Genetic Distance (below diagonal) and genetic differentiation (Φst) of E. longifolia seedling accessions (above diagonal) ...... 92 Table 4.19 Analysis of molecular variance of seedlings from different provenances using RAPD, SCoT and BPS markers ...... 92 Table 4.20 Spearman’s correlation coefficient between eurycomanone content and tree traits (root water content, root diameter, tree diameter, tree height and plant ages) and geographical factors (coordinates and altitude)...... 99

IX List of Appendices LIST OF APPENDICES

Appendix 1. Baseline and plot map for surveying E. Longifolia ...... 156 Appendix 2. Value of soil value: colors, texture, pH-value and moisture ...... 156 Appendix 3. Baseline information and tree, sapling density; tree, sapling height and diameter in mountainous and sandy areas ...... 158 Appendix 4. Summary of the tree and predictor variables in Pearson’s correlation coefficient ...... 159 Appendix 5. Seedling germination across two areas (Mann-Whitney test) ...... 159 Appendix 6. Comparison of seedling leaves, seedling height and seedling collar diameter between moisture and dry areas...... 159 Appendix 7. Coordinates X (axis) and Y (asix) for the highest tree and sapling density in the typical plots ...... 160 Appendix 8. Correlation among genetic diversity, geographical distance and elevation factors and morphological traits by Mantel test ...... 160 Appendix 9. Average ± standard deviation, range of eurycomanone and water content and biological traits in different areas ...... 160 Appendix 10. Average ± standard deviation of log probability of LN for each K (STRUCTURE program) ...... 161 Appendix 11. List of 276 DNA samples for analysing genetic diversity of mature trees ...... 163 Appendix 12. List of 15 E. longifolia mother samples and 269 seedlings analysed from four different sites ...... 166 Appendix 13. Typical trichomes of E. longifolia ...... 170 Appendix 14. Images of DNA electrophoresis ...... 171 Appendix 15. Images of electrophoresis of RAPD and SCoT primers ...... 171 Appendix 16. HPLC chromatogram of eurycomanone extract from two root samples (a, b) ...... 172 Appendix 17. Images of taking the root samples in the field (measuring the root diameter at 20 cm from the surface) ...... 173 Appendix 18. Images of the field work ...... 173 Appendix 19. The trees were illegally harvested by local people ...... 174

X Abbreviations ABBREVIATIONS

asl. Above sea level AMOVA Analysis of Molecular Variance ANOVA Analysis of Variance AL A Luoi BM Bach Ma BMNP Bach Ma National Park bp base paires BPS Branch Point Signal Sequences BSA Bovine Serum Albumin CTAB Cetyl trimethylammonium bromide DBH Diameter at Breast Height DNA Deoxyribonucleic Acid Dw Dry weight EDTA Ethylenediaminetetraacetic acid EMR Effective marker ratio FA Fragment Analyzer Fw Fresh weight GPS Garmin Position System HPLC High Performance Liquid Chromatography IBA Idole-3-butyric acid Lao PDR Lao People's Democratic Republic LC-MS/MS Liquid chromatography-mass spectrometry MI Marker Index min minute(s) MS Murashige and Skoog n.d. no date ND Nam Dong NDG Nam Dong Geographic NS Nam Dong statistics NPF Non Polymorphic Fragment PCoA Principle Component Analysis PCR Polymerase chain reactions PD Phong Dien PIC Polymorphic Information Content RAPD Random Amplified Polymorphic DNA RNase Ribonuclease Rp Resolving power SCoT Start Codon Targeted Polymorphism Stdv Standard deviation TE/TAE Tris-EDTA/Tris-acetate-EDTA UPGMA Unweighted Pair Group Method with Arithmetic Mean UTM Universal Transverse Mercator Wc Water content WHO World Health Organization

General introduction 1 General introduction

1.1 Status of medical plant resources in Vietnam

Vietnam is considered as one of the 25 countries having a high value of biodiversity with many rare tree species, non-timber forest products (NTFPs) and medicinal plants (Hoi, 2013; MonRE, 2014). For centuries, the consideration of traditional medicine (i.e., herbal medicine) has been increasing in both developed and developing countries (WHO, 2019). In addition, medicinal plants play a significant role as an alternative medicine due to the damaging effects of food processing and environment as well as hazardous side effects of prolonged medications (WHO, 2002). Globally, more than 35,000 botanic species are being used for medicinal purposes (Lewington, 1993; Effendy et al., 2012). In Vietnam, at least 4,700 recorded plant species have been utilized for traditional medicine (Chi, 2012).

The widespread medicinal plant usage and the high demand of commercial trade due to unsustainable harvesting have significantly reduced several plant populations (Traffic, 2013). A lack of effective management, the expansion of roads, land-use change and habitat loss have also dramatically reduced the forests and thereby forced an irreplaceable loss of genetic stock of many species (Nishteswar, 2014). Especially, destructive harvesting leads to depletion and scarcity of medicinal plants such as Paris polyphylla Smith, Tetracera scandens Linn. Merr., Stemona pierrei Gagnep, Blumea balsamifera (L.) DC., Ampelopsis cantoniensis (Hook. et Arn.) Planch., Eurycoma longifolia Jack, Morinda citrifolia L., Andenosma caeruleum R.Br., Dioscorea persimili Prain et Burkill, Homalomela pierriana Engl., Smilax glabra Roxb., Tacca integrifolia Ker-Gawl., Streptocaulon juventas (Lour.) Merr. and Acorus macrospadiceus (Yam.) Wei and Li (Thoa et al., 2015). In the case of E. longifolia, the root tissues of E. longifolia is the most widely used among popular critical products.

Several programs and policies of development and conservation of potential and valuable plants have been promoted in the country (Hoi, 2013; Traffic, 2013). Up to now, these solutions could not be implemented effectively due to a lack of information on the distribution, status of conservation, species identification (classification, morphology, genetic diversity) and the composition of chemical compounds. Therefore, it is necessary to conduct several studies of developing cultivating techniques and in understanding traditional medicine for conservation and sustainability.

1 General introduction 1.2 The E. longifolia tree and its importance

Eurycoma longifolia Jack is a famous medicinal plant in the family of Simaroubaceae, which was studied and developed across the globe, including Vietnam (Effendy et al., 2012; Hassan et al., 2012). E. longifolia is a shrub or small timber tree that grows up to 10 m in height (Chi, 2012; Effendy et al., 2012). It is mostly distributed in Asian countries such as Malaysia (‘Tongkat Ali’), Indonesia (Pasakbumi), Indo-China, Philippines, Myanmar, Thailand (Tung saw), Lao and Vietnam (Cay Ba binh) (Osman et al., 2003; Loi, 2006; Chi, 2012; Jagananth and Teik, 2000; Ang et al., 2000). This plant species can be found in lowland mountains (300- 700 m), highland mountains (>700 m), midlands (<300 m) and sandy areas (Cam et al., 2002; Hadiah, 2000; Ang et al., 2002; Husen et al., 2004; Muhamad et al., 2010; Nordin, 2014).

Several studies show that this species has high medicinal values such as anti-malarial, anticancer, anti-pyretic, anti-bacterial, anti-cytotoxic, anti-schistosomal and anti-lytic activities (Morita et al., 1990; Kardono et al., 1991; Ang et al., 2002; Kuo et al., 2004; Jiwajinda et al., 2002; Ang and Cheang, 1999). Primarily, E. longifolia is known for its ability to stimulate the production or action of the androgen hormone (testosterone) in humans (Effendy et al., 2012). The main bioactive compounds including quassinoids (eurycomanone, eurycomanol, eurycomalactone, etc.), alkaloids (9-methoxycanthin-6-one, canthine-6-one, 9-hydroxycanthin- 6-one, etc.) and triterpenes are mainly isolated from the root of E. longifolia (Hong, 2006; Effendy et al., 2012; Hassan et al., 2012; Tee and Azimahtol, 2005). According to Low et al. (2013), eurycomanone, a type of quassinoids with the highest concentration in the root of E. longifolia, can improve fertility by increasing testosterone content and spermatogenesis. Furthermore, 9-methoxycanthin-6-one compound from the roots of this plant displays cytotoxicity against human lung cancer A-549 (Adenocarcinomic human alveolar basal epithelial cells) and human breast cancer MCF-7 (Michigan Cancer Foundation-7) cell lines (Chan et al., 1997; Kuo et al., 2003).

Due to its capacity for medicinal utilization, this species has attracted a flux of attention, leading to the great harvest of wild-grown trees for all parts of the plant and especially roots for medicinal use. As a consequence, the plant has suffered from a rapid decrease in its natural populations which negatively affected genetic diversity in return (Hassan et al., 2012; Osman et al., 2003). At present, molecular techniques are available to analyze the genetic diversity in plants (Gömöry et al., 2001; Nassar et al., 2001; Osman et al., 2003). Although several studies have been conducted in Vietnam, they are quite preliminary with incomplete data sets; thus, there is no detailed information on its genetic systems (Hong, 2006). It is widely known that

2 General introduction populations with high genetic diversity will be well adapted to a variety of environmental conditions such as diseases, pests, or climate changes (Susilowati, 2008). Studies on its genetic diversity are required with a view to providing information for propagation, domestication and breeding programs as well as conservation of genetic resources.

Nowadays, research in Vietnam and other Asian countries mainly focus on species distribution, analysis of bioactive constituents, or genetic diversity in tropical forests, however, hardly any research on E. longifolia in sandy areas can be found. This species' appearance seems differences in leaf morphology between sandy and mountainous areas throughout the preliminary surveys. Furthermore, there is a lack of studies about this species' adaptation regarding leaf characteristics (phenotypic plasticity) to the different site conditions (dry and moist sites). The quality of bioactive constituents also should be taken into account, especially the simulation of the components in the different distribution areas. Among the bioactive compounds, eurycomanone is a major quassinoid with high biological activity and a unique compound in E. longifolia but hardly found in other plants. The studies on eurycomanone concentration in E. longifolia, play a crucial role in defining the quality of this species as well as estimating its quality from different distribution regions. Besides, it is essential to estimate the accumulation of eurycomanone that depends on biological characteristics, ecological distribution areas, or genetic diversity.

Thus, the present research focuses on natural distribution areas, leaf morphology and anatomy, bioactive component (eurycomanone) and genetic diversity of E. longifolia populations in lowland and highland mountains as well as sandy regions. In doing so, this study will contribute to the identification of its genotypes at various sites and to the catalogue of these genotypes in different ecological zones in the province of Thua Thien Hue. With regard to the species distribution, genetic diversity and eurycomanone content of E. longifolia play a crucial role in this study providing information for propagation, domestication and breeding programs as well as conservation of forest genetic and valuable medicinal plant resources.

1.3 Specific approaches 1.3.1 Objectives

The overall goal of this dissertation is to evaluate the status of E. longifolia populations in Vietnam, its leaf morphology and anatomy, its genetic diversity and its eurycomanone content among different ecological areas in order to suggest appropriate species conservation, management strategies and potential breeding approaches.

3 General introduction The specific objectives of the study are to: i) investigate the ecological distribution areas of E. longifolia; ii) determine morphological and anatomical traits related to different distribution sites; iii) evaluate the growth of seedlings in a nursery garden; iv) assess the genetic relationships within and between populations in different ecological areas of E. longifolia; v) estimate the population genetic structure of E. longifolia and vi) analyze eurycomanone content of the root tissues of E. longifolia from different sites.

1.3.2 Scope of study

The scope of the study is to assess the population status, phenotype, genotype and main compound of E. longifolia from different habitat conditions in the province of Thua Thien Hue.

1.3.3 Hypotheses

Several studies have shown that E. longifolia has a high medicinal value such as being antimalarial, anti-cancer, anti-bacterial, anti-cytotoxic, etc. (Kardono et al., 1991; Ang et al., 2002; Kuo et al., 2004). All parts of this plant were abundantly harvested for medicinal purposes, especially in lowland and sandy areas in the central part of Vietnam. This has exerted a significant influence on the natural populations and plant forest genetic resources. During the Vietnam wars (1945-1975), the plant patterns were partly destroyed, particularly in A Luoi district (Thua Thien Hue province). The natural populations of E. longifolia have recently decreased. Against this backdrop, the following hypothesis are proposed:

H1: There are differences in terms of the density of trees, saplings and seedlings between sandy and mountainous areas.

Eurycoma longifolia was listed in several studies that it is a shrub or small timber tree (Chi, 2012; Effendy et al., 2012). Through a preliminary survey, we discovered many big trees in the high mountains which were difficult for local people to access. This species occurs not only in the mountainous area but also in the sandy area. The sandy area and midland mountains (<300 m asl.) have been strongly affected by local people for root harvesting and firewood demand so E. longifolia populations in these areas are decreasing. So far, there are no studies about the adaptation of the leaf traits to the different site conditions (dry and moist sites). Thus, the following hypothesis was checked:

H2: Leaf size and stomatal density of mature trees from moist sites are both larger than those from the dry sites.

Leaves from moist sites or shade conditions are often larger than those from dry sites (Medina, 1983; Pallardy, 2008) and the stomatal density will increase if there is sufficient water for gas 4 General introduction exchange in the photosynthesis process (Schlüter et al., 2003). In the nursery garden, an experiment was conducted to see how seedlings from different provenances adapt to the same environmental condition. Then, the following hypothesis was formed:

H3: Seedlings of the same age and of the same growth condition will have similar growth pattern, even from the different provenances.

Chan and Toh (1984) showed that the same growth pattern of seedlings from many provenances of Carica papaya L. occurred in stable conditions. In addition to the phenotypes of mature trees, seedlings and relevant mother trees, the present study focuses on the genetic variation of the populations from mountainous and sandy areas, which leads to the following hypothesis:

H4: Major part of genetic diversity occurs within populations.

Several studies have shown that genetic diversity of the woody perennial species often occurs within populations, especially in the tropical plants (Hamrick et al., 1992; Bussell, 1999). Genetic structure of the population also should be taken into account as the hypothesis does below:

H5: There are potentially two genetic groups, namely sandy and mountainous.

A preliminarily survey indicated that E. longifolia shows different morphological traits between mountainous and sandy areas. Hence, it can be assumed that they should have two groups of population structure. E. longifolia is a famous medicinal plant, which contains many high valuable bioactive compounds such as eurycomanone and 9-methoxycanthin-6-one. Among them, eurycomanone is one of the major components in E. longifolia (Chan et al., 1997; Kuo et al., 2003). This fact leads to the final hypothesis:

H6: Eurycomanone content (quality and quantity) of the roots depends on ecological distribution areas, genetic diversity and plant age.

So far, there is no research estimating eurycomanone content under natural conditions in the province of Thua Thien Hue, especially in both sandy and mountainous areas. Thus, it is necessary to define eurycomanone accumulation at different sites and its relationship with ecological distribution areas, genetic factor, or tree age.

5 Literature review 2 Literature review 2.1 Species description 2.1.1 Biology

Eurycoma longifolia is a shrub or small evergreen tree, which can grow up to a height of 10 meters. The plant is one of four species in the Eurycoma genus belonging to the family of Simaroubaceae (Chi, 2012; Effendy et al., 2012). However, this family is not large and it has only approximately 50 species (Reveal and Chase, 2011). Beside Eurycoma apiculate Benn, E. longifolia is one of two recorded species of Eurycoma in Malaysia. Most of the research on medicinal properties mentions only E. longifolia, popularly known as “Tongkat Ali”, which is the more common plant name (Nordin, 2014). Up to now, there is no report about E. apiculata in Vietnam. Rosmaina et al. (2015) suggested that the distinctness of both species could be based on the color of petiole and the level of bitter taste. In 2019, Zulfahmi et al. revealed that the scientific name of male eurycoma is E. longifolia Jack and femal eurycoma is E. apiculate Benn through leaf morphometric traits.

Regarding their branch characteristics, E. longifolia trees are generally un-branched or carry just a few branches (Fig. 2.1). An umbrella-like rosette of compound leaves crowns each branch and the long leaves consist of 15-30 pairs of leaflets (20-30 cm), with spirally arranged pinnate leaves and lanceolate to obovate-lanceolate leaflets. After the leaves fall, the scars always remain on the stems. The roots and stem are often yellowish in color and bitter in taste (Ho, 1999). Unlike many other plants, E. longifolia is a deciduous plant with male and female flowers producing pistils and stamens in large panicles on different individuals (Bhat and Karim, 2010; Zanolia et al., 2009). Distinguishing between the vegetative forms of male and female trees without flowers is difficult (Saleh, 1992). According to Susilowati (2008), E. longifolia is predominantly outcrossing occasional self-pollinationing. As the size of its fruit is quite large, the seed dispersal by long distance should be possible by rainwater flow, birds or rodents but impossible to be distributed by wind.

The fruits are ellipsoid or ovoid, 1-2 cm long and 0.5-1 cm broad, green to blackish-red when ripening and they carry only one seed with 0.5-1 cm in diameter (Hadiah, 2000; Zanolia et al., 2009). E. longifolia is a slow-growing plant, which bears fruits after 2-3 years of cultivation and this species is diploid with 2n = 14 chromosomes (Zulfahmi et al., 2018). However, it is believed that the plant may take up to 25 years for completing the maturation (Bhat and Karim, 2010).

6 Literature review

A B

C C

Figure 2.1 Natural occurrence of E. longifolia in a mountainous area in Vietnam (A) and typical flowers (B) and fruits (C)

2.1.2 Ecology and distribution

Eurycoma longifolia is basically found as an understorey tree from moist to wet in the lowland rainforests up to 500 m asl. in the Southeast Asian (Hussein et al., 2005). According to Kuo et al. (2003), this species can grow on hillsides and ridges with sandy or clay soils and can be also commonly distributed in primary and secondary evergreen and mixed dipterocarp forests in Burma, Indo-China, Thailand, Sumatra, Borneo and Philippines (Fig. 2.2). Besides, the plant can only be found in certain forest patches in Singapore and its offshore islands (Chong et al., 2009). Similarly, this tree species appears naturally in Kalimantan and Sumatra (Indonesia) at <700 m asl. (Hadiah, 2000).

Kartikawati et al. (2014) state that E. longifolia has a wide range of topographic distribution which can be found at an altitude of 0-700 m asl. where these populations were approximately estimated to be 114 individuals on 1 ha consisting of 71 seedlings (62.28%), 42 saplings (36.84%) and 1 pole (0.87%). This research claims that the highest density of population was discovered within the altitude range of 320-402 m asl. with 40 individuals and a clumped distribution pattern where the seedlings tend to grow in the close proximity of the mother trees. Furthermore, seedlings require shade throughout the time when they develop their expandable root system (EL, 2016; Hadiah, 2000). In juvenile stages, the plant needs stronger light to develop its vegetative and reproductive organs. Therefore, it takes approximately more than five years to reach the reproductive age. Although E. longifolia has flowers and fruits

7 Literature review throughout the year, its peak bloom period is between June and July and fruiting time is in September in Sumatra, Indonesia (Hadiah, 2000; Corner, 1988). Even though trees produce plenty of fruits and seeds (200-300 seeds per bunch) during the fruiting season, the number of seedlings growing-around adult trees is quite low. It is discussed whether this is the reason why the distribution of seedlings in the forest floor is often poor (Keng et al., 2002).

In Vietnam, E. longifolia occurs widely in the lowland mountains (lower than 1,000 m asl.) which are mainly located in the Central area and Central Highland (Chi, 2012). However in 2000, Vietnamese researchers have discovered numerous E. longifolia populations in the Bai Tu Long National Park, in the Northern part of Vietnam. According to Men et al. (2014), E. longifolia is a light-demanding and drought-tolerant species in Vietnam. The species occurs at an elevation of 200-1,100 m asl. (concentrated mostly on 500-900 m). It can grow in ferrasols, lavisoils, slightly acidic soil of average particle size distribution. In sandy areas, E. longifolia mainly distributes in the natural forest in the high sand dunes or sandy shrub areas. Because of the strict environmental conditions with high precipitation, high sunlight intensity and low clay ratio (<15%) and mainly shoot regeneration, the trees intend to grow in clumps to support each other during their development (Phuong et al., 2013).

Rifai et al. (1975) also showed that this species preferred acid soil with sandy clay texture and silt silica. In Indonesia (Bukit Lawang Forest), Ginting (2010) reported that although E. longifolia can grow in poor nutrient and fertile soil as well as in sandy soil, ash, clay and can survive under the lack of soil nutrients but does not prefer inundated soil. Therefore, the species was basically found in sloping land or hillsides. In the same way, it was noted that the soils in the Riau Province, Indonesia (Hadiah, 2000) were found to be poor in nutrients and E. longifolia distributes normally with other species of Dipterocapaceae in this area. Moreover, Kartikawati et al. (2014) stated that the plants could survive under the poor nutrient conditions of red- yellowpodsol with very acid soil pH which can be normally found in low slopes and ridges so that the taproots of this species are perpendicularly able to penetrate into the soil layers to absorb nutrient and water. In summary, it can be stated that this species is widely distributed on many types of different soils. In addition, E. longifolia populations were observed in areas with an average temperature of between 25oC and 30oC. This species requires an annual rainfall range of 2,000-4,000 mm and 86% humidity (Effendy et al., 2012; EL, 2016).

8 Literature review

(Source: PM, 2012) Figure 2.2 Distribution map of E. longifolia in Asia The yellow dots represent the locations of natural stands, which has been found and described in the above- mentioned literature.

2.2 Importance and use

Until now, a number of scientific research has been carried out on local and international scales to determine bioactive components of E. longifolia. Noticeably, most parts of the plant are well- off in various classes of bio-active constituents including quassinoids, alkaloids, triterpenoids, squalence derivatives, steroid, etc. (Kuo et al., 2004; Ang et al., 2002).

2.2.1 Traditional use

All plant parts of E. longifolia (roots, bark root, stems, flowers, fruits and leaves) have been utilized for traditional medicinal purposes among diverse ethnic groups in Southeast Asian countries for long time that make the plant famous (Hadiah, 2000; Kartikawati et al., 2014). For example in Malaysia, the bark of roots is used to cure fever, mouth ulcers, intestinal worms and as tonic after childbirth and particularly for sexual health of men (Hadiah, 2000). This plant has been considered as Malaysia’s “home-grown Viagra” (Cyranoski, 2005). In Indonesia, local people also use roots as anti-peric, dysentery, while the decoction of bark is used to relieve bone pain that of leaves for washing itches that of roots or stems for curing malaria. In Vietnam, E. longifolia is named as the tree that cures hundreds of diseases and Vietnamese people commonly use its flowers and fruits for treating dysentery (Hadiah, 2000). However, the root is the most valuable plant part and used as folk medicine for the treatment of aches, persistent fever, malaria, sexual insufficiency, dysentery, glandular and health supplements. It is believed 9 Literature review that the plant contains photo chemicals which could stimulate the healing effect of the body and treat a range of frequent diseases (Effendy et al., 2012; Bhat and Karim, 2010).

2.2.2 Modern extracts

In recent years, there has been a significant increase in the demand for this plant species for its tremendous health benefits in both traditional and western medicine (Loc et al., 2018). The process of producing root extracts is based on the traditional method in which roots are cut into smaller pieces, boiled to blood-warm temperature and then used as tea together with honey or sugar to reduce the bitterness. Because of their traditional and scientific benefits more than 200 E. longifolia products are registered by the National Pharmaceutical Control Bureau of Malaysia (Bhat and Karim, 2010). These products made by drying and grinding the roots without adding any other chemicals, are in the form of raw crude powder and now extensively available in health-food markets (Effendy et al., 2012; Aziz et al., 2003). The plant products are also available in form of capsules, pills and tonics, which may either contain raw crude powder or standardized extracts as an additive mixed with coffee or as an alternative for ginseng (Bhat and Karim, 2010; Effendy et al., 2012). The bitter taste from all parts of this plant is the main portion of quassinoid phytoconstituents. The quassinoids are a collection of nortriterpenoids with dynamic pharmacological properties. The concentration and metabolite mechanic in E. longifolia also depends on temperature and geographical factors (Fiascheti et al., 2010). Abubakar et al. (2017) also highlighted the importance of several bioactive components of E. longifolia. Table 2.1 shows some major compounds of this species with their secondary metabolites.

Table 2.1 Summary of several main bioactive compounds and pharmacological effects from E. longifolia

10 Literature review

Bioactive compounds Plant Pharmacological effects References parts Eurycomanone (C20) Roots Increased testosterone production Mahmood et al. (2011)

Eurycomanol (C20) Roots Antimalarial against P. falciparum Kartikawati et al. (2014)

Eurycomalactone (C19) Roots Cytotoxicity against human lung cancer Miyake et al. (2009) Eurycomadilactone (C20) (A-549), breast cancer (MCF- 7) and gastric cancer (MGC- 803) cell lines Cytotoxicity against human HT1080 cells Antimalarial against P. falciparum

Pasakbumin-A, -B, -C, -D Roots Anti-ulcer; Cytotoxicity against human Rehman et al. (2016) (C20) lung cancer (A-549) and human breast cancer (MCF-7) cell lines; Tirucallane- type triterpenes

Longilactone (C19); 6- Leaves, Cytotoxicity against human HT1080; Miyake et al. (2009) Dehydroxylongilactone; Roots Cytotoxicity against human lung cancer 11-Dehydroklaineanone (A-549) and human breast cancer (MCF-7) cell lines

Tirucallane-type triterpenes Stem Anti-cancer activity against ovarian Rehman et al. (2016) leukemia and renal cell lines

Canthin-6-one alkaloids; Plant Cytotoxicity against human lung cancer Mahmood et al. (2011); 9-Methoxycanthin-6-one (barks, (A-549) and human breast cancer Miyake et al. (2009) stems (MCF-7) cell lines; Antimalarial and against P. falciparum; Anti-ulcer roots) activity

Two canthin-6-one Plant Antimalarial against P. falciparum Mahmood et al. (2011); alkaloids; Rehman et al. (2016) 9-Hydroxycanthin-6-one; 9-Methoxycanthin-6-one

2.2.3 Eurycomanone content

Among the above components, eurycomanone is the main compound in improving spermatogenesis by enhancing testosterone production and anti-malaria activity (Chan et al., 1986). However, this compound has not been explored in other plants and it is considered as a unique component of E. longifolia. Moreover, eurycomanone can also support the body against lung cancer cells by inhibiting signaling pathways of proliferation and inflammation (Wong et al., 2012; Hajjouli et al., 2014). Malaysia is a leading country in producing Tongkat Ali products and the quality of freeze-dried Tongkat Ali extract based on several criteria, including eurycomanone level, total protein, total polysaccharide and total glycosaponin. Therefore, the

11 Literature review Malaysian standards (2011) illustrated the concentration of eurycomanone content of E. longifolia in a range between 0.8 and 1.5 w/v (%) (Norhidayah et al., 2015).

In 2013, Mohamad et al. conducted a study by evaluating the impact of extraction process parameters, including extraction time, agitation speed, temperature and root raw material from E. longifolia on the yield of eurycomanone, benzoic acid and gallic acid. Their results showed that the level of eurycomanone reached highest at 45 minutes (extraction time), 400 rpm (agitation speed), 100oC, with around 1.25-1.30 mg/g for the particle sizes (root powder) and nearly 1.20 mg/g for chip roots (E. longifolia at about seven-year-old). Moreover, Nhan and Loc (2017) presented the highest eurycomanone accumulation from cell suspension culture (14th day), which was around 1.70 mg/g dry weight.

Recent research from Vietnam of Dung (2018), who isolated and identified eurycomanone in E. longifolia by LC-MS/MS (liquid chromatography-mass-spectroscopy/mass spectroscopy system) from six different regions presented the different concentrations of eurycomanone. The highest content was found in the Bac Giang province (3.13 mg/g) and the lowest in the Dak Nong province (0.17 mg/g).

2.3 Cultivation and propagation of the species

In recent years, many Asian countries such as Malaysia, Indonesia and Vietnam have been promoting E. longifolia cultivation and propagation programs, which aim to preserve natural habitats, forest genetic resources and wild specimens of this plant. Malaysia is one of the leading countries for these kinds of programs. However, as breeding programs have not been significantly mentioned in the previous studies, the following sub-chapters will mainly discuss the cultivation and propagation of the species.

2.3.1 Cultivation

Malaysia has developed the commercial cultivation of E. longifolia in order to preserve the natural habitats and remaining wild specimens of this plant for its medicinal value. Their roots are often harvested from wild habitats such as rainforests, causing a remarkable decrease in E. longifolia individuals (Idris et al., 2009). However, the land area for propagation programs is also quite limited, so the former plantation programs are reused for E. longifolia plantings (Razi et al., 2013).

12 Literature review Already some plantation trials of E. longifolia have been carried out in the past. In 2001, 2 hectares (3,461 seedlings) were established at 40 m asl. of secondary forest surrounded by oil palm plantations with a population density of 1,750 plants/ha in the site at the Forest Research Institute Malaysia (FRIM) in Pahang. The potential dry yield from the roots was 3.16 ton/ha after five years of cultivation (Idris et al., 2009). In 2003, another trial on the integration of E. longifolia and oil palm was conducted in Sepang. The dry root yield was up to 1 ton/ha after 4 years of planting and a density of 1,900 plants/ha. Noticeably, the integration with oil palms shows a great potential regarding land productivity, land use, farmer’s incomes as well as the sustainable conservation of the local herbal industry and national economy (Khasim et al., 2009). However, the obtained yield does not reflect the actual potential yield of E. longifolia grown under mono cropping systems with a higher density of seedlings. In particular, the investigations also showed that E. longifolia under mono-crop can be grown with 6,700 plants/ha in Pahang. After six months of planting, the mortality rate was approximately estimated to be 10 per cent. The roots were harvested after 5 and 7 years of planting and the potential of dried root yield from this plantation was up to 5 and 8.9 ton/ha (Idris et al., 2009).

Correspondingly, Effendy et al. (2012) showed good results from cultivating E. longifolia in Malaysia. They suggested that the roots could be collected after four years of cultivation and a dry root yield of 3 ton/ha can be expected. In summary, the studies show that the plant can grow well as a monocrop in open planting and as an intercrop together with other plants such as young coconut, fruit trees and in the semi-cleared forests. It is evident that E. longifolia plantations bring many advantages for rural farmers in Malaysia. They are very interested in plantation programs for developing excellent business opportunities and developing the industry of herbal medicinal plants in their country (Idris et al., 2009).

According to the FELDA (Federal Land Development Authority), Malaysia has planted more than 300,000 trees which is one of the major plantation establishments in this country since 2009 (Sajap et al., 2014). Besides, Wan et al. (2010) designed a planting trial of E. longifolia on soils of sandy dunes and beach ridges interspersed (BRIS) in Peninsular (Malaysia). The soils with more than 98% sand showed a low productivity due to a number of physical and chemical limitations such as high soil temperature, low moisture and nutrient content. The previous research showed that the way to green the bare land is to utilize the land for agroforestry and E. longifolia is one of the favorite plants selected for inter-planting among

13 Literature review forest trees. Moreover, fertilizer input had some significant effects on plant height and root biomass yield (Wan et al., 2010).

In Vietnam, the Bai Tu Long National Park is one of the first places where E. longifolia has been propagated successfully by the method of using stem cuttings. Particularly, E. longifolia was researched and cultivated in the land area of approximately one hectare surrounding the National Park (TLNP, 2015). Kartikawati et al. (2014) acknowledged that E. longifolia is one of 41 medicinal plants, which should be regarded with main priority in research, propagation, breeding, development and utilization in Indonesia. As a result, Malaysian researchers have established many E. longifolia plantations in the recent years. Nevertheless, its cultivation activities are still quite unpopular in Vietnam, Indonesia and other Asian countries due to the long period that it takes for cultivating and harvesting their roots.

2.3.2 Propagation and germination of seeds

Most of the methods to propagate E. longifolia are usually conducted through wild seeds, which sprout after approximately one month. However, the propagation of this plant through seeds is difficult. As the embryo is not mature during its dispersion, it takes a long time to germinate and to grow. Seeds with immature embryos do not germinate under normal conditions and they are usually classified as morphologically dormant (Hassan et al., 2012). Therefore, the following sub-chapters are focused on two ways of propagation: by seeds and cuttings.

2.3.2.1 Propagation by seeds

In 2002, Keng et al. (2002) carried out a preliminary study on germination of E. longifolia’s seeds. The authors tested the seed germination in three environmental conditions including: (1) 1:1 soil and sand mixture; (2) jiffy pellets with peat soil and (3) in vitro techniques.

- Condition (1) with a 1:1 soil and sand: The seeds of E. longifolia with intact endocarp sown in the mixture started to germinate after 43 days, then continued to germinate until day 99. During this period, the inhibition and delay of germination could happen due to the high impermeability of the endocarp to either water, oxygen or to both. The proportion of germination was low because of the coat of the seeds was impermeable to water. In other words,

14 Literature review ripe and unripe seeds sown in a soil and sand mixture (1:1) reached 58 per cent and 46 per cent germination rate over a 120 days period.

- Condition (2) with seeds sown in jiffy pellets: The seeds germinated earlier and within a shorter period of time (35 to 85 days) since jiffy pellets consist mostly of peat soil. The pellets are able to maintain higher moisture content with more water absorption by the seeds. As a result, the ripe and unripe seeds germinated with 46 per cent and 29 per cent, respectively.

- Condition (3) in vitro culture: The germination percentage of unripe seeds without endocarp was 53% that of ripe seeds was 30%. Moreover, none of the seeds with intact endocarp germinated and blackening occurred on these non-germinated seeds. The seeds without endocarp did not release any black exudates. It is assumed that the appearing black exudates were mainly phenolic compounds and could contribute to the impermeability of the seed coats to water, which prevents germination of the seeds.

There was the same growth in height pattern in all three above-mentioned environmental conditions. In particular, the seedlings grew rapidly after germination within the first two weeks followed by a steady increase until week 22 after sowing. Between 22 and 28 weeks, the seedlings have a slightly faster growth. The leaf number of seedlings depends on the growth performance of the seedlings (Keng et al., 2002).

Regarding Kartikawati et al. (2014), around 90 per cent of fully ripe seeds will germinate 6 to 8 weeks after being directly sown in the sand. After that, less-than-6-month-old seedlings in the nursery were transferred to poly-bags before being planted in the field. While in vitro propagation is considered as a potential alternative for the propagation of E. longifolia for commercial and conservation purposes, there is only little of research on tissue culture (Hussein et al., 2005). Therefore, these authors focused on developing an effective seedling production by using somatic embryogenesis systems. The results show that the percentage of embryogenic callus formation was moderately low (30%) and the mature embryos required around 31±2 days

15 Literature review of culture to regenerate into a seedling. This period is significantly shorter compared to 120 days for the germination from seeds by the in vitro method (Keng et al., 2002).

2.3.2.2 Propagation by cuttings

Until now, there is little research, which considers E. longifolia propagation by cutting techniques. It is reported that these techniques can be achieved in nursery gardens or cultured in MS (Murashige and Skoog) media (Madke et al., 2014).

In 2005, Hussein et al. (2005) successfully propagated plants in vitro by using bud tip culture. The proportion of plant regeneration is between 30 and 90 percent (5.0 mg l-1 kinetin) with multiple bud formation in the basal MS medium and different concentrations of kinetin. The young roots started to develop after 14 days of culture with a concentration of 0.5 mg l-1 IBA (indole-3-butyric acid). Hassan et al. (2012) carried out another noticeably successful research on in vitro germination. Their study reports that after two weeks, the first new shoots appeared in culture with MS and WPM (woody plant medium) basal media. Then, the first new individual and multiple shoots were observed after three or five weeks in culture.

In 2012, Susilowati propagated E. longifolia by shoot cutting techniques and its adventitious root formation. 7-month-old seedlings were used for propagation by cuttings. The media for propagation were a combination of coconut fiber with chaff rice with the ratio 1:0 (A1 media); 1:1 (A2 media); 2:1 (A3 media). The study showed that after 20 weeks, the length of secondary roots in A3 media was significantly improved (77%) compared to that in other media 44% (A2) and 33% (A1). In 2012, Hassan et al. successfully conducted investigations on propagation of E. longifolia by in vitro root cuttings. This is a trial on producing the individual shoots in MS medium with different IBA (Idole-3-butyric acid) concentrations (0.0, 0.5, 1.0, 2.5, 5.0, 7.5 and 10.0 mg/L). After 8 weeks, the proportion of roots produced and the number of roots per shoot were recorded. The highest number of roots was recorded in media containing 10 mg/L IBA compared to other media. The next trial on acclimatization of plantlets was carried out in a nursery garden. The percentage of plantlets’ survival was observed after three weeks of being transferred into sandy media. Recently, Dasrul et al. (2018) conducted a successful agroforestry model, whereas commercial trees were intercropped with E. longifolia in sandy soil. Their study results revealed that the growth of E. longifolia was nearly similar to the trees planted in enough mineral soil.

2.4 Pests and diseases 16 Literature review The tiger moth (Atteva sciodoxa Meyrick) is one of the major insect pests affecting E. longifolia. The insect feeds gregariously on the shoots, flowers and young leaves (Sajap et al., 2014). Besides, crickets attack the root system. Their number increases at the beginning of the rainy seasons when many young shoots start growing. Another disease called SDS (Sudden Death Syndrome) may also actively affect growing plants, but up to now the pathogen causing the disease has not been identified on E. longifolia (Patahayah et al., 2008; Wan-Muhammad-Azrul et al., 2018).

In Peninsular (Malaysia), E. longifolia was significantly affected by SDS and killed up to 30% of plants in affected populations (Mansor et al., n.d.). Moreover, fungi can attack on young plants and cause the sudden death of some seedlings (Susilowati, 2008). In addition, larvae’s attack of A. sciodoxa on buds is also recorded. It is reported that a biological insecticide (Bacillus thuringienses Berliner) and white oil may be used to control the infestations and scale insects (Effendy et al., 2012; Patahayah et al., 2008). However, there is no specific method introduced before 2014 to control the pest. Sajap et al. (2014) started investigations by using the two indigenous fungi Isaria fumosorosea Wise and Metarhizium anisopliae (Metchnikoff) Sorokin to control lepidopteron (A. sciodoxa) in order to control the harmful insects.

2.5 Genetic knowledge 2.5.1 Assessment of genetic variation

Up to now, varied types of molecular markers have been applied which divide into two major groups, including dominant markers: AFLP, RAPD, RFLP, IRAP, ISSRs, SCoT, BPS, etc. and co-dominant markers: SNP, SSRs, isozymes, microsatellites and so on. A specific marker system will have its advantages and drawbacks compared to another one.

In 2009, Rodrigues and Tam utilized potential molecular markers to isolate sequences containing repeat motifs and to design specific primer pairs in order to describe the genetic diversity and population genetic structure of E. longifolia populations and to use it for QTL (quantitative trait loci) mapping. A total of twelve specific molecular markers were designed for single locus amplifications. Meanwhile, Tnah et al. (2011) developed 18 polymorphic microsatellite markers (Table 2.2). These microsatellites were designed from a microsatellite- enriched library and employed as useful tools for DNA profiling, particularly in cultivar identification and estimating genetic diversity and population genetic structure for this species. The authors identified a total of 157 alleles within a natural population (Semangkok Forest

17 Literature review Reserve, Malaysia) of 32 individuals. Moreover, the number of alleles per locus ranged from four to sixteen whereas the observed and expected heterozygosity ranged from 0.097 to 0.938

(Ho) and 0.095 to 0.916 (He), respectively.

Razi et al. (2013) used RAPD (Random Amplified Polymorphic DNA) markers to investigate the genetic relationships on seven populations of E. longifolia cultivars in order to define whether the originality of these populations may be conserved in uncontrolled cultivation regions. The authors used 60 randomly amplified primers with 10 nucleotides and a total of 320 DNA fragments were amplified with an average of 53.3 fragments per primer. Besides, 71 (22%) out of 320 DNA fragments were found to be monomorphic and the remaining 249 (78%) were polymorphic (Table 2.2). The results detected the original cultivars of this species in uncontrolled fields.

According to the existing literature, there is not much information available on the genetic variation of E. longifolia in Indonesia in comparison to Malaysia. In 2013, Rosmaina and Zulfahmi used RAPD markers to access the genetic diversity and genetic population structure of E. longifolia in Indonesia. However, only 25 individual samples of E. longifolia were harvested from five natural stands of five districts in the Riau province. The results showed that five highly polymorphic primers could be selected from 20 initially tested primers. The five primers produced 44 DNA fragments. In addition, there were 56.80% (PPL); the expected heterozygosity was He = 0.20 (Rosmaina and Zulfahmi, 2013). This result is quite similar to that of Osman’s study et al. (2003) (He = 0.22). However, the genetic diversity of E. longifolia in the study of Osman et al. was lower than that in Susilowati’s research (2008) (He = 0.31 and PPL = 78.57%), which is also based on RAPD markers. From a total of 28 primers, seven random primers (OPY-6, OPY-15, OPY-17, OPY-19 OPY-20 and OPC-7) were found to be highly polymorphic. The authors recommended that all populations included in this study should be preserved because of the high polymorphisms within the populations. Among those populations, the Kuansing population might become the center of diversity for this plant in the Riau province.

So far, the research in Vietnam on genetic diversity of E. longifolia has not been well developed compared to other Asian countries. Currently, Loc et al. (2016) presented the preliminary research on genetic diversity of three populations in Thua Thien Hue province by using RAPD

18 Literature review markers. The results showed that 73 (18.5%) and 321 (81.5%) of the total 394 DNA fragments were estimated to be polymorphic and monomorphic, respectively.

Table 2.2 Summary of the previous studies of genetic diversity in E. longifolia No. of Markers Results populations, References samples 1. SNPs - Identified 51 SNPs - 06 populations Osman et al. (2003) - He = 0.22, PPL = 63.70% - 47 samples

2. RAPD - He = 0.31, PPL = 78.57% - 02 populations Susilowati et al. (2008)

3. Specific molecular - Estimating population diversity - Lack of Rodrigues and Tam markers levels information (2009) - Visualizing QTL mapping

4. Microsatellites - Developed 18 polymorphic - 01 population Tnah et al. (2011) microsatellite markers - 32 samples - He = 0.72

5. RAPD - Total 320 fragments: 71 - 07 populations Razi et al. (2013) monomorphic, 249 polymorphic - 45 samples - PPL = 78%

6. RAPD - He = 0.20, PPL = 56.80% - 05 populations Rosmaina and Zulfahmi - 25 samples (2013)

7. RAPD - 10 RAPD primers - 02 populations Rosmaina et al. (2015) - Shannon index (I) = 0.18, PPL = 38.46

8. RAPD - 10 RAPD primers - 03 populations Loc et al. (2016) - Total 394 fragments: 321 - 13 samples monomorphic, 73 polymorphic

9. AFLP - 12 ALFP primers - 05 populations Sathibut et al. (2016) - Total 262 fragments: 190 polymorphic bands; PPL = 72.52%

10. SSR - 41 SSR primers - 05 populations Lee et al. (2018) - He = 0.45, 232 alleles, 5 alleles - 102 samples per locus

11. IRAP - He = 0.22, PPL = 59.70% - 02 populations Fadilah et al. (2019) - 56 accessions

19 Literature review 2.5.2 Parameters for comparing the utility of markers

The polymorphic information content (PIC), marker index (MI), effective multiplex ratio (EMR) and resolving power (Rp) parameters are useful to estimate the marker information and marker power. The PIC value measures the informativeness of a given DNA marker. The PIC value of each marker ranges from 0 to 1, which is highly informative (PIC > 0.5), reasonably informative (0.5 > PIC > 0.25) and slightly informative (PIC < 0.25). Loci with many alleles and a PIC near 1 are most desirable (Botstein et al., 1980). The EMR of a primer is defined as the product of the polymorphic fragments and the number of polymorphic loci for an individual estimation (Varshney et al., 2007). A given marker system's utility is a balance between the level of polymorphism detected and the extent to which an assay can identify multiple polymorphisms. The marker index (MI) includes a product of information content (PIC) and the effective multiple ratios (EMR). Milbourne et al. (1997) used the Marker Index as a basic parameter for comparing the different primers. Resolving power (Rp) shows the ability of the most informative primers to differentiate between the genotypes. The primers with higher Rp values have a greater capacity to separate genotypes. Resolving power (Rp) is based on the distribution of alleles within the sampled genotypes (Prevost and Wilkinson, 1999).

2.5.3 Genetic variation – Selected Examples

The Malaysian Government has been promoting the development and conservation programs for medicinal plants, especially for E. longifolia (Razi et al., 2013). In 2003, Osman et al. published a study, which showed the genetic diversity of SNP markers for E. longifolia in Peninsular (Fig. 2.3). The study (Table 2.3) presented a high degree of genetic variation among populations. The authors studied six populations of E. longifolia from different parts of Peninsular to assess the level of genetic diversity. As a result, 64% of the loci (P) were polymorphic; expected heterozygosity (He) was 0.22(±0.03); observed heterozygosity H0 corresponds 0.18(±0.03); the total genetic diversity (Ht) and gene diversity within populations at polymorphic loci was 0.29 and 0.22, respectively; a total of 51 SNPs were identified.

Results performed with iso-enzymes, show that the mean heterozygosity of E. longifolia populations was higher compared to other regional, tropical, long-lived tree species. In some populations, there were no significant differences between the observed and expected heterozygosities. A cluster analysis indicated a significant distinction in genetic distance among populations and divided them into two groups. The geographical isolation can be the reason for the limited gene exchange and the differentiation among populations. As only six to ten samples

20 Literature review from one population have been investigated, the number of samples under investigation might have been too small for reliable statistics.

Figure 2.3 Geographic distribution of six E. longifolia populations selected in Peninsular, Malaysia modified after Osman et al. (2003) and https://www.sketchbubble.com/en/powerpoint-malaysia-map.html

Table 2.3 Main indices of genetic diversity in E. longifolia populations studied by Osman et al. (2003)

Population Percentage of Observed Expected polymorphic heterozygosity heterozygosity loci (PPL) (Ho) (He) 1. Johor 70.60 0.18±0.03 0.25±0.03

2. Langkawi 64.70 0.19±0.04 0.23±0.03

3 Terengganu 66.70 0.18±0.03 0.23±0.03

4. Pahang 56.90 0.17±0.04 0.18±0.03

5. Melaka 74.50 0.20±0.04 0.23±0.03

6. Tissue culture 49.00 0.18±0.04 0.18±0.03

Mean 63.70 0.18±0.03 0.22±0.03

21 Material and research methods 3 Material and research methods

3.1 Study Site

3.1.1 Overview

Vietnam has a total area of 331,688 sq.km, including 327,480 sq.km of land and 4,200 sq.km of inland water and 4,000 islands. The country has 63 provinces divided into eight ecological zones, including Northwest, Northeast, Red River Delta, North Central Coast, South Central Coast, Central Highlands, Southeast and Southwest. Regarding the aspect of forestry, its area and quality have decreased for many years due to its unsustainable utilization and management and a high demand of forestry products for socio-economic development. In 1943, Vietnam had 14.3 million hectares of forest, with a forest cover of 43%. However, by 1990, only 9.18 million hectares were left with a forest cover of 27.2%. Then from 1995 to 2005, the forest area was on a constant rise thanks to plantation and natural forest restoration with an average annual increase of 200,000 ha. Up to 2016, the forest area has increased with a forest cover of 41.19%, but in many places, the quality and biodiversity of natural forests is still declining (MARD, 2017).

Thua Thien Hue province is located in the Northern Central Coast region. The total area of forest and forestry land is 288,334.37 ha (211,373.11 ha for forested area and 76,957.26 ha for plantations) of which the forest cover occupies 57.37% (UBND, 2019). There are three types of forest including protection (76,957.28 ha), special-use (93,200.43 ha) and production (11,8176.66 ha) forests (Table 3.1). Meanwhile, protected area management boards, forest protection management boards and forest companies occupy 65% of the forest area. Local communities, households and civil organizations conducted the remaining regions (35%). Moreover, Thua Thien Hue is one of the central provinces with the unique ecological landscape of “interior sandy areas”. It covers 49,289.7 ha and occupies 66.34% of the total area of the sandy land divided into five districts, including Phong Dien (17,010 ha), Quang Dien (8,099 ha), Phu Vang (14,597.7 ha), Huong Thuy (8,536 ha) and Phu Loc (1,047 ha) (Thang and Huy, 2003; Thao et al., 2015).

Against this backdrop, the current research was conducted in Bach Ma National Park (Phu Loc district), Nam Dong (Huong Phu, Thuong Quang, Huong Loc communes), A Luoi (Hong Kim, Huong Lam, Huong Nguyen communes) and Phong Dien (Phong Binh, Phong Chuong, Phong Hoa communes) (Fig. 3.1).

22 Material and research methods

(Source: GFD, 2019 edited by Dien D.). Figure 3.1 Map of the province of Thua Thien Hue located in central Vietnam Including four study sites (A Luoi, Bach Ma, Nam Dong and Phong Dien with different colors). These sites are used for surveying the status of natural populations, estimating leaf morphology and collecting the samples (leaves, seeds and roots).

23 Material and research methods

Table 3.1 Forest status in the province of Thua Thien Hue

Including A Luoi, Nam Dong, Phu Loc (Bach Ma National Park) and Phong Dien districts, which only refer to the study sites. Special-Use forests: conserving nature and its biodiversity; Protection forests: preservation or improvement of a forest and prevention and control of damage to forest by natural or human causes; Production forests: producing both timber and non-timber forest products as well as environmental services.

Forest categories (ha) Management agencies Total % Special-Use Protection Production

Thua Thien Hue province 93,200.43 76,957.28 11,8176.66 288,334.37

Phu Loc district 9,504.90 10,567.00 18,671.88 38,743.92 Bach Ma National Park 37,487.00 96.76

Phong Dien district 35,418.40 7,847.00 23,058.60 66,324.20 Phong Binh commune 326.71 1.92

Phong Chuong commune 1,091.10 6.41

Phong Hoa commune 1,283.00 7.54

A Luoi district 15,322.30 45,933.80 48,430.30 109,686.40 Hong Kim commune 3,804.90 3.46 Huong Lam commune 4,325.90 3.94 Huong Nguyen commune 30,976.00 28.24

Nam Dong district 30,095.88 8,383.74 18,401.00 56,881.40 Huong Phu commune 6,040.80 10.62 Huong Loc commune 6,154.60 10.82 Thuong Quang commune 14,584.39 25.64

Source: Statistics Yearbook of the districts, 2018-2019.

3.1.2 Nam Dong district

Nam Dong district is situated in the south-west of Thua Thien Hue province in central Vietnam. The coordinates of Nam Dong district broaden from 16°00’ to 16°15’ latitude north; 107°27’ to 107°53’ longitude east. It is surrounded in the east by Phu Loc, in the north by Huong Thuy, in the west by A Luoi districts and Quang Nam province and in the south-west by Laos. According to the district statistics (NS, 2018), the total natural area of the district is 64,777.89

24 Material and research methods ha while the forestry land occupies 86% (56,881.4 ha) (Table 3.1). The forests in Nam Dong are classified into three major categories: fractionally logged primary forests, closed secondary forests and open secondary forests (Averyanov et al., 2003).

Nam Dong has a tropical climate with two distinct seasons, a dry season and a rainy season, caused by the monsoon. As can be seen in Fig. 3.2 and Fig. 3.3, there was a significant change in the temperature and precipitation in the district between the years 2013 to 2017. The average annual temperature on the land surface is 25.2°C, with a maximum of 29.6°C. The temperature of the land surface is fluctuating based on the general climatic fluctuation. The land surface receives sun radiation directly and has a higher temperature than that of the surrounding air. The following graph shows the average monthly temperature from 2013 to 2017 (Fig. 3.2).

Nam Dong is a district with the high precipitation in the province of Thua Thien Hue. The largest rainfall appears during the months October, November and December. Fig. 3.3 points out that the average annual precipitation is 310 mm, reaching its peak of 1,047 mm in November while its lowest point of around 63 mm is in March in the period from 2013 to 2017 (NS, 2018). Similarly, the absolute air humidity fluctuates as the temperature does. The annual relative humidity is 86.6% (NDG, 2018).

Topographically, Nam Dong has many streams, which are short and slope in nature. The soil in Nam Dong district is categorized into three main types, including alluvial soil, old and stream or river alluvium; yellow-red soil on magma stone; and yellow-red soil on claystone. The soil characteristics can be seen as favorable conditions for the development of the Nam Dong forest vegetation cover (NDG, 2018).

Two main forest types: evergreen broad-leaved forests and semi-deciduous broad-leaved forests are provided with the forest protection services in Nam Dong. The total forestland areas in Nam Dong can be divided into different categories based on the land purposes. Most of the productive forest areas in Nam Dong belong to protection forest areas, because these forests have a high diversity. The medium and weak woods are used for production purposes (NDG, 2018).

The primary vegetation types comprise of moisture green forest, montane forest and area of scrubby-grass. Unfortunately, Nam Dong forests are seriously disturbed. This disturbance affects the current forest structure and species compositions in the area. Most of the forest products, especially timber products with high economic value, have been extracted from this area.

25 Material and research methods 3.1.3 Bach Ma National Park

Bach Ma National Park (BMNP) is located on the southern edge of Thua Thien Hue province in central Vietnam within the coordinates of 16°05’-16°15’N and 107°43’-107°53’E. It is situated in the central Annamite Mountains and lies on a high mountain ridge that runs western- east from the Laotian border to the East Sea with many high peaks above 1,000 m (An et al., 2018). BMNP was established in 1991, with a total area of 22,031.0 ha. However, in 2008, the park was extended to a total area of 37,487.0 ha, which includes 12,064.8 ha of strict protection zone, 20,234.0 ha of ecological rehabilitation zone and 5,188.2 ha of service and administration zone. This large coverage explains why BMNP has complicated geological and topographical attributes. The park has a high floral diversity, including many endemic species. At present, 2,373 flora species have been recorded from the National Park, representing approximately 17% of the entire flora of Vietnam. Among these, 86 species are listed as endangered in the Red book of Vietnam (BMNP, 2018).

BMNP has two typical primary forests, including tropical evergreen, closed forest (>900 m) and sub-tropical evergreen closed forest (<900 m). Notwithstanding, the former is no longer in its original form due to the pre-1975 war and now is left with only sparse and restored forests. Meanwhile, the latter consists of four types of forests: rich, medium, restored and poor ones. The terrain here is intensively separated and very sloping with upward slopes (more than 40°C), making it quite challenging to access. The tropical climate in BMNP is influenced by two main phenomena, including the northeast monsoon with rain from September to April and southwest drought monsoon from May to September, with an annual temperature of 23.1°C. Bach Ma has the highest rainfall in Thua Thien Hue province. Between 1998 and 2000, the total rainfall reached 10,758 mm/year. The total annual precipitation of the whole region is 3,722 mm, while at the Bach Ma Mount, that figure is up to 7,977 mm. The average humidity in Bach Ma is high, with 86% (BMNP, 2018) (Fig. 3.2, Fig. 3.3).

3.1.4 A Luoi district

A Luoi is a remote mountain district, which is located in the western terrain of North Annamite Mountains. The altitude ranges between 156 and 1,162 m asl. The coordinates of A Luoi district extends from 16°00’ to 16°27’ latitude north and from 107°03’ to 107°30’ longitude east (Toai and Dien, 2017). The community is surrounded in the east by Huong Thuy town, Nam Dong and Huong Tra districts, in the west by Salavan and Se Kong provinces (Lao PDR), in the south by Tay Giang district (Quang Nam province) and in the north by Phong Dien district (Thua

26 Material and research methods Thien Hue) and Dakrong district (Quang Tri province) (ALS, 2017). A Luoi mountains were considered as a major battleground and the areas were seriously affected during the American- Vietnamese war (1955-1975) by herbicide sprayings and bombings (Robert 2016). The total land area is 109,686.40 ha consisting of 81,334.67 ha of natural forest (74%) and 9,461.51 ha of plantation (Table 3.1).

Similar to Nam Dong and Bach Ma, A Luoi has a tropical monsoon climate, which is influenced by the transitions between the North and the South. As can be seen from Fig. 3.2, A Luoi has the lowest average temperature (22.3°C) in comparison to the other regions and it increases rarely up to more than 25°C. A Luoi also has the highest annual average precipitation and the highest total annual rainfall in A Luoi, which are 331.9 mm and 3,983 mm, respectively. The yearly humidity in A Luoi is 89% higher than that in other areas (Fig. 3.3).

3.1.5 Phong Dien

Phong Dien district is located in the north of Hue city with the geographic coordinates of 16°20’ to 16°44’ latitude north and 107°03’ to 107°30’ longitude east. The East Sea connects the district with Quang Dien district in the east, while it borders by Huong Tra town in the southeast, Quang Tri province in the north-west and A Luoi district in the south.

Phong Dien has three terrain types: (1) mountains and hills, (2) delta, lagoons and (3) coastal areas (PDG, 2019). The present study focuses on the interior sandy area, which low natural soil fertility with unique physical properties and light mechanical composition. The amount of clay is less than 15% and the rest is mainly sand. In some places, it is mostly white sand. Therefore, the ability to retain water and nutrients is poor (Cam, 2002). The average annual temperature is around 20°C to 25°C. In the summer, the temperature can reach up to 40.7°C while it may fall down to only 8.8°C in the winter. In comparison to other regions, the temperature fluctuation is quite high (10°C).

27 Material and research methods

35 A Luoi

30 Bach Ma

C) 25 ° 20

10 Temperature ( Temperature

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Months Source: Statistical Yearbook of the districts, 2017-2018.

Figure 3.2 Average temperatures in Nam Dong, A Luoi, Bach Ma and Phong Dien between 2013 and 2017

1200 A Luoi

1000 Bach Ma

800

600

400 Precipitation (mm) Precipitation 200

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Months

Source: Statistical Yearbook of the districts, 2017-2018. Figure 3.3 Average precipitation in different areas between 2013 and 2017

The district experiences a rainy season from September to December, with an annual rainfall during this period accounting for 75% of the yearly precipitation. The rainfall in the remaining eight months only occupies approximately 25% of the annual precipitation, leading to a water shortage in the summer, especially during the years without prolonged rain (sandy soil areas). Compared to other regions such as A Luoi (3,983 mm), Nam Dong (3,722 mm), Bach Ma (3,610

28 Material and research methods mm); the total rainfall in Phong Dien within one year is the lowest with 2,959 mm (including the mountains) (Fig. 3.2, Fig. 3.3).

The air humidity in the mountainous areas in Phong Dien is around 83%, which is similar to that of the mountainous regions. Nevertheless, the rainfall is low in the sandy area during dry periods (hot western winds from Lao PDR) with approximately 65%. The preservation of indigenous forests on the interior sandy regions is critical in the management and protection of forests, as it contributes to the microclimate improvement of the region. In the interior sandy areas of Thua Thien Hue, there are mainly grasslands along water-filled hills and evergreen tropical forests with the following characteristics: small patches of forest on sandy soils highly deposited along water-filled patches. The trees are large and dense, with an average canopy cover of 79% (Cam, 2002).

3.2 Field and lab design

3.2.1 Distributive survey

Transects and plots were designed for estimating the population distribution status and collecting the plant materials of E. longifolia in lowland and highland mountains and sandy area.

3.2.1.1 Baseline survey

Based on the forest status, foresters’ experience and information from local people about the occurrence of E. longifolia in particular sites, the transect survey was carried out with at least five baselines per study site following the randomized and semi-stratified system method:

 21 transects were located on hill mountains (61.86 ha),  6 transects were established in sandy areas (12.38 ha).

The length and the width of the survey baselines are at least 1 km and 10 m, from the lowest to the highest areas (top hill). A GPS tool (Garmin 60CSX, US) was used to record the starting and ending coordinates as well as the tracks in the forests to define the total area of each baseline (Fig. 3.4, App. 3). The information on altitude, longitude and latitude, population status of the species including the number of individuals (trees/saplings), total tree height and DBH (diameter at breast height) were recorded. Other methods including leaf sampling for DNA

29 Material and research methods isolation, fruit sampling for plant propagation, leaf size measurement or collection of natural seedlings were also conducted (Table 3.4). 3.2.1.2 Plot survey

The total number of plots for investigation of seedling density and the distance among E. longifolia (tree/sapling) was 47 rectangular plots (each covering 500 m²-25 m Χ 20 m). From each transect, one to three plots were established with the distance between the plots being at least 300 m from corner to corner, based on the length, topography and the weather condition:

 37 plots were designed on hill mountains (1.85 ha),  10 plots were created in sandy areas (0.50 ha). The plots were located by GPS, which were set out in a North-South direction with the aid of the compass to layout the schemes.

Figure 3.4 Diagram of baseline and plot design Recording baseline N and E UTM co-starting points, baseline N and E UTM co-ending points for estimating the baseline’s length by GPS and the width of each baseline at 10 m.

3.2.2 Plant material 3.2.2.1 Leaf sampling

A number of leaf samples for measuring leaf length, leaf width, number of leaflets, leaf area, leaf stomata and leaf samples for DNA extraction were collected based on the plot and baseline survey. Fully differentiated leaf samples were taken from the oldest main stem in order to avoid an experimental error. For each sample tree, the GPS coordinates, elevation, tree height and 30 Material and research methods DBH were recorded. A number of the samples were used for measuring and extracting representatives for four populations (A Luoi, Bach Ma, Nam Dong and Phong Dien), although there were fewer leaf samples for DNA extraction from Bach Ma because of its DNA quality and narrow population. Moreover, the sampling design and methods were similar for each study site. The list of analysed leaf traits is reported in Table 3.3.

* Leaf area

A total of 440 leaf samples from 88 different mature trees with five replications were randomly harvested to measure the leaf area. In details:

 137 leaf samples (27 trees) were collected from A Luoi (mountain),  70 leaf samples (14 trees) were harvested from Bach Ma (mountain),  125 leaf samples (25 trees) were used from Nam Dong (mountain),  110 leaf samples (22 trees) were taken from Phong Dien (sandy area). Regarding seedlings, 45 leaves from 45 different seedlings (belonging to 9 mother trees) in the nursery garden were collected to estimate leaf area (also used for counting stomatal density) including:  30 seedlings from 6 mother trees of provenances from the moist sites, (1 mother tree from Bach Ma, 2 mother trees from A Luoi, 3 mother trees from Nam Dong);  15 seedlings belonging to 3 mother trees of provenance from the dry site.

Leaf area was measured by using a Flatbed scanner (Canon Line 300, Vietnam) and saved in A4 format together with a scale. The scans are then evaluated by ImageJ 1.51j8 (Wayne Rasband, National Institutes of Health, USA) and leaf area is specified in cm2.

* Leaf stomata

Twelve leaf samples, each collected from a mature individual among 88 trees in four studied areas (3 trees per site), were used to estimate the stomatal density. From each leaf, five leaflets were taken from five different positions (one leaflet in the top, two leaflets in the middle part and two leaflets at the basis, at right and left) (Fig. 3.5). According to Rajmohan (2014), in most cases the lower epidermis has more stomata than the upper surface. Thus, the number of stomata was estimated by preparing slides from a clear nail polish imprint at two opposite middle positions of lower surfaces of each leaflet. Ten images were obtained from each slide with the total area of 0.2378 mm² per picture (at 100 µm scale) with the use of AxioVision SE64 Rel. 4.9.1 Software (Carl Zeiss Microscopy, United State). Finally, the number of stomata was

31 Material and research methods counted for 600 images by ImageJ. Moreover, cross sections of stomata were also obtained in order to estimate the size and shade of stomata and trichrome type. Similarly, 2,250 images were captured from 45 leaves of seedlings to estimate the stomatal density.

Figure 3.5 Five different positions of leaflets from one leaf to estimate the number of stomata through microscope and microscope slides in the paper-clip boxes

Figure 3.6 Measurement of the size of the leaves and the number of leaflets in the field (A) The leaf from mountainous area and (B) the leaf from sandy area (the black and white rule: 15.5 cm, the yellow rule: 20.5 cm)

32 Material and research methods

* Leaf length, leaf width and number of leaflets A total of 505 leaf samples from 101 different trees was collected to estimate leaflet numbers and leaf dimensions (leaf length and leaf width). To be more detailed, the samples were divided as following:  140 leaf samples (28 plants) from A Luoi (mountain),  40 leaf samples (8 plants) from Bach Ma (mountain),  205 leaf samples (41 plants) from Nam Dong (mountain),  120 leaf samples (24 plants) from Phong Dien (sandy area).

Leaf width was measured from tip to tip at the most extended position of the leaf and leaf length was extended from lamina tip to the basis of the petiole (Fig. 3.6).

* Leaf sampling for DNA isolation

A total of 276 individuals from four natural populations of E. longifolia was collected in Thua Thien Hue province during the summer season between July and September 2016. The four populations were A Luoi, Bach Ma, Nam Dong (mountainous areas) and Phong Dien (sandy area) and their geographical locations and sample size are provided in Table 3.2 and Fig. 3.8. The trees which leaves were taken for DNA extraction had been marked with paint before the samples were harvested. From each genotype, around ten leaflets were harvested (from mature trees) and kept in falcon tubes (50 mL) and ice-cold down by liquid nitrogen before being stored at -80°C in the laboratory for further DNA isolation.

Table 3. 2 Population name, geographical location and sample size of DNA samples

DNA samples A Luoi Bach Ma Nam Dong Phong Dien Total

Number of leaf samples (mature 74 23 84 95 261 trees)

Number of leaf samples 71 42 77 79 269 (seedlings)

Number of leaf samples (mother 4 2 4 5 trees) 15

33 Material and research methods 3.2.2.2 Root sampling

Roots were collected from 30 E. longifolia individuals (ca. 2.5 cm in DBH), including 14 mother trees, with 8 samples from A Luoi, 6 samples from Bach Ma, 7 samples from Nam Dong and 9 samples from Phong Dien. Based on the findings of the genetic analysis, the root tissues were collected from different sites that include the polymorphic and non-polymorphic fragments in order to ascertain the representation of samples in the correction assessment between the genetic diversity and eurycomanone content.

Besides, GPS coordinates elevation, tree height, tree diameter, root diameter and tree age were also estimated. Root samples were carefully extracted from the depth of 30 cm downward and placed in plastic bags. The root samples were harvested with starting at root bark up to the center of the root (Fig. 3.7, App. 12). Depending on the tree location, the weight of these root samples varied from 20 to 175 grams. a)

Figure 3.7 a) Before taking the root sample and b) after digging the hole and taking the root sample

34 Material and research methods 3.2.3 Propagation

Eurycoma longifolia fruits were collected from three sites of the secondary forests, namely A Luoi, Bach Ma, Nam Dong and one place from the sandy area (Phong Dien). There were about 150-350 seeds, each from one fruit, in every bunch. These seeds, including young, unripe and ripe ones, were marked with their provenances. A total of 60 fresh and dry fruits from both locations were measured for the fruit dimensions (fruit length and diameter) for further comparison between moist and dried sites. In addition, a minimum of 50 ripe seeds from a minimum of 15 mother trees from all sites was harvested to use for germination. These seeds were labeled with the number of their corresponding mother tree before they were used for the propagation process in the nursery garden. According to our previous research (Duc et al., 2018), for optimal germination conditions of E. longifolia seeds, they should be treated 8 hours at 40-45°C in a drying oven. Then, the seeds were transferred to polybags filled with sand soil mixture and fertilized (10% organic manure: 1% phosphate) before they were covered by soil (1 cm) and placed in the nursery garden. This procedure guarantees a germination rate of 89% after 15 days. One month after germination, the number of seedlings was counted and the plant development was recorded by measuring plant height, collar diameter and the number of leaves.

As can be seen from Table 3.3, there are two groups of seedlings (propagated and natural seedlings): 448 seeds from sandy and 400 seeds from mountainous areas. They were cultivated in a nursery garden and used for the propagation trials. Finally, 309 propagated seedlings from sandy areas (6 mother trees) and 181 propagated seedlings from mountainous regions (4 mother trees) were used to estimate the seedling’s growth. Especially, because there were not enough seedlings from propagation in the nursery garden, we had to collect 297 natural seedlings (<20 cm in height and within 1 m radius from the same mother tree). They were collected and grown in the nursery garden for further DNA isolation. After one year, around ten leaflets from one seedling out of 269 seedlings, were harvested to be used for total DNA isolation.

35 Material and research methods

Table 3.3 Number of different provenances, propagated and natural seedlings and seedlings, for estimating germination, growth pattern and total DNA isolation PSD: Propagated seedlings; NSD: Natural seedlings; SD1: seedlings for estimating germination; SD2: seedlings for growth estimation; SD3: seedlings for DNA isolation and the number of samples

2 sites 4 areas Mother trees PSD NSD SD1 SD2 SD3 1. PB2-1 57 x x 15 2. PB2-3 40 x x 17 Phong Dien 3. PB2-5 61 x x - Dry 4. PB2-6 57 x x 13 5. PB2-9 21 x x 15 6. PB2-10 33 x x 19 Total 269 79

7. M29 76 x x 19 8. M33 18 x x - Nam Dong 9. M89 46 x x 21 10. M61 38 x 19 11. ND6-1 53 x 18 12. ND6-6 24 x -

Moist 13. BM1-3 34 x 22 Bach Ma 14. BM1-6 49 x 20

15. AL5-13 49 x 17 16. AL6-7 24 x 17 A Luoi 17. AL9-1 23 19 18. AL9-2 27 18

Total 164 297 190 Total 433 297 269

3.3 Data collection 3.3.1 Distributional data

Eurycoma longifolia’s population status was identified through the evaluation of species density (trees, saplings and seedlings) in the mountains and sandy areas. With those data on species per hectare (tree/sapling/seedling counts), the distance among the trees inside each plot was measured, the structural patterns of this species was analysed and its population dynamics was investigated. Furthermore, the leaf, root and fruit samples were harvested in both transects and plots together, while recorded by a camera.

The identification of the potential distribution of this species among different areas was performed by the measurement of DBH, the height of trees and saplings and the height of seedlings based on IFRI methods (IFRI, 2004). Trees and saplings were counted by height and

36 Material and research methods diameter at breast height (DBH); that means for trees DBH > 6 cm, for saplings 2.5 cm ≤ DBH ≤ 6 cm, ≥1.5 m in height and <1.5 m in height or DBH < 2.5 cm for seedlings. Measuring tape was used to measure the circumference at breast height (C) of the tree. Diameter (D) was calculated by the following formula: D = C/Л.

The identification of physical soil characteristics was conducted in a plot survey, including measuring the distance from one tree to another tree and defining soil pH-value, soil moisture (relative moisture %), slope steepness by SDL-series data loggers 400 Light meter machine (Extech, Japan) directly (Table 3.4).

Table 3.4 Variables measured of baselines, plots and tools for these measurements

Tools used to Description of variables measured Baselines Plots No. measure variables Dependent variables Independent variables 1 Tree density Field survey x (x) (individuals/ha) 2 Sapling density Field survey x (x) (individuals/ha) 3 Seedling density Field survey x

(individuals/m2 ) 4 Tree height (m) Inclinometer x x

5 Tree DBH (cm) Measuring Tape x x

6 Sapling height (m) Inclinometer x x

7 Sapling DBH (cm) Measuring Tape x x

8 Seedling height (cm) Measuring Tape x

9 Number of leaves Field survey x x

Leaf length/leaf width Measuring Tape x x

10. Leaf area (cm²) Scanner, ImageJ x x

11 Stomatal density Microscope, (stomata/mm²) AxioVision SE64, x ImageJ 12. Slope steepness (%) x

13. Coordinates GPS x x

14. Elevation (m) GPS x x

15. Soil pH-value Extech SDL 400 x

16. Soil moisture, soil texture Extech SDL 400 x

17. Forest status x x

37 Material and research methods 3.3.2 Eurycomanone quantification

Fresh weight was determined from 30 root samples. Then E. longifolia root was dried at 55oC within 5 days until constant weight. After that the root moisture content of each sample was ��−�� estimated by the following formula: Wc (%) = ∗ 100%. ��

The steps of the extraction process are as below: First, the dried samples with approximately between 8 and 17 years old were cut into small pieces by an electrical machine (Power-saw, Makita 4350CT, Japan) and ground to fine powder in liquid nitrogen by using pestle and mortar. The powder was stored at -80oC until used. 0.5 g of this powder was soaked in 10 mL methanol at 60oC with a shaking speed of 100 rpm for 8 h, repeating this step for three times. Next, the extract (30 mL) was filtered entirely and concentrated at 50oC. The precipitate was dissolved in 5 mL methanol (eurycomanone extract) and filtered through Minisart 0.2 µm membrane (Sartorius, Goettingen, Germany) to prepare the sample for HPLC (High Performance Liquid Chromatography).

The HPLC analysis was carried out at ambient temperature with a C18 column (Xbridge, 5 µm, 4.6 Χ 250 mm) (Shimadzu, Japan), flow rate: 0.8 mL/min, run time: 17.5 min, detector wavelength: 254 nm. The stationary phase was silica gel and the mobile step was acetonitrile:

H2O (15:85). An amount of 20 µL aliquot of the sample was injected into the column using a Hamilton syringe. The HPLC was performed on an LC-20 Prominence system (Shimadzu, Kyoto, Japan) with an SPD-20A UV-VIS detector using LC-Solution software (Shimadzu, Japan). All solvents were of analytical grade and were purchased from Merck and Co. Inc. (Darmstadt, Germany). A standard curve of eurycomanone (Santa Cruz, California, USA) with different concentrations from 30 to 250 µg/mL was used for the measurement of the eurycomanone content in the samples (Nhan and Loc, 2018).

3.3.3 DNA isolation

A total of 276 leaf samples from mature trees (including 15 mother tree samples) and 269 leaf samples of seedlings from the nursery garden were used for extracting DNA. Genomic DNA was isolated from leaf tissues using the modified CTAB method (Doyle and Doyle, 1990). 5 g of each mature leaf was assimilated utilizing liquid nitrogen by mortar and pestle into a fine powder. This powder was transferred to Falcon tube (50 mL), added 5 mL of 1.5Χ CTAB and mixed for 30 s, added 2% β-mercaptoethanol (100 µL) and mixed well for 10 s. The samples

38 Material and research methods were incubated at 60oC (1h), infrequent mixing (4-5 times) and then centrifuged for 10 min at 13,500 rpm/4oC to obtain the supernatant. An unbiased volume of phenol: chloroform: isoamyl alcohol (25:24:1) was added to each sample, mixed briefly and centrifuged at 13,000 rpm/4oC for 10 min. The DNA supernatant was transferred to a new eppendorf tube and precipitated by an equal amount of isopropanol at -20oC for 1 h. Next, these tubes were centrifuged at 13,000 rpm at 4oC for 20 min and discarded the supernatant. The DNA pellet was washed by 700 µL ethanol (70%), incubated for 3 min and centrifuged 13,000 rpm/4oC for 2 min, discarded the washing solution, air-dried for 30 min. The DNA precipitate was dissolved with a double distilled water (20 µL) and stored at -20oC. The quality of DNA isolation was determined by agarose gel electrophoresis of 0.8% (w/v) (App. 14). The banding patterns of the gel were observed under an UV light apparatus and the gel documentation was carried out by using the Gel Doc system (Bio-Rad, USA).

39 Material and research methods

(Source: GFD, 2019 edited by Dien D.). Figure 3.8 Geographical distribution map of DNA samples, mother trees and root samples of four (wild) E. longifolia populations in the province of Thua Thien Hue

40 Materials and research methods 3.3.4 RAPD analysis and primer screening

Several research laboratories have been selected the marker systems based on the available equipment, technical knowledge and the ability of the research funding (Collard and Mackill, 2009). In the present study, RAPD, SCoT and BPS markers are selected for evaluating the genetic variation of E. longifolia. They are widely utilized because of cost effectiveness, technical simplicity and high numbers of DNA products per accession. Unlike SSRs, these markers do not require earlier information of DNA sequence. Despite some limitations, e.g. low reproducibility of RAPD, a combination of SCoT, BPS and RAPD markers can reduce the drawbacks (Collard and Mackill, 2009; Panda et al., 2015; Arumugam et al., 2019).

The sequence of the RAPD primers is indicated in Table 3.6. The PCR amplification were performed in 25 μL reactions, containing 1.0 mL of 50 ng DNA template, 12.5 μL master mix solution (2Χ GoTaq®) Green Master Mix, Promega, Madison, USA), 2 μL primer (10 µM), 8.5

μL water-free RNase, 1.0 μL MgCl2 (1.5 mM). The PCR conditions were 95°C for 2 min, 95°C for 1 min, TA (Table 3.6) for 1 min, 72°C for 2 min, for 37 cycles and a final extension at 72°C for 5 min (Table 3.5). After that, the Fragment Analyzer (5200 Fragment Analyzer System, Agilent, USA) was used to separate the PCR products.

3.3.5 SCoT and BPS analysis and primer screening

The SCoT primers depend on the short conserved region in genes surrounding ATG translation initiation codon, while the BPS are single primers with designing from conserved consensus branch point signal sequences within introns. Because SCoT and BPS markers are nearly similar in the length of the oligonucleotides with single 18-mer and 16-mer, respectively. In the present study, the SCoT and BPS primers are combined in one group of the markers that is the SCoT and BPS markers (Collard and Mackill, 2009; Xiong et al., 2011).

The sequence of the SCoT and BPS primers is indicated in Table 3.6. The amplification protocol for SCoT primers was similar to that of RAPD marker. The BPS primers showed reproducible fragments with TopTaq Master Mix with long cycles. Thus, the PCR amplification for BPS primers were conducted in 25 μL reactions, using 1.0 μL of 50 ng genomic DNA, 12.5 μL master mix (TopTaq Mater Mix Kit, QIAGEN GmbH, Hilden, Germany), 2 μL primer (10 µM),

6.5 μL water-free RNase, 1.0 μL MgCl2 (1.5 mM), 2.0 μL BSA. The PCR conditions were 94°C for 3 min, 94°C for 1 min, 49°C for 1 min, 72°C for 2 min, for 45 cycles and a final extension

41 Material and research methods at 72°C for 6 min (Table 3.5). After the PCR, the Fragment Analyzer performed the PCR products.

Table 3.5 PCR pipetting instructions and cycling protocol for GoTaq® Green Master Mix used for SCoT21, SCoT36, OPC02, OPB05 primers and TopTaq Master Mix used for LA2a primer

TA (annealing temperature) is a placeholder for the different annealing temperatures in Table 3.6. “Hold” is the final temperature after all cycling steps, which is held by the cycler until removal. Per sample Temperature Time Cycles (µL) (0C) (min) PCR reaction Cycle step Top Go Top Go Top Go Top Go Taq Taq Taq Taq Taq Taq Taq Taq Initial Distilled water 6.5 8.5 94 95 3 2 1 1 denaturation Master Mix 12.5 12.5 Denaturation 94 95 1 1

MgCl2 1.0 1.0 Annealing TA TA 1 1 43 35

Primer (10 µM) 2.0 2.0 Extension 72 72 2 2 Final BSA 2.0 0 72 72 6 5 1 1 extension

Template DNA 1.0 1.0 Hold 10 10 ∞ ∞ (25 ng/µL)

Total 25.0 25.0 45 37

42 Materials and research methods

Table 3.6 List of RAPD, SCoT and BPS primer sequences tested In which, five primers are selected in the present study, including OPB05, OPC02 (William et al., 1990), LA2a

(Xiong et al., 2011), SCoT21 and SCoT36 (Collard and Mackill, 2009), TA: annealing temperature.

o Primers Sequence 5’-3’ TA ( C) 1. OPA02 TGCCGAGCTG 36 2. OPA05 AGGGGTCTTG 36 3. OPA11 CAATCGCCGT 36 4. OPA17 GACCGCTTGT 36 5. OPB05 TGCGCCCTTC 36 6. OPB10 CTGCTGGGAC 36 7. OPC02 GTGAGGCGTC 36 8. OPO06 CCACGGGAAG 34

9. BPS1 GCGACGGTGTACTGAC 54 10. BPS2 GCGACGGTGTACTAAT 49 11. BPS3 TGAGTCCAAACTAAC 41 12. BPS4 TGAGTCCAAACTGAC 46 13. BPS5 TGAGTCCAAACTAAT 41 14. BPS6 TGAGTCCAAACTGAT 41 15. BPS9 TGAGTCCAAACTAACATA 46 16. BPS10 TGAGTCTAAACTGAC 41 17. BPS11 TGAGTCTAGACTGAC 46 18. BPS13 TGAGTATAGACTGAC 41 19. LA1a GCGACGGTGTACTAAC 52 20. LA2a CGTGCAGGTGTTAGTA 49

21. SCoT3 CAACAATGGCTACCACCG 57 22. SCoT9 CAACAATGGCTACCAGCA 63 23. SCoT15 ACGACATGGCGACCGCGA 61 24. SCoT21 ACGACATGGCGACCCACA 57 25. SCoT27 ACCATGGCTACCACCGTG 57 26. SCoT33 CCATGGCTACCACCGCAG 61 27. SCoT35 CATGGCTACCACCGGCCC 63 28. SCoT36 GCAACAATGGCTACCACC 57

3.4 Data analyses

The data were analysed and assessed using the different methods of statistics in forestry and biotechnology. Particularly, descriptive statistics, including averages, standard deviation, range (minimum and maximum) etc., were applied. All data were imported to Excel 2016 (v16, included XLSTAT v2019.1); Minitab (v17) and XLSTAT were used to analyze the collected data as well as several programs used in the genetic analysis including PAST3, GenAlex 6.5 and STRUCTURE (Fig. 3.9, Table 3.7).

3.4.1 Data of distribution, phenotype, propagation and eurycomanone component

A Kolmogorov-Smirnov (n ≥ 30) or Shaprio-Wilk (n < 30) test was selected to express the difference of a normal distribution of all statistical data. F-test (Fisher) was applied to estimate

43 Material and research methods the variances. One-way ANOVA was utilized to compare the mean differences of the density of saplings and seedling height from different distribution areas, leaf area of seedlings from different provenances. The mean of the density, diameter and height of trees; sapling diameter and height; leaf traits and eurycomanone content among four distribution areas were compared using a Kruskal-Wallis test (nonparametric test). Mann-Whitney test was used to define the differences of germination rate, height and collar diameter, leaf number of seedlings in arboretum condition and for the paired tests after analyzing Kruskal-Wallis test. Based on the sample size and independent samples, z-test or student’s t-test were chosen to define the paired samples.

The association among variables was tested by Pearson’s or Spearman’s correlation coefficient based on the parametric or nonparametric tests. Environmental and geographical factors including longitude, latitude, elevation, slope steepness, forest status, soil pH-value and soil moisture variables were considered as predictor variables (independent variables) while tree diameter, tree height, sapling height, sapling diameter, tree density, seedling density and seedling height as dependent variables (Cochard, 2010). Regression model was included to express the relationship between root water content and elevation, between eurycomanone content with root water content and root diameter. Moreover, the Mantel test was used to test the relationship among (1) morphological parameters and eurycomanone content and (2) between them and geographical and biological factors. All hypothesis tests were performed at α = 0.05 significant level.

3.4.2 Analysis of genetic variation and population structure

Quantitative analyses of data were stored and processed in the Excel database. The RAPD, SCoT and BPS products (bands) were scored by the R program in a binary matrix as either presence (1) or absence (0) of a DNA band in individual lanes. The replication of DNA fragments was considered at 10 percent of random samples, which can be used to assess the reproducibility and to create RAPD, SCoT and BPS matrices (Bonin et al., 2004, Sloop and Ayres, 2011). Molecular data were analysed by GenAlex 6.5 (Blyton and Flanagan, 2012) and POPGENE 1.31 (Yeh et al., 1999). After scoring of DNA bands, some basis terminologies and descriptive statistics should be reported to estimate genetic variation within and among

44 Materials and research methods populations. Several measures characterized genetic variation; hence, some of them can be used for the dominant markers.

3.4.2.1 Measurement of genetic variation within populations

RAPD, SCoT and BPS primers are the dominant markers. The basis of most population genetic analyses is the assumption of Hardy-Weinberg equilibrium (HWE). In a dominant marker system, given two alleles M and m for a specific locus, genotypes MM and Mm show the same phenotype, but the allele frequencies of each sample could not be correctly calculated (Ng and Tan, 2015). The relative frequency of the recessive allele (scored as 0) and the relative frequency of the dominant allele (scored as 1) at a locus is calculated after PCR amplification (Finkeldey and Hattermet, 2007).

The following measures identified genetic variation within a population: The proportion of polymorphic loci, expected heterozygosity and Shannon’s information index.

The proportion of polymorphic loci (PPL): �� = ∗ 100 �� + ��

Where: NP: the number of polymorphic loci;

NM: the number of monomorphic loci. PPL will be calculated regardless of allele frequencies for RAPD. If one locus has two phenotypes, namely presence (1) and absence (0) of a band, then this locus is considered polymorphic.

Expected heterozygosity (gene diversity He, Nei, 1978):

1 − ∑�2 �ⅇ = � �

Where pi: the frequency of the i-th allele at one locus; L: the total number of loci studied.

Shannon’s information index (I, Lewontin, 1972):

� = − ∑ � � � log2 �

Where pi: the frequency of the presence or absence of a given RAPD phenotype (band).

3.4.2.2 Measurement of genetic variation among populations

45 Material and research methods Genetic variation among populations should be estimated through genetic distance and genetic differentiation.

Genetic distance (Nei, 1978):

�̂�� D̂ = −�� [ ] √�̂��̂�

Where Gx: the averages of (2nxJx-1)/ (2nx-1) over the r loci studied;

Gy: the averages of (2nyJy-1)/ (2ny-1) over the r loci studied;

Gxy = Jxy; nx: the numbers of individuals sampled from population X; ny: the numbers of individuals sampled from population Y; 2 Jx: the averages of ∑xi over the r loci studied; 2 Jy: the averages of ∑yi over the r loci studied;

Jxy: the averages of ∑xiyiover the r loci studied; xi: the corresponding sample allele frequencies;

Genetic differentiation among populations:

Nei (1978) and Wright (1951) suggested that Gst and Fst could estimate the genetic differentiation among populations. Wright (1951) developed Fst (F statistics) to summarize a population's genetic structure and defined this statistic for one locus with two alleles. Fst describes the distribution of allele frequencies among populations, while Gst can be derived from multiple loci (Nei, 1978). However, dominant markers do not estimate allele frequencies (Ouborg et al., 1999). Therefore, Weir and Cockerham (1984) offered the alternative estimators ΦPT and Φst that can be calculated from the variance among populations and the variance within-population components. ΦPT and Φst are analog of Fst for dominant markers, generated from an AMOVA (based on genetic distance) (Ouborg et al., 1999; Christina and Debra, 2010; Satya et al., 2015; Ng and Tan, 2015): ΦPT (Φst) = AP / (WP + AP) = AP / TOT.

Where: AP = Estimated variance among populations, WP = Estimated variance within populations and TOT = Total.

3.4.2.3 Calculation of polymorphic information content (PIC), effective multiplex ratio (EMR), marker index (MI) and resolving power (Rp)

For each primer, measures were calculated by the following formula:

46 Materials and research methods PIC = 2Pi (1-Pi) (Roldán-Ruiz et al., 2000), where: Pi = the frequency of occurrence of polymorphic bands in different primers;

EMR = n Χ ß (Varshney et al., 2007), where: n = the average number of amplified fragments by each genotype to a specific system (multiplex ratio); ß is estimated from the number of PB polymorphic loci (PB) and the number of non-polymorphic loci (MB); ß = ; PB+MB

MI = EMR Χ PIC;

Rp = ƩIb (Prevost and Wilkinson, 1999), where Ib represents the informative fragments on a scale of 0/1 by formula: Ib = 1 - (2 Χ ǀ0.5-piǀ), where pi is the percentage of accessions containing the ith band.

3.4.2.4 Analysis of Molecular Variance (AMOVA), Principle Coordinate Analysis (PCoA), Cluster Analyses (STRUCTURE and dendrogram)

The percentage of polymorphic loci (PPL), Shannon index (I) and He (expected heterozygosity) were calculated for genetic diversity within each population. An Analysis of Molecular Variance (AMOVA) approach (Excoffier et al., 1992) was used to observe variation among and within-population components. PhiPT (ΦPT) and Φst for pairwise genetic differentiation are relatively unbiased estimator and independent of the population numbers. Genetic distances among individuals were also calculated based on Nei and Li (1979). All analyses were performed using GenAlex 6.5 (AMOVA) software. Principle Coordinate Analysis, Principle Component Analysis (PCoA, PCA) and cluster analyses (STRUCTURE and dendrogram) are the multivariate analytical techniques, which are popularly used in the analysis of genetic variation irrespective of the molecular marker dataset. PCoA is recommended over PCA for a lot of missing data from dominant markers. PCoA is a scaling or ordination method that begins with a binary matrix of similarities or dissimilarities among a set of genotypes (Rohlf, 1972; Mohammadi and Prasanna, 2003). Thus, PCoA was performed to highlight the genetic relationship among populations based on the similarity coefficient of matrix data.

Two types of clustering approaches, namely distance-based (a specific clustering algorithm - dendrogram) and model-based methods (standard statistical methods - Bayesian approach) have been utilized for estimating genetic structures (Johnson and Wichern, 1992). Several limitations of distance-based methods from the model-based clustering methods were improved and innovated based on Bayesian statistics. Thus, the population genetic structure was assessed by

47 Material and research methods using the Bayesian clustering approach implemented in the software STRUCTURE v2.3.1 (Pritchard, 2007). The genetic matrix was specified for dominant markers in the program instructions. All 276 DNA samples from mature trees and 269 DNA samples from seedlings were analysed without prior population information; STRUCTURE is expected to produce unbiased findings. K values were tested from one to ten with ten independent runs for each K. The length of the burn-in was 500,000 steps followed by 106 iterations. The K value with the highest posterior probability was identified in this way and by using the ΔK statistics that quantify the second-order rate of change of the likelihood function concerning K (Evanno et al., 2005).

Among hierarchical clustering methods, Ward’s minimum variance method was used to present genetic similarity relationships among individuals in the present study. Ward’s method depends on the squared Euclidean distance, also known as the sum of squares, divided into two groups, like the PCoA and STRUCTURE program. Furthermore, the Mantel test using PAST3 v3.22 (Hammer, 1999) was implemented to define the correlation between geographical distances among populations, elevation factor, morphological traits and genetic matrix, generated 9999 random permutations. We used the Jaccard similarity index for the genetic matrix, the Gower similarity index for morphological matrix and the Euclidean similarity index for geographical and elevational matrices.

48 Materials and research methods

Figure 3.9 Research Design Framework

49 Material and research methods

Table 3.7 Summary of the methodologies of data collection Data types, data analysis and tools/software, which are relevant to the four objectives.

Objectives Data Types Data Analysis Tools/ Software (O) - Primary and derived data: - Analysis in Lab/ Field - Excel + Population density and structure - Multivariate data analyses - XLSTAT (trees, saplings and seedlings) (Kruskal-Wallis test, One-way - Minitab v17 + Soil moisture, soil pH-value, slope ANOVA, Mann-Whitney test, z-test - ImageJ steepness and t-test) - Axio SE64 + Field observation (insert, storm, - Pearson or spearman’s correlation - MapInfo logging...) coefficient O1-O3 + Leaf morphology and anatomy: - Coordinates, location, individual mature trees and seedlings numbers + Seedlings‘ growth + Distribution map - Secondary data: data of forest status in the past, related studies

- Derived data: - Banding patterns of gel from DNA - Modified + DNA extraction isolation/PCR reactions CTAB/PCR + RAPD, SCoT and BPS amplification - Polymorphic Information - Gel Doc system + Genetic diversity Component (PIC), Effectiveness (BioRad) + Genetic variation across generations Marker Ratio (EMR), Marker index - Fragment O4 (MI), Resolving power (Rp) Analyzer - Genetic diversity parameters: - GenAlex 6.5 genetic variation (He), percentage - Excel of polymorphic loci (P), genetic distance, Shannon index (I), genetic differentiation (Φst, ΦPT ) - Population genetic structure - Structure clustering analysis - STRUCTURE Variation within population and - Principal coordinate analysis software between populations (PCoA) - PAST3 - Genetic distances/ relationships - Analysis of Molecular Variance O5 - Correlation between genetic and (AMOVA) environmental and morphological - Dendogram clustering factors - Mantel test

- Primary and derived data: - Multivariate data analysis - Minitab/ Excel + Root water content, root diameter, tree (Regression model, Kruskal-Wallis - PAST3 age, tree diameter, tree height test, Spearman’s correlation O6 + Eurycomonanone concentration coefficient) + Correlation between genetic matrix - Mantel test and eurycomanone content

50 Results 4 Results

4.1 Population distribution status of E. longifolia

4.1.1 Terrain and soil

As the investigated area in the mountains was larger than that in the sandy area, 21 transects were located in the mountains and six transects in the sandy area (App. 1). The highest elevation where E. longifolia appears is in transect AL5 (1,005 m asl.) and the lowest altitude belongs to all the sandy baselines with around 30 m asl. The mountains are topographically steep with high slope steepness, especially in baselines ND3 and AL5, which have a steepness with an average of 40o and the lowest 20o in baseline ND5 and AL6. Among three types of forests (rich, medium and poor), there are only poor forests with shrubs and small trees in the sandy area.

Table 4.1 and App. 2 show the average of the soil moisture content, soil pH-value through soil field tools and forest status (secondary data sources). Soil differs by soil pH-value and soil moisture between mountain and sandy areas and is highly different between baselines with different steepness and forest status. The average soil moisture is 39.42% with decreasing along the elevation gradient (46.49% in the mountains and only 14.67% in sandy soil sites), while soil pH-value is 6.8 and 5.87, respectively.

Table 4.1 Soil factors and forest status MA: mountainous area, SA: Sandy area, R: Rich, M: Medium and P: Poor

Slope steepness Soil pH-value Soil moisture Forest status Areas (0) (%) (%) Average Range Average Range Average Range R M P ± Stdv ± Stdv ± Stdv

MA 30.0±5.9 20.0-0.5 6.8±0.7 6.0-8.7 46.5±13.7 28.0-76.7 9.5 76.2 14.3

SA 5.0±0.0 5.0-5.0 5.9±0.2 5.5-6.1 14.7±4.9 10.0-23.0 100.0

51 Results 4.1.2 Population distribution

4.1.2.1 The topographic distribution of E. longifolia

Eurycoma longifolia has an extensive range of topographic distribution areas based on the environmental conditions in each baseline (Fig 4.2, Fig. 4.7). This species is found from the ground up to an elevation of 1,000 m asl., but it mostly appears at the altitude range of 300-700 m (Fig. 4.1). While most plants are found at an elevation of between 300 and 700 m asl. and they are extremely rare at above 1,000 m.

10 Percentage Percentage individuals of 5

0 30-40 75-300 300-500 500-700 700-1000 Elevation classes (m, asl.)

Figure 4.1 Topographic distribution of E. longifolia based on elevation (m) In total N = 1,546; blue color: sandy area (N = 215) and green color: mountainous area (N = 1,331)

4.1.2.2 The density of trees and saplings

Based on the gathered data from 27 baselines with 74.24 ha in total, E. longifolia trees are mostly found in these transects (excluding some baselines without any individuals) with 1,546 individuals ranging from 2-62.50 individuals ha-1 to an average of 23.02±14.60 individuals ha- 1. The density of trees ranges from 0.48-39.18 trees ha-1 with an average of 9.49±10.63 trees ha- 1 only appearing in the mountains (Fig. 4.2). Sapling density in the mountainous and sandy areas ranges from 0.77 to 34.93 and from 12.98 to 26.64 saplings ha-1 with an average of 14.96±9.17, 18.00±5.49 saplings ha-1 (Fig. 4.3). There is no statistical difference in the sapling density between the mountains and the sandy area.

52 Results In addition, the densities correlate positively with some factors, i.e., soil moisture and slope steepness. For example, the tree density increases with the slope steepness and the soil moisture. The seedling density correlates positively with the pH of the soil (App. 4).

In general, the tree density is lower than the sapling density along each baseline. Almost all baselines from Bach Ma have a much higher number of saplings compared to the number of trees. In contrast, only one baseline AL4 in A Luoi and two baselines ND3 and ND7 in Nam Dong out of the 21 baselines have a tree density, higher than the sapling density (Fig. 4.3).

Among the four different areas investigated, Nam Dong is the leading site with the highest tree density (DBH ≥ 6 cm) and tree height at 14.49 counts/ha and 8.5 m, followed by A Luoi at 9.10 counts/ha and 7.6 m. Bach Ma has the lowest tree density, with only 1.25 counts/ha and 7.1 m tree height (p < 0.05, Kruskal-Wallis test). There is no significant difference in the tree diameters among these regions (H = 3.59, p > 0.05, Kruskal-Wallis test). Although the number of saplings across the four areas is the same (F = 0.99, p = 0.416, one-way ANOVA), the sapling diameters in Nam Dong and A Luoi (ca. 3.9 cm) are much higher than that in Bach Ma and Phong Dien (ca. 3.3 cm) (H = 73.33, p = 0.000, Kruskal-Wallis test). Nam Dong has the tallest saplings with an average of 4.02±1.77 m whereas the lowest mean sapling height is accounted for by Phong Dien (H = 200.06, p = 0.000, Kruskal-Wallis test) (Table 4.2).

Regarding elevation, Nam Dong and Bach Ma have the similar percentages of individuals at an altitude of 75-400 m (around 44%), while A Luoi trees are only found at >400 m asl. At the elevation of 400-700 m, the tree density (including saplings) of these three mountainous areas reached approximately 50%. Conversely, there is a significant difference compared to the 700 to 900 m altitude. A Luoi occupies 32% of the total number of individuals in this region, while 1.3% of that is recored in Bach Ma (2 individuals at 715 m). Surprisingly, there are no individuals to be discovered at the elevation above 730 m in Bach Ma. Even at the 900-1,000 m elevation, 14.4% of E. longifolia individuals are found and only one individual (0.28%) could be detected above 1,000 m in A Luoi (Table 4.2).

53 Results

(Source: GFD, 2019 edited by Dien D.).

Figure 4.2 Distribution map of E. longifolia Jack in the province of Thua Thien Hue Red points: individuals at tree class; blue points: individuals at sapling class

70 Sapling density Tree density 60

Density (counts/ha) Density 10

AL1 AL2 AL3 AL4 AL5 AL6 AL8

PC6 PC7

PB1 PB3

PH4 PH5

ND2 ND3 ND4 ND5 ND6 ND7

BM1 BM3 BM4 BM5 BM6

ND1.2 ND1.1 ND1.3 Baselines Figure 4.3 E. longifolia occurencies along the observation baselines A Luoi: AL1 - AL8; Bach Ma: BM1 - BM6; Nam Dong: ND1 - ND7; Phong Dien: PB1 - PH5

54 Results

Table 4.2 Tree density (including saplings), tree size and the frequency of individuals based on elevation in four different areas Indl: Individuals; (*) Kruskal-Wallis test; (**) One-way ANOVA; a, b, c: Mann-Whitney test

Areas A Luoi Bach Ma Nam Dong Phong Dien p-value Tree density 0.002 9.10±10.08b 1.25±0.82c 14.39±11.67a - (counts/ha) (*) Sapling density 0.416 15.84±11.87 9.96±9.55 17.05±6.06 17.98±5.49 (counts/ha) (**) Tree diameter 0.166 9.13±2.73 9.21±4.4 9.65±3.24 - (cm) (*) Tree height 0.003 7.61±3.23b 7.19±2.7b 8.47±3.4a - (m) (*) Sapling diameter 0.000 3.90± 1.05a 3.26±0.91b 3.81±1.02a 3.27±0.82b (cm) (*) Sapling height 0.000 3.88±1.57a 3.16±1.29b 4.02±1.77a 2.41±0.93b (m) (*) Elevation (m) Indl % Indl % Indl % Total % 75-200 - - 30 19.74 39 4.85 69 5.18

201-400 - - 42 27.63 302 36.61 344 25.85

401-700 191 53.80 78 51.32 403 48.85 672 50.49

701-900 112 31.55 2 1.32 80 9.70 194 14.58

901-1,000 51 14.37 51 3.83

>1,000 1 0.28 1 0.08

Total 355 152 824 1,331 100

4.1.2.3 Morphometric parameters

The diameter-based class distribution among all trees and saplings with a DBH ≥ 2.5 cm in the surveyed baselines is summarized in Table 4.3. The frequency of distribution of tree and sapling DBH and height is roughly normally distributed (Fig. 4.4). The average tree diameter is 9.5±3.2 cm and it ranges from 6.0 cm to 24.8 cm. The distribution of trees in the 6-15 cm-diameter class is 29.98% (Table 4.3). Trees with a DBH > 15 cm represent only 2.01% of the total number of individuals. The two largest trees with a diameter of around 25 cm, grow in the baselines of BM6 (Bach Ma) and ND1.3 (Nam Dong). The mean tree diameter correlates positively with the mean tree height and the mean sapling diameter with the mean sapling height with r2 = 77%, 54% and 48%, respectively (Fig. 4.6, App. 4).

Regarding tree distribution in the height classes, nearly half of the trees are in the <5 m height class (44%), whereas trees in >15 m height class are found to be rare (1.4%). The mean tree height is 8.2±3.4 m, ranging from 1.5 m to 23 m. The tallest trees are discovered in the baselines

55 Results of ND7 (23 m, Nam Dong) and AL4 (22 m, A Luoi). Trees at >6 cm diameter class are absent in the sandy area and there is only 1% out of the total number of 510 trees (5 individuals, 5.5- 6.5 m in height) growing in this area (Fig. 4.4, Table 4.3). Other significant predictors of the mean tree height and diameter are variables including the soil moisture, the soil pH-value, the slope steepness, the elevation and the forest status (App. 4).

Although there is no difference in the sapling density between the mountains and the sandy area, the height and diameter from the former (at 3.9 m and 3.8 cm), are significantly higher than those from the latter (at 2.4 m and 3.3 cm) (p = 0.000, Mann-Whitney test, Fig. 4.5). Moreover, there is a strong correlation between the mean sapling height and the mean sapling diameter (Fig. 4.6).

a) b)

Figure 4.4 Distribution of trees and saplings in diameter (a) and height classes (b)

56 Results

(cm) sapling diameter (m) sapling height 6 8 p = 0.000 p = 0.000 7 5 6

5 4 4

3 3 2

2 1 Mountainous area Sandy area Mountainous area Sandy area

Figure 4.5 Differences of sapling diameter and height between sandy and mountainous areas

(m) (m) mean tree height*mean tree diameter mean sapling height*mean sapling diameter 16 5 12 4 8

2 3 4 r = 0.77 r2 = 0.53 p = 0.0001 2 p = 0.004 0 0,0 2,5 5,0 7,5 10,0 (cm) 3,0 3,5 4,0 4,5 (cm) (m) (cm) mean tree height*slope steepness mean tree diameter*mean elevation 16 10,0 12 7,5 8 5,0 r2 = 0.83 r2 = 0.75 4 2,5 p = 0.000 p = 0.000 0 0,0 0 10 20 30 40 (%) 0 200 400 600 800 (m)

Figure 4.6 The increase of tree height and diameter with elevation, slope steepness and the positive correlation of trees and saplings between height and diameter

57 Results

A B

Figure 4.7 E. longifolia in mountainous area (A) and sandy area (B)

Table 4.3 The diameter and height classes of trees and saplings MA: mountainous area, SA: sandy area

Diameter % individuals Height class % individuals Individuals Individuals class (cm) MA SA (cm) MA SA

2.5-5.9 1,060 54.7 13.9 <5.0 886 43.7 23.7

6.0-9.9 316 20.4 0.0 5.9-9.9 510 32.7 1.0

10.0-4.9 139 9.0 0.0 10.0-14.9 129 8.3 0.0

>15.0 31 2.0 0.0 >15.0 21 1.4 0.0

Total 1,546

4.1.2.4 Natural regeneration

Seedling density (plants ≤1.5 m in height) is surveyed in 47 plots within an area of 23,500 m2. Among these, 37 plots are located in the mountains and the other 10 plots in the sandy area. 1,752 seedlings in the mountainous area and 99 seedlings in the sandy area are recorded throughout the survey. The medium density of seedlings in the mountainous area is 0.095 seedlings/m2 and only 0.02 seedlings/m2 in the sandy area. Therefore, the seedling counts in the mountains are higher than those in the sandy area despite the principal (positive) predictor in the soil pH-value regression model (W = 10023,0, p = 0.0005, Mann-Whitney test). The average height of the seedlings from the mountains is with 0.46 m significantly higher than that of the seedlings from the sandy area with 0.35 m (F = 6.26, p = 0.016, one-way ANOVA, Fig. 4.8). It is lowest in the plots of P4, P39, P42, P46, P47 (the sandy area) with 0.01 seedlings/m2 and highest in P25 (mountains) with 0.56 seedlings/m2 (App. 2). Figure 4.9 shows the number of individuals in different frequency classes. The highest number of individuals is recorded in the

58 Results height class of 0-20 cm with 596 seedlings, followed by the 20-40 cm, 40-60 cm, 60-80 cm, 120-149cm and 80-100 cm height classes with 380, 355, 235, 118 and 109 individuals. The lowest amount is recorded in the height class of 100-120 cm with only 58 seedlings.

mean seedling height (cm) seedling density (count/m2) 0,14 0,50 p = 0.016 0,12 p = 0.0005

0,45 0,10

0,08 0,40 0,06

0,35 0,04

0,02 0,30

0,00 Mountainous area Sandy area Mountainous area Sandy area

Figure 4.8 Comparison of seedling density and height between mountains (N = 1,752) and sandy areas (N = 99)

With the tests of Mann-Whitney (seedling density) and one-way ANOVA (seedling height)

er of seedlings per ha per seedlings of er

Num

Height classes (cm)

Figure 4.9 Distribution of seedling number along the height classes (cm)

59 Results 4.2 Phenotype 4.2.1 Leaf morphology and anatomy of mature trees 4.2.1.1 Leaf area

This study investigates four areas, including A Luoi, Bach Ma, Nam Dong from mountains and Phong Dien from sandy area. The mountainous area is often moister, cooler and shows a lower radiation than the arid environment in sandy area. Thus, we may use the terminologies “moist site” to refer to the mountainous area and “dry site” for sandy area. Several studies have been reported that leaf temperature and water availability are two main variables correlating with leaf size (Poorter and Rozendaal, 2008; Wright et al., 2017). However, the present research was so far unable to estimate these two parameters.

A total of 440 leaves from 88 different trees was collected in four regions. The leaf area of mature trees differs across sites (H = 90.896, p = 0.000, Kruskal-Wallis test). A Luoi has the largest area of the leaves, at 501.7±115.1 cm2 on average, followed by Bach Ma and Nam Dong with the average areas of 476.7±118.7 cm2 and 469.2±109.4 cm2, while Phong Dien has the smallest leaf area (at 355.7±95.5 cm2) (Fig. 4.10, Fig. 4.11). Nam Dong has the highest tree density and the most abundant tree size and Bach Ma has the lowest tree density, while Phong Dien has no individual at the tree size class. Regarding sapling size, Bach Ma and Phong Dien have a smaller size than A Luoi and Nam Dong (Table 4.2). There is a positive correlation between leaf area and altitude (r2 = 0.38, p = 0.000), meaning that the leaf area increases along the elevation gradient. With regard to moist and dry sites, Fig. 4.12 shows that the leaf area from the moist site are much larger than that from the dry location, which ranges between 253.2 cm2 and 813.4 cm2 (humid) and from 184.0 cm2 to 525.8 cm2 (dry) (W = 13573,5 , p = 0.0000, Mann-Whitney test).

60 Results

700 a a 600 b

) 500 2 c 400

300 Leaf area (cm area Leaf

200

100

0 A Luoi Bach Ma Nam Dong Phong Dien

Figure 4.10 Leaf area of E. longifolia mature trees Different letter(s) represent a statistically significant difference with p < 0.05 (Kruskal-Wallis and Mann- Whitney tests). Error bars indicate the mean ± standard deviation.

(4) (1) (1) (1)

(2) (2) (3) (3)

Figure 4.11 Leaf morphology in different areas: A Luoi (1), Nam Dong (2), Bach Ma (3) and Phong Dien (4)

61 Results

Boxplot of leaf area of mature trees 800

700

) 600

(

a 500

e 400

300

200

Moist site Dry site

Figure 4.12 Leaf area differences in moist (mountainous) and dry (sandy) sites (p < 0.05, Mann-Whitney test)

4.2.1.2 Stomatal density

The different environmental conditions, such as water availability, temperature, light exposure or CO2 concentration, may affect stomatal density and size (Xu and Zhou, 2008; Salisbury, 1927). Besides, stomatal density declines with the height to maintain the minimum of leaf water potential. The lower stomatal density in tall trees may restrict the carbon uptake (Yoder et al., 1994).

A total of 600 images were captured from 60 leaflets (12 individuals), which were harvested from four different sites. Within each of the sites, stomatal densities present a wide range of 88.3-201.9 stomata/mm2 in A Luoi, 46.3-214.5 stomata/mm2 in Bach Ma, 105.1-214.5 stomata/mm2 in Nam Dong and 197.6-386.9 stomata/mm2 in Phong Dien (Fig. 4.13, Fig. 4.14). In comparison, Phong Dien has twice as many stomata as the remaining sites, with an average of 284.46±38.9 stomata/mm2 (H = 352.55, p = 0.000, Kruskal-Wallis test). Particularly, the number of stomata/mm2 from A Luoi and Bach Ma are the same but they are less than that from Nam Dong, which is a site in the mountains.

62 Results The stomatal density of the mature trees has a negative relationship with tree height, stem diameter and elevation. It means that the mature trees have larger stem diameters and heights, which are located in high elevation those trees show lesser the number of stomata. Stomatal densities (stomata/mm2) are significantly higher (W = 78814,0, p = 0.0000, Mann-Whitney test) in dry-grown leaves than in moist-grown leaves and there are significant differences in total number of stomata per leaf between dry- and moist-grown leaves as well (Fig. 4.15, Table 4.4).

Beside the densities of stomata, its shape also reflects the adaptation of stomata to different habitat conditions. In the study of Ichie et al. (2016), the stomata occur with different growth forms namely, flat, mound and pit types. Flat and mound types appear in the understorey, sub- canopy, canopy, or emergent layers, while pit type mainly appears in canopy and sub-canopy of the tropical forests. Additionally, in the present study the cross-sections also show trichomes, which may play an important role in the defense mechanisms of the plants (Levin, 1973) and which are epidermal specializations consisting of one or more cells that protrude from the epidermis (Bock and Norris, 2016). Therefore, besides counting the number of stomata per mm2, the study also checks some samples from moist and dry sites for the stomata types and its trichomes. The results show that the leaves from the moist site present the flat-type (Fig. 4.16b), whereas those from the dry site show the pit-type with the stomata tending to be more inside the epidermal cells (Fig. 4.16a). In the lower leaf surface, E. longifolia leaves also have some trichomes or small hairs measuring around 5-10 µm (App. 13). The trichomes appear both in mountains and in the sandy area.

Table 4.4 Stomatal density (stomata/mm2) and the number of stomata per leaf for moist- and dry-grown E. longifolia trees

*: Mann-Whitney test

Average leaf area Habitats Stomata per mm2 Stomata per leaf (cm2)

Mountains 137.94*a 6,113,200*a 444±89*a

Sandy areas 284.4b 7,771,875b 278±73b

63 Results

350

a )

2 300

250

200 b c c 150

100

Stomatal density Stomatal(stomata/mm density 0 A Luoi Bach Ma Nam Dong Phong Dien

Figure 4.13 Stomatal density differences of mature trees at the four distribution areas

Different letter(s) represent a statistically significant difference with p < 0.05 (Kruskal-Wallis and Mann- Whitney tests). Error bars indicate the mean ± standard deviation.

Boxplot of stomatal density of mature trees

350

)

m 300

o 250

(

n 200

m 150

100 Moist site Dry site

Figure 4.14 Stomatal density differences of mature trees in moist and dry sites

64 Results

Figure 4.15 Different distribution of stomatal density in moist (mountainous) (a) site and dry (sandy) site (b) The red arrows indicate stomata and the yellow arrow indicate epidermal cells.

65 Results

Figure 4.16 Cross section of the leaves: a) pit-type stomata in dry site and b) flat-type stomata in moist site

4.2.1.3 Correlation between leaf area and stomatal density of mature trees

There is a generally considered strong correlation between the leaf area and the number of stomata per mm2 with r2 = -0.59 (p = 0.000). The direction of the relationship is negative, meaning that greater leaves come along with lower stomatal density (Fig. 4.17).

350

) ² 300 2 m r = -0.59

/ p = 0.000

a t 250

s 200

(

n 150

t 100

S 50

0 200 300 400 500 600 700 800 Leaf area (cm²)

Figure 4.17 Negative correlation of leaf area with stomatal density of mature trees Scatter plot between leaf area and stomatal density and Spearman’s correlation coefficient for nonparametric test

66 Results 4.2.1.4 Leaf length, leaf width and number of leaflets

505 leaves were collected from 101 different trees to measure the leaf sizes in the field. Leaves from Bach Ma are the longest with 71.0±7.4 cm followed by A Luoi (64.4±10.8 cm) while the leaf length from Phong Dien and Nam Dong are the same (61.6±9.2 cm and 59.4±11.6 cm). Bach Ma has the highest number of leaflets with an average of 36 leaflets per leaf while A Luoi and Phong Dien have around 31 leaflets per leaf. However, the average leaf width from A Luoi is the most comprehensive (20.2±4.1 cm). For the remaining collection sites, leaf width ranges from 16.8 cm (Phong Dien) to 17.7 cm (Nam Dong) (Table 4.5).

Table 4.5 Average ± standard deviation, range of leaf length, leaf width and number of leaflets of E. longifolia in four different sites and statistical differences H-value: the test statistic for the Kruskal-Wallis test; a, b, c: Mann-Whitney test)

Sites A Luoi Bach Ma Nam Dong Phong Dien p-value H-value Traits Leaf length 64.4±10.8b 70.5±7.4a 61.6±9.2c 59.4±11.6c 0.000 46.40 (cm) (39.5-95.0) (56.0-91.0) (38.0-87.0) (39.0-92.0)

Leaf width 20.2±4.1a 17.3±2.6bc 17.7±2.7b 16.8±3.5c 0.000 61.47 (cm) (13.0-33.0) (13.0-24.0) (11.0-26.0) (10.0-28.0)

Leaflet 31.3±5.2c 35.9±5.1a 33.2±5.0b 31.3±3.9c number per 0.000 29.27 (18.0-39.0) (26.0-47.0) (20.0-50.0) (18.0-42.0) leaf

Stem 6.5±3.7a 4.9±1.6b 6.4±3.7a 3.8±1.2c diameter 0.837 0.04 (2.4-17.8) (2.9-7.7) (2.2-20.4) (2.5-6.6) (cm)

Height 5.4±2.9b 4.8±1.4b 6.5±3.3a 2.6±0.8c 0.001 10.64 (m) (1.7-14.0) (2.2-6.4) (2.0-17.0) (0.8-4.0)

Similar to the results of leaf area and stomatal density, leaf length, leaf width and the number of leaflets also differ between the sites in the mountain and in the sandy area (p = 0.0000; Mann- Whitney test). Leaf width ranges from 11 cm to 33 cm for the moist site while plants from the dry places form leaves that are 10-28 cm large. Besides, the number of leaflets of these sites varies between 31 (in the dry location) and 33 (in the moist site) leading to an average leaf length of 63.6±10.0 cm from the moist site of being longer than that from the dry area with 59.4±11.6 cm (Table 4.6).

The length of leaves shows significant positive correlations with the leaf width, the number of leaflets, the stem diameter and the tree height (Table 4.7). Particularly, there is a strong relationship between the leaf length and the leaf width with r2 = 0.69 (p = 0.000, Fig. 4.18). The

67 Results leaf width does not correlate with the leaflet numbers, but it demonstrates relationships with the diameter and the height of the plant. Regarding the geographical factors, the elevation has an impact on the leaf size, with larger leaf size at the higher altitude, whereas the number of leaflets is not affected by this parameter.

Table 4.6 Average ± standard deviation, range of leaf length, leaf width and number of leaflets of E. longifolia in moist and dry sites p < 0.05, Mann-Whitney test Moist site Dry site Leaf traits (mountain) (sandy area) 63.6±10.0 59.4±11.6 Leaf length (cm) (38.0-95.0) (39.0-9.02) 18.6±3.5 16.8±3.5 Leaf width (cm) (11.0-33.0) (10.0-28.0) 32.8±5.3 31.3±3.9 Leaflet numbers per leaf (18.0-50.0) (18.0-42.0) 6.3±3.6 3.8±1.2 Stem diameter (cm) (2.2-20.4) (2.5-6.6) 5.9±3.1 2.4±0.8 Height (m) (1.7-17) (0.8-4.0)

Table 4.7 Spearman’s correlation coefficient (r2) and p of correlation among leaf characteristics (N = 505) **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

Leaf traits Leaf Number of DBH Height E UTM N UTM Elevation width leaflets coordinates coordinates (m) Leaf length 0.69** 0.36** 0.27** 0.20** 0.07 -0.07** 0.25** (cm) Leaf width - 0.07 0.35** 0.25** -0.14** -0.24 0.34** (cm) Number of 0.03 - 0.1 0.13* 0.20** -0.03 0.02 leaflets

68 Results

)

( 70

g r2 = 0.69

p = 0.000 f 60

40 10 15 20 25 30 Leaf width (cm)

Figure 4.18 Positive correlation of leaf length (cm) with leaf width (cm)

4.2.2 Leaf characteristics of seedlings and seedling’s growth performances

The phenotypic plasticity of the plant species is a potential mechanism as a response to changes in the living conditions (Nicotra et al., 2010; Krabel, 2016a). Plants can behave with environmental conditions by adapting their phenotypes and alternating those of the next offspring through changes in seed provisioning, leaf morphological and anatomical traits, seedling size, etc. (Sultan, 2000). In the present study, the leaves of the mature trees were harvested from different sites (mountains and sandy areas), while the seedling leaves were collected from the cultivation under the same conditions in order to estimate the phenotypic plasticity of this tree species. The leaf traits from seedlings also were evaluated and their growth performances as well.

4.2.2.1 Leaf traits

Stomatal density and leaf area of seedlings

A total of 45 seedling leaves from different individuals (age between one to two years) were harvested to estimate the leaf area and stomatal density. There is not too much difference in the leaf morphological traits of the seedlings, which were cultivated in the nursery garden from several provenances. In particular, the average leaf area of different origins is the same at around 110 cm2 (F = 0.30, p = 0.827, one-way ANOVA). Regarding the stomatal density, although the seedlings were cultivated in the same environmental conditions, their stomatal density differs

69 Results across four different provenances. A Luoi has the lowest number of stomata, with an average of 202.1±4.0 stomata/mm2, followed by Phong Dien with an average of 247.0±52.4 stomata/mm2 which is similar to Bach Ma (261.2±9.9 stomata/mm2) and fewer than Nam Dong (280.1±53.1 stomata/mm2) (H = 489.18, p = 0.000, Kruskal-Wallis test). In general, the number of stomata mm2 of the seedlings originating from the dry site is lower than that of the moist site (239.2±43.1 and 250.9±56.7 stomata/mm2) (W = 1737338.0, p = 0.0000, Mann-Whitney test) (Table 4.8).

No significant correlations are observed between the leaf area with stomatal density and the seedling age, seedling height, seedling collar diameter (p > 0.05), but the stomatal density of the seedlings has positive relationships with the seedling age and seedling collar diameter (r2 = 0.44 and r2 = 0.30 with p = 0.002 and 0.045). The seedling stomata are not affected by the seedling height (p = 0.703).

Table 4.8 Average ± standard deviation of seedling characteristics from four different provenances H-value: the test statistic for the Kruskal-Wallis test; a, b, c: Mann-Whitney test

Seedling traits Provenances p-value H-value A Luoi Bach Ma Nam Dong Phong Dien Leaf area 114.3±32.2 100.6±32.2 111.5±24.0 109.4±25.8 0.686 1.482 (cm2/leaf) Stomatal density 202.1±4.0c 261.2±9.8ab 280.1±53.1a 247.0±52.4b 0.003 14.011 (stomata/mm2) Height 35.0±18.2a 36.5±3.7a 42.2±10.0a 21.5±14.0b 0.000 18.840 (m) Collar diameter 0.1±0.4b 0.9±0.06b 1.2±0.2a 0.9±0.1b 0.006 12.446 (cm)

Leaf area and stomatal density in comparison between mature trees and seedlings

As can be seen from Fig. 4.19, the patterns of leaf traits differ between the seedlings and the mature trees. The leaf area of the mature trees is nearly 3-5 times larger than that of the leaves of the seedlings. Regarding the stomatal density, the mature trees show a wide range of 131.3- 284.4 stomata/mm2 in comparison to the leaves of the seedlings with 202.1-247.0 stomata/mm2. Although the seedlings from the different sites have a similar leaf area, the number of stomata is not the same as the mature trees. The average number of stomata per mm2 from A Luoi, Bach Ma and Nam Dong provenances (247.0 stomata/mm2) is nearly twice as high as that of the

70 Results mature trees (137.9 stomata/mm2). In contrast to stomatal density, the mature trees from the dry site have a smaller leaf area than those from the moist site.

The data for mature trees from the dry site (284.4 stomata/mm2) are significantly higher than that of the seedlings from the same provenance site (247.0 stomata/mm2) (W = 95238.5, p = 0.000, Mann-Whitney test). a) 700 Mature trees a Seedlings

600 ab b )

2 500 c 400

300 Leaf area (cm area Leaf

200 d d d d 100

0 A Luoi Bach Ma Nam Dong Phong Dien

b) Mature trees

Seedlings )

2 a 350 a b 300 c

250 d

200 f f e 150

100

Stomatal density (stomata/mm density Stomatal 50

0 A Luoi Bach Ma Nam Dong Phong Dien

Figure 4.19 Differences between mature trees and seedlings in aspects of a) leaf area and b) stomatal density in different sites a) Leaf area of seedlings: One-way ANOVA (p > 0.05); b) Mann-Whitney test

71 Results 4.2.2.2 Seedlings’ growth * Fruit traits

The fruits of E. longifolia often have the shape of an ovoid or ellipsoid with a thin shiny exocarp, fleshy mesocarp and hard stony endorcarp (Keng et al., 2002). The fruits are formed in sizeable auxiliary bunches with 140 and 350 seeds per bunch. Table 4.9 shows the length and width of fresh and dry fruits from moist and dry sites. The fruit size is larger on the moist site than on the dry location (p < 0.05, t-test). The fresh and dry fruit lengths in the mountains are 1.58±0.14 cm and 1.40±0.12 cm while in the sandy area they are measured for 1.37±0.10 cm and 1.07±0.07 cm. Similarly, the fresh and dry fruit diameter from the dry site (0.97 cm and 0.71 cm) are smaller than those from the moist site (0.92 cm and 0.67 cm) (Fig. 4.20).

Table 4.9 Average ± standard deviation, range of fruit morphology of moist and dry sites p < 0.05, t-test

Fruit traits Moist site Dry site 1.58±0.14 1.37±0.10 Fruit length (fresh, cm) (1.3-1.8) (1.2-1.6) 0.97±0.10 0.92±0.09 Fruit diameter (fresh, cm) (0.8-1.2) (0.8-1.1) 1.40±0.12 1.07±0.07 Fruit length (dry, cm) (1.2 1.6) (1.0-1.2) 0.71±0.04 0.67±0.04 Fruit diameter (dry, cm) (0.6-0.8) (0.6-0.8) a)

Figure 4.20 Typical examples of fruit from a) sandy area (dry site) and b) mountainous area (moist site)

72 Results Germination rate of seeds

A total of 448 ripe seeds from the sandy area and 400 mature seeds from the mountains were used for the seed germination evaluation. The seeds of E. longifolia have started to germinate 15 days and continued to grow for 40 days after being sown. The duration of seed germination of the provenances from the sandy area (between 15 and 36 days) is much shorter than that of the mountainous origins (approximately from 15 to 51 days). The percentages of seedling germination of the two areas is significantly different (W = 133,500, p = 0.001, Mann-Whitney test) (App. 5). The seedling germination rate of sandy area stands with 73.7% higher than the mountainous seedling germinate rate, which is merely 45.3% (Table 4.10).

Figure 4.21b presents the proportion of seed germination from a total of ten different mother trees, including one provenance from the dry site (PB2-1, PB2-3, PB2-5, PB2-6, PB2-9, PB2- 10) and two origins from the moist places (M29, M33, M89, AL6-7). As far as the sandy site is concerned, seeds from PB2-9 have the lowest percentage of germination (31%) and this value is even smaller than that of three mother trees from the mountainous area. In contrast, seeds from M29 show a germination proportion of 79%, this is higher than that of the three mother trees belonging to the mountainous provenances. Although seeds from the two different collections (sites) have started to germinate at the same time (after 15 days), the seeds from the dry site have a higher percentage of germination (74%) than those from moist site (45%) (Fig 4.21a).

Seedling survival rate Generally, the average percentage of seedling survival rate from the different provenances is high. The survival rate of seedlings from the moist and dry sites is with 87% and 85% nearly the same after 22 months (t-test, p = 0.414) (Table 4.10).

Seedling growth dynamic The plant height, collar diameter and the number of leaves were recorded during the period of one to twenty-two months after germination (Fig. 4.22, Fig. 4.24). Although all the seedlings were planted in similar environmental conditions in the nursery garden, they perform different growth patterns in terms of plant height, collar diameter and number of leaves. In general, the mountainous provenances have a stronger growth in height, collar diameter and number of leaves than those of the sandy origin (p < 0.05, z-test, App. 6).

73 Results

Table 4.10 Seedling’s provenances and seed maturity from the seed germination (days) within a 55-day period and survival rate >22 months Nam Dong provenance (M29, M33, M89), A Luoi provenance (AL6-7). Germ: Germination, SD: number of seedlings 22 months Seed germination (days) (Apr 2018) Seeds for Germ Sites Mother Germ propagation (%) trees (days) Survival (Jun 2016) 15 20 25 40 55 SD rate (%)

PB2-1 103 41 56 58 64 64 15-40 62.14 57 89 PB2-3 50 33 43 45 45 45 15-27 90.00 40 89 PB2-5 70 58 60 61 65 65 15-40 92.86 61 94 Dry PB2-6 75 40 59 63 63 63 15-27 84.00 57 90 PB2-9 100 15 20 25 31 31 15-40 31.00 21 68 PB2-10 50 29 35 40 41 41 15-40 82.00 33 80 Total 448 309 15-36 73.67 269 85

M29 100 61 71 75 76 79 15-55 79.00 76 96 M33 100 8 8 9 25 28 15-55 28.00 18 64 Moist M89 100 17 20 38 42 48 15-55 48.00 46 96 AL6-7 100 14 20 25 26 26 15-40 26.00 24 92 Total 400 181 15-51 45.25 164 87

Seedlings of one month illustrate a rapid growth with an average of 5.6 cm at the moist site and 5.1 cm at the dry site; especially the average height of the seedlings from M29 (Nam Dong provenance) reaches 7.6 cm. Then the seedlings have a slight increase up until ten months before growing faster from 10 to 22 months after germination. The average seedling height from the mountainous provenances nearly doubles that of the sandy origins (39 cm and 22 cm) (Fig. 4.22). In particular, the seedlings from mother trees in the mountains (M29, M89 and AL6-7) increase sharply in term of the height after 10 months at 38.9 cm, 43.6 cm and 44.2 cm on average (M33 performs the same pattern as the seedlings from the sandy area). Regarding the growth of seedlings from the sandy area, seedlings from PB2-1 present a faster growth after 10 months and they reach a height of 31.6 cm after 22 months whereas the seedlings from PB2- 6 show the weakest growth performance pattern reaching only 16 cm at 22 months after germination.

The collar diameter increases very slowly from germination to the age of 5 months, followed by a slightly faster increase in both groups (seedlings from moist and dry sites). The seedlings from the moist site have a considerably larger collar diameter than those from the dry origin,

74 Results measuring on average 1.09 cm and 0.86 cm after 22 months. Notably, the seedlings belonging to AL6-7, M29 and M89 show larger collar diameter in comparison with other seedlings. Seedlings from PB2-1 not only have a fast growth in height but also in collar diameter compared to other seedlings from the sandy area.

The number of leaves produced by the seedlings seems to follow the same growth pattern as that of the seedlings’ height and collar diameter, excluding the phase during the leaf sampling between 5 and 9 months and from 14 to 20 months. The average leaf numbers of seedlings from the moist and dry sites are approximately the same up to 5 months (from 8 to 10 leaves), followed by a sharp decrease after 9 months before a slight increase after 14 months. Then the number of leaves is reduced between 14 and 20 months before it reaches a number of 7 leaves for seedlings from the dry place and 11 leaves from the moist site.

The Pearson’s correlation coefficient shows a strong correlation between the seedling height and the collar diameter from dry and moist sites (r2 = 0.93, p = 0.000) (Fig. 4.23).

a) b)

Figure 4.21 Germination response of the ripe seeds of E. longifolia from different provenances in the period between 15 and 55 days a) From different mother trees and b) from dry (PB2-1, PB2-3, PB2-5, PB2-6, PB2-9, PB2-10) and moist sites (M29, M33, M89, AL6-7)

75 Results a) b)

Figure 4.22 The growth pattern of E. longifolia seedlings in terms of height, collar diameter and number of leaves a) From dry and moist sites and b) from different mother trees (provenances)

76 Results

r2 = 0.93 p = 0.000

Figure 4.23 Correlation between seedling height and seedling collar diameter, from different sites

77 Results

Figure 4.24 Images of seedlings from different stages (A1-A3) Seedlings 1-2 months after germination; (B1-B2) one year old seedlings; (C1-C2) two years old seedlings in nursery garden; (D1-D3) Seedlings (2-3 years old) planted with indigenous species in the field

78 Results 4.3 Genetic diversity

4.3.1 RAPD, SCoT and BPS amplification

Two RAPD primers (OPC02 and OPB05) and three SCoT and BPS (SCoT 21, SCoT36 and LA2a) primers display clear, without water contaminants and reproducible bands and were obtained by the screening of a total of 8 RAPD and 20 SCoT and BPS primers. The five primers are therefore selected for the analysis of genetic diversity in 276 E. longifolia samples. The selected primers amplify a total of 118 bands, which range from 159 to 3,787 bp. In total, 118 polymorphic bands and only 1 monomorphic band (OPB05, 721 bp) are obtained. Thus, most primers show 100% polymorphism, except the primer OPB05 (92.3%) (Fig. 4.25, App. 15). As can be seen in Table 4.11, an average of about 24 fragments per primer is found with results ranging from 11 (SCoT36) to 37 fragments (SCoT21). Each band is observed at an average of around 23% of the examined samples. Furthermore, 17% of the bands are kept in a maximum of 10 samples, 42% of the fragments in between 11 and 50 samples and 41% in more than 50 samples.

To analyze the suitability of the marker systems in the evaluation of the genetic profiles of E. longifolia, the performance of the markers is estimated using four estimators: polymorphic information content (PIC), effective multiplex ratio (EMR), marker index (MI) and resolving power (Rp). The mean of PIC values is analysed for all bands to define the PIC values of each primer. The overall PIC values for individual primer combinations range from 0.116 (OPB05) to 0.204 (LA2a), with an average of 0.154±0.035. For the SCoT and BPS markers, PIC values are in the range of 0.144-0.204, while the RAPD marker is only 0.116-0.133.

79 Results

Figure 4.25 A gel image generated by SCoT21 primer (12 samples) and one sample investigated with five different primers SCoT21, LA2a, OPC02, SCoT36 and OPB05 running by Fragment Analyzer System

80 Results The results of EMR depend on the average of amplified fragments (n) and the number of polymorphic and non-polymorphic bands (β), which were obtained from one primer by all examined genotypes (Varshney et al., 2007). In the present study, SCoT and BPS markers reveal 100% polymorphic fragments, the n and β are calculated as 5.8 and 1, while those of RAPD markers are defined as 4.9 and 0.96. Thus, the SCoT and BPS markers have a higher EMR value in comparison with the RAPD marker at 5.8±1.7 and 4.7±0.7 on average. SCoT21 occupies the highest EMR value (7.29), followed by LA2a (6.08) and the lowest value is accounted for SCoT36 (4.03).

In order to test the general utility of the given marker system, the MI is evaluated for these marker systems analysed. The MI value from the SCoT and BPS markers (1.03) is approximately twice as high as that of the RAPD marker (0.59). Notably, LA2a and SCoT21 have similar MI values at 1.24 and 1.27.

The Rp is a parameter that indicates the discriminatory potential of the chosen primers. The highest value is captured with SCoT21 (14.58), followed by LA2a (12.17) and the lowest with SCoT36 (8.05) at an Rp average of 10.87±2.58 (Table 4.11).

Table 4.11 Details of banding pattern revealed through RAPD, SCoT and BPS markers FSR: fragment size range, TNF: total number of fragments, NPF: number of polymorphic fragments, PPF: percentage of polymorphic fragments, PIC: polymorphic information component, EMR: effective marker ratio, MI: marker index, Rp: resolving power

Markers Primers FSR (bp) TNF NPF PPF PIC EMR MI Rp SCoT and LA2a 364-3787 35 35 100 0.20 6.08 1.24 12.17 BPS SCoT21 159-1821 37 37 100 0.17 7.29 1.27 14.58 SCoT36 654-1735 11 11 100 0.14 4.03 0.58 8.05 Mean 27.67 27.67 100 0.17 5.80 1.03 11.60

Total 83 83

RAPD OPB05 285-1573 13 12 92.3 0.12 3.91 0.45 9.12 OPC02 300-1912 22 22 100 0.13 5.21 0.69 10.43 Mean 17.50 17.00 96.15 0.12 4.56 0.57 9.77 Total 34 35

Total 118 117 Mean 23.60 23.40 98.08 0.15 5.30 0.85 10.87

81 Results 4.3.2 Genetic diversity in E. longifolia population

The 276 accessions are divided into four major populations, including A Luoi, Bach Ma, Nam Dong, Phong Dien based on the geographical locations, topography, ecological distribution, etc. from which they were collected. Regarding SCoT and BPS profiling, the genetic diversity of the species across all the populations is at an average of 0.16. The Shannon index is on average 0.27 (ranging from 0.25 to 0.28) (Table 4.12). Although the region of Nam Dong which presents the highest degree of polymorphic fragments (74.7%) as compared to Bach Ma (55.4%), the genetic diversity indices show that there is insignificantly difference among the populations; even.

Similar to the SCoT and BPS analysis, the RAPD analysis shows nearly the same with the proportion of polymorphic fragments (66.43%). The highest percentage of polymorphic fragment occurs in Nam Dong (80%) and the lowest in Bach Ma (48.57%). The genetic variation observed shows a significant difference among the populations with RAPD markers at an average of 0.18 (ranging from 0.15 to 0.21) and the Shannon index is, on average 0.28±0.02.

Combining the data of RAPD, SCoT and BPS markers, 118 bands for four populations are detected, whereas the number of polymorphic fragments of SCoT and BPS markers is much higher than that of RAPD maker (83 vs 35). The selected primers amplify 90 fragments occupying the highest polymorphic fragment (76%) in the Nam Dong population. However, the Kruskal-Wallis test does not show significant genetic indices among the four populations (p > 0.05). The average of genetic diversity of E. longifolia is 0.17(±0.007) and its Shannon index is 0.28(±0.011) which is relatively low (Table 4.12).

82 Results

Table 4.12 Summary of the genetic variation as revealed through RAPD, SCoT and BPS markers among four populations of E. longifolia NPF: number of polymorphic fragments, PPL %: percentage of polymorphic fragments at population level, I: Shannon index, He: expected heterozygosity

Markers Populations NPF PPL % I He A Luoi 59 71.08 0.27 0.16 Bach Ma 46 55.42 0.25 0.16 62 74.70 0.28 0.17 SCoT and Nam Dong BPS Phong Dien 54 65.06 0.27 0.17 Mean 55 66.57 0.27 0.16 Overall fragments 83

A Luoi 25 71.43 0.31 0.20 Bach Ma 17 48.57 0.23 0.15 Nam Dong 28 80.00 0.33 0.21 RAPD Phong Dien 23 65.71 0.27 0.17 Mean 23 66.43 0.28 0.18 Overall fragments 35

A Luoi 84 71.19 0.28 0.17 Bach Ma 63 53.39 0.25 0.16 Nam Dong 90 76.27 0.29 0.18 Total Phong Dien 77 65.25 0.27 0.17 Mean 79 66.53 0.27 0.17 Overall fragments 118

4.3.3 Genetic differentiation

The analysis of molecular variance (AMOVA) was used to evaluate the variation within and among the examined populations. The SCoT and BPS dataset for the investigated accessions reveals through the nested AMOVA a high variance of 87% which occurs within populations and a variance of 13% which occurs among populations with p < 0.05 (ΦPT = 0.131). The distribution of total genetic variation for the RAPD dataset shows that almost all the total variance is attributable to the within population variance (94%) and only 6% of the difference is partitioned among populations (ΦPT = 0.06). In combining the RAPD, SCoT and BPS marker data, the AMOVA analysis reveals that 11% (p < 0.05) of the total genetic variation is among populations (ΦPT = 0.112) and 89% is within populations (Table 4.13).

83 Results

Table 4.13 Analysis of molecular variance (AMOVA) using RAPD, SCoT and BPS markers

Sources of Markers Among Populations Within Populations Total variation Degree of 3 272 275 freedom

SCoT and BPS 308.45 2575.16 2883.61 Sum of RAPD 59.30 1046.43 1105.73 squares Overall 367.75 3621.58 3989.34

SCoT and BPS 102.82 9.47 RAPD 19.77 3.85 Mean of squares Overall 122.59 13.32

SCoT and BPS 1.43 9.47 10.90 Variance RAPD 0.24 3.85 4.09 components Overall 1.67 13.32 14.99

SCoT and BPS 13 (ΦPT = 0.131) 87 100 Percentage of RAPD 6 (ΦPT = 0.060) 94 100 variance (%) Overall 11 (ΦPT = 0.112) 89 100

p-value = 0.001

The matrix of pairwise population PhiPTP (Φst) values and Nei’s genetic distance determined between pairs of populations for the total dataset are shown in Table 4.14. The genetic differentiation is medium among the populations with Φst ranging from 0.070 to 0.164. Nam Dong and Bach Ma populations present the largest genetic differentiation (Φst = 0.164). Nei’s genetic distance reveals that Nam Dong and Bach Ma populations have comparable high genetic distance (0.038), while Phong Dien and Bach Ma populations have smallest genetic distance (0.014).

Table 4.14 Pairwise genetic differentiation (Φst) of E. longifolia populations (above diagonal) and pairwise population matrix of Nei’s genetic distance (below diagonal)

Populations A Luoi Bach Ma Nam Dong Phong Dien A Luoi 0.000 0.141 0.079 0.109

Bach Ma 0.026 0.000 0.164 0.070

Nam Dong 0.017 0.038 0.000 0.131

Phong Dien 0.014 0.014 0.024 0.000

4.3.4 Principle Coordinate Analysis (PCoA)

The two-dimensional PCoA for 276 individuals across four populations based on RAPD, SCoT and BPS dataset accounts for 8.49% (coordinate 1) and 5.48% (coordinate 2) of the total

84 Results variance (Table 4.15). Thus, the grouping of individuals using two coordinates is indicated in Fig. 4.26. PCoA analysis also categorizes genotypes into two main groups, including Group 1 (Nam Dong and A Luoi) and Group 2 (Bach Ma and Phong Dien); even the pattern of genetic similarity is not as clear as has been expected.

Table 4.15 Analysis of Principle Coordinates (PCoA) based on RAPD, SCoT and BPS markers

Principle 1 2 3 coordinates Eigenvalues 338.79 218.79 167.15

% variance 8.49 5.48 4.19

Figure 4.26 Principle Coordinate Analysis (PCoA) based on genetic relationship of four subpopulations of 276 E. longifolia accessions Blue dots: A Luoi, turquoise dots: Nam Dong; purple dots: Bach Ma; orange dots: Phong Dien

4.3.5 Population structure

The population structure was analysed using Bayesian cluster analysis as implemented in STRUCTURE v2.3 (Pritchard, 2007). The calculation was carried out for the RAPD, SCoT and BPS dataset. The real K value with the highest amount of ΔK for the 276 accessions was recorded. The mean and standard deviation of log probability of data Ln over 10 independent runs for each K (App. 10) and plot of ΔK statistics with respect to the number of genetic clusters K (from 1 to 5). Thus, the values of ΔK indicates that K = 2 is the most likely figure of the

85 Results genetic groups of E. longifolia (Fig. 4.27), according to the ad hoc statistic ΔK (Evanno et al., 2005). This finding suggests the cluster of the four populations into two groups, which is consistent with the result from Ward’s method and PCoA.

The vertical lines partitioned into two different colors (K) that represent the admixture of genetic ancestry in the corresponding genetic groups presenting of each individual. Group 1 includes A Luoi and Nam Dong populations in the dark violet color, whereas group 2 includes Bach Ma and Phong Dien populations in the pink color.

2000

1500

1000

K Δ 500

0 0 1 2 3 4 5 6 Number of genetic groups (K)

Figure 4.27 STRUCTURE analyses ΔK values for each assumed number of populations (K) for E. longifolia (Evanno et al., 2005)

Figure 4.28 STRUCTURE Q plots generated utilizing the maximum value of ΔK E. longifolia samples perform two genetic groups, demarcated by labeled bars (i.e., A Luoi). A vertical line represents each individual. Group 1 includes A Luoi and Nam Dong - dark violet color; Group 2 includes Bach Ma and Phong Dien - pink color. Horizontal axis (from 0.0 to 1.0) is the proportion of alleles from each subpopulation.

86 Results 4.3.6 Genetic cluster analysis

Although the present dataset does not present well-separated groups in Principle Coordinate Analysis (PCoA), the result of PCoA indicates two main genetic groups. It is considerably more challenging to assess the findings of the clustering analysis and the correct number of clusters. Thus, it is necessary to use the criteria for evaluating the quality of cluster analysis. These criteria may be used to compare the adequacy of specific algorithms and dissimilarity measures or determine the best number of clusters (Jäkel and Nöllenburg, 2004; Carvalho et al., 2019). Regarding the structure clustering analysis result, there are mainly two genetic groups of E. longifolia (K = 2). Therefore, PCoA and Structure clustering analysis also confirm two main groups, including group 1 (Nam Dong and A Luoi) and group 2 (Bach Ma and Phong Dien). Several hierarchical methods have been applied to express the genetic data, including single linkage, complete linkage, average linkage (UPGMA), Ward’s method, etc. When comparing the results of different clustering methods for the population genetic data, Ward’s method shows a more straightforward pattern than others do. Because the clusters' internal cohesion can reflect the point scatter within the cluster, Ward’s approach tends to create compact groups with the same objects (Ward, 1963).

The dendrogram of Ward’s method, based on Euclidean genetic distance, illustrates the relationships among individuals, which reveals two main clusters. The first cluster contains Phong Dien and Bach Ma (90 accessions), while the second cluster is comprised of the two remaining populations. Each cluster is again grouped into sub-clusters with few outliers. All the samples from Sub-cluster Ia belong to Phong Dien, while Sub-cluster Ib comprises the samples from both Phong Dien and Bach Ma. Sub-cluster IIa, IIb involve all the samples from A Luoi and Nam Dong (including some outliers from Phong Dien). Sub-cluster IIa2 and IIb1 on the other hand, only contain the accessions from Nam Dong while Sub-cluster IIa1 and IIb2 cover not only A Luoi but also some genotypes from Phong Dien (Fig. 4.29, App. 11).

The Mantel test result exposes a positive correlation between the geographical and genetic distances, between the elevation and genetic matrices, between the morphological traits (height and diameter) among populations and the genetic matrix with r2 = 0.19, 0.17 and 0.12 respectively, p < 0.05, permutation = 9999 (App. 8).

87 Results

Ia Ib IIb1 IIb2

IIa1 IIa2

IIa IIb

I II

Figure 4.29 Dendrogram of agglomerative clustering using Ward’s method and Euclidean distances among individuals Ward’s method is calculated basing on the sum of squares, which start at zero (each point is in its own cluster) and grows as clusters are merged. Ward’s method keep this growth to minimum (Ward, 1963).

88 Results 4.3.7 Genetic variation across different generations

4.3.7.1 Comparison of genetic variation between mother trees, seedlings and mature trees

269 seedlings (in particular 71, 42, 77, 79 samples from A Luoi, Bach Ma, Nam Dong and Phong Dien) and 15 samples of mother trees were used for assessing their genetic variation (Table 3.4, App. 12). Similar to the genetic diversity assessment of the mature trees, five of RAPD, SCoT and BPS primers were also applied to the seedlings and could produce reproducible fragments. The sizes of the fragments range from 246 bp to 2.325 bp. In total, 112 polymorphic bands with an average of 22.4 bands for each primer are observed in the seedlings that is a higher number than that of the mother trees (15.8 bands). However, the percentage of polymorphic fragments from mother trees is higher than the one of their offspring (71.43% and 65.18%). The Shannon index and the expected heterozygosity are nearly the same among offspring, mother trees and mature trees (Table 4.16a).

Table 4.17 presents that genetic diversity from Nam Dong provenance reduces remarkably at the propagated and natural seedling level (He = 0.14) compared to He = 0.18 for mother trees. Conversely, the genetic variation of seedlings from A Luoi is greatly increased in comparison to the mature trees, while it remains unchanged in the next generation from Phong Dien and Bach Ma. In addition, the genetic differentiation between the seedlings and the mother trees is as small as the genetic distance among them (Table 4.16b).

Table 4.16 a) Genetic diversity across different generations and b) Pairwise Matrix of Nei Genetic Distance (above diagonal) and genetic differentiation (Φst) of E. longifolia mother and seedling populations (below diagonal) Shannon index (I), expected heterozygosity (He), number of polymorphic fragments (NPF), percentage of polymorphic loci (PPL) a) Number Generations I He NPF PPL of samples Mother trees 0.25 0.16 82 71.43 15 Seedlings 0.27 0.17 112 65.18 269 Mature trees 0.27 0.17 118 66.53 276 b) Generations Mother trees Seedlings Mother trees 0.000 0.116 Seedlings 0.024 0.000

89 Results 4.3.7.2 Genetic variation between propagated and natural seedlings from different provenances

The seedlings are used to estimate the genetic diversity among the propagated seedlings in A Luoi, Nam Dong and Phong Dien and among the natural ones in A Luoi, Nam Dong and Bach Ma. Particularly, the comparison in terms of genetic diversity between the propagated and the natural seedlings focuses on A Luoi and Nam Dong provenances. Regarding the propagated seedlings, A Luoi shows a higher expected heterozygosity and Shannon index than those of Nam Dong and Phong Dien while the less polymorphism is attributed to Nam Dong provenance. Similarly, A Luoi presents higher genetic indices compared to Bach Ma and Nam Dong for natural seedlings. However, the Kruskal-Wallis test does not show a significant difference of genetic indices among the four origins (p = 0.061). Thus, the mean genetic diversity in terms of E. longifolia is 0.17(±0.008) and its Shannon index is 0.27(±0.010) (Table 4.17).

Similar to the mature accessions, E. longifolia provenances at the nursery garden are moderately differentiated with Φst ranging from 0.081 to 0.182, whereas Bach Ma and Nam Dong provenances have the largest differentiation (Φst = 0.182). However, the genetic distance among the provenances shows no large difference (Table 4.18). The AMOVA analysis indicates that 14% (p < 0.05) of the total genetic variation is among the provenances (ΦPT = 0.141) and 86% is within the populations (Table 4.19).

90 Results

Table 4.17 Genetic diversity at seedling generation in comparison with mature trees and the differences of its variation between propagated and natural seedlings from different provenances Shannon index (I), expected heterozygosity (He), number of polymorphic fragments (NPF), percentage of polymorphic loci (PPL)

Mature Propagated seedlings Natural seedlings All seedlings trees Provenances Number Number Number I He PPL NPF of I He PPL NPF of I He PPL NPF of I He samples samples samples

A Luoi 0.33 0.21 71.28 68 17 0.33 0.21 82.57 92 54 0.33 0.21 81.25 93 71 0.28 0.17

Bach Ma - - - - - 0.28 0.18 63.30 71 42 0.27 0.17 61.61 70 42 0.25 0.16

Nam Dong 0.26 0.17 63.83 60 40 0.22 0.14 54.13 60 37 0.23 0.14 54.46 63 77 0.29 0.18

PhongDien 0.32 0.20 75.53 74 79 - - - - - 0.27 0.17 63.39 72 79 0.27 0.17

Mean 0.30 0.19 70.21 0.28 0.18 66.67 0.27 0.17 65.18 0.27 0.17

Overall 93 136 107 133 112 269

91 Results

Table 4.18 Pairwise Matrix of Nei Genetic Distance (below diagonal) and genetic differentiation (Φst) of E. longifolia seedling accessions (above diagonal)

Provenances A Luoi Bach Ma Nam Dong Phong Dien A Luoi 0.000 0.139 0.168 0.145 Bach Ma 0.028 0.000 0.182 0.141 Nam Dong 0.031 0.030 0.000 0.081 PhongDien 0.027 0.026 0.014 0.000

Table 4.19 Analysis of molecular variance of seedlings from different provenances using RAPD, SCoT and BPS markers df: degree of freedom, SS: sum of squares, MS: mean squares, Est. Var: Estimate of variation, P%: percentage of total variance, p-value: based on 1,000 permutations

Source df SS MS Est. Var. P% p-value

Among Pops 3 425.37 141.79 1.96 14 < 0.001 (ΦPT = 0.141) Within Pops 265 3169.09 11.96 11.96 86

Total 268 3594.46 13.92 100

The dendrogram of Ward’s method based on Euclidean genetic distance illustrates that the genetic relationships among individuals is devided into two main clusters. The first cluster contains only the provenance of Bach Ma (30 accessions) and some outliers, while the second cluster comprises three groups. The Sub-cluster IIa comprises the samples from both Phong Dien and Nam Dong, which are separated into different patterns for each group. Mainly the 61 accessions from the provenance of A Luoi forms the Sub-cluster of IIb with few outliers (Fig. 4.30).

92 Results

IIa IIb I II

Figure 4.30 Dendrogram of 15 E. longifolia mother tree samples and 269 seedlings analysed from four different sites Obtained by Ward’s method based on the similarity of Jaccard’s coefficient (63.36%).

93 Results 4.4 Eurycomanone content

4.4.1 Biological traits and root water content * Root characteristics

The fresh and dry weights of the roots were measured to determine the root water content in the different distributional areas. Figure 4.31 presents the differences between the fresh and dry weights of the root tissues of 30 samples from four different sites. In general, the water loss in the roots of E. longifolia after drying from the mountainous area is higher than from the sandy soil area. The average of the fresh root weight of the sandy soil sites was lost nearly halves compared to those of the moist sites (20 g and 41 g).

250

Fresh weight (g) 200 Dry weight (g)

150

100

50 The weight of the root plant materials (g) materials plant root the of weight The

PD1 PD2 PD3 PD4 PD5 PD6 PD7 PD8 PD9

AL1 AL2 AL3 AL4 AL5 AL6 AL7 AL8

ND5 ND1 ND2 ND3 ND4 ND6 ND7

BM1 BM2 BM3 BM4 BM5 BM6

Root samples in four sites

Figure 4.31 The weight of all the root samples from four sites A Luoi: AL1 - AL8; Bach Ma: BM1 - BM6; Nam Dong: ND1 - ND7; Phong Dien: PD1 - PD9

* Root water content

The average water content of the roots is ca. 42%, ranging from 17% to 70% (App. 9). The root water content from the moist site is significantly higher than that from the dry place at an average of 46.4±12.2 and 30.5±8.2, respectively (F = 12.51, p = 0.001, one-way ANOVA). However, among the smaller areas, the three mountainous regions are more or less similar in terms of root water content; even one of them is the same as the sandy area. Moreover, the

94 Results eurycomanone content could be predicted from the elevation by the following formula: y = 32.30 + 0.01945x, R2 = 25.5% (Fig. 4.32). Besides, the tree height affects the water content of the roots; the taller the trees, the higher the root water content (Table 4.20).

Interval Plot of Root water content vs two sites 95% CI for the Mean 55

) 45

(

e 40

t a 30

20 Moist site Dry site The pooled standard deviation was used to calculate the intervals. b)

Fitted Line Plot Root water content (%) = 32,30 + 0,01945 Elevation (m)

70 S 11,8746 R-Sq 25,5% R-Sq(adj) 22,8% 60

)

(

n 50

r 40

t 30

R 20

10 0 100 200 300 400 500 600 700 800 900 Elevation (m)

Figure 4.32 a) Difference of water content in moist and dry areas and b) positive correlation between root water content and elevation factor

95 Results 4.4.2 Eurycomanone content of E. longifolia root tissues 4.4.2.1 Standard line and eurycomanone standard A total of 30 root tissue samples were examined for the eurycomanone content using HPLC. The concentration of eurycomanone in all samples was calculated based on the calibration curve of eurycomane standard (Fig. 4.33). In each of the root samples, the retention time is detected between 5.73 min and 7.28 min. A standard curve (y = 70308x - 1E+06, R2 = 0.9834) of eurycomanone is used to quantify the level of the compound in 30 samples. Fig. 4.34 shows (in turns) the chromatograms of the eurycomanone standard and Fig. 4.35 presents one root sample of E. longifolia as an example.

20 18 16 14 12 10

8 Peak area area Peak 6 4 y = 70308x - 1000000 2 R² = 0,983 0 0 50 100 150 200 250 300

Eurycomanone standard content (µg/mL)

Figure 4.33 Standard curve of eurycomanone

mAU

Figure 4.34 HPLC chromatogram of eurycomanone standard at 250 µg/µL

96 Results

mAU

Figure 4.35 HPLC chromatogram of eurycomanone extract from root

4.4.2.2 Eurycomanone content

Figure 4.36a and App. 12 show the average concentrations of eurycomanone and the relative parameters for different distribution areas. The average eurycomanone content is 1.1 mg/g dry weight. The highest level of eurycomanone in the root can be observed in AL5-13 (A Luoi) at 2.34 mg/g and the lowest eurycomanone content is found in PB2-10 (Phong Dien) at only 0.29 mg/g. Besides, there is a significant difference in the eurycomanone content among the four areas, in which A Luoi occupies the highest content at 1.39±0.44 mg/g on average, followed by 1.16±0.36 mg/g, 1.11±0.47 mg/g and 0.74±0.29 mg/g in Nam Dong, Bach Ma and Phong Dien (H = 10.30, p = 0.016, Kruskal-Wallis test). In general, the eurycomanone content of the samples from the moist sites is higher than that of dry site at 1.22±0.42 mg/g and 0.74±0.29 mg/g (W = 388.0, p = 0.005, Mann-Whitney test) on average, respectively (Fig. 4.36b).

a) 97 Results

2.00 a 1.80

1.60 ab ab 1.40 1.20 b 1.00 0.80 0.60

0.40 Eurycomanone content (mg/g) content Eurycomanone 0.20 0.00 A Luoi Bach Ma Nam Dong Phong Dien

Boxplot of Eurycomanone component 2,0

)

/ 1,5

(

n a 1,0

E 0,5

Moist site Dry site

Figure 4.36 a) Eurycomanone content of the root of E. longifolia in four different areas and b) eurycomanone content in moist and dry sites

21 samples from mountains and 9 samples from sandy areas; a) Kruskal-Wallis test; a, ab, b: Mann-Whitney test; b) Mann-Whitney test (p < 0.05)

98 Results 4.4.3 Correlation between the eurycomanone content and plant traits and geographical factors

The eurycomanone content shows a positive correlation with the root water content, meaning that the content of eurycomanone increases with the higher water content in the roots (r2 = 0.51, p = 0.02) (Table 4.20, Fig. 4.37). Moreover, the diameter of the roots (8-17 year-old) also affects the eurycomanone content. A larger root diameter is relatively associated with a higher eurycomanone content (r2 = 0.37, p = 0.046) (Fig. 4.38). Besides, the altitude exerts an impact not only on the root water content but also on the eurycomanone content, which refers to the fact that the content depends on the ecological conditions or toporgraphical factors, e.g. elevation above the sea level at a higher altitude and with a greater root water content. Moreover, there is no significant correlation between the genetic diversity and the eurycomanone content (p > 0.05, Mantel test).

Table 4.20 Spearman’s correlation coefficient between eurycomanone content and tree traits (root water content, root diameter, tree diameter, tree height and plant ages) and geographical factors (coordinates and altitude) **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

Predictor variables

Descriptive Statistics Dependent

variables

average standard deviation minimum maximum Water content (%) Root diameter (cm) Stem diameter (cm) Height (m) UTME coordinates coordinatesN UTM Altitude (m) Year levels Eurycomanone 1.09 0.45 0.29 2.34 0.51* 0.37* -0.46** 0.47* (mg/g)

Water content 41.77 13.52 16.75 70.43 0.47** -0.46** 0.48**

Root diameter 9.51 4.02 3.50 21.00 0.61** 0.47** 0.40* -0.52* 0.75**

Stem diameter 6.70 3.67 2.50 20.40 0.76** -0.50** 0.39 0.41**

Tree height 5.78 3.02 1.30 12.00 0.45* -0.66** 0.49**

99 Results

2,5 r2 = 0.51

)

g p = 0.02 / 2,0 g

(

t 1,5

o 1,0

r 0,5

0,0 10 20 30 40 50 60 70 Root water content (%)

Figure 4.37 Positive correlation between eurycomanone and root water content

2,5

2 ) r = 0.37

g / 2,0 p = 0.046

(

t 1,5

o 1,0

r 0,5

0,0 5 10 15 20 Root diameter (cm)

Figure 4.38 Root diameter dependence on eurycomanone content

100 Discussion 5 Discussion

Eurycoma longifolia is a popular medicinal tree species in most Asian countries, including Vietnam (Hassan et al., 2012; Chi, 2012). This plant's deterioration tends to increase significantly due to unsustainable harvesting and habitat loss, leading to the reduction of the natural populations and genetic resources. A few studies have also been conducted in Vietnam, but mainly related to general information on biological characteristics and components of high valuable bioactive compounds (Nhan and Loc, 2018) and rarely the investigations focused on natural populations' distribution, genetic systems, or phenotypic plasticity characteristics (Loc et al., 2016; Loc et al., 2018). In addition, no research was found about this plant species on sandy areas. To close this gap of knowledge, there is a need to research the occurrences of both mountains and sandy area in terms of phenotype, genotype and some main components, e.g. eurycomanone. The bioactive compound concentration often depends on several factors such as the plant age, ecological or genetic parameters. Thus, the present study aims to evaluate the plant's distribution status, the adaptability of leaf morphology and anatomy, genetic diversity and one main bioactive compound in plants from both mountains and sandy area. This study plays an essential role in providing useful information for potential breeding approaches and the conservation of this valuable medicinal species. The study was conducted in four areas, including A Luoi, Bach Ma, Nam Dong and Phong Dien in the province of Thua Thien Hue, Vietnam. With regard to the results, the four main parts (ecological distribution of E. longifolia population; leaf traits and seedling growth; genetic diversity and eurycomanone content) are discussed as follows.

5.1 Ecological distribution of E. longifolia population

The ecological distribution of a population is indicated by two criteria, namely population density and size. Estimating the density and size of a population can help to describe the current status of that particular population (Oliver, 2013).

5.1.1 Population density

Tree density (DBH ≥ 6 cm) with 9.5 individuals per ha is rarely found in the mountains but not in the sandy area. The trees of the mountainous area often occurs in the form of a single stem while they have several stems or branches in sandy area. As the sandy inner regions are close to the residential areas and as they are not well protected, the growth of E. longifolia trees is harmfully affected by cutting them off for firewood purposes at a young age. This fact makes 101 Discussion the trees in this area tend to grow like small shrubs and create a clumped distribution type. Also, the plants in the sandy area have to compete among themselves for the habitats like soil moisture, light intensity or slope steepness, etc.

As can be seen from Table 4.1, the average soil moisture in the sandy area is only 14.7% in comparison with 46.5% in mountainous area. Although E. longifolia can adapt to sandy clay texture and acid soil (Rifai, 1975), these environmental factors also limit its growth. Therefore, there is no tree with a diameter higher than 6 cm in this region and the number of saplings (2.5- 6 cm in DBH) is still similar compared to the mountainous area. Loc et al. (2016) also estimated the distribution of E. longifolia by using selected plots containing mainly this plant species. It is a preliminary study in a narrow area that may partially explain why the authors observed a high density of E. longifolia, especially in sandy area. In the present research, 74.2 ha (27 baselines) were randomly surveyed, covering a large area, in comparison with only 2.8 ha (56 plots) in the study of Loc et al. (2016). Therefore, the present findings indicate that the population density of E. longifolia is relatively low.

Kartikawati et al. (2014) presented only one individual per ha for tree density of this species and the sapling density with 42 individuals per ha in the Kalimantan forest (Indonesia). A similar finding was also obtained by Hasibuan et al. (2016) in the Rumbio Forest and by Susilowati et al. (2019) in North Sumatra (Indonesia). They reported that E. longifolia was abundant in the seedling stage but the numbers decrease drastically in the mature stage. These authors stated that there is no tree stage at diameter ≥6 cm for E. longifolia. This plant performs mostly as a shrub and other tree species dominated their research location.

As shown in Fig. 4.3, among the mountains, A Luoi and Nam Dong include more large trees than Bach Ma; especially the plots AL3, ND3 and ND7 have a higher number of trees than saplings. Although the Bach Ma mountains' highest elevation is 1,440 m in asl., only two individuals are found at the altitude higher 700 m (Table 4.2). In contrast, A Luoi and Nam Dong still contain this species, even at an elevation up to 1,000 m. As Bach Ma National Park is a strictly protected area compared to Nam Dong and A Luoi, in our point of view, the possible reasons for the absence of E. longifolia mentioned above may relate to environmental factors, like soil condition or forest status. For example, while this species prefers open areas, the forests in Bach Ma above 700 m in elevation mostly contain high-value tree species of the Dipterocapaceae family, which form a very dense and high forest cover.

102 Discussion In the mountains, E. longifolia individuals are found from approximately 80 m up to 1,000 m along the ridges (70-80%), downhill (<10%), top-hill (10-20%) and very rare along the stream or at the forest with high density. However, depending on geographical and ecological factors and forest status, E. longifolia occurs at different elevations. For example, Susilowati et al. (2019) concluded that this species could not be discovered below the altitude of 250 m in North Sumatra (Indonesia).

Based on the results in Table 4.3, there is a tendency that the larger the tree diameter and height are, the lower the population density of E. longifolia would be. A large amount of studies indicates that this species belongs to shrubs or small trees (Chi, 2012); it is even listed as non- timber forest products (NTFPs) (Kartikawati et al., 2014). Nevertheless, the present study shows that among the total number of individuals found in the mountains, there is up to 31.4 percent of tree diameter class >6 cm in which trees with diameter >10 cm occupy 11 percent. Similarly, trees in the group of height class >5 m occupy up to 43.4% of the total individuals, whereas those trees higher than 10 m count 9.7% (Table 4.3). Under undisturbed growth conditions in the sandy areas, E. longifolia can grow nearly as high as in the mountains. Based on this result, this plant should be considered and listed as a medium size of timber tree.

Although the sapling and tree stage of E. longifolia occur in the research sites, the sapling and tree density are remaining low. The main reason for this low density is the excessive harvesting of its mature roots for medicinal purposes (field observation and local community), as local people prefer to collect the roots at the sapling or tree stage. Interviews with local people revealed that, from 2013 to 2015, this plant was gravely exploited in order to respond to the high demand. By pulling out the roots, local people destroyed many parts of the forest, especially in the lowland area. Among the mountainous forests, Nam Dong is one area that was seriously affected by human activities, such as land encroachment in order to plant acacia, road and hydroelectric dam constructions and illegal logging. Consequently, a large number of holes was left behind after root digging which also prevents that the species is the most abundant in this stage (sapling and tree stage).

5.1.2 Natural regeneration

Natural regeneration is an important measure for the rehabilitation of forests, the status of exiting population and species conservation. Natural regeneration of E. longifolia in the study

103 Discussion sites may depend on several factors such as the status of the populations, the level of disturbances, reproductive rate and seedling survival rate, geographic and ecological conditions, etc. The density of 0.095 seedlings per m2 for the mountain and 0.02 seedlings per m2 for the sandy area is low compared to Hutchings' research (1997), in which the medium density is at 0.144 seedlings per m2. Kartikawati et al. (2014) counted only 0.0071 seedlings per m2 of E. longifolia, so the density occurs in between these two studies.

Although the sapling density is the same between mountainous and sandy areas, the seedling density in the former area is higher than that in the latter. There are two reasons for this. At first, the big trees (DBH > 6 cm), which have a high chance of providing fruits and leading to a high density for seedlings, mainly grow in the hills. Second, the living conditions in mountainous area, including factors like elevation, location, soil moisture, soil pH-value, slope steepness, etc. are more favorable compared to that in the sandy area. Therefore, the mountains provide an ideal environment for seed germination.

According to Van and Cochard (2016), the location and elevation are essential factors for seedling survival and tree density or crown cover. They can also affect seedling recruitment and early survival. The present study shows that seedling density depends significantly on soil pH-value, as it tends to increase with the soil pH-value. Moreover, the offspring often competes with each other as well as with their neighbors. The number of seedlings tends to decrease gradually corresponding to their growing stages. Sandy areas with low soil humidity and pH- value and a soil texture with a high proportion of sand do not offer the optimal growing conditions for seedlings in comparison with grasses, herbaceous plants, or other tree species (Table 4.1, Cam et al., 2002). Therefore, this might be the main reason for the higher number of seedlings in the mountains compared to that in the sandy area. Besides, the topographical conditions of mountainous area may partially discourage local people from destroying mature trees for their daily needs, while in sandy areas; firewood activities are quite popular for them. This fact is one of the reasons why E. longifolia mainly occurs at the sapling form in the sandy area, but the number of saplings in this area is still as high as in the mountains.

The present results indicate that seedlings of E. longifolia can be found in clumps and they commonly distribute close to the mother trees (App. 7 - E. longifolia seedlings distribute surrounding its mother). This characteristic is also similar to the study of Susilowati et al. (2019) for this species who observed the seedlings with clumped distribution patterns. The

104 Discussion capacity of seed dispersal, a lower number of seed and seedling predators, or high survival rates (Okuda et al., 1997) may cause the clumping of seedlings. Because E. longifolia is a dioecious plant species, insects or wind often, carry out the cross-pollination process (Padua et al., 1999). In general, the seed dispersal depends on the pollinator condition, pollinators' movement, distance, etc. (García et al., 2007). Wind, insects, or water flow probably disperse them.

Eurycoma longifolia seeds are quite heavy (around 0.25-0.30 g) and the fruit diameter is between 0.5 cm and 0.7 cm so it is not easy to be dispersed across a long distance. Besides, the limitation of seed dispersal of E. longifolia might be caused by the bitter endocarp surrounding the seed, because most of animals do not prefer the bitter taste from the seeds or leaves. As a result, the seeds are dispersed close to their mother trees and the surviving seeds will develop into seedlings.

5.2 Morphological and anatomical adaptability of E. longifolia to different distribution areas 5.2.1 Leaf morphology and anatomy of mature trees under different site conditions

The leaf morphology, anatomy, or physiology are essential parameters in taxonomic studies to reflect the plants’ adaptive mechanisms to different living conditions (plant phenotypic plasticity). The present research focuses on estimating the main parameters, including leaf area, stomatal density, leaf length, leaf width and the number of leaflets of E. longifolia autochthone mature trees from contrasting sites (dry to humid site conditions).

Based on the geographical locations where the leaf samples were harvested for estimating the leaf traits, the study sites are divided into four areas. Also, E. longifolia from the mountains (A Luoi, Bach Ma and Nam Dong) often distributes in the understorey in the lowland rainforests while in the sandy area (Phong Dien), the canopy is more open with full sunlight. Thus, the former can be considered the moist or humid site and the latter as a dry site. Additionally, the leaves from the moist site develop more in the shadow than those of the dry site. So anatomically, “shade-grown leaves” and “sun-grown leaves” can be differentiated within this study. The present results show that the leaf morphology (leaf area, leaf length, leaf width, number of leaflets) and anatomy (stomatal density) traits differ among the sampled sites.

105 Discussion 5.2.1.1 Leaf area, leaf size and number of leaflets

The leaves from moist places are more prominent in the area, length, width and leaflet numbers per leaf compared with dry sites. The findings reveal that three ecological distribution areas in the mountain have remarkably larger area of the leaves (484 cm2 per leaf) than that of the sandy soil area (356 cm2 per leaf). It has been shown in Table 4.1 that the average soil pH-value and soil moisture from the mountains are much higher than those of sandy areas. Besides, the mountains have remarkably more precipitation and a lower temperature than sandy regions (Fig. 3.2 and Fig. 3.3). Hong et al. (2018) also concluded that the geographical and climate conditions might influence the photosynthetic and transpiration processes. Thus, the environmental factor is the primary influence in the distribution and development of plant growth, specifically the leaf area. It is clearly that the forest status such as crown cover or forest density in mountains and sandy soil sites which also plays a very important role in plant growth. Several studies concluded that shade-grown leaves are frequently larger, thinner and less deeply loped and have fewer conducting tissues, probably in order to position their leaves at the top of the vegetation gap for capturing more light and outcompete their neighbors (Medina and Mooney, 1983; King, 1998; Pallardy, 2008).

As mentioned above, the understorey plant species in the mountains are surely higher in humidity (soil and air) and lesser radiation (sunlight) and hence, the leaf size strongly responds to this factor by differentiating larger leaves. Schimper (1898) noticed that plants in dry areas have small leaves to mitigate their transpiration surface and overall water loss. Parkhurst and Loucks (1972) suggested that sun leaves are smaller than shade because it is probably to mitigate the boundary layer thickness, enhance convective heat loss, reduce transpiration and prevent the leaf from overheating. The present study also shows that the trichomes of the leaves appear both in sandy soil and mountainous areas. Thus, this characteristic may partly support for the above explanations. In dry conditions, the leaf area was severely reduced due to a reduction in leaf growth (Rowshanaie et al., 2014). As Mitchell (1998) stated, Taxus brefolia Nutt has longer leaves in the shade and shorter leaves in the sun. However, the leaf size is not always larger in the shade-grown sites. For instance, the leaves of a conifer like Picea sitchensis (Bong.) Carriere are short and narrow when shade-grown (Leverenz and Jarvis, 1980).

The number of leaflets from the moist sites is higher than that from the dry sites (33 and 31, respectively). However, there is no relationship between the number of leaflets and the plant

106 Discussion size (stem diameter and height). This indicates that this parameter has environmental flexibility independent of the size of plants. On the other hand, the leaf length and width not only display a positive strong relationship but also show a strong correlation with stem circumference and plant height. Besides, leaves of Bach Ma (moist site) are 11 cm longer and formed five leaflets more than those of Phong Dien (dry site), while the leaves’ size of the dry site are 2 cm narrower and 5 cm shorter than those of the moist site. It is well known that the width, or length of leaves, or the number of leaflets tend to increase correspondingly with moisture gradients, gloominess, rainfall or nutrient-high habitats (Webb, 1968; Walter, 1973; Hall and Swaine, 1981). For example, the average leaf width increases with the logarithm of annual rainfall on the relatively fertile and well-drained sites in the tropical lowland forests (Webb, 1968). Walter (1973) argued that the leaf size decreases in areas with droughts such as limited water capacity and lack of minerals because these conditions prevent cell growth and affect leaf growth. Limited water availability, rather than light availability, may limit plant growth in dry conditions. Thus, leaf morphology can be well adapted to moist and dry conditions and it can play different ecological roles in these two environments.

5.2.1.2 Stomatal density

Stomatal density or frequency (the number of stomata per unit area of one leaf surface) may fluctuate significantly within leaves and plant species to respond to the changes of the environmental factors, leaf morphology or genetic aspects (Colin and Mark, 1996). The results of the present study show that the average figures of stomatal density in moist and dry sites are 284 and 138 stomata/mm2, respectively. These numbers are lower than the findings from the research of Ichie et al. (2016) in the Malaysian rainforest with up to 301.4 stomata/mm2. This difference may be caused by ecological conditions, which has an influence on the physiological processes or in the all occurrence of stomata (Pekşen et al., 2006). However, the investigation by Ichie et al. (2016) was conducted among 176 individuals of 136 different tree species with three to five leaves for each plant, so that the average number of samples was only around 1.2 individuals per species. Their study is a general observation for many tree species at once and hence, the collected data were remarkably reduced for a single species.

It is crucial to have more research, which focuses on leaf anatomy for this species, as hardly any research could be found relating to this field of study, excluding the study of Ichie et al. (2016). The authors also provided general information about the stomatal frequency for

107 Discussion understorey species that are around 162 stomata/mm2 and for canopy gaps that mean strongest light intensity with up to 559 stomata/mm2. In comparison with this finding, the stomatal density, which was determined in the context of the present study in lowland mountains and sandy areas, is quite similar.

Stomatal frequency can fluctuate within leaves and plants and can be influenced by environmental factors such as water availability, light intensity, temperature and CO2 concentration (Colin and Mark, 1996). It is predicted that the leaves from those environments with lower air humidity would have a lower number of stomata per mm2 in order to reduce evapotranspiration. This hypothesis is based on some previous studies, stating that plants under insufficient soil water conditions can increase their stomatal frequency, leading to an increase in their transmission for gas exchange in photosynthesis (Schlüter et al., 2003). The higher density requires more water to be transpired and more CO2 to be taken up. Rajmohan (2014) also shows that when water is limited in a dry environment, photosynthesis can be harmed and the plants are going to dry out. Besides, the stomatal density of Euonymus europaeus L. from the sunny site is higher than that from the shady site.

The changes in stomatal density and size depend on different ecology conditions. The stomata type fluctuates from the understorey layer to the canopy layer in the forests and sandy areas. The transpiration of stomata affects the stomata types. Thus, stomata types may be necessary to understand how this plant's stomata can deal with the sunlight in the sandy area and the less light in the understorey canopy (Ichie et al., 2016). The anatomical investigation of the leaves shows that E. longifolia has mainly flat-type stomata in the mountains and pit-type in the sandy area. Ichie et al. (2016) also reported the predominantly flat or mound-type stomata in understored or canopy species, including this plant species, but they did not observe any species with pit-type stomata for understorey and canopy layers. It is concluded that understored and sub-canopy species have a small number of stomata and hairless, thin leaves because of their adaptation to shadow, humidity and high CO2 conditions in the forest. Chapman (1976) confirmed that many species have pit-type stomata in mangrove forests due to the lack of water capacity in the dry period. As our findings presented pit-type stomata in the sandy area, they could probably be adapted to dry conditions. As mentioned by Roth-Nebelsick (2007), the plant has to develop an active boundary layer due to the high pressure of water vapor together with

108 Discussion the limitation of hydraulic leaf conductance by great tree size and thereby decreasing stomatal transpiration under dry conditions.

In term of phenotypic plasticity, the leaves of Phong Dien reveal impressively a much larger range of the stomatal density compared to A Luoi, Bach Ma and Nam Dong (moist places). As observed from a field survey in the sandy area, the leaf size is smaller than that of the mountains and the leaves’ thickness from the former is thicker than the latter. According to several studies of McClendo (1962), Nobel et al. (1975) and Jurik (1986), the leaves in dry conditions are thick and the amount of chlorophyll per unit of leaf area increases because they have to arrange all chloroplasts along the cell surface of the mesophyll. Thus, there are strong correlations among photosynthetic capacity, leaf thickness and mesophyll's cell surface area. Pallardy (2008) confirmed that shade leaves have lower stomatal density, superior interveinal areas and a lower internal to external surface ratio.

A study on Taxus brevifolia of Mitchell (1998) presented a higher stomatal density (stomata per mm2) in the sun than in shade-grown leaves. Moreover, many studies in the past also concluded that leaves are often smaller and thicker from individuals grown in full sunlight or dry soil conditions and low air humidity than those in the shade (Salisbury, 1927; Penfound, 1931; Cooper and Qualls, 1967). Salisbury (1927) recognized that the values of stomatal frequency at the top of Mercuria lisperennis L. tends to reduce or disappear when this plant grows in humid conditions. In addition, two Abies species (Abies alba Mill. and Abies grandis Lindl.) grown in the shade had a lower stomatal frequency than those grown in the sun (Magnussen, 1983). Dökren and Hubertus (2018) also concluded that leaves of most Abies species in xeric conditions are smaller but they have a higher number of stomata per mm2 than those in moist conditions. As a result, humid conditions and drought have a high impact on the distribution and frequency of stomata.

5.2.1.3 The relationship between leaf area, stomatal density and tree size

The present study shows inverse relationships between leaf area and stomatal density. As explained above, the leaf area tends to be small in dry conditions, but it does not mean that the tiny leaves always have low transpiration and fewer stomata. Thoday (1931) reported that the plants with small leaves in dry conditions have a higher rate of transpiration surface compared to the plants with the larger leaves in moist sites. In humid understories, transpiration of stomata and leaf temperature decrease because the leaf size increases. It means that the leaves from

109 Discussion environment with low temperatures can drive up to mitigate the stomatal frequency (Medina et al., 1983).

Generally speaking, the lack of water in dry conditions can influence the leaf size by declining the cell size and causing an increase in stomatal frequency (Martínez et al., 2007; Xu and Zhou, 2008). According to Franks and Farquhar (2007), the number of stomata can rise when the leaf area declines because it needs to fit sufficient stomata units for each unit of leaf area surface and to respond to the CO2 flux as desired and to support the photosynthetic capacity. Salisbury (1927) discovered that the stomatal frequency is negatively correlated to the leaf size because of the high number of cells per unit area in the smaller leaves. Gay and Hurd (1975) explained that stomata are often produced in the early stage of leaf expansion. This means that the stomata can reach the highest density in juvenile not fully differentiated leaves and then decline to a relatively stable density in matured leaves.

The inverse correlation between the leaf area and the stomatal density was also discovered for Taxus brevifolia (Mitchell, 1998). Humidity and dry conditions have a high effect on the occurrence of stomata numbers and leaf size. However, stomatal density has a negative correlation with the leaf area and the tree size. It may be explained by the fact that the leaf area is reduced in a dry area which affects photosynthesis and finally tree growth. The influence of tree height on stomatal density is associated with the effect of light level, which depends on the tree’s position in the canopy (Schäfer et al., 2000). They found that stomatal frequency decreases as the tree height increases in Fagus sylvatica L. Recently, Peel et al. (2017) also confirmed that the stomatal density of Rhizophora mangle L. in mangrove forests is inversely correlated to tree height and diameter. Therefore, stomatal density depends on the plant growth (e.g. the leaf size and tree size), while plant growth also depends on habitat conditions.

5.2.2 Leaf characteristics of seedlings and mature trees of E. longifolia

It is important to observe the phenotypic plasticity of plants, because it is a promising mechanism to respond to the changes in the habitat conditions (Nicotra et al., 2010). Recently, Zulfahmi et al. (2019) revealed that E. longifolia has high variation among individuals in morphometric traits because of phenotypic flexibility. The pattern of leaf traits differs between mature trees and seedlings. In the present study, E. longifolia seedlings from several provenances were grown in the same conditions in the nursery garden to reduce the environmental variation. Thus, the leaf area of the seedlings from different provenances is nearly

110 Discussion the same, whereas a difference in stomatal frequency is found in the same environment among the investigated provenances.

The present study indicates that leaves of around two-year-old seedlings are much smaller than leaves from mature plants. The seedlings in the nursery garden capture more light than the natural seedlings. E. longifolia is an understorey species (Dasrul et al., 2018), which has the smaller leaves in seedling and sapling stages than the mature trees (field observation). Several species in emergent layer or deciduous species have larger leaves in juvenile stage compared to those of the mature trees, such as Scaphium macropodum Miq., or Campnosperma auriculatum (Blume) Hook. f. in Malaysia, because of the shade in the forest (Whitmore 1975; Hall and Swaine, 1981). In juvenile plants or in the case of heteroblasty, the leaves of seedlings or saplings tend to be larger and thinner and develop without any order, compared to leaves from mature individuals of the same species (Jones, 1999).

As there are no data on leaf traits for natural seedlings, it is impossible to compare their leaf traits to those from adult plants. In the actual case, the comparison of seedlings and mature trees shows unexpected results, namely that seedlings have a smaller leaf area but a higher number of stomata than the adult trees (246 and 175 stomata/mm2). To be more precise, the stomatal density of seedlings is much higher than that of mature trees from the moist site but lower than that of the dry site. Because the nursery garden is an open area with intense sunlight compared to its origins. This result indicates that stomata is a highly adaptive trait, which allows the plant to react to different environmental conditions. Recently, a study of Loc et al. (2018) also confirmed that some physiological characteristics (photosynthetic rate, chlorophyll a, b and relative water content) of E. longifolia seedlings grown in the nursery garden are relatively strong. These traits may support the plant in overcoming in drought, which present adaptation of this species under water stress conditions such as in a sandy soil area. It shows the high phenotypic plasticity of E. longifolia. Thus, the present study implies that E. longifolia is a tree species with extremely high adaptability to contrasting habitat conditions.

5.2.3 Seedling growth 5.2.3.1 Germination

Propagation by seeds is the most popularly utilized method to multiply plants because of its high efficiency. Seed germination requires several criteria, such as the quality of embryo, appropriate environmental conditions and primary dormancy status (Kozlowski and Pallardy, 1997). The

111 Discussion research shows that seeds from dry provenances germinate in a shorter period of time (15-36 days) and have a higher germination rate as compared to those from moist origins (15-51 days). Perhaps, the seed coats from the moist site are thicker and harder than those from the dry places, so the germination performance will be more difficult for these seeds. As the imbibition process will occur until the tissues have adequate moisture, the mechanism of water absorption of the seeds from dry sites will be faster than those from a moist site (Lars, 2007).

Rudrapal et al. (1992) showed a low percentage of germination rate due to the impermeability of the seed coat to water. Rolston (1978) also stated that the impermeability of the hard seed coat was typical for legume seeds. However, the imbibition rate also depends on the size, morphology, internal structure of the seeds, as well as temperature. Many species, including Acacia tortilis Hayne, Acacia mellifera Benth, Acacia hockii De Wild and Diospyros scabra Cufod from dry zones present a high rate of imbibition when suitable moisture is obtained. In legume species from dry places, for instance, their seeds are fully imbibed within a few hours. Smaller seeds often produce mucilage and seeds with smooth coats tend to be most proficient in water absorption (Bewley and Black, 1994).

In the present study, the length and diameter of fresh and dry fruits of E. longifolia from moist conditions are larger than those from dry conditions. In comparison with previous studies, the germination duration of E. longifolia seeds from the present research is faster than that in the study of Keng et al. (2002) who conducted their research with seeds collected from the secondary forest in Penang, Malaysia. It took 45 to 99 days for these seeds to ripe in a mixture of soil and sand. The inhibition and postponement of germination can happen due to the high endocarp impermeability to water or oxygen. In general, seeds of E. longifolia have a long period of germination, especially seedlings from mountainous area. This, however, is similar to the findings from another study by Keng et al. (2002). Marbach and Mayer (1974) highlighted that black exudates release phenolic compounds, which may contribute to water absorption of seed coats and hence, affect the seed germination positively.

Although there is a difference in the germination rate between seeds from moist and dry conditions, their survival rate of more than 80 percent was similar. A reason might be that all evaluated seedlings (age 22 months) were still kept in the nursery garden under optimal conditions. The seedlings’ survival capacity is not affected by unfavorable living conditions for both groups. Therefore, the germination of seeds depends on the provenances, while in similar habitats, the survival rate is the same among provenances.

112 Discussion 5.2.3.2 The morphological characteristics of seedlings in the nursery garden

Based on previous trials, it was able to determine that the optimal growing media for E. longifolia’s seedlings is sand: manure: phosphate (89%: 10%: 1%) and 50-75% of shade. Thus, this method was utilized to compare the growth pattern of seedlings from different provenances. Seedlings from moist provenances have lower germination rates but greater height and collar diameter and an increased number of leaves compared to those from xeric origins.

The growth of seedlings from xeric provenances is slower than that from moist derivations. The growth pattern of height and collar diameter of the seedlings - from one to age >22 months - from both site conditions are slightly different. However, the number of leaves produced by seedlings does not seem to follow the same growth pattern as that of seedling size. Initially, seedlings from dry conditions tend to have more leaves than those from humid sites. From five to nine months of age, the number of leaves from the xeric site uncommonly reduces before they experience a slight increase, while the leaf number from the moist site slightly decreased during this period. After twenty-two months, the seedling leaf numbers from xeric conditions are still lower than those from humid conditions. A similar finding is found for the seedling height and diameter patterns.

It is quite interesting that the leaves tend to fall in the periods of 5 to 9 months and 14 to 20 months. Thus, the seedlings need to develop the new leaves within a period of 5 to 6 months. The reason for the reduction of leaves may be correlated to different growing periods. Under natural conditions, the leaf age and fall of whereas the plants grow. Even under the optimal conditions as arboretum, the older leaves do not receive enough light or water, therefore these leaves will slowly yellow and fall from the plant.

Keng et al. (2002) found that the height, collar diameter and the leaf numbers of E. longifolia’s seedlings follow the same growth patterns, which slightly increased when they are from two to twenty-eight months in the same environmental conditions. However, their study also showed that the growth pattern of leaf numbers reduce slightly or keep stable in some period. Chan and Toh (1984) also suggested that Carica papaya L. grown in constant living conditions from different provenances in Malaysia have a similar growth pattern. In conclusion, the germination rate of E. longifolia is low, but it is different between dry and moist sites. The seedlings' growth patterns in the nursery garden are quite similar and seedlings from mountainous provenances growing faster than sandy provenances.

113 Discussion 5.3 Genetic diversity of E. longifolia 5.3.1 Comparison of the utility of different marker systems

The PPF, PIC, MI, EMR and Rp criteria are mainly used to evaluate the informativeness of the applied marker system or the marker power (Varshney et al., 2007; Kumar et al., 2014; Etminan et al., 2016). The present study used two techniques to compare the efficiency of the applied markers and to generate polymorphisms across 276 mature accessions and 269 seedling samples and 15 mother trees of E. longifolia. Regarding the number of polymorphisms detected (PPF, PIC, EMR, MI, Rp), SCoT and BPS markers are slightly more informative than RAPD markers in the assessment of genetic diversity for this plant. However, based on the findings of the Shannon index and expected heterozygosity, both marker systems show nearly the same amount of polymorphisms detected.

In the present study, the three SCoT and BPS markers show 100 percent polymorphisms (PPF), which is similar to the finding from several previous investigations, including those of Etminan et al. (2016) and Morgenstern et al. (2016). Besides, the average PIC and Rp values are comparable to those, found by Etminan et al. (2016) and Hamidi et al. (2014) who used SCoT marker in wheat, or by Kumar et al. (2014) who studied Justicia adahatoda L. However, none of those previous studies has focused on the effectiveness of these two marker systems by studying E. longifolia. The only exception is a study by Fadilah et al. (2019), who published an estimation of IRAP (Inter-Retrotransposon Amplified Polymorphism) markers’ potential via their obtained PIC, EMR, MI parameters in E. longifolia. The IRAP markers are also dominant markers like RAPD, SCoT and BPS which make the results comparable to the present investigation. The PIC value from the IRAP marker in their study was higher, but EMR and MI values are lower than those obtained from RAPD, SCoT and BPS markers.

When comparing PIC values between RAPD and SCoT and BPS markers, the latter PIC is smaller than that of the former, which might be due to the lower number of polymorphic fragments of the RAPD marker. The marker index (MI) can be considered as a standard measure of efficiency in defining polymorphisms (Varshney et al., 2007). MI of SCoT and BPS (1.03) doubles that of RAPD (0.57). This character makes the SCoT and BPS marker system applicable to fingerprinting approaches or assessing genetic variation for breeding programs. For example, Varshney et al. (2007), Etminan et al. (2016) and Milbourne et al. (1997) also used the marker index (MI) as a fundamental parameter to compare different marker techniques. According to Etminan et al. (2016), an appropriate marker system's feature is the capacity to

114 Discussion differentiate among several accessions. In general, the present research exposes that the resolving power of SCoT and BPS primers is higher than that of RAPD primers, except for SCoT36. MI and Rp are suggested to be the most popular marker parameters for choosing informative primers. Therefore, the SCoT21 primer is defined as the best primer for distinguishing between the 276 E. longifolia genotypes tested.

5.3.2 Genetic diversity between different distribution areas

The genetic diversity of species in small population sizes is often lower than that of larger populations due to the genetic drift and inbreeding (Willi et al., 2006; Li et al., 2012). Thus, those species with narrow geographical distributions achieve lower diversity than the species with widespread geographical distances (Hamrick and Godt, 1996). Wu et al. (2015) concluded that some rare and endangered species (e.g. Rhododendron) could maintain a high level of genetic diversity even at a small population size. Thus, as the distribution area of E. longifolia populations in Thua Thien Hue province is quite narrow, the unsustainable harvesting and illegal logging of this tree species have caused its depletion. Hence, it is necessary to get detailed information on the genetic structure and population genetics for appropriate management and conservation programs.

The two marker systems (RAPD, SCoT and BPS) were adopted in the study to assess the genetic structure and diversity of E. longifolia populations. The SCoT and BPS markers were selected for this investigation as they can help to mitigate the drawbacks of the available RAPD markers. Additionally, the SCoT and BPS primer sequence is much longer than the sequence of RAPD primers, thus, they promise a high level of polymorphism and reproducibility of the findings. Based on Culley’s study (2005), it was decided not to calculate observed heterozygotes (Ho), genetic differentiation (Fst, Gst), or Nm (gene flow). This is because these two dominant marker systems (RAPD, SCoT and BPS) have certain limitations that do not reflect the real results.

The Shannon index (the percentage of polymorphic loci) varies from 0 to 1. The closer the values are to zero, the lower is the genetic diversity (Silva et al., 2015). However, Soares et al. (2016) and Silva et al. (2015) considered this parameter as an essential measure of genetic diversity. Several studies such as Nei (1987) and Culley (2005) concluded that the expected heterozygosity is more appropriate to describe genetic diversity than the Shannon index. Based on these debates, the present research took both, expected heterozygosity (He) and the Shannon index into account to get a full understanding of the genetic diversity. The previous studies on

115 Discussion the genetic diversity of this species mainly focuses on estimating expected heterozygosity, rather than Shannon index (Tnah et al., 2011; Osman et al., 2003; Rosmaina and Zulfahmi 2013).

The Shannon informative index in the present study is 0.27, which can be reflected as low due to it is below 0.5. The result of the present study for this index is nearly the same as that of the study of Fadilah et al. (2019) with I = 0.33 (IRAP markers). The level of diversity of E. longifolia in the present study (He = 0.17) is nearly equal or slightly less than that of some other reviews for this species, e.g. Osman et al. (2003): He = 0.22 (SNPs), Rosmaina and Zulfahmi (2013): He = 0.20 (RAPD), Fadilah et al. (2019): He = 0.22 (IRAP). These authors used a small number of accessions to estimate the genetic diversity. For example, Osman et al. (2003) collected 47 samples from six populations and only six to nine samples per population, while Rosmaina and Zulfahmi (2013) gathered twenty-five accessions for five populations within the province of Riau (Indonesia). Genetic diversity does not only depend on the marker techniques but also on the number of samples under investigation. Therefore, it was tried to ensure a sufficient number of samples to increase the reliability and validity of the study.

Monfared et al. (2018) concluded that genetic characteristics are influenced by both polymorphic and monomorphic fragments. Thus, a small percentage of polymorphic bands (66.5%) will cause a low genetic diversity in E. longifolia. Regarding polymorphic fragments, other studies also found numbers between 57 and 64 percent, which are not too high for genetic diversity (Osman et al., 2003; Rosmaina and Zulfahmi, 2013; Fadilah et al., 2019). The genetic diversity in E. longifolia is quite low, but it is within the limits observed in other studies. So, our result is similar to that of Salvadora persia L. by using ISSR markers (He = 0.17; Monfared et al., 2018) or other long-lived perennial plants investigated by allozyme markers (He = 0.18; Hamrick and Godt, 1996) and it is higher than that of other widerspread tree species like Ulmus laevis Pall., He = 0.09 (Vakkari et al., 2009) and Prunus mahaleb (L.) Mill., He = 0.14 (Jordano and Godoy, 2000).

The findings of the AMOVA analysis explain that the major part of genetic diversity occurs within the surveyed populations, with 87% based on SCoT and BPS markers and 94% based on RAPD markers. Therefore, this result indicates that E. longifolia genotypes within populations show a greater genetic diversity than those among populations. The high genetic variance within-population is caused by the high level of gene flow and population mixing (Hague and Routman, 2016). Zulfahmi (2013) also reported that this species maintains much

116 Discussion more significant genetic changes within populations than among them. Hamrick et al. (1992) stated that the genetic variation of the woody perennial species significantly occurs within populations, in particular in tropical trees. For out-crossing species, the genetic differentiation among-population ranges between 15 and 38% (Bussell, 1999).

Based on the two marker techniques, the genetic differentiation among-populations (11%) shows that E. longifolia has a comparable low genetic diversity, even lower than the findings reported by Susilowati et al. (2012), who studied five E. longifolia populations (22.34%) in Indonesia. The main reason might be that this plant is an outcrossing species and has unisexual flowers, in which male flowers have only one sterile pistil and female flowers have several sterile stamens. Although gene flow is a crucial factor for identifying the plant genetic variation, this parameter is not used to assess genetic diversity with dominant markers due to their limitation. However, the authors collected only five-leaf samples per population, so the sample size may affect the results of their research and should be taken into account.

5.3.3 Genetic structure of different distribution areas

The genetic structure is influenced by many factors, including population size, genetic drift, gene flow, seed dispersal, breeding, or natural selection (Hamrick and Godt, 1996). Depending on Nei’s genetic distance, AMOVA and genetic differentiation among populations (ΦPT), the genetic structure can be characterized.

The genetic structure of E. longifolia populations shows a significant subdivision among populations, but the pattern of genetic structure, especially genetic differentiation, is weak (ΦPT = 0.112, p < 0.01). In comparison with the results obtained by SNPs and RAPD (Osman et al., 2003; Rosmaina and Zulfahmi, 2013), the genetic differentiation in the present study is lower, which indicates an increased gene flow, possibly due to a large distance of seed and/or pollen dispersal. These authors mentioned that a higher genetic differentiation might have resulted from the small number of examined accessions per population.

Hamrick and Godt (1996) reported that wind or insect-pollinated out-crossing species display a higher genetic variation within populations with a low level of genetic differentiation among- populations. Several tropical tree species have high levels of genetic variation but low or moderate levels of genetic differentiation, even in populations with a distance of only ten kilometers (Lacerda et al., 2001; Winkler, 2011). Winkler (2011) explained that the relationship between the pollinators and seed dispersal plays a vital role in contributing to the genetic

117 Discussion diversity within and among populations. E. longifolia seeds are quite large, but can be dispersed over long distances by rainwater flow, birds, or rodents (Susilowati, 2008). When seed dispersal is carried out by these vectors, the plants tend to have a high level of genetic variation within the populations (Hamrick et al., 1992).

The genetic distance specifies the genetic relationships among populations (Finkeldey and Hattermet, 2007). According to the present results of genetic distance, Nam Dong and A Luoi are regarded as one group, while Bach Ma and Phong Dien represent a second group. These conclusions have nearly a similar range of genetic distance among populations. This result is supported by the PCoA analysis, which also divides these four populations into two groups without a complete clear-cut distinction. The findings of the Bayesian clustering analysis of the genetic structure using STRUCTURE as software presents that the examined E. longifolia populations are best to fit into two potential genetic groups (K = 2), consisting of Bach Ma and Phong Dien (group 1) and Nam Dong and A Luoi (group 2). Evanno et al. (2005) proposed to use an ad hoc statistic ΔK depending on the rate of change in the log probability of data among the true number of clusters (K values). Furthermore, this approach has the capacity to use the genetic information to define the population relationship between the samples without any supposedly predetermined populations (Chen et al., 2017).

The results of the PCoA and STRUCTURE program are generally corresponding to the genetic relationships described in the dendrogram of the Ward’s method. These genetic relationships are not clearly shown in the dendrogram of UPGMA. The Ward’s method was applied in order to calculate the genetic differentiation between two of the above potential genetic groups by estimating the Euclidean distance from 276 analysed samples. The populations of Nam Dong and A Luoi have comparatively higher levels of genetic diversity in comparison to those of Bach Ma and Phong Dien. As Nam Dong, Bach Ma and A Luoi are geographically closer to each other than to Phong Dien, it is hypothesized that the populations in the mountains should have the same genetic diversity. However, Bach Ma and Phong Dien unexpectedly, create one group separate from Nam Dong and A Luoi.

The reason for these unexpected results may be firstly: three populations are located in the mountains and one in the sandy area. The populations are not so far away from each other in a maximum of around 140 km. Although Bach Ma is more adjacent to Nam Dong than to Phong Dien (sandy area), it is close to deltas and sandy regions. Besides, E. longifolia seeds can be

118 Discussion transferred by the lagoons and stream river (water flow) along the province through the sandy area districts, which are close to Phong Dien district (sandy area). Several previous studies have reported that the pollination process often occurs in a limited distance (few hundred meters), but Kremer et al. (2012) and Hague and Routman (2016) observed a movement of pollen further than 100-km distance. The seed dispersal of this species might also be possible by wild birds. Even though E. longifolia fruits are bitter, their colorful layer is the main attractive trait for wild birds. When these birds fly further, the seed dispersal distance will increase (Fadilah et al., 2019). Therefore, water flow and wild birds might possibly be the reasons why the fruits were transferred from Bach Ma to Phong Dien. The isolation pattern of a population by a distance is usually found in those tree species with their pollination by insects. That means, the seed dispersal at the long-distance within the natural range should be taken into account (Bekessy et al., 2002; Juchum et al., 2007).

Secondly, the results highlight that the Bach Ma population has the lowest genetic diversity in comparison with the other populations. The number of analysed accessions in this study may influence the low level of genetic variation of this population as Bach Ma has the lowest quantity of examined individuals (24 DNA samples). According to Egbadzor et al. (2014) and Kapoor et al. (2000) the higher the number of samples is, the more significantly the genetic variation is detected. Therefore, as Bach Ma has a low number of polymorphic loci and genetic diversity, this may affect the findings of the genetic structure. This is in contradiction to the finding that Phong Dien which is the site with the highest number of analysed samples, shows lower variation parameters in comparison to Nam Dong and A Luoi. This result suggests that the number of individuals may also not have any influence on diversity factors, which is confirmed by other studies, e.g. Chen et al. (2017).

Thirdly, during the Vietnamese-American war (1955-1975), Agent Orange weapons seriously destroyed the mountainous area, especially A Luoi and Bach Ma (Robert, 2016). The United States sprayed herbicides in Vietnam to defoliate 1.1 million ha with ca. 10% of the country (Westing, 1984). The war has dramatically influenced the canopy cover, flora and fauna biodiversity and forest genetic resources. Once the forests were destroyed, its fragmented areas also changed the future forest succession. Regarding geographic distance, Nam Dong is close to A Luoi. The pollination (gene flow) may occur from Nam Dong to A Luoi. After the war, A Luoi surely had some new invasion plant species. The former species (e.g. E. longifolia) continue to develop by support from other areas (Nam Dong or Tay Giang area of Quang Nam

119 Discussion province). Thus, the war factor and geographic distance are probably one reason to create the same genetic group containing of A Luoi and Nam Dong.

Finally, the field survey in Bach Ma indicated that it is not easy to find E. longifolia individuals in areas higher than 700 m in elevation. Thus, the geographical and ecological factors might limit the distribution of this species and could probably contribute to the isolation of the populations of Nam Dong and A Luoi. This is confirmed by the Mantel test, which indicates a significant correlation between geographic distances (including elevation factor) and genetic diversity (r2 = 0.19, p < 0.05). It shows considerable isolation by distance and elevation patterns for E. longifolia, indicating that genetic isolation has a significant effect on the genetic variation and distribution structure of this plant species. The Mantel test also revealed that genetic distance and morphological factors have a significant correlation, but their relationship is not strong enough. Together with the limited morphological traits (only tree height and tree diameter), this finding indicates that morphological traits may not provide sufficient information for genetic diversity.

5.3.4 Comparison of the genetic variation between mature trees and seedlings

The differences within or among populations (provenances) can be detected in seeds, seedlings and spatial distribution patterns (Stihl and Persson, 1991; Varelides et al., 2001; Ivetíc et al., 2012). Also, genetic diversity across the different generations is assessed through embryo and pollen donors in the study of Kassa et al. (2018). The present study focuses on assessing the genetic diversity in different generations through 269 seedling accessions from 15 mother trees (four provenances). Currently, deforestation and land-use changes have intensely occurred in many places within the province, especially in Nam Dong forest and Phong Dien (sandy area). The present study focuses on detecting the mature trees' genetic variation, the offspring and the corresponding mother trees; precisely the genetic diversity of seedlings from different provenances, between propagated and natural seedlings.

Overall, there are slightly differences in genetic variation between E. longifolia mature trees and their offspring. The values for A Luoi offspring's genetic parameters are much higher than that of the adult population, while Bach Ma's offspring is slightly higher than the mature trees (Table 4.17). A Luoi district is a remote area of Thua Thien Hue province located at a high altitude (average of 600-800 m asl.). The illegal logging or land-use change activities in this area are low compared to lower areas like Nam Dong or Bach Ma National Park's buffer zone.

120 Discussion As previously mentioned, A Luoi was seriously affected by the Vietnamese-American war (1955-1975). The activities going along with the war destroyed the forests, which may change in plant population patterns (forest composition). Some new species dominated, or the previously dominant species had to establish their development again. Perhaps, E. longifolia is one of the examples reflecting this situation. This fact is usually a form of the selection, which reduces genetic diversity of that population, but in the present study shows, the next generation from A Luoi has a greater genetic diversity compared to Nam Dong. Thus, the war factor may not directly affect the genetic diversity in the next generation of E. longifolia in A Luoi.

According to Krabel (2016b), the survival and adaptability of a plant population depend on the number of antagonistic genotypes available in that current population. An adaptation facing inconvenient environmental factors can occur several decades. Hague and Routman (2016) also revealed several factors that could affect genetic variation such as population size, geographic location, population history and subdivision. According to Wright (1951), bottlenecked populations are presumed to decrease additive genetic variation, but in some cases, such population can occur an increase in additive genetic variation. The variation in a trait under selection depends on the additive genetic diversity (Ledig, 1992). Genetic variation included in adaptation to new environments is properly additive, while the strength variation in the environment to the population which have been adapted long-term. Thus, there is often a predominant expression of non-additive variation such as the increase in additive genetic variation in population bottlenecks (Ledig, 1992; Rollin et al., 2013).

Based on the results of leaf traits, E. longifolia has a high phenotypic plasticity that can adapt to a wide range of habitat conditions. The adaptive capacity of species to new environments may depend on the species’ ability to respond to natural selections (Kaňuch et al., 2014). Lewig (1992) explained that inbreeding might increase phenotypic plasticity by exosing recessive alleles. Total variance will be raised because of gene drift or migration among sub-populations (Hague and Routman, 2016). In the circumstance of E. longifolia in A Luoi, the gene flow can be supported from the adjacent populations after the war, e.g. Nam Dong or two areas of Quang Nam province (Tay Giang and Dong Giang).

Conversely, the value of genetic variation for the Nam Dong offspring is lower, while the value of the seedlings of Phong Dien is nearly the same as that of both the mother and mature trees. These results suggest that the genetic diversity of E. longifolia has not changed in the whole population, but on a smaller scale, significant changes can be observed. The genetic diversity

121 Discussion of the Nam Dong population considerably declined in the offspring generation, so it is highly recommended that there should be some protection measures for this population to maintain its genetic resources for the future.

In this research, propagated and natural seedlings are used to estimate the genetic changeability in regeneration compared to the mature trees or the corresponding mother trees. Because Bach Ma National Park is a protected area, it is not allowed to visit the area several times for seed collection. As we could not find enough seeds from this area for propagation in the nursery garden, we collected the wildlings instead of seeds. Although there are four provenances, the results can be compared only between two provenances due to the lack of the data from Bach Ma and Phong Dien sites.

The findings show that there is almost the same genetic variation between naturally regenerated and propagated seedlings in A Luoi, but the pattern recorded in Nam Dong provenance is different. The genetic variation from Nam Dong naturally regenerated seedlings is considerably lower compared to that from propagated seedlings. Therefore, the reduction of genetic diversity in Nam Dong provenance is mainly caused by the low genetic diversity in natural seedlings. However, it is not so certain to conclude that the values of genetic parameters from propagated regenerations are much higher than naturally regenerated seedlings due to the result of only one provenance. Therefore, it is necessary to conduct more research related to this aspect.

The genetic diversity from the mother trees is lower in comparison to their corresponding seedlings. Possibly, this is because of the small accession size for mother trees (Egbadzor et al., 2014; Kapoor et al., 2000). It was impossible to harvest the seeds from many mother trees; thus, it has been decided to increase the number of seedlings per mother tree. Based on the findings of Nei’s genetic distance and differentiation, a small genetic difference between mother trees and regeneration is also observed. Thus, genetic variation between mother trees and seedlings and between the mature trees and offspring are not large.

5.4 The correlation between eurycomanone content and the distribution areas

Eurycomanone content was selected in this study because it has been reported as one of the main quassinoids in the root of E. longifolia and the concentration of this compound largely contributes to the biological activities of this plant (Kardono et al., 1991; Low et al., 2013). In addition, eurycomanone is considered as a unique constituent of this plant species that can enhance the testosterone and anti-malaria activity (Chan et al., 1986).

122 Discussion The moist site's root water content is significantly higher than that of the dry place because it is relatively positively correlated with elevation and tree height. The growth of plants often relies on ecological, geographical or topographical factors. Thus, the plants growing at sufficient conditions habitats will vigorously develop in biomass. The plant species with long-live ages and in hard living conditions will have a high content of metabolites. The result shows that eurycomanone has low content, which may be due to less tree age (8-17 years old.). The present study notes that the accumulation of eurycomanone content depends on the ecological distribution sites, which has a higher content in the moist places (1.22 mg/g on average) compared to that in the dry site (0.74 mg/g on average). Also, eurycomanone content positively correlates with the root water content, the root diameter and the elevation factor. The secondary data show much lower precipitation from the sandy soil sites than those of the mountains. Although the present study does not focus on estimating the precipitation, this factor may also be considered one of the significant parameters that can affect the accumulation of eurycomanone.

In the more humid place, the amount of moisture in the roots can increase the metabolic process in the roots; thus, it may raise the accumulation of this compound. This result is quite similar to the findings of Mohamad et al. (2013), who investigated the eurycomanone component of E. longifolia from Muar - a mountainous area called Johor in Malaysia with nearly 1.4 mg/g. The present content is much higher than that found by Osman et al. (2016), who extracted eurycomanone from E. longifolia’s roots by SPE-HPLC from various types of rainforests in Sarawak, Teman Negara, Jengka in Malaysia. These authors found that the obtained amount of eurycomanone is very low, within a range from 0.18 to 0.44 mg/g. Recent research has revealed that the average amount of eurycomanone content of E. longifolia is 1.19 mg/g from six provinces in Vietnam (except Thua Thien Hue province) by LC-MS/MS (Dung, 2018). However, up to now, none of the previous studies report this compound extracted from a plant grown in sandy soil areas in Vietnam and other countries.

It is hypothesized that bioactive compounds might be influenced by genetic factors, the ecological distribution of the plant species and tree traits. Nevertheless, the Mantel test does not show significant correlations between eurycomanone content and genetic diversity; thus, it might be concluded that genetic factors do not influence the concentration of eurycomanone. The data do not show any significant correlations between eurycomanone content and tree age. It more or less indicates a relationship between this compound and root diameter, which is

123 Discussion strongly influenced by tree age, tree diameter and plant height. The tree age was estimated via an approach of indigenous knowledge, which could probably cause some inaccuracies together with a small accession size (30 trees).

Zhang et al. (2019) showed that the moist content of the root had wide variability in Heterocarpus altaicus (Willd.) and Poa sphodylodes L. Their research also suggested that diverse plant roots may have different capacities or requirements to receive the moisture from the soil, so similar plants may still have different root mechanics and qualities. Hales et al. (2013) concluded that the change in moist content along roots might influence the mechanical hydration of root tissues. Several studies (Zhang et al., 2019; Guo et al., 2013) also showed that root moisture content affects the root diameter and tree size. However, the present study does not find any effects of root water content on root diameter but tree height (Table 4.21). In conclusion, there is no evidence for a correlation between the eurycomanone content and factors of the genetic diversity. Instead, it seems to depend on the ecological distribution areas and the diameter of the root.

124 Conclusions and recommendations 6 Conclusions and recommendations

6.1 Conclusions

The distribution status of E. longifolia populations has been decreasing considerably due to the overharvesting of this species. Therefore, this research is necessary to provide basic information for propagation, domestication, breeding programs and the conservation of this valuable medicinal plant. This study is also a small contribution to enhance the understanding of this important forest genetic resource since the genetic diversity of this species is reducing. The research findings will contribute to species information on the adaptation ability of phenotype and genotype and bioactive component accumulation in the mountains (moist site) and sandy areas (dry location). The study was conducted in Nam Dong, A Luoi, Phong Dien districts and Bach Ma National Park, Thua Thien Hue province in central Vietnam. It focuses on several aspects of the tree species E. longifolia, including its ecological distribution, leaf morphology, seedling growth, genetic diversity and eurycomanone content in mountainous and sandy areas. The findings can be summarized as follows:

This tree species is mainly distributed at an elevation from 300 to 700 m asl. The density of trees and saplings increases with altitude. E. longifolia at tree size (≥6 cm in DBH) is only discovered in the mountains with 9.49 counts per ha and Nam Dong has the highest tree density and tree height. Tree density is lower than sapling density, especially in Bach Ma site with only 1.29 trees per ha compared to 9.96 saplings per ha. However, the sapling and tree density is still low. Seedling density and tree height from mountainous area are higher than those from the sandy area. The offspring of E. longifolia commonly appears in clumps and distributes in the surrounding of its mother trees. There is no significant difference in the tree diameter among mountainous areas. The sapling density from Nam Dong and A Luoi is much higher than those from Bach Ma and Phong Dien. Therefore, it can be predicted that in the next generation, the tall trees (at tree size) of E. longifolia may concentrate on these two regions.

The survey results show a difference in tree and the seedling density between mountainous and sandy areas, but the sapling density is nearly identical. Therefore, the first hypothesis is partly accepted. The population density of E. longifolia is relatively low. A single stem of E. longifolia mainly occurs in mountainous area, while several stems or branches occur in sandy area. This plant species can be listed as a medium timber tree. In Bach Ma, this species does not distribute at an elevation higher than 700 m, while it appears up to 1,000 m in A Luoi and Nam Dong areas.

125 Conclusion and recommendations Most of the leaf traits (leaf area, leaf size, number of leaflets) and stem sizes from plants of the mountainous area are larger than those from the sandy area. However, the stomatal density from the dry site is twice as much as the moist site. Thus, the hypothesis of H2 is partly received. The results confirm that the leaf area is inversely related to the stomatal density and dry conditions may increase the number of stomata, whereas the leaf size decreases. Leaf length has a positive correlation with leaf width, leaflet number, stem diameter and tree height. In addition, the plants at higher elevation appear to have a larger leaf size but the number of leaflets is not affected by this parameter.

The fresh and dry fruit sizes from dry sites are smaller than those from plants in moist places. The germination rate of seedlings from the sandy area (73.7%) is higher than that of mountainous regions (45.4%). However, the survival rates of seedlings between the two areas are not significantly different. The seedlings from mountainous provenances have a more robust growth performance in height, collar diameter and leaf number than those from the sandy area. However, under the same growth conditions, the seedlings from the same seed lot show different growth pattern and seedlings from different provenances grow nearly in the same way. Accordingly, the third hypothesis is rejected. The leaf area of seedlings from different provenances is similar, but show a contrasting pattern of the stomatal density. The comparison of the seedlings and the mature trees shows that the former has a smaller leaf area but a higher stomatal frequency than the latter. In conclusion, E. longifolia is a tree species with vastly high adaptability (phenotypic plasticity) to contrasting environmental conditions.

Five random RAPD, SCoT and BPS primers (OPB05, OPC02, SCoT21, SCoT36 and LA2a) were selected to give reliable amplification results of the E. longifolia DNA samples. Among those, SCoT and BPS markers are more informative than RAPD markers, whereas SCoT21 primer performs as the best primer for detecting the genetic variation of E. longifolia. The total genetic diversity of the four selected E. longifolia populations is found to be low (He = 0.17 and I = 0.27). A low polymorphism (66.5%) may cause low genetic diversity of this species. In the present study, molecular variances within populations (89%) are higher than those among populations. Thus, the hypothesis H4 is accepted. The investigated traits and the mating system could induce high genetic variation within populations. In contrast, the lower genetic differentiation among populations could be caused by effective gene exchange, characteristics of outcrossing species and the particular geographical distribution.

126 Conclusions and recommendations The genetic structure analysis of E. longifolia populations presents a significant subdivision among the groups, but the genetic differentiation is not vigorous (ΦPT = 0.112). Based on the result of the Bayesian clustering analysis, the genetic structure is presented as two potential genetic groups of E. longifolia, which are reflected by the K value. Group 1 consists of two populations from the mountains (A Luoi and Nam Dong) and group 2 includes two populations from mountainous and sandy areas (Bach Ma and Phong Dien). The results of PCoA and clustering analysis by Ward’s method also reflect this result. These findings lead to the acceptance of hypothesis H5. There is a significant correlation between genetic diversity and geographical distances including elevation. It implies that genetic isolation has a relative effect on genetic variation and distribution structure of E. longifolia in Vietnam. Although there is a significant relationship between genetic diversity and morphological traits, the limited morphological features may not fully explain genetic variation.

Regarding the genetic variation between the mature trees and the offspring, there are slightly differences in genetic diversity among them. The genetic diversity of this plant species has not differed in the whole population, but significant changes can be detected on a smaller scale. In specific, the genetic variation of the Nam Dong population decreases in the offspring compared to both naturally and propagated seedlings; thus, its genetic diversity may be predicted to be low in the future growth stages. A Luoi and Bach Ma increase relatively in the level of the genetic diversity in the next generation.

The root water content from mountains is significantly higher than that of the sandy areas and it has a positive relationship with elevation and the tree height. The present study notes that the eurycomanone content has a relatively positive correlation with the water content in the root tissues of E. longifolia. It means that the concentration of eurycomanone depends on the ecological distribution sites, which increases this compound in moist sites than that of a dry place (1.22 mg/g and 0.74 mg/g on average). There is no correlation between the genetic factor and eurycomanone compound. However, this content is affected by some biological and geographical factors such as root diameter, the water content of the root tissues and elevation. Thus, the final hypothesis is partly received.

6.2 Suggestions for further research

Based on the results, the following suggestions are provided:

127 Conclusion and recommendations The most massive trees (ca. 23 m in height and 24 cm in DBH) are found in Nam Dong and A Luoi mountains. It indicates that E. longifolia is not a shrub or a small timber tree as stated in several previous studies. Therefore, it is highly recommended to list this species as a medium size timber tree.

The field survey results report that E. longifolia is seriously exploited, leaving a large number of holes in the lowland areas, especially in Nam Dong forests and sandy areas. Accordingly, the protection of these two areas is essential for the sake of its preservation.

The results show that this species appears mostly in the areas below 700 m in Bach Ma due to the habitat conditions such as species composition, soil characteristics, climatic factors, etc. Thus, future detailed studies are crucial to reveal those aspects.

The relationships between leaf morphological and anatomical characteristics of E. longifolia plants from the mountainous and sandy areas should be investigated in more detail. For example, the previous studies showed that not only the density of stomata but also its size might correlate to the leaf size. In addition, further studies of the adaptation process of leaf stomata from seedlings to mature trees should be considered for a better understanding of how this species adapts to different environmental conditions.

Regarding the limitations of investigating the genetic diversity in this study, there is a need to collect samples from different regions and countries in the future. The present findings suggest that the degree of logging activities in our study sites in A Luoi, Bach Ma and Phong Dien does not affect the genetic diversity of the next regeneration. However, the genetic variation found in Nam Dong is reduced in both naturally and propagated offspring. It is necessary to conduct further detailed research to estimate the genetic diversity of the seedling generation. In addition, this aspect should be compared with that of mature trees, especially in this region.

Further studies should also focus on root characteristics of E. longifolia in moist and dry sites to explain why the eurycomanone content in the tree from mountainous area is higher than that from sandy areas. In addition to the evaluation of eurycomanone content, other components of this species should be considered for further analyses. Finally, the study area should be expanded and more samples from other regions should be collected to provide the comparative patterns of this compound throughout different ecological regions of Vietnam.

128 Summary 7 Summary

Introduction and Objectives

Eurycoma longifolia is a valuable medicinal plant species distributed in tropical forests and dry sandy regions in most Asian countries, including Vietnam. The plant has a rapid decrease in its natural populations due to the unsustainable harvesting for medicinal purposes and high demand for commercial trade in Vietnam. It is crucial to research species information, such as natural distribution, species identification (phenotype and genotype), or bioactive compounds. Thus, the present research is necessary to provide the basic information for propagation and breeding programs and to contribute to conservation of the species. In detail, the present study enhances the grasp of population status, adaptability of phenotypic plasticity through morphological and anatomical traits. Besides, the research findings will contribute to distinguish the genetic relationship within and among populations, population genetic structure and define the impact factors on the accumulation of eurycomanone - one of the main compounds in E. longifolia tissues.

Material and Methods

The study was conducted in four areas, including Nam Dong, Bach Ma, A Luoi (mountains) and Phong Dien (sandy area) in Thua Thien Hue, located in the Northern Central Coast region of Vietnam. Data were collected from the field survey and laboratory works. Regarding field work, 27 transects (74.24 ha) were designed to estimate the density and size of trees and saplings, while the natural regeneration was surveyed within 47 plots (500 m²), additionally soil factors and leaf traits were recorded. Leaves of the mature trees, mother trees and seedlings were used for evaluating morphological and anatomical traits, extracting DNA. Based on the results of the genetic analyses, 30 root samples were also selected to estimate the eurycomanone content.

Data analyses was conducted by using several tools/programs and software. ImageJ and Axio SE64 programs were applied to estimate leaf area and stomata, while a Fragment Analyzer was used for separating the PCR products. The softwares of STRUCTURE, PAST3, AMOVA, GenAlex were utilized in genetic analysis. Excel v2016 (XSTAT v2019 and GenAlex 6.5 included) and Minitab 17.0 were included for statistical analyses. Particularly, descriptive statistics such as averages, standard deviation and range and so on were applied. The Kruskal- Wallis, Mann-Whitney tests and one-way ANOVA were used to check the mean differences between groups. The association among variables was tested by Pearson’s or Spearman’s correlation coefficient based on the normality test. The Mantel test was considered to check the

129 Summary relationship between factors, e.g. eurycomanone content and genetic diversity or genetic diversity and geographical factors.

Results and Discussion

After data analysis, several important findings were obtained as follows:

Hypothesis H1 “There are differences in terms of the density of trees, saplings and seedlings between sandy and mountainous areas” is partly accepted. E. longifolia distributes around 100 m to 1,000 m along the ridges, downhill, top-hill and rarely along the stream or high forest cover. Although this species occurs at an elevation up to 1,000 m in Nam Dong and A Luoi forests, it has not been discovered in Bach Ma at higher than 700 m. The main reason might be because of the high forest cover and habitat conditions in Bach Ma National Park. Several stems or branches are found in the sandy area, while single stems usually occur in mountainous regions. The trees larger than 6 cm in diameter only appear in the mountains with ca. 10 trees per ha. Among the total number of individuals in the mountains, 31.4% of trees at diameter >6 cm and 11% of trees with diameter >10 cm are found. The sapling density (2.5 to 6 cm at diameter) is similar between mountains and sandy areas, with an average of ca. 16 saplings per ha. In contrast, the seedling density in the former site is significantly higher than that in the latter. The present results show that the density of trees and saplings is relatively low due to the overharvesting of the mature roots for medicinal purposes. E. longifolia can grow nearly as tall as in the mountains under the controlled growth conditions in sandy areas.

Hypothesis H2 “Leaf size and stomatal density of mature trees from moist sites are both larger than those from the dry sites” is partly received. The morphological and anatomical traits of the leaves vary between moist and dry sites. The area, the length, the width and the leaflet numbers of the leaves from the humid places are more extensive than those from the dry zone (484 and 356 cm2 per leaf). The geographical and climatic conditions such as soil characteristics, humidity, water availability, temperature and so on may affect the morphological traits. In contrast, the stomatal density in the moist site is remarkably lower than that of the dry place (284 and 138 stomata/mm2). The present study also shows the inverse correlation between leaf area and the number of stomata. The plants with small leaves in dry conditions have a higher transpiration surface rate than larger leaves in humid sites. Thus, the limited water capacity under drought may affect the leaf size through the cell size decreases and stomata frequency increases. It demands sufficient stomata units for each unit of leaf area surface to respond to photosynthetic capacity.

130 Summary Hypothesis H3 “Seedlings of the same age and of the same condition will have similar growth pattern, even from the different provenances” is rejected. The germination duration of seeds from dry provenances is shorter and the rate of germination is higher than that of the moist derivations. The growth patterns of seedling height and diameter from one to twenty-two months increase together, but the development of seedlings from moist provenances is faster than that of dry derivations. However, the number of leaves does not always have the same growth pattern as seedling size. It is impressive that the leaves tend to fall between five and nine months after germination and from 14 to 20 months. This result gives hints to the fact that the seedlings need approximately five months to develop the new leaves and the falling of the old leaves.

Hypothesis H4 “Major part of genetic diversity occurs within populations” is accepted. SCoT and BPS markers perform more information than RAPD markers. In comparison with other studies, the level of genetic variation and the Shannon informative index of E. longifolia are relatively low (He = 0.17, I = 0.27). Genetic diversity depends on both the marker techniques and the number of accessions used. Most of the previous studies utilized a small number of samples to estimate genetic diversity compared to the large sample size in the present study. Based on the AMOVA analysis, a major part of genetic diversity occurs within populations, with 87% from SCoT and BPS markers and 94% from RAPD markers. E. longifolia is an outcrossing species with unisexual flowers; the genetic differentiation among-populations is low (ΦPT = 0.112).

Hypothesis H5 “There are potentially two genetic groups, namely sandy and mountainous” is received. With regard to the analysis of Bayesian clustering of genetic diversity by STRUCTURE program, the result presents two potential genetic groups (K = 2). This is similar to the findings of Principle Coordinates Analysis (PCoA) and Ward’s method. Phong Dien and Bach Ma (sandy area and mountain) create one group separated from A Luoi and Nam Dong (mountains). Several factors can explain this result. At first, water flow and birds can be the reason why the fruits are transferred from Bach Ma Mountain to the sandy areas. Secondly, Bach Ma has the lowest quantity of examined accessions, which can affect the low level of genetic variation of this population leading to influence the findings of the genetic structure. Thirdly, the Vietnamese-American war (1955-1975) and geographical distance might perhaps two other reason to create the same genetic group of A Luoi and Nam Dong. Lastly, it is difficult to find any individuals in Bach Ma from an elevation higher than 700 m, so geographical and ecological factors might limit the distribution of this plant species.

131 Summary Hypothesis H6 “Eurycomanone content (quality and quantity) of the roots depends on ecological distribution areas, genetic diversity and plant age” is partly accepted. The eurycomanone component of E. longifolia relies upon the ecological distribution sites. This component’s concentration in the moist sites is much higher than that in the dry site (1.22 mg/g and 0.74 mg/g on average). Nevertheless, there is no significant relationship between eurycomanone content and genetic diversity, meaning that genetic factors do not influence the eurycomanone accumulations. The Spearman’s coefficient does not show a remarkable correlation between the eurycomanone content and the plant age, but a significant relationship between this component and root diameter is addressed.

Conclusions

The medicinal plant resources play an essential role in healthcare needs as well as to maintain the sustainability of biodiversity and ecosystems. E. longifolia is a well-known medicinal tree species, but under serious threats in Vietnam. In the province of Thua Thien Hue, the status of this species is still relatively positive, but it is threatened by overharvesting and the lack of species information. Therefore, it is essential to contribute to the basic information of this species to preserve this valuable medicinal plant resource. This study has obtained critical findings. The population density is quite low. No large trees (≥6 cm in diameter) are found in the sandy area, while the natural regeneration from mountains is higher than that from the sandy area. The leaf size of the plants from mountains is larger than that from the sandy soil area. The leaf area shows significant inverse correlation with stomatal density. Under similar habitat conditions, the seedlings from different provenances show contrasting growth patterns. The genetic diversity level of this plant species is relatively low and mainly occurs within populations. The analysis of the genetic structure of E. longifolia populations indicates two significant different groups, but the genetic differentiation is small. A significant correlation between genetic diversity and geographical distances is also indicated. The eurycomanone content of the root tissues depends on the ecological distribution sites, but it is not affected by genetic factors and plant age. These findings address the status of E. longifolia population distribution and provide basic information for future propagation and breeding programs to manage the medicinal plant resources sustainably.

132 Zusammenfassung 8 Zusammenfassung

Titel: "Verbreitung, genetische Vielfalt und Eurycomanon-Gehalt von Eurycoma longifolia [Jack] in der Provinz Thua Thien Hue, Vietnam"

Einführung und Ziele

Eurycoma longifolia ist eine wertvolle Heilpflanzenart, die in tropischen Wäldern und auf trockenen sandigen Böden in den meisten asiatischen Ländern, einschließlich Vietnam, verbreitet ist. Die natürlichen Vorkommen der Pflanze nehmen aufgrund der nicht nachhaltigen Nutzung für medizinische Zwecke und der hohen kommerziellen Nachfrage in Vietnam rapide ab. Es ist daher wichtig, Informationen über die Art, wie beispielsweise deren natürliche Verbreitung, die Chrakterisierung von Phänotyp und Genotyp oder deren Gehalt an bioaktiven Verbindungen zu gewinnen und diese Informationen dann für Vermehrungs- und Züchtungsprogramme zu nutzen, um auf diese Weise zur Erhaltung der Art beizutragen.

Material und Methoden

Die Studie wurde in vier Gebieten durchgeführt, darunter Nam Dong, Bach Ma, A Luoi (Berge) und Phong Dien (sandiges Gebiet) sowie in Thua Thien Hue an der nördlichen Küstenregion Vietnams. Die Daten wurden im Rahmen von Felduntersuchungen und Laborarbeiten gesammelt. In Bezug auf die Freilanduntersuchungen wurden 27 Transekte (74,24 ha) entworfen, um die Dichte und Größe von Bäumen und von Setzlingen abzuschätzen. Die natürliche Verjüngung wurde in 47 Parzellen (500 m²) ermittelt. Des Weiteren wurden Bodenbeschaffenheit und Blattmerkmale untersucht. Die Blätter der alten Bäume, Mutterbäume und Sämlinge wurden zur Bewertung morphologischer und anatomischer Merkmale und zur Extraktion von DNA verwendet. Basierend auf den Ergebnissen der genetischen Analysen wurden 30 Wurzelproben ausgewählt, um den Eurycomanone-Gehalt zu ermitteln.

Die Datenanalysen wurden mit verschiedenen Tools/Programmen und Software durchgeführt. Die Programme ImageJ und Axio SE64 wurden angewendet, um die Blattfläche und die Stomatadichte zu ermitteln, während der Fragment Analyzer zur Trennung der PCR-Produkte verwendet wurde. Die Software STRUCTURE, PAST3, AMOVA, GenAlex wurde für die genetische Analyse verwendet. Für statistische Analysen wurden Excel v2016 (einschließlich XSTAT v2019 und GenAlex 6.5) und Minitab 17.0 angewendet. Insbesondere wurden deskriptive Statistiken wie u.a. Durchschnittswerte, Standardabweichung ermittelt. Die Kruskal-Wallis-, Mann-Whitney-Tests und die einfaktorielle-ANOVA wurden verwendet, um die Unterschiede zwischen den verschiedenen Gruppen zu überprüfen. Die Assoziation zwischen Variablen wurde anhand des Pearson- oder Spearman-Korrelationskoeffizienten, basierend auf dem Normalitätstest, ermittelt. Der Mantel-Test wurde verwendet, um die

133 Zusammenfassung Beziehung zwischen Faktoren, wie zum Beispiel Eurycomanone-Gehalt und genetische Vielfalt oder genetische Vielfalt und geografische Faktoren zu überprüfen.

Ergebnisse und Diskussion

Die Analyse der Daten ergab folgende wichtige Ergebnisse:

Die Hypothese H1 "Es zeigen sich Unterschiede in Bezug auf die Dichte von Bäumen und Keimlingen zwischen sandigen und bergigen Gebieten" wird teilweise akzeptiert. E. longifolia Vorkommen verteilen sich etwa in einem Bereich zwischen 100 m bis 1.000 m entlang der Bergkämme, selten entlang von Flüssen oder Gebieten mit einem hohen Kronen-Deckungsgrad. Obwohl diese Baumart in den Wäldern von Nam Dong und A Luoi in einer Höhe von bis zu 1.000 m ü. NN vorkommt, wurde sie in Bach Ma in mehr als 700 m Höhe nicht entdeckt. Der Hauptgrund könnte in dem hohen Deckungsgrad und den übrigen Umweltbedingungen im Bach-Ma-Nationalpark liegen. Auf sandigen Böden findet man häufig Klumpungen von Pflanzen, während in Bergregionen normalerweise Einzelpflanzen zu finden sind. Die Bäume mit einem größeren Durchmesser von mehr als 6 cm kommen nur in den Bergen, mit 10 Individuen pro ha vor. Unter der Gesamtzahl der Individuen finden sich hier 31,4% der Bäume mit einem Durchmesser >6 cm und 11% der Bäume mit einem Durchmesser >10 cm. Die Pflanzendichte (2,5 bis 6 cm Durchmesser) ist zwischen Berg- und sandigen Gebieten mit durchschnittlich 16 Pflanzen pro ha ähnlich. Im Gegensatz dazu, ist die Keimlingsdichte in den Bergen signifikant höher als in den sandigenRe gionen. Die vorliegenden Ergebnisse zeigen, dass die Dichte von Bäumen und jungen Pflanzen aufgrund der Übernutzung der Wurzeln für medizinische Zwecke relativ gering ist. E. longifolia kann unter kontrollierten Wachstumsbedingungen in sandigen Gebieten fast so groß werden, wie in den Bergen.

Die Hypothese H2 "Die Blattfläche und die Stomatendichte von alten Bäumen an feuchten Standorten sind größer/ höher als die an trockenen Standorten" wird teilweise angenommen. Die morphologischen und anatomischen Merkmale der Blätter variieren zwischen feuchten und trockenen Standorten. Fläche, Länge, Breite und Anzahl der Blätter der Pflanzen von feuchten Gebieten sind größer als die der trockenen Regionen (484 und 356 cm2 pro Blatt). Klimatische Bedingungen sowie Bodeneigenschaften können die Ausprägung morphologischer Merkmale beeinflussen. So ist die Stomatendichte der Pflanzen vom feuchten Standort deutlich niedriger als die von Pflanzen auf trockenen Standorten (284 und 138 Stomata/mm²). Die vorliegende Studie zeigt auch die inverse Beziehung zwischen Blattfläche und Anzahl der Stomata. Pflanzen mit geringer Blattfläche unter trockenen Bedingungen haben eine höhere Stomataanzahl als Individuen mit großer Blattfläche von gut wasserversorgten Standorten. Es erfordert ausreichend Stomata-Einheiten pro Einheit Blattflächenoberfläche, um die erforderliche Photosyntheseleistung zu erbringen.

134 Zusammenfassung Die Hypothese H3 "Sämlinge gleichen Alters und gleichen Entwicklungszustands aus unterschiedlichen Herkunftsgebieten weisen ein ähnliches Wachstumsmuster auf" wird abgelehnt. Die Keimdauer von Samen aus trockenen Gebieten ist kürzer und die Keimrate höher als die der Samen aus gut wasserversorgten Regionen. Pflanzengröße und -durchmesser nehmen für beide Standort-Gruppen (Alter 1 - 22 Monate) gleichmäßig zu, jedoch entwickeln sich Keimlinge von feuchten Gebieten schneller, als die von trockenen Standorten. Die Anzahl der Blätter variiert trotz zunehmender Keimlingsgröße, denn zwischen dem fünften und neunten Entwicklungsmonat sowie dem vierzehnten und zwanzigsten Monat ist ein Blattfall zu beobachten. Dieses Ergebnis gibt Hinweise darauf, dass die Pflanzen ungefähr fünf Monate für die Entwicklung neuer Blätter benötigen.

Die Hypothese H4 "Ein Großteil der genetischen Vielfalt findet sich innerhalb der Populationen" wird akzeptiert. SCoT- und BPS-Marker liefern mehr Informationen als RAPD- Marker. Im Vergleich zu anderen Studien, sind der Grad der genetischen Variation und der Shannon-Informationsindex von E. longifolia relativ niedrig (He = 0,17; I = 0,27). Die genetische Vielfalt hängt sowohl von den Markertechniken als auch von der Anzahl der untersuchten Akzessionen ab. Die meisten früheren Studien untersuchten eine geringere Anzahl an Proben, um die genetische Vielfalt zu ermitteln. Im Vergleich dazu ist der Probenumfang der vorliegenden Studie deutlich höher. Basierend auf der AMOVA-Analyse findet sich ein Großteil der genetischen Vielfalt innerhalb der Populationen, wobei 87% von SCoT- und BPS- Markern und 94% von RAPD-Markern stammen. E. longifolia ist eine sich auskreuzende Art mit unisexuellen Blüten. Die genetische Differenzierung zwischen den Populationen ist vergleichsweise gering (ΦPT = 0,112).

Die Hypothese H5 "Es gibt zwei genetische Gruppen, nämlich eine solche auf sandigen Standorten und eine auf Gebirgsstandorten" wird bestätigt. Die Analyse der genetischen Vielfalt durch das Programm STRUCTURE zeigt zwei potenzielle genetische Gruppen (K = 2). Dies ähnelt den Ergebnissen der PCoA und der Ward-Methode. Phong Dien und Bach Ma bilden eine Gruppe, die von A Luoi und Nam Dong (Gebirge) getrennt ist. Mehrere Faktoren können dieses Ergebnis erklären. Gewässer und Vögel können ein Grund sein, weshalb Samen und Pollen von Bach-Ma- in das Küstengebiet gelangt sind. Zweitens weist Bach Ma die geringste Anzahl an untersuchten Akzessionen auf, was die geringe genetische Variation dieser Vorkommen beeinflussen und die Ergebnisse der genetischen Struktur beeinflussen kann. Drittens sind der Vietnam- Krieg (1955-1975) sowie die geografische Entfernung möglicherweise weitere Faktoren, weshalb A Luoi und Nam Dong dieselbe genetische Gruppe zu bilden. Schließlich ist es schwierig, in Bach Ma Individuen aus einer Höhe von mehr als 700 m ü. NN zu finden. Deshalb ist anzunehmen, dass die geografischen und ökologischen Faktoren

135 Zusammenfassung die Verbreitung dieser Pflanzenart einschränken können. Eine signifikante Korrelation zwischen genetischer Vielfalt und geografischer Entfernung, könnte die oben erwähnte Erklärung widerspiegeln.

Die Hypothese H6 "Der Eurycomanone-Gehalt (Qualität und Quantität) der Wurzeln hängt von den Verbreitungsgebieten, der genetischen Vielfalt und dem Pflanzenalter ab" wird teilweise akzeptiert. Der Eurycomanone-Gehalt von E. longifolia hängt von den ökologischen Verbreitungsgebieten ab. Die Konzentration dieser Komponente in Wurzeln von feuchten Standorten ist deutlich höher, als die von Wurzeln trockener Standorte (durchschnittlich 1,22 mg/g und 0,74 mg/g). Dennoch gibt es keinen signifikanten Zusammenhang zwischen dem Eurycomanone-Gehalt und der genetischen Vielfalt, was bedeutet, dass genetische Faktoren die Eurycomanone-Akkumulation nicht zu beeinflussen scheinen. Der Spearman-Koeffizient zeigt keine deutliche Korrelation zwischen dem Eurycomanone-Gehalt und dem Pflanzenalter. Eine signifikante Beziehung zwischen dieser Komponente und dem Wurzeldurchmesser kann beobachtet werden.

Schlussfolgerungen

Eurycoma longifolia ist eine bekannte Heilpflanze, deren Vorkommen in Vietnam jedoch ernsthaft gefährdet ist. In der Provinz Thua Thien Hue ist der Status dieser Art noch relativ positiv, er ist jedoch durch Übernutzung gefährdet. Um geeignete Managementkonzepte dieser wertvollen Heilpflanzenressource zu entwerfen und Züchtungsprojekte auf den Weg zu bringen, sind deshalb grundlegende Informationen zu dieser Art von Bedeutung. Die Populationsdichte der Pflanze ist generell recht gering. In trockenen Gebieten finden sich keine großen Bäume (≥6 cm Durchmesser). Die Naturverjüngung im Gebirge ist besser als die der trockenen Standorte. Die Blattfläche zeigt eine signifikant inverse Korrelation mit der Stomatendichte. Unter ähnlichen Standortbedingungen zeigen die Keimlinge unterschiedlicher Herkunft unterschiedliche Entwicklungsmuster. Die genetische Vielfalt dieser Pflanzenart ist vergleichsweise gering und tritt hauptsächlich innerhalb der Populationen auf. Die Analyse der genetischen Struktur von E. longifolia-Populationen läßt zwei signifikant unterschiedliche Gruppen erkennen. Die genetische Differenzierung ist jedoch gering. Eine signifikante Korrelation zwischen genetischer Vielfalt und geografischer Entfernungen ist ebenfalls gegeben. Der Eurycomanone-Gehalt des Wurzelgewebes hängt von den ökologischen Standortfaktoren ab, wird jedoch nicht von genetischen Faktoren und dem Pflanzenalter beeinflusst.

136 Tóm tắt 9 Tóm tắt

Tên đề tài: “Phân bố, đa dạng di truyền và hàm lượng eurycomanone trong cây Bách bệnh (Eurycoma longifolia Jack) ở tỉnh Thừa Thiên Huế, Việt Nam”

Giới thiệu và Mục tiêu nghiên cứu

Bách bệnh (Eurycoma longifolia Jack) là loài cây thuốc có giá trị, thường phân bố ở vùng núi và vùng cát, ở một số nước Châu Á, trong đó có Việt Nam. Sự phân bố quần thể của loài này ngoài tự nhiên đang giảm xuống đáng kể do việc thu hái không bền vững cho mục đích làm thuốc nên nhu cầu buôn bán thương mại tăng cao. Việc nghiên cứu thông tin về loài như phân bố tự nhiên, đánh giá loài (kiểu gen và kiểu hình) hay thành phần hoạt chất sinh học là vô cùng cần thiết. Bởi vậy, nghiên cứu này nhằm cung cấp các thông tin cơ bản cho việc gây trồng nhằm đóng góp vào công tác bảo tồn loài. Tăng cường sự hiểu biết về tình trạng phân bố quần thể Bách bệnh, khả năng thích nghi hay tính mềm dẻo về kiểu hình thông qua các đặc điểm về hình thái và giải phẫu của lá. Bên cạnh đó, kết quả nghiên cứu thu được sẽ giúp phân biệt mối quan hệ di truyền của loài bên trong và giữa các quần thể, cấu trúc di truyền của quần thể và xác định các yếu tố ảnh hưởng đến sự tích lũy eurycomanone - một trong những hoạt chất chính của rễ cây Bách bệnh.

Nguyên liệu và Phương pháp nghiên cứu

Nghiên cứu được thực hiện ở bốn vùng sinh thái, A Lưới, Bạch Mã, Nam Đông (vùng núi) và Phong Điền (vùng cát), tỉnh Thừa Thiên Huế, miền Trung Việt Nam. Số liêu nghiên cứu được thu thập thông qua điều tra thực địa và phân tích thí nghiệm. Đối với điều tra thực địa, chúng tôi đã thiết lập 27 tuyến điều tra (74,24 ha) để đánh giá mật độ và cấp kính cây trưởng thành và cây nhỡ; 47 ô tiêu chuẩn (500 m2) để điều tra cây tái sinh, yếu tố đất đai và đặc điểm lá. Lá của cây trưởng thành, cây mẹ (tự nhiên) và lá các cây con (vườn ươm) được dùng để đánh giá đặc điểm hình thái và giải phẫu, tách chiết DNA (mẫu lá từ 276 cây trưởng thành và 269 cây con ở vườn ươm). Từ kết quả phân tích di truyền, 30 mẫu rễ được chọn để đánh giá hàm lượng eurycomanone.

Phần mềm ImageJ và Axio SE64 được dùng để đo đếm diện tích lá và mật độ khí khổng, máy phân tích (Fragment Analyzer) dùng để phân tách sản phẩm PCR. Các phần mềm như STRUCTURE, PAST3, AMOVA, GenAlex dùng cho phân tích di truyền. Phân tích thống kê thông qua Excel v2016 (XSTAT v2019 và GenAlex 6.5) và Minitab v17.0 . Để kiểm tra sự sai khác nhau về giá trị trung bình các nhóm mẫu, kiểm định phi tham số như Kruskal-Wallis, Mann-Whitney và kiểm định phân tích phương sai một yếu tố (one-way ANOVA) được dùng để đánh giá. Xác định mối liên hệ giữa các biến bởi hệ số tương quan Pearson hay hệ số tương

137 Tóm tắt quan hạng Spearman tùy vào giả định về phân phối chuẩn. Phép thử Mantel được sử dụng để kiểm tra mối liên quan giữa các yếu tố (bởi các ma trận) như giữa hàm lượng eurycomanone và yếu tố di truyền, giữa yếu tố di truyền và yếu tố địa hình...

Kết quả và Thảo luận

Từ phân tích số liệu thống kê, chúng tôi thu được các kết quả quan trọng tương ứng với từng mục tiêu và giả thuyết nghiên cứu như sau:

Giả thuyết H1 “Có sự khác nhau về mật độ cây trưởng thành, cây nhỡ, cây con giữa vùng cát và vùng núi”được chấp nhận một phần. Bách bệnh phân bố ở độ cao khoảng từ 100 m đến 1.000 m dọc theo các vùng đồi núi, các sườn dốc và cả ở vùng núi cao, tuy nhiên hiếm gặp ở các khu vực ven suối hoặc nơi có độ che phủ rừng lớn. Mặc dù loài này xuất hiện ở độ cao gần 1.000 m ở khu vực Nam Đông và A Lưới nhưng lại không thấy hoặc hiếm gặp các cá thể Bách bệnh ở độ cao trên 700 m ở khu vực Bạch Mã. Lý do có thể là độ che phủ rừng lớn và các điều kiện sinh thái đặc thù ở đây. Ở vùng cát, xuất hiện cây với nhiều thân và nhánh, trong khi đó ở vùng núi, cây này thường có một thân. Bách bệnh với đường kính trên 6 cm chỉ thấy ở vùng núi với mật độ khoảng 10 cây/ha. Trong tổng số cá thể ở vùng núi, có 31.4% cây với đường kính >6 cm và 11% cây với đường kính >10 cm. Mật độ cây nhỡ (đường kính từ 2.5 cm đến 6 cm) là tương tự giữa vùng núi và vùng cát, với mật độ trung bình là gần 16 cây/ha. Ngược lại, mật độ cây tái sinh thuộc vùng núi lại cao hơn so với vùng cát. Các kết quả này cho thấy, mật độ cây trưởng thành và cây nhỡ khá thấp có thể do việc khai thác quá mức loài cây này ngoài tự nhiên với mục đích làm thuốc. Do đó, Bách bệnh ở vùng cát có thể phát triển thành cây gỗ với cấp kính trên 6 cm như vùng núi nếu được bảo vệ nghiêm ngặt.

Giả thuyết H2 “Diện tích và mật độ khí khổng của lá ở vùng núi lớn hơn so với vùng cát” được chấp nhận một phần. Đặc điểm hình thái và giải phẫu của lá có sự thay đổi giữa vùng cát và vùng núi. Diện tích lá, chiều dài lá, chiều rộng lá và tổng số lá chét trên một lá từ vùng núi lớn hơn so với vùng đất cát (484 và 356 cm2/lá). Yếu tố khí hậu và địa hình như đặc điểm đất đai, độ ẩm đất, yếu tố nước, yếu tố nhiệt độ có thể ảnh hưởng đến hình thái giải phẫu của lá. Tuy nhiên, mật độ khí khổng trên một đơn vị diện tích lá từ vùng núi lại thấp hơn nhiều so với vùng cát (284 và 138 khí khổng/mm2). Kết quả nghiên cứu cho thấy, giữa diện tích lá và mật độ khí khổng có mối quan hệ ngược chiều tương đối mạnh. Bách bệnh thường có lá nhỏ ở vùng khô lại có tỉ lệ hô hấp ở bề mặt lá cao hơn so với vùng ẩm. Do vậy, khả năng yếu tố nước bị giới hạn trong điều kiện khô có thể ảnh hưởng đến kích cỡ của lá bằng cách giảm kích thước tế bào và tần suất xuất hiện mật độ khí khổng tăng lên. Đây cũng là cách để thực vật đáp ứng đủ lượng khí khổng trên mỗi đơn vị bề mặt diện tích lá nhằm thỏa mãn khả năng quang hợp.

138 Tóm tắt Giả thuyết H3 “Cây con vùng độ tuổi và cùng điều kiện sống sẽ có sự sinh trưởng tương tự nhau cho dù có xuất xứ khác nhau” bị loại bỏ. Thời gian nảy mầm của hạt xuất xứ từ vùng khô là ngắn hơn và tỉ lệ nảy mầm cũng cao hơn so với vùng ẩm. Xu hướng phát triển chiều cao và đường kính gốc từ 1 tháng đến 22 tháng tuổi tăng lên tương ứng nhưng khả năng sinh trưởng của cây con có nguồn gốc từ vùng ẩm cao hơn so với vùng khô. Tuy nhiên, tổng số lá trên một cây con lại không tuân theo quy luật như sự phát triển cấp kính của chúng. Cây con thay một số lá già trong giai đoạn từ 5 đến 9 tháng và từ 14 đến 20 tháng sau khi nảy mầm. Kết quả này cho thấy, cây con cần khoảng 5 tháng để phát triển lá non mới và lá già rụng đi.

Giả thuyết H4“Đa dạng di truyền cây Bách bệnh chủ yếu diễn ra bên trong các quần thể” được chấp nhận. Các chỉ thị SCoT và BPS khi áp dụng trên đối tượng Bách bệnh cho thông tin nhiều hơn so với chỉ thị RAPD. So sánh với các nghiên cứu trước đây, mức độ đa dạng di truyền và chỉ số đa dạng Shannon của 276 cá thể Bách bệnh là khá thấp (He = 0.17, I = 0.27). Tính đa dạng di truyền thường phụ thuộc vào loại chỉ thị khác nhau và tổng số mẫu dùng để phân tích. Hầu như các nghiên cứu trước đây áp dụng trên đối tượng này, sử dụng kích cỡ mẫu khá thấp để đánh giá đa dạng di truyền so với nghiên cứu của chúng tôi. Kết quả phân tích AMOVA cho thấy, đa dạng di truyền diễn ra bên trong các nhóm quần thể là chính và chiếm 87% khi sử dụng chỉ thị SCoT và BPS, chiếm đến 94% đối với chỉ thị RAPD. Bách bệnh là loài lai xa với hoa đơn tính khác gốc, sự khác nhau về di truyền giữa các quần thể là thường thấp (ΦPT = 0.112).

Giả thuyết H5 “Có hai nhóm di truyền tiềm năng là nhóm vùng cát và nhóm vùng núi”được chấp nhận. Kết quả phân tích đa dạng di truyền theo nhóm Bayesian bằng phần mềm STRUCTURE cho thấy có hai nhóm di truyền tiềm năng (K = 2). Kết quả này khá tương đồng với các kết quả phân tích thành phần chính (PCoA) và phân tích cây di truyền phả hệ (Ward’s method). Các nhóm quần thể Phong Điền và Bạch Mã tạo thành một nhóm tách biệt với nhóm quần thể A Lưới và Nam Đông. Một số yếu tố có thể giải thích cho kết quả này. Thứ nhất, hạt Bách bệnh có thể di chuyển theo dòng nước hay các loài chim ăn hạt có thể mang vác từ vùng núi Bạch Mã đến các vùng cát. Thứ hai, số mẫu dùng để phân tích di truyền của quần thể Bạch Mã là khá thấp có thể ảnh hưởng đến mức độ đa dạng di truyền của quần thể này và cũng có thể là nguyên nhân ảnh hưởng đến cấu trúc di truyền. Thứ ba, yếu tố chiến tranh và khoảng cách địa lý có thể là một lý do khác tạo ra nhóm di truyền giống nhau giữa quần thể A Lưới và Nam Đông. Cuối cùng, quá trình điều tra thực địa cho thấy rất hiếm gặp các cá thể Bách bệnh xuất hiện ở độ cao trên 700 m ở vùng Bạch Mã, nên yếu tố địa lý và sinh thái có thể giới hạn sự phân bố của loài này.

Giả thuyết H6 “Hàm lượng eurycomanone (chất lượng và số lượng) của rễ phụ thuộc vào các vùng phân bố sinh thái khác nhau, yếu tố di truyền và độ tuổi của cây” được chấp nhận một

139 Tóm tắt phần. Kết quả nghiên cứu cho thấy hàm lượng eurycomanone của rễ cây Bách bệnh phụ thuộc vào vùng phân bố sinh thái. Hàm lượng trung bình của hoạt chất này trong rễ ở vùng núi cao hơn nhiều so với vùng cát (1.22 và 0.74 mg/g). Tuy nhiên, kết quả phân tích Mantel cho thấy, không có mối liên hệ nào giữa hàm lượng eurycomanone và yếu tố đa dạng di truyền, nghĩa là yếu tố di truyền có thể không ảnh hưởng đến sự tích lũy eurycomanone. Kết quả tính toán hệ số tương quan hạng Spearman cho thấy, không có mối liên quan giữa hàm lượng eurycomanone và độ tuổi của cây, nhưng lại có mối quan hệ với đường kính rễ.

Kết luận

Nguồn tài nguyên cây thuốc đóng vai trò quan trọng đối với sức khỏe cũng như với việc duy trì tính bền vững của hệ sinh thái và đa dạng sinh học. Bách bệnh là một loài cây thuốc phổ biến nhưng hiện bị đe dọa nghiêm trọng ở Việt Nam. Mặc dù ở tỉnh Thừa Thiên Huế, tình trạng loài này đang ở mức độ tương đối nhưng nó cũng đang bị đe dọa do việc khai thác quá mức và thiếu thông tin về loài cũng như các khía cạnh khác. Do vậy, nghiên cứu một số thông tin cơ bản đối với loài Bách bệnh nhằm góp phần vào bảo tồn nguồn tài nguyên cây thuốc là vô cùng cấp bách. Nghiên cứu này đạt được một số kết quả quan trọng. Mật độ phân bố quần thể Bách bệnh ngoài tự nhiên là khá thấp. Không có cá thể Bách bệnh nào với đường kính trên 6 cm được phát hiện ở vùng đất cát, trong khi đó tỉ lệ tái sinh tự nhiên của vùng núi lại cao hơn vùng cát nhiều. Đặc điểm hình thái lá cây Bách bệnh ở vùng núi lớn hơn so với vùng cát. Diện tích lá có mối tương quan ngược với mật độ phân bố của khí khổng. Với điều kiện sống như nhau, cây con có nguồn gốc khác nhau cho thấy xu hướng phát triển không giống nhau. Đa dạng di truyền của loài này khá thấp và chủ yếu đa dạng bên trong mỗi quần thể. Phân tích cấu trúc di truyền quần thể cho thấy, có hai nhóm di truyền khác nhau, tuy sự khác nhau này là không lớn lắm. Có mối liên quan giữa yếu tố di truyền và khoảng cách địa lý. Hàm lượng eurycomanone của rễ cây Bách bệnh phụ thuộc vào các vùng phân bố sinh thái khác nhau nhưng không bị ảnh hưởng bởi yếu tố di truyền hay độ tuổi cây. Như vậy, từ các kết quả nghiên cứu cho thấy tình trạng phân bố của các quần thể Bách bệnh ngoài tự nhiên đang đe dọa và cung cấp các thông tin cơ bản về loài, hỗ trợ cho việc gây trồng, tách chiết nhằm để đảm bảo tính bền vững nguồn tài nguyên cây thuốc.

140 References

References Abubakar, B.M., Salleh, F.M. and Wagiran, A. 2017. Chemical Composition of Eurycoma longifolia (Tongkat Ali) and the Quality Control of its Herbal Medicinal Products. J. Appl. Sci. 17 (7): 324-338. doi:10.3923/jas.2017.324.338.

An, L.T., Markowski, J., Bartos, M., Thoai, T.Q., Tuan, T.H. and Rzenca, A. 2018. Tourist and Local Resident Preferences for the Northern Yellow-Cheeked Gibbon (Nomascus annamensis), Conservation Program in the Bach Ma National Park, Central Vietnam. Trop. Conserv. Sci. 11. doi:10.1177/1940082918776564.

ALS. 2018. Statistical Yearbook of A Luoi district (ALS). [Vietnamese]

Ang, H.H and Cheang, H.S. 1999. Studies on the anxiolytic activity of Eurycoma longifolia Jack roots in mice. Jpn J. Pharmacol: 79: 497-500. doi:org/10.1254/jjp.79.497.

Ang, H.H., Cheang, H.S. and Yusof, A.P. 2000. Effects of Eurycoma longifolia Jack (Tongkat Ali) on the Initiation of Sexual Performance of Inexperienced Castrated Male Rats. 49 (7): 35-38. doi:10.1538/expanim.49.35.

Ang, H.H., Lee, K.L. and Jack, E. 2002. Effect of Eurycoma longifolia Jack on orientation activities in middle- aged male rats. 16: 479-483. doi:10.1046/j.1472-8206.2002.00106.x.

Arumugam, T., Jayapriya, G. and Sekar, T. 2019. Molecular fingerprinting of the Indian medicinal plant Strychnos minor Dennst. Biotechnol. Rep. 21, e00318. doi.org/10.1016/j.btre.2019.e00318.

Averyanov, L.V, Loc, P.K., Hiep, N.T. and Harder, D.K. 2003. Phytogeographic review of Vietnam and adjacent areas of Eastern Indochina. Komarovia 3: 1-83. Retrieved from https://www.semanticscholar.org/paper/Phytogeographic-review-of-Vietnam-and-adjacent-of-Petersburg- Averyanov/dc43a5c09499e005fa6b4514c952d5afba6d00f3. [Accessed in Dec. 2018].

Aziz, R.A., Sarmidi, M.R., Kumaresan, S., Taher, Z.M. and Yee, F.C. 2003. Phytochemical processing: The next emerging field in chemical engineering - aspects and opportunities. J. Chem. Eng. Malaysia 3: 45-60. Retrieved from http://www.geocities.ws/foodominic/IJKM_Phytochemical.pdf. [Accessed in Nov. 2020]

Bekessy, S.A., Allnutt, T.R., Premoli, A.C., Lara, A., Ennos, R.A., Burgman, M.A., Cortes, M. and Newton, A.C. 2002. Genetic variation in the Monkey Puzzle tree (Araucaria araucana (Molina) K. Koch), detected using RAPD. Heredity (Edinb). 88: 243-249. doi:10.1038/sj/hdy/6800033.

Bewley, J. D. and Black, M. 1994. Seeds. Physiology of development and germination. Springer US: XV, 445, 2nd ed. Retrieved from https://www.springer.com/gp/book/9780306447471 [Accessed in Feb. 2020]

Bhat, R. and Karim, A.A. 2010. Tongkat Ali (Eurycoma longifolia Jack): A review on its ethnobotany and pharmacological importance. Fitoterapia 81 (7): 669-679. Elsevier B.V. doi:10.1016/j.fitote.2010.04.006.

BMNP (Bach Ma National Park). 2018. The Forest Resource of Bach Ma National Park. Retrieved from http://www.bachmapark.com.vn/gioi-thieu/tai-nguyen-rung_108.html. [Vietnamese, accessed in Jan. 2019]

Blyton, D.J.M and Flanagan, S.F.N. 2012. A Comprehensive Guide to: GenAlEx 6.5 - Genetic Analysis in Excel. Australian National University. Retrieved from http://biology.anu.edu.au/GenALEx/. [Accessed in Oct. 2019]

Bock, J.H. and Norris, D.O. 2016. Forensic Plant Anatomy. In Forensic Plant Science. doi:10.1016/b978-0-12- 801475-2.00004-x.

Bonin, A., Bellemain, E., Eidesen, P.B., Pompanon, F., Brochmann, C. and Taberlet, P. 2004. How to track and assess genotyping errors in population genetics studies. Mol. Ecol. 13 (11): 3261-3273. doi:10.1111/j.1365- 294X.2004.02346.x. Botstein, D., White, R.L., Skolnick, M. and Davis, R.W. 1980. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 32 (3): 314-331. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1686077/. [Accessed in Sep. 2019]

141 References

Bussell, J.D. 1999. The distribution of random amplified polymorphic DNA (RAPD) diversity among populations of Isotoma petraea (Lobeliaceae). Mol. Ecol. 8 (5): 775-789. doi:10.1046/j.1365-294X.1999.00627.x.

Cam, D.X. 2002. Survey on the status of flora and propose the sustainable development and recovery solutions for sandy area belonging to Thua Thien Hue province. Agriculture Publishing House. [Vietnamese]

Carvalho, F.J., Maciel, G.M., Marques, O.J., Silva, I.G., Braga, G.D., Marquez, G.R., Siquieroli, A.C.S. and Neto, L.D.C. 2019. Analysis of genetic divergence in sweet corn genotypes through hierarchical optimization methods. Genet. Mol. Res. 18 (3): 1-10. doi:10.4238/gmr18384.

Chan, Y.K. and Toh W.K. 1984. Growth studies on some vegetative characteristics of papaya (Carica papaya L.). Mardi Res. Bull. 12 (1): 46-54. Retrieved from https://agris.fao.org/agris- search/search.do?recordID=MY19850109097. [Accessed in Dec. 2019]

Chan, K.L., O’Neill, M.J., Phillipson, J.D. and Warhurst, D.C. 1986. Plants as sources of antimalarial drugs. Part 3. Eurycoma longifolia. Planta Med. No. 2: 105-107. doi:10.1055/s-2007-969091.

Chan, H.Y.E., Zhang, Y., Hoheisel, J.D. and O’Kane, C.J. 1997. Identification and characterization of the gene for Drosophila l3 ribosomal protein. Gene 212 (1): 119-125. doi:10.1016/S0378-1119(98)00145-0.

Chapman, V.J. 1976. Mangrove vegetation. J Cramer, Leuterhausen Charles E, Basset Y (2005) Vertical stratification of leaf-beetle assemblages (Coleoptera: Chrysomelidae) in two forest types in Panama. J Trop Ecol 21: 329-336.

Chen, H., Chen, H., Hu, L., Wang, L., Wang, S., Wang, M.L. and Cheng, X. 2017. Genetic diversity and a population structure analysis of accessions in the Chinese cowpea [Vigna unguiculata (L.) Walp.] germplasm collection. Crop J. 5 (5): 363-372. Elsevier B.V. doi:10.1016/j.cj.2017.04.002.

Chi, V.V. 2012. The Dictionary of Medicinal plants in Vietnam. Medicinal Publishing House.

Chong, K.Y., Tan, H.T.W. and Richard, T.C. 2009. A checklist of the total vascular plant flora of Singapore Native, Naturalised and Cultivated Species. National University of Singapore Publishing House. Retrieved from https://lkcnhm.nus.edu.sg/nus/pdf/publication/lkcnh Museum Books/LKCNHM Books/flora_of_singapore_tc.pdf. [Accessed in May 2016]

Cochard, R. 2010. Applied Statistics Guide using Minitab. Lecturer of Research Design for Natural Resource Management, ED76.15. School of Environment, Resources and Development. Asian Institute of Technology. Thailand.

Colin, W. and Mark, F. 1996. Stomata. Published by Chapman and Hall, 2-6 Boundary Row, London SE1 8HN, UK. 2nd ed.

Collard, B.C.Y. and Mackill, D.J. 2009. Start Codon targeted (SCoT) polymorphism: a simple, novel DNA marker technique for generating gene-targeted markers in plants. Plant Mol. Biol. Rep. 27 (1): 86-93. doi:10.1007/s11105-008-0060-5.

Corner, E.J.H. 1988. Wayside trees of Malaya. Kuala Lumpur, Malaysia: United Selangor Press. 698-699, 3rd ed.

Cooper, C.S. and Qualls, M. 1967. Morphology and chlorophyll content of shade and sun leaves of two legumes. Cro. Si, 7, 672-673. doi.org/10.2135/cropsci1967.0011183X000700060036x.

Cyranoski, D. 2005. Malaysian researchers bet big on home-grown Viagra Complaints of gender bias compel NIH to revise a scheme. Nat. Publ. Gr. 11 (9). Retrieved from http://www.nature.com/nm/journal/v11/n9/pdf/nm0905-912a.pdf. [Accessed in Jun. 2019] Culley, T.M. 2005. Population Genetic Analysis of ISSR Data. Retrieved from https://scholar.google.com/scholar?oi=bibsandhl=enandcluster=9143417728509536306. [Accessed in Sep. 2019]

142 References

Dasrul, I. D., Lok, E.H., Faridah A.A., Rosdi, K. and Amir, S. 2018. Growth performance of Eurycoma longifolia on two different bris soil series at Setiu, Terengganu, 5 (11): 1-5.

Dörken, V.M. and Hubertus, N. 2018. A monograph of leaf characters in the genus Abies (Abietoideae, Pinaceace). Remagen-Oberwinter:Kessel. ISBN 978-3-945941-40-9.

Doyle, J.J. and Doyle J.L. 1990. Isolation of plant DNA from fresh tissue. Focus 12 (1): 13-15. Retrieved from https://scholar.google.com/scholar_lookup?title=Isolation%20of%20plant%20DNA%20from%20fresh%2 0tissue&journal=Focus&volume=12&pages=13- 15&publication_year=1988&author=Doyle%2CJ.J.&author=Doyle%2CJ.L. [Accessed in July 2016]

Duc, T.M., Yen, V.T., Hoi, N. and Lan, P.T.N. 2018. In vivo propagation of Eurycoma longifolia Jack. Hue Univ. J. Sci. Agric. Rural Dev. 127 (3A): 81. doi:10.26459/hueuni-jard.v127i3a.4438. [Vietnamese]

Dung, D.M. 2018. Isolation and identification eurycomanone content in Eurycoma longifolia J. (Simaroubaceae) by LC-MS/MS. Hue Univ. J. Sci. Nat. Sci. 127 (1B): 153. doi:10.26459/hueuni-jns.v127i1b.5000. [Vietnamese]

Effendy, N.M., Mohamed, N., Muhammad, N., Mohamas, I.N. and Shuid, A.N. 2012. Eurycoma longifolia: Medicinal plant in the prevention and treatment of male osteoporosis due to androgen deficienc. Hindawi Publ. Corp.: 9. doi:10.1155/2012/125761.

Egbadzor, K.F., Ofori, K., Yeboah, M., Aboagye, L.M., Opoku-Agyeman, M.O., Danquah, E.Y. and Offei, S.K. 2014. Diversity in 113 cowpea [Vigna unguiculata (L) Walp] accessions assessed with 458 SNP markers. J. Korean Phys. Soc. 3 (1): 1-15. doi:10.1186/2193-1801-3-541.

EL. 2016. Eurycoma longifolia - Tongkat Ali. Retrieved from https://taxo4254.wikispaces.com/Eurycoma+longifolia. [Accessed in Oct. 2017]

Etminan, A., Pour-Aboughadareh, A., Mohammadi, R., Ahmadi-Rad, A., Noori, A., Mahdavian, Z. and Moradi, Z. 2016. Applicability of start codon targeted (SCoT) and inter-simple sequence repeat (ISSR) markers for genetic diversity analysis in durum wheat genotypes. Biotechnol. Biotechnol. Equip. 30 (6): 1075-1081. Taylor and Francis. doi:10.1080/13102818.2016.1228478.

Evanno, G., Regnaut, S. and Goudet, J. 2005. Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Mol. Ecol. 14 (8): 2611-2620. doi:10.1111/j.1365-294X.2005.02553.x.

Excoffier, L., Smouse, P.E. and Quattro, J.M. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics, 131 (2): 479- 491. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1205020/. [Accessed in Oct, 2020]

Fadilah, W.N., Mohamad, O., Zaki, A.M. and Shamsiah, A. 2019. Assessment of IRAP Markers to evaluate the genetic diversity of Eurycoma longifolia. Pertanika J. Trop. Agric. Sci. 42 (3): 921-937. Retrieved from http://web.a.ebscohost.com/ehost/detail/detail?vid=0&sid=493634f3-f126-43c1-aa6d- 70d1bb0c4e2b%40sdc-v- sessmgr03&bdata=JnNpdGU9ZWhvc3QtbGl2ZQ%3d%3d#AN=141162888&db=afh. [Accessed in Aug. 2020]

Fiaschetti, G., Grotzer, M., Shalaby, T., Castelletti, D. and Arcaro, A. 2010. Quassinoids: From traditional drugs to new cancer therapeutics. Curr Med Chem 18: 316-328. doi: 10.2174/092986711794839205.

Finkeldey, R. and Hattermet, H.H. 2007. Tropical forest genetics. In Nature. doi:10.1038/255578a0. Franks, P.J. and Farquhar, G.D. 2007. The Mechanical Diversity of Stomata and Its Significance in Gas-Exchange Control [OA]. 143 (1): 78-87. doi:10.1104/pp.106.089367.

García, C., Jordano, P. and Godoy, J.A. 2007. Contemporary pollen and seed dispersal in a Prunus mahaleb population: patterns in distance and direction. Mol Ecol. 16 (9):1947-55. doi: 10.1111/j.1365- 294X.2006.03126.x. 143 References

Gay, A.P. and Hurd, R.G. 1975. the Influence of Light on Stomatal Density in the Tomato. New Phytol. 75 (1): 37- 46. doi:10.1111/j.1469-8137.1975.tb01368.x.

GFD (Green Field Consulting and Development). 2019. Retrieved from https://www.environmental- expert.com/companies/greenfield-consulting-development-gfd-31072. [Accessed in Jan. 2020]

Ginting, B.R.A. 2010. Ecological assessment of Pasak Bumi (Eurycoma longifolia Jack) and its utilization by local community around the Bukit Lawang Forest [Thesis]. Medan: University of North Sumatera.

Gömöry, D., Yakovlev, I., Zhelev, P., Jedináková, J. and Paule, L. 2001. Genetic differentiation of oak populations within the Quercus robur/Quercus petraea complex in central and eastern Europe. Heredity (Edinb). 86 (5): 557-563. doi:10.1046/j.1365-2540.2001.00874.x.

Guo, L., Lin, H., Fan, B.H., Cui, X.H. and Chen, J. 2013. Impact of root water content on root biomass estimation using ground penetrating radar: evidence from forward simulations and field controlled experiments. Plant Soil 371, 503‐520. doi.org/10.1007/s11104-013-1710-4.

Hadiah, J.T. 2000. Eurycoma longifolia Jack (Pasak Bumi). Retrieved from http://www.bestwebmaster.com/Nutraceutical/LongJackNews_1.html. [Accessed in Oct. 2019]

Hague, M.T.J. and Routman, E.J. 2016. Does population size affect genetic diversity? A test with sympatric lizard species. Heredity (Edinb). 116 (1): 92-98. Nature Publishing Group. doi:10.1038/hdy.2015.76.

Hajjouli, S., Chateauvieux, S., Teiten, M.H., Orlikova, B., Schumacher, M., Dicato, M., Choo, C.Y. and Diederich, M. 2014. Eurycomanone and eurycomanol from Eurycoma longifolia jack as regulators of signaling pathways involved in proliferation, cell death and inflammation. Molecules 19 (9): 14649-14666. doi:10.3390/molecules190914649.

Hamidi, H., Talebi, R. and Keshavarzi, F. 2014. Comparative efficiency of functional gene-based markers, start codon targeted polymorphism (SCoT) and conserved DNA-derived polymorphism (CDDP) with ISSR markers for diagnostic fingerprinting in wheat (Triticum aestivum L.). Cereal Res. Commun. 42 (4): 558- 567. doi:10.1556/CRC.2014.0010.

Hammer, H.W. 1999. PAST - PAieontological STatistics Ver. 3.22. Reference manual. Natural History Museum, University of Oslo. 1999-2018.

Hamrick, J.L., Godt, M.J.W. and Sherman-Broyles, S.L. 1992. Factors influencing levels of genetic diversity in woody plant species. New For. 6 (1-4): 95-124. doi:10.1007/BF00120641.

Hamrick, J.L. and Godt, M.J.W. 1996. Effects of life history traits on genetic diversity in plant species. Philos. Trans. R. Soc. B Biol. Sci. 351 (1345): 1291-1298. doi:10.1098/rstb.1996.0112.

Hales, T.C., Cole‐Hawthorne, C., Lovell, L., Evans, S.L., 2013. Assessing the accuracy of simple field based root strength measurements. Plant Soil 372, 553‐565. doi.org/10.1007/s11104-013-1765-2.

Hall, J.B. and Swaine, M.D. 1981. Distribution and ecology of vascular plants in a tropical rain forest. The Hague, Dr. W. Junk Publishers. doi.org/10.1007/BF00047109.

Hasibuan, S., Suhesti E. and Insusanty, E. 2016. Larangan Adat Rumbio, Kabupaten Kampar Provinsi Riau. 11 (2): 112-126.

Hassan, N.H., Abdullah, R., Kiong, L.S., Ahmad, A.R., Abdullah, N., Zainudin, F., Ismail, H. and Rahman, S.S.A. 2012. Micropropagation and production of eurycomanone, 9-methoxycanthin-6-one and canthin-6-one in roots of Eurycoma longifolia plantlets. African J. Biotechnol. 11 (26): 6818-6825. doi:10.5897/AJB11.3414.

Ho, P.H. 1999. An Illustrated flora in Vietnam, Part I, II, III. Young Publishing House. [Vietnamese]

144 References

Hoi, N. 2013. The thesis submitted in partial fulfilled of the requirement for degree of Master of Forestry. Hue University of Agriculture and Forestry, Vietnam. [Vietnamese]

Hong, P.T.N. 2006. Research on chemical constituents of Eurycoma longifolia Jack. National University of Ho Chi Minh. [Vietnamese]

Hong, G., Pachter, R. and Ritz, T. 2018. Coupling Drosophila melanogaster Cytochrome Light Activation and Oxidation of the Kvβ Subunit Hyperkinetic NADPH Cofactor. J. Phys. Chem. B 122 (25): 6503-6510. doi.org/10.1021/acs.jpcb.8b03493.

Husen, R., Hawariah, A., Pihie, L. and Nallappan, M. 2004. Screening for antihyperglycaemic activity in several local herbs of Malaysia. 95: 205-208. doi:10.1016/j.jep.2004.07.004.

Hussein, S., Ibrahim, R., Kiong, A.L.P., Fadzillah, N.M. and Daud, S.K. 2005. Micropagation of Eurycoma longifolia Jack via Formation of Somatic Embryogenesis. Asian J. Plant Sci. 4 (5): 472-485. doi:10.3923/ajps.2005.472.485.

Hutchings, M.J. 1997. The Structure of Plant Populations. In Plant Ecology. doi:10.1002/9781444313642.ch11.

Ichie, T., Inoue, Y., Takahashi, N., Kamiya, K. and Kenzo, T. 2016. Ecological distribution of leaf stomata and trichomes among tree species in a Malaysian lowland tropical rain forest. J. Plant Res. 129 (4): 625-635. Springer Japan. doi:10.1007/s10265-016-0795-2.

Idris, A., Kadir, A.A., Ali, R.M., Arshad, S. and Hidayu, M. 2009. Responses of Eurycoma longifolia Jack to open planting area and fertilizer applications: Study of the effects on growth performance and root yield. J. Trop. For. Sci. Retrieved from https://agris.fao.org/agris-search/search.do?recordID=MY2015000432. [Accessed in Mar. 2017]

IFRI (International Forestry Resources and Institutions). 2004. Asian Institute of Technology. Thailand.

Ivetić, V., Isajev, V., Nikolić, A., Krstić, M., Ristić, D. and Kostadinović M. 2012. Delineation of beech provenance regions in Serbia by spatial analysis of genetic diversity. Genetika 44 (1): 101-108. doi: 10.2298/GENSR1201101I.

Jagananth, I.B. and Teik, N.L. 2000. Herb: The Green Pharmacy of Malaysia. In VinpressSdn. Bhd. and Malaysian Agriculture Research and Development Institute (MARDI). Kuala Lumpur, Malaysia. 45-46.

Jäkel, J. and Nöllenburg, M. 2004. Validation in the Cluster Analysis of Gene Expression Data. Work. Fuzzy- Systeme Comput. Intell.: 13-32. Retrieved from https://scholar.google.com/scholar?gs_lcp=CgZwc3ktYWIQA1Dx9QZY8fUGYKb6BmgBcAF4AIABSog BSpIBATGYAQCgAQGgAQKqAQdnd3Mtd2l6sAEAwAEB&uact=5&um=1&ie=UTF- 8&lr&cites=12407680416147574278. [Accessed Oct. 2020]

Jiwajinda, S., Santisopasri, V., Murakami, A., Kawanaka, M., Kawanaka, H., Gasquet, M., Eilas, R., Balansard, G. and Ohigashi, H. 2002. In vitro anti-tumor promoting and anti-parasitic activities of the quassinoids from Eurycoma longifolia, a medicinal plant in Southeast. 82: 0-3. doi:10.1016/s0378-8741(02)00160-5.

Johnson, R.A. and Wichern, D.W. 1992. Applied Multivariate Statistical Analysis. Prentice Hall, Englewood Cliffs.

Jones, C.S. 1999. An Essay on Juvenility, Phase Change and Heteroblasty in Seed Plants. International Journal of Plant Sciences. 160 (S6): 105-111. doi:10.1086/314215.

Jordano, P. and Godoy, J.A. 2000. RAPD variation and population genetic structure in Prunus mahaleb (Rosaceae), an animal-dispersed tree. Mol. Ecol. 9 (9): 1293-1305. doi:10.1046/j.1365-294X.2000.01009.x.

145 References

Juchum, F.S., Leal, J.B., Santos, L.M., Almeida, M.P., Ahnert, D. and Correa, R.X. 2007. Evaluation of genetic diversity in a natural rose - wood population (Dalbergia nigra Vell. Allemaoex Benth.) using RAPD markers. Genetics and Molecular Research 6 (3): 543-553. Retrieved from https://pubmed.ncbi.nlm.nih.gov/17985307/. [Accessed in Nov: 2019]

Jurik, T.W. 1986. Temporal spatial patterns of specific leaf weight in successional northern hardwood tree species. Am J. Bot 73 (8): 1083-1092. doi.org/10.2307/2443788.

Kaňuch, P., Berggren, Å. and Cassel-Lundhagen, A. 2014. Genetic diversity of a successful colonizer: Isolated populations of Metrioptera roeselii regain variation at an unusually rapid rate. Ecol. Evol. 4 (7): 1117-1126. doi:10.1002/ece3.1005.

Kapoor, T.M., Mayer, T.U., Coughlin, M.L. and Mitchison, T.J. 2000. Probing spindle assembly mechanisms with monastrol, a small molecule inhibitor of the mitotic kinesin, Eg5. J. Cell Biol. 150 (5): 975-988. doi:10.1083/jcb.150.5.975.

Kardono, L.B.S., Angerhofer, C.K., Tsauri, S., Padmawinata, K., Pezzuto, J.M. and Kinghorn, A.D. 1991. Studies on Indonesian medicinal plants. IV. Cytotoxic and antimalarial constituents of the roots of Eurycoma longifolia. J. Nat. Prod. 54 (5): 1360-1367. doi:10.1021/np50077a020.

Kartikawati, S.M., Zuhud, E.A.M., Hikmat, A., Kartodihardjo, H. and Fuadi, M. 2014. Habitat Preferences, Distribution Pattern and Root Weight Estimation of Pasak Bumi (Eurycoma longifolia Jack). J. Trop. For. Manag. 20 (1): 43-50. doi:10.7226/jtfm.20.1.43.

Kassa, A., Konrad, H. and Geburek, T. 2018. Mating pattern and pollen dispersal in the wild olive tree (Olea europaea subsp. cuspidata). Tree Genet. Genomes 14 (1). Tree Genetics and Genomes. doi:10.1007/s11295- 017-1215-z.

Keng, C.L., Sai, S.. and Teo, C.K.H. 2002. A Preliminary Study on the Germination of Eurycoma longifolia Jack (Tongkat Ali) Seeds. Pertanika J. Trop. Agric. Sci 25 (1): 27-34. Retrieved from http://www.pertanika.upm.edu.my/Pertanika%20PAPERS. [Accessed in Apr. 2016]

Khasim, N., Omar, R.Z.R., Ismail, S. and Omar, W. 2009. Integration of Tongkat Ali with Oil Palm. Minist. Plant. Ind. Commod. Malaysia No. 423: 3-6. Retrieved from https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKE wjVjIfaqcvtAhWIAWMBHSbbBrMQFjAAegQIARAC&url=http%3A%2F%2Fpalmoilis.mpob.gov.my% 2Fpublications%2FTOT%2FTT-423.pdf&usg=AOvVaw0gnMzo0TB--QSS3sdFETHy. [Accessed in Sep. 2019]

Kozlowski, T.T. and Pallardy, S.G. 1997. Seed Germination and Seedling Growth. Growth Control Woody Plants: 14-72. doi:10.1016/b978-012424210-4/50002-4.

Krabel, D. 2016a. Fundamentals of tree biology for urban trees. In: A. Roloff (ed.): Urban Tree Management - for a Sustainable Development of Green Cities. Wiley-VCH, Oxford, 20-35.

Krabel, D. 2016b. Genetic aspects. In: A. Roloff (ed.): Urban Tree Management - for a Sustainable Development of Green Cities. Wiley-VCH, Oxford, 211-219.

Kremer, A., Ronce, O., Robledo-Arnuncio, J.J., Guillaume, F., Bohrer, G., Nathan, R., Bridle, J.R., Gomulkiewicz, R., Klein, E.K., Ritland, K., Kuparinen, A., Gerber, S. and Schueler, S. 2012. Long-distance gene flow and adaptation of forest trees to rapid climate change. Ecol. Lett. 15 (4): 378-392. doi:10.1111/j.1461- 0248.2012.01746.x. Kumar, A., Mishra, P., Singh, S.C. and Sundaresan, V. 2014. Efficiency of ISSR and RAPD markers in genetic divergence analysis and conservation management of Justicia adhatoda L., a medicinal plant. Plant Syst. Evol. 300 (6): 1409-1420. doi:10.1007/s00606-013-0970-z.

Kuo, P.C., Shi, L.S., Damu, A.G., Su, C.R., Huang, C.H., Ke, C.H., Wu, J. Bin, Lin, A.J., Bastow, K.F., Lee, K.H. and Wu, T.S. 2003. Cytotoxic and Antimalarial β-Carboline Alkaloids from the Roots of Eurycoma longifolia. J. Nat. Prod. 66 (10): 1324-1327. doi:10.1021/np030277n. 146 References

Kuo, P., Damu, A.G., Lee, K. and Wu, T. 2004. Cytotoxic and antimalarial constituents from the roots of Eurycoma longifolia. 12: 537-544. doi:10.1016/j.bmc.2003.11.017.

Lacerda, D.R., Acedo, M.D.P., Lemos Filho, J.P. and Lovato, M.B. 2001. Genetic diversity and structure of natural populations of Plathymenia reticulata (Mimosoideae), a tropical tree from the Brazilian Cerrado. Mol. Ecol. 10 (5): 1143-1152. doi:10.1046/j.1365-294X.2001.01264.x.

Lars, S. 2007. Tropical Forestry. J. Sustain. For. 3 (2-3): 91-156. doi:10.1300/j091v03n02_06.

Ledig, F.T. 1992. Human Impacts on Genetic Diversity in Forest Ecosystems. Oikos 63 (1): 87. doi:10.2307/3545518.

Lee, C.T., Norlia, B., Tnah, L.H., Lee, S.L., Ng, C.H., Ng, K.K.S., Nor-Hasnida, H., Nurul-Farhanah, Z., Suryani, C.S. and Nur-Nabilah, A. 2018. Isolation and characterisation of SSR markers in tongkat ali (Eurycoma longifolia) using next-generation sequencing approach. J. Trop. For. Sci. 30 (3): 279-291. doi:10.26525/jtfs2018.30.3.279291.

Leverenz, J.W. and Jarvis, P.G. 1980. Photosynthesis in Sitka Spruce (Picea sitchensis (Bong.) Carr.): VI. Response of Stomata to Temperature. Retrieved from http://www.jstor.org/stable/2402096 Access. J. Appl. Ecol. 17 (3): 697-708. [Accessed in Sep. 2020]

Levin, D.A. and Levin, B.Y.D.A. 1973. of BIOLOGY. 48 (1): 3-15.

Lewington, A. 1993. Medicinal Plants and Plant Extracts: A review of their importation into Europe, traffic international, Cam-bridge, UK. Retrieved from https://portals.iucn.org/library/node/6651 [Accessed in Aug. 2020]

Lewontin, R.C. 1972. The Apportionment of Human Diversity. In: Dobzhansky T., Hecht M.K., Steere W.C. (eds) Evolutionary Biology. Springer, New York. doi.org/10.1007/978-1-4684-9063-3_14.

Li, W., Guo, Q., Jakubowski, M. and Kelly, M. 2012. A New Method for Segmenting Individual Trees from the Lidar Point Cloud. Photogrammetric Engineering and Remote Sensing 78: 75-84. doi:10.14358/PERS.78.1.75.

Loc, N.H., Lan, P.T.N., Thanh, L.T.H., Thang, N.V., Luong, N.N., Duc, T.M., Yen, V.T., Hoi, N., Tu, H.T.N. and Doanh, P.H. 2016. An investigation on the distribution and genetic diversity of Eurycoma longifolia Jack and in vitro conservation of this valuable medicinal tree in Thua Thien Hue, Vietnam. Plant Cell Biotechnol. Mol. Biol. 17 (5-6): 226-234.

Loc, N.H., Ngoc Lan, P.T., Yen, V.T. and Dat, H.T. 2018. Some physiological and biochemical characteristics of Eurycoma longifolia Jack tree grown in the arboretum. Plant Cell Biotechnol. Mol. Biol. 19 (5-6): 249-255. Retrieved from https://www.ikprress.org/index.php/PCBMB/article/view/1354. [Accessed in Oct. 2020]

Loi, D.T. 2006. Vietnamese medicinal plants and herbs. Medicine Publication. [Vietnamese]

Low, B.S., Choi, S.B., Abdul Wahab, H., Kumar Das, P. and Chan, K.L. 2013. Eurycomanone, the major quassinoid in Eurycoma longifolia root extract increases spermatogenesis by inhibiting the activity of phosphodiesterase and aromatase in steroidogenesis. J. Ethnopharmacol. 149 (1). doi:10.1016/j.jep.2013.06.023. Madke, S.S., Cherian, K.J. and Badere, R.S. 2014. A modified Murashige and Skoog media for efficient multipleshoot induction in Gmelina arborea Roxb. Journal of Forestry Research 25, 557-564. doi.org/10.1007/s11676-014-0449-y.

Magnussen, S. 1983. Investigation of the influence of humidity on the transpiration resistance of young, shaded and non-shades silver and grand firs (Abies alba and Abies grandis). Flora 173: 279-291.

147 References

Mahmood, M., Normi, R. and Subramaniam, S. 2011. Distribution of 9-methoxycanthin-6-one from the intact plant parts and callus cultures of Eurycoma longifolia (Tongkat Ali). Aust. J. Crop Sci. 5 (12): 1565-1569.

Mansor, P., Lee, S., Farid, A.M. and Parid, M.M. (n.d.). Disease survey of Eurycoma longifolia Jack in Peninsular, Malaysia. Retrieved from https://www.researchgate.net/publication/237641849_Disease_survey_of_Eurycoma_longifolia_Jack_in_ Peninsular_Malaysia. [Accessed in Apr. 2016]

Marbach, I. and Mayer, A.M. 1974. Permeability of Seed Coats to Water as Related to Drying Conditions and Metabolism of Phenolics. Plant Physiol. 54 (6): 817-820. doi:10.1104/pp.54.6.817.

MARD. 2017. Annual decisions by the Minister of Ministry of Agriculture and Rural Development (MARD) on annual forest cover in Vietnam. [Vietnamese]

Martínez, J.P., Silva, H., Ledent, J.F. and Pinto, M. 2007. Effect of drought stress on the osmotic adjustment, cell wall elasticity and cell volume of six cultivars of common beans (Phaseolus vulgaris L.). Eur. J. Agron. 26 (1): 30-38. doi:10.1016/j.eja.2006.08.003.

McClendon, J.H. 1962. The relationship between the thickness of deciduous leaves and their maximum photosynthetic rate. Am J. Bot 49: 320-322. doi.org/10.1002/j.1537-2197.1962.tb14944.x.

Medina, E., Mooney, H.A. and Yánes, C.V. 1983. Phisiological ecology of plants of the wet tropics. In Tasks for Vegetation Science 12. Dr. W. Junk Publishers.

Men, N.T., Truong, H.T., Trang, H.T.M. and Thong, N.D. 2014. Distributional and ecological characteristics of Mohonia nepalensis DC. and Eurycoma longifolia Jack in Lam Dong. J. For. Sci. 3: 3424-3432. Retrieved from http://www.vafs.gov.vn. [Vietnamese, accessed in May 2016].

Milbourne, D., Meyer, R., Bradshaw, J.E., Baird, E., Bonar, N., Provan, J., Powell, W. and Waugh, R. 1997. Comparison of PCR-based marker systems for the analysis of genetic relationships in cultivated potato. Mol. Breed. 3 (2): 127-136. doi:10.1023/A:1009633005390.

Mitchell, A.K. 1998. Acclimation of Pacific yew (Taxus brevifolia) foliage to sun and shade. Tree Physiol. 18 (11): 749-757. doi:10.1093/treephys/18.11.749.

Miyake, K., Tezuka, Y., Awale, S., Li, F. and Kadota, S. 2009. Quassinoids from Eurycoma longifolia. J. Nat. Prod. 72 (12): 2135-2140. doi:10.1021/np900486f.

Mohamad, M., Ali, M.W., Ripin, A. and Ahmad, A. 2013. Effect of extraction process parameters on the yield of bioactive compounds from the roots of Eurycoma longifolia. J. Teknol. Sciences Eng. 60 (4): 51-57. doi:10.11113/jt.v60.1441.

Mohammadi, S.A. and Prasanna, B.M. 2003. Analysis of genetic diversity in crop plants - Salient statistical tools and considerations. Crop Sci. 43 (4): 1235-1248. doi:10.2135/cropsci2003.1235.

Monfared, M.A., Samsampour, D., Sharifi-Sirchi, G.R. and Sadeghi, F. 2018. Assessment of genetic diversity in Salvadora persica L. based on inter simple sequence repeat (ISSR) genetic marker. J. Genet. Eng. Biotechnol. 16 (2): 661-667. Academy of Scientific Research and Technology. doi:10.1016/j.jgeb.2018.04.005.

MonRe (Ministry of Natural Resources and Environment). 2014. Vietnam National Biodiversity strategy to 2020, vision to 2030. Retrieved from https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKE wiFgpCT8s3tAhUS_BQKHUoYDFgQFjAAegQIBxAC&url=https%3A%2F%2Fwww.cbd.int%2Fdoc%2 Fworld%2Fvn%2Fvn-nbsap-v3-en.pdf&usg=AOvVaw0nTnXScDsr6ZdopQBU_hP2. [Accessed in Jun. 2019]

148 References

Morgenstern, K., Polster, J. and Krabel, D. 2016. Genetic variation between and within two populations of Rhabdocline pseudotsugae in Germany. 724 (3): 716-724. doi.org/10.1139/cjfr-2015-0430.

Morita, H., Kishi, E., Takeya, K., Itokawa, H. and Tanaka, O. 1990. New quassinoids from the roots of Eurycoma longifolia. Chemi. Lett 749-752. doi.org/10.1246/cl.1990.749.

Muhamad, A.S., Keong, C.C., Kiew, O.F., Abdullah, M.R. and Lam, C.K. 2010. Effects of Eurycoma longifolia Jack Supplementation on Recreational Athletes’ Endurance Running Capacity and Physiological Responses in the Heat. 22 (2): 1-19. doi:10.24985/ijass.2010.22.2.1.

Nassar, J.M., Hamrick, J.L. and Fleming, T.H. 2001. Genetic variation and population structure of the mixed- mating cactus, Melocactus curvispinus (Cactaceae). Heredity (Edinb). 87 (1): 69-79. doi:10.1046/j.1365- 2540.2001.00910.x.

NDG (Nam Dong National Geographic). 2018. Retrieved from https://thuathienhue.gov.vn/vi-vn/Trang- ch%E1%BB%A7/Th%C3%B4ng-tin-chung/D%C6%B0-%C4%91%E1%BB%8Ba- ch%C3%AD/Th%C3%B4ng-tin-chi-ti%E1%BA%BFt/tid/Huyen-Nam-Dong/newsid/9FB78D15-D352- 4971-B9B7-1703A629A8A1/cid/4F1EE63E-3C7C-4E3D-AB1F-908301E9AA1C. [Vietnamese, accessed in Jun. 2020]

Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583-590.

Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University Press, New York.

Nei, M. and Li, W.H. 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sct. USA. 76 (10): 5269-5273.

Ng, W.L. and Tan, S.G. 2015. Inter-Simple Sequence Repeat (ISSR) Markers: Are We Doing It Right? ASM Science Journal, 9 (1): 30-39.

Nhan, H.N. and Loc, N.H. 2017. Production of eurycomanone from cell suspension culture of Eurycoma longifolia. Pharm. Biol. 0 (0): 2234-2239. Informa Healthcare USA, Inc. doi:10.1080/13880209.2017.1400077.

Nicotra, A.B., Atkin, O.K., Bonser, S.P., Davidson, A.M., Finnegan, E.J., Mathesius, U., Poot, P., Purugganan, M.D., Richards, C.L., Valladares, F. and van Kleunen, M. 2010. Plant phenotypic plasticity in a changing climate. Trends Plant Sci. 15 (12): 684-692. Elsevier Ltd. doi:10.1016/j.tplants.2010.09.008.

Nishteswar, K. 2014. Depleting medicinal plant resources: A threat for survival of Ayurveda. AYU (An Int. Q. J. Res. Ayurveda) 35 (4): 349. doi:10.4103/0974-8520.158972.

Nobel, P.S., Zaragoza, L.J. and Smith, W.K. 1975. Relationship between mesophyll surface area, photosynthetic rate and illumination level during development for leaves of Plectranthus parviflorus Henkcle. Plant Physiol. 55: 1067-1070. doi:10.1104/pp.55.6.1067.

Nordin, M.S. 2014. Medicinal and Aromatic Plants Distribution of the Population of Tongkat Ali (Eurycoma spp.) in Malaysia Based on Data Taken from Herbarium Records. Med. Aromat. Plant 3 (2): 3-5. doi:10.4172/2167-0412.1000155.

Norhidayah, A., Vejayan, J. and Yusoff, M.M. 2015. Detection and Quantification of Eurycomanone Levels in Tongkat Ali Herbal Products. J. Appl. Sci. 15 (7): 999-1005. Science Alert. doi:10.3923/jas.2015.999.1005.

NS. 2018. Statistical Yearbook of Nam Dong district (NS). [Vietnamese]

Okuda, T., Kachi, N., Yap, S.K. and Manokaran, N. 1997. Tree distribution pattern and fate of juveniles in a lowland tropical rain forest - Implications for regeneration and maintenance of species diversity. Plant Ecol. 131 (2): 155-171. doi:10.1023/A:1009727109920.

149 References

Oliver, J. 2013. Biology. In Journal of Chemical Information and Modeling. doi:10.1017/CBO9781107415324.004.

Osman, A., Jordan, B., Lessard, P.A., Muhammad, N., Haron, M.R., Riffin, N.M., Sinskey, A.., Rha, C. and Housman, D. 2003. Genetic Diversity of Eurycoma longifolia Inferred from Single Nucleotide Polymorphisms. J. Plant Physiol. 131: 1294-1301. doi:10.1104/pp.012492. Molecular.

Osman, R., Saim, N., Saaid, M. and Zaini, N.N. 2016. An experimental design method for the extraction of eurycomanone from Tongkat Ali (Eurycoma longifolia) roots using pressurised liquid extraction (ple). Malaysian J. Anal. Sci. 20 (2): 342-350. doi:10.17576/mjas-2016-2002-17.

Ouborg, N.J., Piquot, Y. and Van Groenendael, J.M. 1999. Population genetics, molecular markers and the study of dispersal in plants. J. Ecol. 87 (4): 551-568. doi:10.1046/j.1365-2745.1999.00389.x.

Padua, L.S. De, Bunyapraphatsara, N. and Lemmens, B. 1999. Plant Resources of South-East Asia. Backhuys Publishers, Leiden.

Pallardy, S.G. 2008. Physiology of Woody Plants - 3rd ed. School of Natural Resources University of Missouri Columbia. Academic Press is an imprint of Elsevier.

Panda, S., Naik, D. and Kamble, A. 2015. Population structure and genetic diversity of the perennial medicinal shrub Plumbago. AOB PLANTS, 7: 1-12. doi:10.1093/aobpla/plv048.

Parkhurst, D.F. and Loucks, O.L. 1972. Optimal leaf size in relation to environment. J Ecol. 60: 505-537. Retrieved from https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKE wiizp- Gt8vtAhVb5eAKHWSgB7AQFjAAegQIARAC&url=https%3A%2F%2Fwww.jstor.org%2Fstable%2Fpd f%2F2258359.pdf&usg=AOvVaw2zUmjv3Jd3OytMRU3wm9jC. [Accessed in Sep. 2020]

Patahayah, M., Lee, S.S., Farid, A.M. and Parid, M.M. 2008. Disease survey of Eurycoma longifolia Jack in Peninsular Malaysia. Retrieved from https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKE wj1stS9t8vtAhURdBQKHbimDBEQFjAJegQICxAC&url=https%3A%2F%2Fwww.frim.gov.my%2Fv1% 2Fpdf%2FAnualReport%2Far2005.pdf&usg=AOvVaw1BmE8aNUQ6f2pwlzw8q37N. [Accessed in Mar.2016]

PDG (Phong Dien National Geographic). 2018. Retrieved from https://phongdien.thuathienhue.gov.vn/?gd=60andcn=1134andtc=2482. [Vietnamese, accessed in Jan. 2019]

Peel, J.R., Mandujano, S., M.C., Lopez Portillo, J. and Golubov, J. 2017. Stomatal density, leaf area and plant size variation of Rhizophora mangle (Malpighiales: Rhizophoraceae) along a salinity gradient in the Mexican Caribbean. Rev. Biol. Trop. 65 (2): 701-712. doi:10.15517/rbt.v65i2.24372.

Penfound, W.T. 1931. Plant anatomy as conditioned by light intensity and soil moisture. Am. J. Bot. 18: 197-209. doi.org/10.1002/j.1537-2197.1931.tb09609.x.

Pekşen, E., Pekşen, A. and Artik, C. 2006. Comparison of leaf and stomatal characteristics in faba bean (Vicia faba L.). J. Biol. Sci. 6 (2): 360-364. doi:10.3923/jbs.2006.360.364.

Phuong, V.T.M., Dien, L.T. and Duong, T.T.K. 2013. Biological characteristics and distribution status of Eurycoma longifolia Jack in sandy areas in Hai Lang district, Quang Tri province, Vietnam. J. Rural Dev. Agric. 256- 260. [Vietnamese]

PM (Point Map). Discover Life: Point Map of Eurycoma longifolia. 2012. © Discover Life and original sources. Retrieved from

150 References

http://images.google.de/imgres?imgurl=http%3A%2F%2Fwww.discoverlife.org%2Fmp%2F20m%253Fm ap%253DEurycoma%252Blongifoliaandimgrefurl=http%3A%2F%2Feol.org%2Fdata_objects%2F212350 19andh=360andw=720andtbnid=hCEoElwlEuvj_M%3Aanddocid=zZVxh830NblFXManditg=1andei=Ao X7VvG4GcKaOPm_hJgNandtbm=ischandiact=rcanduact=3anddur=342andpage=1andstart=0andndsp=18 andved=0ahUKEwjxwcCS9ufLAhVCDQ4KHfkfAdMQrQMIJDAC. [Accessed in Mar. 2017]

Poorter, L. and Rozendaal, D.M.A. 2008. Leaf size and leaf display of thirty-eight tropical tree species. Oecologia 158 (1): 35-46. doi:10.1007/s00442-008-1131-x.

Prevost, A. and Wilkinson, M.J. 1999. A new system of comparing PCR primers applied to ISSR fingerprinting of potato cultivars. Theor. Appl. Genet. 98 (1): 107-112. doi:10.1007/s001220051046.

Pritchard, J.K. 2007. Documentation for structure software: V2.3.1 Statistics (Ber). Retrieved from https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKE wiLsozMuMvtAhUB5uAKHTjgCbsQFjAAegQIAhAC&url=https%3A%2F%2Fweb.stanford.edu%2Fgro up%2Fpritchardlab%2Fsoftware%2Fstructure22%2Freadme.pdf&usg=AOvVaw1rhIx5rqzL02FcZMGvIrv 5. [Accessed in Aug. 2019]

Rajmohan, A. 2014. Comparing stomatal densities in sun and shade. Department of Biological Sciences, University of Toronto. Retrieved from https://fr.scribd.com/document/248523446/Comparing-stomatal-densities-in- sun-and-shade. [Accessed in Feb. 2020]

Razi, A.R.M., Aziz, A.A., Alwee, S.S.B.S. and Aziz, R. 2013. Relationships between Malaysians Cultivars of Tongkat Ali (Eurycoma longifolia Jack) Obtained through RAPD Analysis. J. Biotechnol. Wellness Ind. 2: 45-50. Retrieved from https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKE wjHtZCfucvtAhXUDmMBHbThATEQFjAAegQIBxAC&url=http%3A%2F%2Fwww.lifescienceglobal.c om%2Fpms%2Findex.php%2Fijbwi%2Farticle%2Fdownload%2F873%2F546%2F&usg=AOvVaw05ROj 9TbNRYY7B7zMkhBxZ. [Accessed in Feb. 2016]

Rehman, S.U., Choe, K. and Yoo, H.H. 2016. Review on a Traditional Herbal Medicine, Eurycoma longifolia Jack (Tongkat Ali): Its Traditional Uses, Chemistry, Evidence-Based Pharmacology and Toxicology. J. Mol. 21 (3): 331. doi:10.3390/molecules21030331.

Reveal, J.L. and Chase, M.W. 2011. APG III: Bibliographical information and synonymy of Magnoliidae. Phytotaxa 19: 71-134. doi:10.1111/j.1095-8339.2009.00996.x.

Rifai, M.A. 1975. Botanic Data of Pasak Bumi Plant in Symposium of Medicinal Plants Research I. Bogor: Department of Pharmacology FKH-IPB.

Ritz, C., Ansorge, H., Wesenberg, J., Seifert, B., and Wesche, K. 2019. Morphometry, Genetics, Heredibility, Biometry of Analysis Collection Data. Senckenberg Museum of Natural History Görlitz, Germany.

Robert, A. 2016. At the Heart of the Vietnam War: Herbicides, Napalm and Bulldozers Against the A Lưới Mountains. Rev. géographie Alp. 104 (1): 0-17. doi:10.4000/rga.3266.

Rodrigues, K.F. and Tam, H.K. 2009. Short communication molecular markers for Eurycoma longifolia and orthosiphon stamineus. 10 (3): 559-562.

Rohlf, F.J. 1972. An empirical comparison of three ordination techniques in numerical taxonomy. Syst. Zool. 21: 271-280. doi.org/10.2307/2412165.

Roldán-Ruiz, I., Dendauw, J., Van Bockstaele, E., Depicker, A. and De Loose, M. 2000. AFLP markers reveal high polymorphic rates in ryegrasses (Lolium spp.). Mol. Breed. 6 (2): 125-134. doi:10.1023/A:1009680614564.

Rollins, L.A., Moles, A.T., Lam, S., Buitenwerf, R., Buswell, J.M., Brandenburger, C.R., Flores-Moreno, H., Nielsen, K.B., Couchman, E., Brown, G.S., Thomson, F.J., Hemmings, F., Frankham, R. and Sherwin, W.B. 2013. High genetic diversity is not essential for successful introduction. Ecol. Evol. 3 (13): 4501-4517. doi:10.1002/ece3.824. 151 References

Rolston, P. 1978. Water impermeable dormancy. Society 44 (4754): 365-396. Retrieved from https://www.jstor.org/stable/4353936. [Accessed in Oct. 2020]

Rosmaina, R. and Zulfahmi, Z. 2013. Genetic Diversity of Eurycoma longifolia Jack Based on Random Amplified Polymorphic DNA Marker. J. Manaj. Hutan Trop. Journal Trop. For. Manag. 19 (2): 138-144. doi:10.7226/jtfm.19.2.138.

Rosmaina, Ashari, R. and Zulfahmi. 2015. Genetic diversity of Eurycoma longifolia Jack using random amplified polymorphic DNA (RAPD) marker in forest reserve of Kenegerian Rumbio, Indonesia. Malaysian Appl. Biol. 44 (4): 73-80. Retrieved from http://repository.uin-suska.ac.id/id/eprint/23467. [Accessed in Dec. 2020]

Roth-Nebelsick, A. 2007. Computer-based studies of diffusion through stomata of different architecture. Ann. Bot. 100 (1): 23-32. doi:10.1093/aob/mcm075.

Rowshanaie, H., Jaafar, H., Halim, M., Wahab, P. and Rowshanaie, O. 2014. Impact of Different Water Levels on Growth, Plant Water Relations and Photosynthesis Parameters in Seedling of Tongkat Ali (Eurycoma longifolia Jack). Open J. Water Pollut. Treat. 1 (4): 11-20. doi:10.15764/wpt.2014.03002.

Rudrapal, D., Okubo, H., Uemoto, S. and Fujieda, K. 1992. Comparison of the anatomy and physiology of seeds of two varieties of winged bean (Psophocarpus tetragonolobus). Sci. Hortic. (Amsterdam). 51 (1-2): 13-24. doi:10.1016/0304-4238(92)90099-X.

Sajap, A.S., Rozihawati, Z., Omar, D. and Lau, W.H. 2014. Isaria fumosorosea and Metarhizium anisopliae for controlling Atteva sciodoxa (Lepidoptera: Yponomeutidae), a pest of Eurycoma longifolia. J. Trop. For. Sci. 26 (1): 84-91. Retrieved from https://www.jstor.org/stable/23617017. [Accessed in May 2019]

Saleh, A.N. 1992. Differences in form between the reproductive organs of similar species of male and female Tongkat Ali (Eurycoma longifolia Jack). Retrieved from http://malcat.uum.edu.my/kip/Record/um611366/Description. [Accessed in May 2016]

Salisbury, E.J. 1927. On the causes ecological significance of stomatal frequency with special reference to woodland flora. Phil. Trans. Roy. Soc. Lond. Ser B. 216: 1-65.

Sathitbut, V., Keeratinijakal, V., and Peyachoknagul, S. 2016. Genetic Diversity of Eurycoma longifolia Jack in Thailand Detected by AFLP Marker. Agric. Innov. Glob. value Chain. Proc. 54th Kasetsar Univ. Annu. Conf. Thailand. 1: 358-365. Retrieved from https://www.cabdirect.org/cabdirect/abstract/20163376007. [Thai, accessed in Dec. 2020]

Satya, P., Karan, M., Jana, S., Mitra, S., Sharma, A., Karmakar, P.G. and Ray, D.P. 2015. Start codon targeted (SCoT) polymorphism reveals genetic diversity in wild and domesticated populations of ramie (Boehmeria nivea L. Gaudich.), a premium textile fiber producing species. Meta Gene 3: 62-70. Elsevier B.V. doi:10.1016/j.mgene.2015.01.003.

Schäfer, K.V.R., Oren, R. and Tenhunen, J.D. 2000. The effect of tree height on crown level stomatal conductance. Plant, Cell Environ. 23 (4): 365-375. doi:10.1046/j.1365-3040.2000.00553.x.

Schimper, A.F.W. 1898. PXanzengeographie auf Physiologischer Grundlage. Fischer, Jena.

Schlüter, U., Muschak, M., Berger, D. and Altmann, T. 2003. Photosynthetic performance of an Arabidopsis mutant with elevated stomatal density (sdd1-1) under different light regimes. J. Exp. Bot. 54 (383): 867-874. doi:10.1093/jxb/erg087. Silva, A.V.C., Muniz, E.N., Almeida, C.S., Da Vitória, M.F., Da Silva Ledo, A., De Vasconcelos Melo, M.F. and Carregosa Rabbani, A.R. 2015. Genetic diversity and sex identification in Genipa americana L. Trop. Subtrop. Agroecosystems 18 (1): 81-86.

152 References

Soares, A.N.R., Vitória, M.F., Nascimento, A.L.S., Ledo, A.S., Rabbani, A.R.C. and Silva, A.V.C. 2016. Genetic diversity in natural populations of mangaba in Sergipe, the largest producer State in Brazil. Genet. Mol. Res. 15 (3). doi:10.4238/gmr.15038624.

Sloop, C.M. and Ayres, D.R. 2010. Conservation Genetics of Three Endangered Vernal Pool Plants of the Santa Rosa Plain, Sonoma County, California and D Ebra R. (2): 17-19. Retrieved from https://www.semanticscholar.org/paper/CONSERVATION-GENETICS-OF-THREE- ENDANGERED-VERNAL-OF-Christina-M./4817a474b7e66f641a40062c6031fd02e87f2099. [Accessed in Sep. 2020]

Stihl, E.G. and Persson, B. 1991. Provenance variation in early growth and development of Piceu rnuriarla (Mill) B.S.P. Studia Forestalia Suecica. 187: 1-17. Retrieved from https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKE wjdi8T5u8vtAhUj8- AKHRVuBLkQFjAAegQIBBAC&url=https%3A%2F%2Fpub.epsilon.slu.se%2F4064%2F1%2FSFS187. pdf&usg=AOvVaw2Q8FStnaOU8jG_Q0rABdO9. [Accessed in Nov. 2019]

Sultan, S.E. 2000. Phenotypic plasticity for plant development, function and life history. Trends Plant Sci. 5 (12): 537-542. doi:10.1016/S1360-1385(00)01797-0.

Susilowati, A. 2008. Propagation Technique and Genetic Relatedness of Pasak Bumi (Eurycoma longifolia Jack). Under. Bogor Agricultural University. Retrieved from http://repository.ipb.ac.id/handle/123456789/8539. [Accessed in Mar. 2016]

Susilowati, A., Supriyanto, A., Siregar, I.Z. and Subiakto, A. 2012. Propagation Technique of Pasak Bumi (Eurycoma longifolia Jack) Via Shoot Cutting. J. Indones. For. 1: 25-29. Retrieved from https://www.neliti.com/publications/15362/propagation-technique-of-pasak-bumi-eurycoma-longifolia- jack-via-shoot-cutting#cite. [Accessed in Mar. 2016]

Susilowati, A., Rachmat, H.H., Elfiati, D. and Hasibuan, M.H. 2019. The composition and diversity of plant species in pasak bumi’s (Eurycoma longifolia) habitat in Batang Lubu Sutam Forest, North Sumatra, Indonesia. Biodiversitas 20 (2): 413-418. doi:10.13057/biodiv/d200215.

Tee, T.T. and Azimahtol, H.L.P. 2005. Induction of apoptosis by Eurycoma longifolia Jack extracts. J. Anticancer Res. 25 (3B): 2205-2213. Retrieved from https://pubmed.ncbi.nlm.nih.gov/16158965/. [Accessed in May 2016]

Thang, L.V. and Huy, N.D. 2003. Landscape characteristics of the sandy area in the province of Thua Thien Hue, Vietnam. Retrieved from https://tailieu.vn/doc/dac-diem-canh-quan-vung-dat-cat-tinh-thua-thien-hue- 2103494.html. [Vietnamese, accessed in Nov. 2020].

Thao, H., Lan, N.K and Hoang, H.D.T. 2015. Characteristics of plant communities at inner sandy areas in Thua Thien Hue province. Journal of Science, Hue University. 108 (9): 269-278. Retrieved from http://jos.hueuni.edu.vn/index.php/TCKHDHH/article/view/2079. [Vietnamese, Accessed in Dec. 2019]

Thoa, P.T.K., Son, H.T. and Yen N.T.K. 2015. Biodiversity of medicinal plants in Ba Na Nui Chua Nature Reserve, Da Nang city, Vietnam. Journal of Biodiversity and Environmental Sciences. 7 (6): 216-221. Retrieved from https://www.academia.edu/19998306/BIODIVERSITY_OF_MEDICINAL_PLANTS_IN_BA_NA_NUI_ CHUA_NATURE_RESERVES_DA_NANG_CITY_VIETNAM. [Accessed in Sep. 2020] Thoday, D. 1931. The significance of reduction in the size of leaves. J. Ecol. 14: 297-303. doi:10.2307/2255823.

TLNP. 2015. Tu Long National Park: Conservation and development a medicinal gene resource. Retrieved from http://www.baoquangninh.com.vn/kinh-te/201501/vuon-quoc-gia-bai-tu-long-bao-ton-phat-trien-nguon- gen-duoc-lieu-quy-2255661/. [Accessed in Mar. 2016] Tnah, L.H., Lee, C.T., Lee, S.L., Kevin, K.S., Chin, H.N. and Hwang, S.S. 2011. Microsatellite markers of an important medicinal plant, Eurycoma longifolia (Simaroubaceae) for DNA profiling. Am. J. Bot. 98 (5): 130-132. doi:10.3732/ajb.1000469.

153 References

Traffic. 2013. Overview of the use of plants and animals in traditional medicine systems in Vietnam. A Traffic Southeast Asia Report. Greater Mekong Programme. Retrieved from https://portals.iucn.org/library/node/9616. [Accessed in Oct. 2020]

Toai, P.M. and Dien, P.V. 2017. Original Research Paper Forestry Science homogeneity of natural forests at the stand scale: evidences fromA Luoi district, Thua Thien Hue province, Vietnam: Vietnam National University of Forestry. Global Journal for Research Analysis. 2 (11): 24-28.

UBND. 2019. Thua Thien Hue People's committee (UBND). Forest status of Thua Thien Hue province. Avaivable from https://thuathienhue.gov.vn/vi-vn/Thong-tin-dieu-hanh-cua-ubnd-tinh/tid/Cong-bo-hien- trang-rung-tinh-Thua-Thien-Hue-nam-2019/newsid/E6BFDC71-B473-46B5-83A1- AB6C00EB240E/cid/B2893D90-84EA-452E-9292- 84FE4331533D?fbclid=IwAR2jvwSbVez_V8MA5ssGts70t9qaiTu3RALkEkZ7Tn7e1kpmY8Jk0ofxSuU. [Vietnamese, accessed in Feb. 2020]

Vakkari, P., Rusanen, M. and Kärkkäinen, K. 2009. High genetic differentiation in marginal populations of European White Elm (Ulmus laevis). Silva Fenn. 43 (2): 185-196. doi:10.14214/sf.205.

Van, T.Y. and Cochard, R. 2017. Chapter 5 - Structure and Diversity of a Lowland Tropical Forest in Thua Thien Hue Province, Editor(s): Tran Nam Thang, Ngo Tri Dung, David Hulse, Shubhechchha Sharma, Ganesh P. Shivakoti, Redefining Diversity & Dynamics of Natural Resources Management in Asia, Elsevier, 3: 71-85. doi.org/10.1016/B978-0-12-805452-9.00005-9.

Varelides, C., Brofas, G. and Varelides, Y. 2001. Provenance variation in Pinus nigra at three sites in Northern Greece. Ann for Sci. 58: 893-900. doi:10.1051/forest:2001103.

Varshney, R.K., Chabane, K., Hendre, P.S., Aggarwal, R.K. and Graner, A. 2007. Comparative assessment of EST- SSR, EST-SNP and AFLP markers for evaluation of genetic diversity and conservation of genetic resources using wild, cultivated and elite barleys. Plant Sci. 173 (6): 638-649. doi:10.1016/j.plantsci.2007.08.010.

Wan, R.K., Rosazlin, A. and Rozita, A. 2010. Growth of Tongkat ali (Eurycoma longifolia) on a sandy beach ridges soil in Malaysia. 19th World Congr. Soil Sci. Publ. DVD. (8): 104-107.

Wan-Muhammad-Azrul, W. A., Mohd-Farid, A., Lee, S. Y., Sajap, A. S., Omar, D. and Mohamed, R. 2018. A survey on the occurrence of pests and diseases in tongkat ali (Eurycoma longifolia) plantations in Peninsular Malaysia. Journal of Tropical Forest Science, 30 (3): 362-375. doi.org/10.26525/jtfs2018.30.3.362375.

Walter, H. 1973. Vegetation of the earth. New York. Springer.

Ward, J.H. 1963. Hierarchical grouping to optimize an objective function, Journal of the American Statistical Association, 58: 236-244.

Webb, L.J. 1968. Environmental relationships of structural types of Australian rain forest vegetarian, Ecol. 49: 296-311.

Weir, B.S. and Cockerham, C.C. 1984. Estimating F - statistics for the analysis of population structure. Evolution, 38: 1358-1370.

Westing, A.H. 1984. Herbicides in war: past and present. - In: Westing, A.H. (ed.), Herbicides in war: the long- term ecological and human consequences. Stockholm Intern. Peace Res. Inst., Stockholm, Swed. 3-24.

Whitmore, T.C. 1975. Tropical rain forests of the Far East, Oxford, Clarendon.

WHO. 2002. World Heal. Organ. 16 (2): 172. Retrieved from http://apps.who.int/medicinedocs/pdf/s4950e/s4950e.pdf. [Accessed in Jul. 2019]

WHO. 2019. Who global report on traditional and complementary medicine 2019.

154 References

Willi, Y., Van Buskirk, J. and Hoffmann, A.A. 2006. Limits to the adaptive potential of small populations. Annu. Rev. Ecol. Evol. Syst. 37: 433-458. doi:10.1146/annurev.ecolsys.37.091305.110145.

Williams, J.G.K., Kubelik, A.R., Livak, K.J., Rafalski, J.A. and Tingey, S. V. 1990. DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Res. 18 (22): 6531–6535. doi:10.1093/nar/18.22.6531.

Winkler, M., Koch, M. and Hietz, P. 2011. High gene flow in epiphytic ferns despite habitat loss and fragmentation. Conserv Genet 12: 1411-1420. doi.org/10.1007/s10592-011-0239-4.

Wong, H., Marie-Nelly, H., Herbert, S., Carrivain, P., Blanc, H., Koszul, R., Fabre, E. and Zimmer, C. 2012. A predictive computational model of the dynamic 3D interphase yeast nucleus. Curr. Biol. 22 (20): 1881-1890. doi:10.1016/j.cub.2012.07.069.

Wright, I.J., Dong, N., Maire, V., Prentice, I.C., Westoby, M., Díaz, S., Gallagher, R. V., Jacobs, B.F., Kooyman, R., Law, E.A., Leishman, M.R., Niinemets, Ü., Reich, P.B., Sack, L., Villar, R., Wang, H. and Wilf, P. 2017. Global climatic drivers of leaf size. Science 357: 917-921. doi:10.1126/science.aal4760.

Wright, S. 1951. The genetic structure of populations. Ann. Eugen. 15: 323-354. doi.org/10.1038/166247a0.

Wu, F.Q., Shen, S.K., Zhang, X.J., Wang, Y.H. and Sun, W.B. 2015. Genetic diversity and population structure of an extremely endangered species: The world’s largest Rhododendron. AoB Plants 7 (1): 1-9. doi:10.1093/aobpla/plu082.

Xiong, F., Zhong, R., Han, Z., Jiang, J., He, L., Zhuang, W. and Tang, R. 2011. Start codon targeted polymorphism for evaluation of functional genetic variation and relationships in cultivated peanut (Arachis hypogaea L.) genotypes. Mol. Biol. Rep. 38 (5): 3487-3494. doi:10.1007/s11033-010-0459-6.

Xu, Z. and Zhou, G. 2008. Responses of leaf stomatal density to water status and its relationship with photosynthesis in a grass. J. Exp. Bot. 59 (12): 3317-3325. doi:10.1093/jxb/ern185.

Yeh, F.C., Yang R.C. and Boyle, T. 1999. POPGENE Software for Population Genetic Analysis Version 1.32. Quick User Guide.

Yoder, B.J., Ryan, M.G., Waring, R.H., Schoettle, A.W. and Kaufmann, M.R. 1994. Evidence of reduced photosynthetic rates in old trees. For. Sci. 40 (3): 513-527. doi:10.1093/forestscience/40.3.513.

Zanolia, P., Zavattib, M., Montanaria, C. and Baraldi, M. 2009. Influence of Eurycoma longifolia on the copulatory activity of sexually sluggish and impotent male rats. J. Ethnopharmacol. 126 (2): 308-313. doi:10.1016/j.jep.2009.08.021.

Zhang, C., Zhou, X., Jiang, J., Wei, Y., Ma, J. and Hallett, P.D. 2019. Root moisture content influence on root tensile tests of herbaceous plants. Catena 172 (10): 140-147. Elsevier. doi:10.1016/j.catena.2018.08.012.

Zulfahmi, Aryanti, E., Rosmaina, Suherman and Nazir, M. 2019. Differentiation of two species of Pasak bumi (Eurycoma spp) based on leaf morphometric. Plant Arch. 19 (1): 265-271. Retrieved from http://repository.uin-suska.ac.id/id/eprint/23469. [Accessed in Jan. 2021]

155 Appendices

Appendices

Appendix 1. Baseline and plot map for surveying E. Longifolia

Appendix 2. Value of soil value: colors, texture, pH-value and moisture MA: Mountainous area, SA: Sandy area, BA: Baselines

Soil Soil Seedling Plot Soil pH- moisture density Areas BA no. Soil colors texture value (%) Forest status (counts/m2) MA AL1 1 yellow, black. white loam 6.2 64 Medium 0.016 MA AL1 2 yello, black. white loam 5.8 64 Medium 0.032 MA AL1 3 black, white loam 6.0 62 Medium 0.082 MA AL2 4 black, white clay loam 6.4 60 Medium 0.014 MA AL2 5 black, white clay loam 6.3 58 Medium 0.05 MA AL3 6 black, white loam 6.0 60 Rich 0.048 MA AL3 7 Black, white loam 6.2 55 Rich 0.046 MA AL4 8 black clay loam 6.1 85 Rich 0.13 MA AL4 9 black, yellow clay loam 6.1 80 Rich 0.016 MA AL4 10 black, yellow clay loam 6.0 65 Rich 0.05 MA AL5 11 black, yellow clay loam 6.2 75 Medium 0.026

156 Appendices

Soil Soil Seedling Plot Soil pH- moisture density Areas BA no. Soil colors texture value (%) Forest status (counts/m2) MA AL5 12 black clay loam 6.0 70 Medium 0.048 MA AL6 13 black loam 6.4 48 Poor 0.044 MA AL8 14 black loam 6.9 40 Medium 0.258 MA AL8 15 black loam 7.0 45 Medium 0.016 MA BM1 16 red, black clay 6.5 35 Poor 0.028 MA BM1 17 red, black clay 7.5 25 Poor 0.15 MA BM3 18 red clay 7.0 28 Poor 0.034 MA BM4 19 red, black clay 7.1 35 Medium 0.018 MA BM5 20 black clay 7.2 30 Poor 0.06 MA BM5 21 black clay 7.0 35 Poor 0.144 MA ND1.1 22 black clay loam 6.8 32 Medium 0.046 MA ND1.1 23 yellow clay loam 6.9 60 Medium 0.024 MA ND1.2 24 yellow clay loam 6.9 50 Medium 0.024 MA ND1.2 25 black, yellow clay loam 6.8 35 Medium 0.564 MA ND1.3 26 black, yellow clay loam 6.4 50 Medium 0.052 MA ND1.3 27 black, white clay loam 5.5 55 Medium 0.062 MA ND2 28 black, white clay loam 8.2 25 Medium 0.044 MA ND2 29 black, white clay loam 8.5 45 Medium 0.124 MA ND4 30 black, white clay loam 8.7 40 Medium 0.296 MA ND4 31 black, white clay loam 8.7 30 Medium 0.114 MA ND5 32 black, white clay loam 6.5 50 Medium 0.046 MA ND6 33 black clay 6.7 45 Medium 0.408 MA ND6 34 black, white clay loam 6.6 40 Medium 0.29 MA ND7 35 black clay loam 6.7 30 Medium 0.034 MA ND7 36 black clay loam 6.7 40 Medium 0.048 MA ND7 37 black, white clay loam 6.6 45 Medium 0.018 SA PB1 38 black, white loamy sand 5.5 10 Poor 0.034 SA PB1 39 black, white loamy sand 6.4 15 Poor 0.01 SA PB3 40 black, white loamy sand 6.3 10 Poor 0.016 SA PB3 41 black, white loamy sand 5.9 12 Poor 0.022 SA PC6 42 black, white loamy sand 6.2 25 Poor 0.008 SA PC6 43 black, white loamy sand 6.0 20 Poor 0.02 SA PC7 44 black, white loamy sand 5.5 10 Poor 0.024 SA PC7 45 black, white loamy sand 5.9 10 Poor 0.046 SA PH4 46 black, white loamy sand 5.7 25 Poor 0.006 SA PH4 47 black, white loamy sand 6.2 10 Poor 0.012

157 Appendices

Appendix 3. Baseline information and tree, sapling density; tree, sapling height and diameter in mountainous and sandy areas

E UTM N UTM E UTM N UTM Sapling Tree (co- (co- (co- (co- Baseline's density density Mean Mean Mean Mean Baseline starting starting ending ending Forest length Baseline's (counts/ (counts/ sapling sapling tree tree label points) points) points) points) status (m) area (ha) ha) ha) diameter height height DBH 1. AL1 0767078 1778621 0768377 1779495 Medium 3,469 3.47 27.96 5.48 3.91 3.99 7.93 6.69 2. AL2 0768372 1779494 0765366 1778955 Medium 2,222 2.22 6.75 4.05 4.37 4.99 7.79 8.44 3. AL3 0766905 1778596 0766261 1777872 Rich 2,551 2.55 3.53 5.10 3.7 4.06 9.78 8 4. AL4 0766084 1779857 0766906 1778250 Rich 1,317 1.32 16.70 31.89 3.5 3.53 10.62 8.76 5. AL5 0736543 1803870 0736368 1805104 Medium 2,449 2.45 15.11 6.12 4.63 4.46 7.89 6.05 6. AL6 0751700 1782894 0752111 1783080 Poor 3,034 3.03 5.93 4.61 3.94 3.84 8.18 6.66 7. AL8 0736454 1804106 0737397 1804586 Medium 1,088 1.09 34.93 6.43 3.4 2.77 8.51 6.64 8. BM1 0806057 1794501 0805427 1795972 Poor 5,142 5.14 10.70 2.53 3.43 3.28 7.64 7.09 9. BM3 0806830 1797086 0807204 1796824 Poor 1,665 1.67 24.62 0.60 2.99 3.16 10.83 5 10. BM4 0805285 1796090 0805298 1796293 Medium 3,381 3.38 2.07 1.48 3.86 3.53 8.82 7.3 11. BM5 0806638 1797613 0806133 1797357 Poor 2,065 2.07 11.62 0.48 3.04 2.92 15.11 4 12. BM6 0809644 1797533 0810577 1795717 Medium 2,581 2.58 0.77 1.16 4.46 2.15 12.79 9.17 13. ND1.1 0796062 1791116 0793464 1790303 Medium 4,521 4.52 12.83 4.42 3.64 3.9 8.02 8.13 14. ND1.2 0797278 1790569 0793506 1791034 Medium 4,590 4.59 16.78 8.93 3.8 3.82 9.58 7.76 15. ND1.3 0794499 1792303 0792473 1792268 Medium 4,172 4.17 19.89 8.63 3.52 3.71 9.41 8.65 16. ND2 0773599 1822134 0792270 1797142 Medium 5,234 5.23 10.70 12.04 3.91 4.57 9.75 9.25 17. ND3 0803894 1786770 0804364 1786717 Medium 1,072 1.07 23.32 39.18 4.47 3.88 9.18 6.71 18. ND4 0778625 1784527 0778055 1784740 Medium 1,871 1.87 25.12 28.33 3.85 4.2 9.56 7.82 19. ND5 0777969 1784476 0777588 1784234 Medium 1,159 1.16 18.12 8.63 4.02 4.38 8.97 6.55 20. ND6 0795598 1787107 0796342 1789246 Medium 4,408 4.41 20.19 5.44 3.9 3.78 8.68 7.05 21. ND7 0791610 1794522 0973180 1796038 Medium 3,871 3.87 6.46 13.95 3.75 4.96 11.37 11.09 22. PB1 0750543 1842920 0751385 1841997 Poor 1,402 1.40 22.82 0.00 3.75 4.96 0 0 23. PB3 0753527 1840553 0752100 1841302 Poor 1,727 1.73 26.64 0.00 3.15 2.4 0 0 24. PC6 0753909 1842827 0753966 1842888 Poor 2,895 2.90 14.16 0.00 3.03 1.79 0 0 25. PC7 0753985 1843168 0753513 1843450 Poor 3,097 3.10 16.79 0.00 3.17 1.89 0 0 26. PH4 0749958 1841104 0750778 1840118 Poor 2,235 2.24 12.98 0.00 3.7 3.44 0 0 27. PH5 0753821 1838960 0753677 1838191 Poor 1,026 1.03 14.62 0.00 2.85 2.14 0 0

158 Appendices

Appendix 4. Summary of the tree and predictor variables in Pearson’s correlation coefficient **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

Mean Mean Mean Mean tree Mean Soil Slope Forest Parameters sapling sapling seedling Soil pH diameter elevation moisture steepness status diameter height height Tree density .454* .359 .333 .322 -.173 .500** .352 .403* .547** (counts/ ha) .017 .066 .089 .101 .389 .008 .072 .037 .003 Mean tree .768** .283 .247 .520** .133 .548** .542** .833** .476* height .000 .152 .214 .005 .508 .003 .004 .000 .012 Mean tree 1 .543** .479* .751** .155 .702** .503** .809** .743** diameter .003 .011 .000 .442 .000 .007 .000 .000 Sapling -.279 -.248 -.033 -.151 -.047 -.056 -.020 -.141 -.116 density .159 .212 .870 .453 .815 .781 .923 .482 .564 Mean sapling .543** 1 .530** .462* -.050 .569** .239 .455* .501** diameter .003 .004 .015 .806 .002 .231 .017 .008 Mean sapling .479* .530** 1 .548** .460* .482* .199 .256 .401* height .011 .004 .003 .016 .011 .319 .197 .038 Seedling .225 -.002 .113 .120 .142 .035 .391* .178 .162 density .258 .991 .574 .550 .479 .863 .044 .373 .421

Appendix 5. Seedling germination across two areas (Mann-Whitney test)

Test Statisticsa Seedling Ranks germination Areas N Mean Rank Sum of Mann-Whitney U 133.500 Ranks Wilcoxon W 343.500 1.00 30 31.05 931.50 Z -3.300 2.00 20 17.18 343.50 Asymp. Sig. (2-tailed) .001 Total 50

Appendix 6. Comparison of seedling leaves, seedling height and seedling collar diameter between moisture and dry areas

z-Test: Two Sample for Means SL 1 SL 2 SH 1 SH 2 CD 1 CD 2 Mean 6.069 8.913 10.80 19.94 0.453 0.549 Known Variance 10.16 20.23 53.3 197.4 0.066 0.09 Observations 3833 1930 3832 1930 3834 1929 Hypothesized Mean Difference 0 z -24.8135 -26.8169 -11.9722 P(Z<=z) one-tail 0 z Critical one-tail 1.644854 P(Z<=z) two-tail 0.000 z Critical two-tail 1.959964

159 Appendices

Appendix 7. Coordinates X (axis) and Y (asix) for the highest tree and sapling density in the typical plots Plot 09 (mountain) and Plot 046 (sandy area) and the distance among the trees (discussion part)

trees Scatterplot of Y vs X t1 t10 plot no. = 009 t11 t12 20 t13 t14 t15 t16 t17 t18 15 t19 t2 t20 t21 t22 t23 10 t24

Y t25 t26 t27 t28 t29 5 t3 t30 t31 t32 t33 t34 t4 0 t5 t6 0 5 10 15 20 25 t7 t8 X t9

Appendix 8. Correlation among genetic diversity, geographical distance and elevation factors and morphological traits by Mantel test

Matrices Mantel test results Correlation R: 0.1909 1. Genetic matrix and Similarity index 1: Jaccard for genetic matrix Geographic distance Similarity index 2: Euclidean distance for Geographic distance Permutation N: 9999 2. Genetic matrix with Correlation R: 0.1681 p (one-tailed) = 0.0001 elevation Similarity index 1: Jaccard for genetic matrix

Similarity index 2: Euclidean for elevation factor 3. Genetic matrix with Correlation R: 0.1195 morph Similarity index 1: Jaccard for genetic matrix Similarity index 2: Gower for morphological traits

Appendix 9. Average ± standard deviation, range of eurycomanone and water content and biological traits in different areas

Stem Eurycomanone Water content Root diameter Height Sites Areas diameter (mg/g) (%) (cm) (m) (cm) Moist 1.39 ±0.44a 46.85±13.99a 8.51±2.92 6.69±2.59 6.05±3.1 A Luoi (0.94-2.34) (31.57-70.43) (5.5-13.5) (3.3-11.1) (3.3-11.5) 1.11±0.47ab 41.51±15.23ab 10.45±4 6.12±2.33 6.72±1.73 Bach Ma (0.58-1.92) (16.75-62.22) (6.0-15.6) (3.5-10.2) (4.0-9.0) 1.16±0.36ab 50.68±7.38a 12.35±4.22 10.4±5.01 8.57±2.25 Nam Dong (0.82-1.82) (42.08-58.04) (8.3-21.0) (4.8-20.4) (5.6-12.9 Average 1.22±0.42 46.35±12.20 10.44±3.71 7.74±3.31 7.11 ±2.36 Dry 0.74±0.29b 30.49±8.17b 7.56±3.84 4.21±1.27 2.73±1.03 PhongDien (0.28-1.19) (20.00-42.06) (3.0-15.0) (2.5-5.8) (1.3-4.3)

160 Appendices

Appendix 10. Average ± standard deviation of log probability of LN for each K (STRUCTURE program) T1-T40: Run identities with 10 replications of each proposed K (from K1 to K5) (1000000 iterations of each K with a burn-in length of 500000 iterations. To define the optimal K

Run ln_prob Mean Mean No. K mean_llh var_llh _id _data alpha LnPD L'K L''K absL''K Stdv L''K/Stdv

1 T1 2 -13616 -13471.4 289.2 0.4231

2 T2 2 -13616.6 -13471.4 290.4 0.4249

3 T3 2 -13615.9 -13471.4 289.2 0.423

4 T4 2 -13616.3 -13471.6 289.4 0.4248

5 T5 2 -13615.4 -13471.4 288.1 0.4235

6 T6 2 -13616 -13471.4 289.3 0.4243

7 T7 2 -13615.9 -13471.6 288.4 0.4243

8 T8 2 -13615.6 -13471.5 288.2 0.4236

9 T9 2 -13616 -13471.4 289.3 0.4241 10 T10 2 -13616.7 -13471.6 290.3 0.4253 -13616 -383.2 765.41 765.41 0.403 1897.78

11 T11 3 -13232.8 -12995.6 474.4 0.2556

12 T12 3 -13232.8 -12995.4 474.8 0.2538

13 T13 3 -13232.8 -12995.4 474.8 0.2543

14 T14 3 -13233.7 -12995.3 476.8 0.2555

15 T15 3 -13234 -12995.3 477.4 0.2551

16 T16 3 -13234 -12995.6 477 0.2546

17 T17 3 -13235.3 -12995.5 479.6 0.2544

18 T18 3 -13234.5 -12995.5 478.1 0.2536

19 T19 3 -13233.7 -12995.5 476.4 0.2549 20 T20 3 -13235.1 -12995.4 479.4 0.2544 -13233.9 382.17 -205.9 205.9 0.91 225.84

21 T21 4 -13050.2 -12730.9 638.6 0.1796

22 T22 4 -13053.8 -12731.1 645.6 0.1787

23 T23 4 -13053 -12730.9 644.2 0.1783

161 Appendices

Run ln_prob Mean Mean No. K mean_llh var_llh _id _data alpha LnPD L'K L''K absL''K Stdv L''K/Stdv

24 T24 4 -13055 -12731.4 647.3 0.1812

25 T25 4 -13075.5 -12734.8 681.3 0.1975

26 T26 4 -13052.4 -12731 642.9 0.1788

27 T27 4 -13054.8 -12731.3 647 0.1778

28 T28 4 -13048.9 -12731 635.9 0.1793

29 T29 4 -13080.6 -12740.2 680.6 0.2198 30 T30 4 -13051.8 -12730.7 642.1 0.1789 -13057.6 176.27 -5.54 5.54 11.01 0.50

31 T31 5 -12882.4 -12498.7 767.4 0.1425

32 T32 5 -12883.2 -12498.6 769.3 0.1422

33 T33 5 -12906.4 -12512.9 787.1 0.1941

34 T34 5 -12884 -12498.2 771.7 0.143

35 T35 5 -12886.4 -12498.7 775.5 0.1428

36 T36 5 -12884.7 -12498.3 772.9 0.1432

37 T37 5 -12889 -12499 780 0.1422

38 T38 5 -12884.3 -12512 744.5 0.1508

39 T39 5 -12885.3 -12498.7 773.3 0.1422 40 T40 5 -12883 -12498.7 768.6 0.1418 -12886.9 170.73 -170.7 170.73 7.12 23.97

162 Appendices

Appendix 11. List of 276 DNA samples for analysing genetic diversity of mature trees AL: A Luoi, BM: Bach Ma, ND: Nam Dong, PD: Phong Dien

Elevation Elevation No. Areas Samples DBH H No. Areas Samples DBH H (m) (m) 1 AL AL1'-1 5.1 7.0 507 139 ND ND2-20 8.6 13.0 847 2 AL AL1-1 9.9 10.2 495 140 ND ND2-3 7.6 7.5 480 3 AL AL1'-2 3.8 4.5 507 141 ND ND2'-4 8.3 17.0 678 4 AL AL1-2 2.5 2.0 444 142 ND ND2-5 5.4 9.0 601 5 AL AL1'-3 6.7 7.5 450 143 ND ND2'-7 10.2 16.0 583 6 AL AL1'-4 7.3 9.0 452 144 ND ND2-7 19.7 11.0 679 7 AL AL1'-5 5.7 5.0 450 145 ND ND2'-8 6.7 13.0 253 8 AL AL1'-6 12.1 5.5 453 146 ND ND2-8 10.5 9.0 768 9 AL AL2-1 3.5 5.0 684 147 ND ND2-9 13.7 8.0 744 10 AL AL2-7 7.0 6.0 763 148 ND ND3'-1 4.1 4.0 335 11 AL AL2-8 2.9 4.5 761 149 ND ND3'-2 8.8 8.0 361 12 AL AL3'-13 3.8 3.0 786 150 ND ND3'-3 6.7 6.0 452 13 AL AL3-18 3.2 2.0 796 151 ND ND3'-4 9.2 11.5 511 14 AL AL3-19 6.7 6.0 777 152 ND ND3'-5 4.8 4.5 432 15 AL AL3'-9 2.5 2.0 808 153 ND ND3-5 6.1 6.5 402 16 AL AL4'-2 8.0 6.5 725 154 ND ND3-8 8.9 11.0 464 17 AL AL4-2 11.8 4.5 650 155 ND ND4-1 4.6 2.8 428 18 AL AL4-3 8.9 7.0 642 156 ND ND4-5 14.0 12.0 505 19 AL AL4'-5 7.3 7.0 748 157 ND ND4-7 12.4 11.0 561 20 AL AL4-6 9.6 7.0 663 158 ND ND4-8 17.5 11.0 560 21 AL AL5-10 3.5 4.0 921 159 ND ND4-9 11.1 9.0 556 22 AL AL5-11 4.8 3.2 776 160 ND ND6'-1 7.3 8.5 411 23 AL AL5-12 8.0 3.5 827 161 ND ND6-1 4.8 5.6 166 24 AL AL5-13 11.1 6.0 841 162 ND ND6'-2 7.6 8.0 457 25 AL AL5-14 3.2 3.5 884 163 ND ND6-2 4.1 4.8 167 26 AL AL5-15 4.5 4.5 887 164 ND ND6'-4 6.1 7.0 354 27 AL AL5-16 7.0 3.5 888 165 ND ND6-6 11.8 6.0 426 28 AL AL5-17 5.4 3.0 887 166 ND ND6-7 7.3 6.0 473 29 AL AL5-18 5.7 4.8 891 167 ND ND7'-10 8.6 9.0 375 30 AL AL5-19 6.7 7.5 897 168 ND ND7-10 7.0 8.0 702 31 AL AL5-20 9.2 7.0 898 169 ND ND7'-11 7.0 12.0 336 32 AL AL5-21 3.8 3.3 902 170 ND ND7-11 7.6 13.0 728 33 AL AL5-22 5.7 6.5 909 171 ND ND7'-12 5.7 9.0 210 34 AL AL5-24 4.0 3.5 923 172 ND ND7'-13 6.1 10.0 170 35 AL AL5-25 3.3 3.6 924 173 ND ND7'-14 6.1 8.0 156 36 AL AL5-26 5.1 7.2 927 174 ND ND7'-15 8.0 6.0 128

163 Appendices

Elevation Elevation No. Areas Samples DBH H No. Areas Samples DBH H (m) (m) 37 AL AL5-27 8.9 8.5 924 175 ND ND7'-2 8.0 8.0 584 38 AL AL5-28 3.8 4.3 923 176 ND ND7'-4 13.7 13.0 598 39 AL AL5-31 4.8 7.0 915 177 ND ND7'-5 6.1 7.0 600 40 AL AL5-32 5.1 6.5 916 178 ND ND7'-6 8.0 10.0 392 41 AL AL5-34 4.5 6.5 915 179 ND ND7'-7 7.0 12.0 392 42 AL AL5-35 3.8 4.0 920 180 ND ND7'-8 5.7 5.0 393 43 AL AL5-37 4.5 6.2 902 181 ND ND7'-9 6.7 11.0 381 44 AL AL5-38 4.5 5.9 900 182 PD PB1'-1 8.3 4.5 5 45 AL AL5-7 5.1 5.8 920 183 PD PB1-10 4.1 5.7 6 46 AL AL5-8 3.8 3.0 925 184 PD PB1-14 4.1 4.9 4 47 AL AL5-9 8.3 5.5 908 185 PD PB1-16 4.5 4.0 6 48 AL AL6'-2 4.8 3.5 747 186 PD PB1-17 5.6 5.0 7 49 AL AL6-7 6.1 6.7 726 187 PD PB1-19 3.0 3.0 5 50 AL AL6'-8 6.7 5.0 705 188 PD PB1'-2 5.7 28.0 5 51 AL AL6'-9 13.1 7.0 689 189 PD PB1'-4 7.3 3.5 5 52 AL AL8-1 2.9 2.0 731 190 PD PB1-4 2.7 1.2 4 53 AL AL8-10 2.9 1.5 821 191 PD PB1-8 3.9 3.2 8 54 AL AL8-11 2.7 1.5 826 192 PD PB1-9 2.9 3.3 6 55 AL AL8-12 2.5 2.0 868 193 PD PB2-1 5.3 2.5 8 56 AL AL8-13 8.9 5.0 902 194 PD PB2-10 5.7 4.3 8 57 AL AL8-14 3.2 2.5 938 195 PD PB2-11 4.9 2.7 4 58 AL AL8-16 2.9 1.7 925 196 PD PB2-12 6.1 3.5 4 59 AL AL8-17 7.0 5.0 938 197 PD PB2-14 4.4 2.9 4 60 AL AL8-19 3.2 5.0 990 198 PD PB2-3 4.5 1.9 8 61 AL AL8-2 2.5 2.5 769 199 PD PB2-4 5.1 1.9 8 62 AL AL8-21 5.7 6.0 990 200 PD PB2-5 3.8 3.5 8 63 AL AL8-22 6.4 6.5 990 201 PD PB2-6 5.8 3.0 8 64 AL AL8-23 4.8 5.0 999 202 PD PB2-7 5.7 2.6 8 65 AL AL8-24 2.9 1.5 1005 203 PD PB2-8 4.7 3.4 8 66 AL AL8-25 2.9 1.5 967 204 PD PB2-9 4.1 4.3 8 67 AL AL8-27 3.7 1.5 861 205 PD PB3-1 2.5 1.3 6 68 AL AL8-29 2.5 2.5 786 206 PD PB3-10 3.5 3.1 6 69 AL AL8-4 2.9 2.5 778 207 PD PB3-11 2.8 1.9 6 70 AL AL8-6 2.5 2.0 770 208 PD PB3-12 3.8 2.4 6 71 AL AL8-7 2.5 2.5 788 209 PD PB3-14 2.6 1.7 6 72 AL AL8-8 2.2 1.7 783 210 PD PB3-16 3.1 1.7 6 73 AL AL9-1 8.0 7.5 781 211 PD PB3-17 3.8 2.3 6 74 AL AL9-2 24.8 11.5 798 212 PD PB3-19 3.2 3.0 6 75 BM BM1-1 2.9 2.2 717 213 PD PB3'-2 3.8 2.8 6

164 Appendices

Elevation Elevation No. Areas Samples DBH H No. Areas Samples DBH H (m) (m) 76 BM BM1-11 7.0 7.0 608 214 PD PB3-2 2.7 1.6 6 77 BM BM1-12 4.8 4.0 578 215 PD PB3-20 6.1 3.8 6 78 BM BM1-2 5.1 6.3 721 216 PD PB3-22 5.1 3.8 6 79 BM BM1-3 6.1 6.0 722 217 PD PB3-23 2.5 2.2 6 80 BM BM1-5 4.8 5.0 636 218 PD PB3-24 2.5 1.9 6 81 BM BM1-6 10.2 8.0 646 219 PD PB3-25 2.9 2.4 6 82 BM BM1-9 6.1 6.0 649 220 PD PB3'-3 3.5 3.5 6 83 BM BM3-1 3.8 5.0 149 221 PD PB3-3 2.8 2.3 6 84 BM BM3-2 3.0 4.0 149 222 PD PB3'-4 4.8 2.7 6 85 BM BM3-3 10.8 5.0 169 223 PD PB3-4 4.8 2.9 6 86 BM BM3-4 3.5 4.0 202 224 PD PB3-5 2.6 2.4 6 87 BM BM3-5 2.9 2.5 206 225 PD PB3-6 2.6 2.2 6 88 BM BM3-6 5.1 6.0 205 226 PD PB3-7 2.7 2.3 6 89 BM BM3-7 4.5 3.5 202 227 PD PB3-9 2.5 2.3 6 90 BM BM4-1 5.0 5.5 494 228 PD PC6-1 5.8 3.1 6 91 BM BM4-2 8.0 6.0 488 229 PD PC6-10 2.5 1.5 6 92 BM BM4-3 2.9 2.2 514 230 PD PC6-12 2.5 1.5 6 93 BM BM5-1 3.5 2.0 178 231 PD PC6-13 4.2 1.2 6 94 BM BM5-6 7.7 4.0 275 232 PD PC6-14 4.8 1.7 6 95 BM BM6-2 6.8 6.0 80 233 PD PC6-3 5.0 3.0 6 96 BM BM6-3 6.7 7.0 221 234 PD PC6-4 2.5 1.4 6 97 BM BM6-4 24.8 14.5 314 235 PD PC6-5 2.8 1.4 6 98 ND M1 6.8 8.2 582 236 PD PC6-6 2.8 1.3 6 99 ND M10 8.4 8.1 565 237 PD PC6-7 3.0 1.4 6 100 ND M17 3.2 3.0 384 238 PD PC6-8 2.5 2.1 6 101 ND M19 2.5 1.6 246 239 PD PC6-9 2.5 1.0 6 102 ND M24 4.1 2.5 744 240 PD PC7-10 2.7 1.2 7 103 ND M26 7.3 8.2 730 241 PD PC7-11 3.1 1.2 7 104 ND M27 6.4 5.0 678 242 PD PC7-12 2.9 1.6 7 105 ND M29 10.2 10.2 669 243 PD PC7-13 3.8 2.0 7 106 ND M33 7.3 10.0 673 244 PD PC7-14 4.1 1.7 7 107 ND M34 2.9 1.5 674 245 PD PC7-15 2.8 1.6 7 108 ND M35 8.8 5.8 618 246 PD PC7-16 4.8 3.1 7 109 ND M39 10.8 9.0 486 247 PD PC7-18 3.2 2.4 7 110 ND M41 4.5 3.5 456 248 PD PC7-20 5.0 3.8 7 111 ND M44 3.5 4.0 457 249 PD PC7-22 2.5 1.7 7 112 ND M49 13.7 12.0 325 250 PD PC7-3 2.6 2.2 7 113 ND M5 6.8 8.2 634 251 PD PC7-4 4.5 2.4 7 114 ND M50 2.5 2.5 297 252 PD PC7-6 4.4 3.2 7

165 Appendices

Elevation Elevation No. Areas Samples DBH H No. Areas Samples DBH H (m) (m) 115 ND M51 2.9 2.0 276 253 PD PC7-8 2.4 1.9 7 116 ND M53 3.5 4.0 185 254 PD PC7-9 2.9 1.6 7 117 ND M54 4.3 4.8 186 255 PD PH4'-1 8.0 3.0 10 118 ND M56 2.5 3.0 264 256 PD PH4-1 5.3 4.7 15 119 ND M61 20.4 12.0 535 257 PD PH4-11 2.8 3.6 13 120 ND M63 1.3 1.0 633 258 PD PH4-12 4.5 3.8 13 121 ND M64 5.7 5.0 608 259 PD PH4'-2 10.2 5.4 10 122 ND M71 2.9 2.5 422 260 PD PH4-2 5.3 2.6 15 123 ND M72 7.6 7.5 385 261 PD PH4'-3 7.3 6.0 10 124 ND M83 5.7 4.5 351 262 PD PH4-3 2.5 2.0 15 125 ND M89 9.9 13.0 241 263 PD PH4'-4 6.7 4.5 10 126 ND ND1.1'-2 10.2 13.5 658 264 PD PH4-4 5.1 2.9 15 127 ND ND1.1'-5 8.6 10.0 610 265 PD PH4-6 2.9 2.2 15 128 ND ND1.2'-2 7.0 9.0 500 266 PD PH4-7 3.7 12.0 13 129 ND ND2'-1 9.2 14.0 357 267 PD PH4-8 3.8 3.0 13 130 ND ND2-10 7.2 9.0 725 268 PD PH4-9 2.5 2.7 13 131 ND ND2-11 12.4 11.0 770 269 PD PH5-2 5.1 3.2 9 132 ND ND2-12 11.1 12.0 775 270 PD PH5-3 2.5 2.5 9 133 ND ND2-13 8.6 11.0 788 271 PD PH5-4 3.7 2.4 9 134 ND ND2-14 8.3 9.0 791 272 PD PH5-5 3.2 2.0 9 135 ND ND2-15 3.8 7.0 798 273 PD PH5-6 3.6 2.7 9 136 ND ND2-16 12.1 12.0 804 274 PD PH5-7 2.7 2.6 9 137 ND ND2-18 8.6 10.0 816 275 PD PH5-8 2.5 1.9 9 138 ND ND2-19 12.1 10.5 875 276 PD PH5-9 2.9 2.3 9

Appendix 12. List of 15 E. longifolia mother samples and 269 seedlings analysed from four different sites

Seedlings/ Seedlings/ No. Seedlings Provenances No. Seedlings Provenances Mother trees Mother trees 1 AL5-13 MT AL 143 M61.13 SDL ND 2 AL5-13.10 SDL AL 144 M61.14 SDL ND 3 AL5-13.18 SDL AL 145 M61.15 SDL ND 4 AL5-13.19 SDL AL 146 M61.17 SDL ND 5 AL5-13.2 SDL AL 147 M61.19 SDL ND 6 AL5-13.23 SDL AL 148 M61.2 SDL ND 7 AL5-13.24 SDL AL 149 M61.22 SDL ND 8 AL5-13.29 SDL AL 150 M61.25 SDL ND 9 AL5-13.31 SDL AL 151 M61.28 SDL ND 10 AL5-13.32 SDL AL 152 M61.30 SDL ND

166 Appendices

Seedlings/ Seedlings/ No. Seedlings Provenances No. Seedlings Provenances Mother trees Mother trees 11 AL5-13.36 SDL AL 153 M61.32 SDL ND 12 AL5-13.41 SDL AL 154 M61.34 SDL ND 13 AL5-13.43 SDL AL 155 M61.37 SDL ND 14 AL5-13.44 SDL AL 156 M61.39 SDL ND 15 AL5-13.6 SDL AL 157 M61.40 SDL ND 16 AL5-13.7 SDL AL 158 M61.41 SDL ND 17 AL5-13.8 SDL AL 159 M61.45 SDL ND 18 AL5-13.9 SDL AL 160 M89 MT ND 19 AL6-7 MT AL 161 M89.15 SDL ND 20 AL6-7.10 SDL AL 162 M89.17 SDL ND 21 AL6-7.11 SDL AL 163 M89.2 SDL ND 22 AL6-7.14 SDL AL 164 M89.20 SDL ND 23 AL6-7.15 SDL AL 165 M89.21 SDL ND 24 AL6-7.16 SDL AL 166 M89.22 SDL ND 25 AL6-7.17 SDL AL 167 M89.24 SDL ND 26 AL6-7.18 SDL AL 168 M89.27 SDL ND 27 AL6-7.19 SDL AL 169 M89.29 SDL ND 28 AL6-7.2 SDL AL 170 M89.30 SDL ND 29 AL6-7.21 SDL AL 171 M89.33 SDL ND 30 AL6-7.24 SDL AL 172 M89.34 SDL ND 31 AL6-7.3 SDL AL 173 M89.36 SDL ND 32 AL6-7.4 SDL AL 174 M89.37 SDL ND 33 AL6-7.5 SDL AL 175 M89.38 SDL ND 34 AL6-7.6 SDL AL 176 M89.39 SDL ND 35 AL6-7.7 SDL AL 177 M89.4 SDL ND 36 AL6-7.8 SDL AL 178 M89.44 SDL ND 37 AL9-1 MT AL 179 M89.6 SDL ND 38 AL9-1.1 SDL AL 180 M89.8 SDL ND 39 AL9-1.11 SDL AL 181 M89.9 SDL ND 40 AL9-1.13 SDL AL 182 ND6-1 MT ND 41 AL9-1.14 SDL AL 183 ND6-1.11 SDL ND 42 AL9-1.16 SDL AL 184 ND6-1.13 SDL ND 43 AL9-1.17 SDL AL 185 ND6-1.14 SDL ND 44 AL9-1.18 SDL AL 186 ND6-1.15 SDL ND 45 AL9-1.19 SDL AL 187 ND6-1.27 SDL ND 46 AL9-1.2 SDL AL 188 ND6-1.34 SDL ND 47 AL9-1.20 SDL AL 189 ND6-1.35 SDL ND

167 Appendices

Seedlings/ Seedlings/ No. Seedlings Provenances No. Seedlings Provenances Mother trees Mother trees 48 AL9-1.21 SDL AL 190 ND6-1.38 SDL ND 49 AL9-1.23 SDL AL 191 ND6-1.40 SDL ND 50 AL9-1.3 SDL AL 192 ND6-1.41 SDL ND 51 AL9-1.4 SDL AL 193 ND6-1.42 SDL ND 52 AL9-1.5 SDL AL 194 ND6-1.43 SDL ND 53 AL9-1.6 SDL AL 195 ND6-1.46 SDL ND 54 AL9-1.7 SDL AL 196 ND6-1.47 SDL ND 55 AL9-1.8 SDL AL 197 ND6-1.48 SDL ND 56 AL9-1.9 SDL AL 198 ND6-1.49 SDL ND 57 AL9-2 MT AL 199 ND6-1.5 SDL ND 58 AL9-2.1 SDL AL 200 ND6-1.9 SDL ND 59 AL9-2.10 SDL AL 201 PB2-1 MT PD 60 AL9-2.12 SDL AL 202 PB2-1.10 SDL PD 61 AL9-2.13 SDL AL 203 PB2-1.12 SDL PD 62 AL9-2.15 SDL AL 204 PB2-1.15 SDL PD 63 AL9-2.16 SDL AL 205 PB2-1.16 SDL PD 64 AL9-2.19 SDL AL 206 PB2-1.17 SDL PD 65 AL9-2.2 SDL AL 207 PB2-1.19 SDL PD 66 AL9-2.20 SDL AL 208 PB2-1.20 SDL PD 67 AL9-2.21 SDL AL 209 PB2-1.23 SDL PD 68 AL9-2.22 SDL AL 210 PB2-1.29 SDL PD 69 AL9-2.23 SDL AL 211 PB2-1.3 SDL PD 70 AL9-2.3 SDL AL 212 PB2-1.30 SDL PD 71 AL9-2.4 SDL AL 213 PB2-1.38 SDL PD 72 AL9-2.5 SDL AL 214 PB2-1.48 SDL PD 73 AL9-2.6 SDL AL 215 PB2-1.5 SDL PD 74 AL9-2.8 SDL AL 216 PB2-1.8 SDL PD 75 AL9-2.9 SDL AL 217 PB2-10 MT PD 76 BM1-3 MT BM 218 PB2-10.11 SDL PD 77 BM1-3.1 SDL BM 219 PB2-10.12 SDL PD 78 BM1-3.10 SDL BM 220 PB2-10.13 SDL PD 79 BM1-3.11 SDL BM 221 PB2-10.14 SDL PD 80 BM1-3.12 SDL BM 222 PB2-10.18 SDL PD 81 BM1-3.15 SDL BM 223 PB2-10.25 SDL PD 82 BM1-3.18 SDL BM 224 PB2-10.27 SDL PD 83 BM1-3.19 SDL BM 225 PB2-10.29 SDL PD 84 BM1-3.2 SDL BM 226 PB2-10.30 SDL PD

168 Appendices

Seedlings/ Seedlings/ No. Seedlings Provenances No. Seedlings Provenances Mother trees Mother trees 85 BM1-3.20 SDL BM 227 PB2-10.32 SDL PD 86 BM1-3.21 SDL BM 228 PB2-10.33 SDL PD 87 BM1-3.22 SDL BM 229 PB2-10.34 SDL PD 88 BM1-3.23 SDL BM 230 PB2-10.37 SDL PD 89 BM1-3.27 SDL BM 231 PB2-10.38 SDL PD 90 BM1-3.28 SDL BM 232 PB2-10.39 SDL PD 91 BM1-3.29 SDL BM 233 PB2-10.4 SDL PD 92 BM1-3.30 SDL BM 234 PB2-10.40 SDL PD 93 BM1-3.31 SDL BM 235 PB2-10.5 SDL PD 94 BM1-3.34 SDL BM 236 PB2-10.6 SDL PD 95 BM1-3.4 SDL BM 237 PB2-3 MT PD 96 BM1-3.7 SDL BM 238 PB2-3.10 SDL PD 97 BM1-3.8 SDL BM 239 PB2-3.11 SDL PD 98 BM1-3.9 SDL BM 240 PB2-3.12 SDL PD 99 BM1-6 MT BM 241 PB2-3.16 SDL PD 100 BM1-6.1 SDL BM 242 PB2-3.17 SDL PD 101 BM1-6.10 SDL BM 243 PB2-3.18 SDL PD 102 BM1-6.12 SDL BM 244 PB2-3.20 SDL PD 103 BM1-6.18 SDL BM 245 PB2-3.22 SDL PD 104 BM1-6.2 SDL BM 246 PB2-3.23 SDL PD 105 BM1-6.23 SDL BM 247 PB2-3.24 SDL PD 106 BM1-6.24 SDL BM 248 PB2-3.25 SDL PD 107 BM1-6.25 SDL BM 249 PB2-3.26 SDL PD 108 BM1-6.28 SDL BM 250 PB2-3.36 SDL PD 109 BM1-6.3 SDL BM 251 PB2-3.4 SDL PD 110 BM1-6.36 SDL BM 252 PB2-3.44 SDL PD 111 BM1-6.37 SDL BM 253 PB2-3.5 SDL PD 112 BM1-6.38 SDL BM 254 PB2-3.6 SDL PD 113 BM1-6.4 SDL BM 255 PB2-6 MT PD 114 BM1-6.40 SDL BM 256 PB2-6.10 SDL PD 115 BM1-6.44 SDL BM 257 PB2-6.13 SDL PD 116 BM1-6.45 SDL BM 258 PB2-6.15 SDL PD 117 BM1-6.46 SDL BM 259 PB2-6.20 SDL PD 118 BM1-6.48 SDL BM 260 PB2-6.22 SDL PD 119 BM1-6.49 SDL BM 261 PB2-6.23 SDL PD 120 M29 MT ND 262 PB2-6.25 SDL PD 121 M29.1 SDL ND 263 PB2-6.27 SDL PD

169 Appendices

Seedlings/ Seedlings/ No. Seedlings Provenances No. Seedlings Provenances Mother trees Mother trees 122 M29.10 SDL ND 264 PB2-6.41 SDL PD 123 M29.13 SDL ND 265 PB2-6.42 SDL PD 124 M29.14 SDL ND 266 PB2-6.43 SDL PD 125 M29.17 SDL ND 267 PB2-6.55 SDL PD 126 M29.19 SDL ND 268 PB2-6.7 SDL PD 127 M29.22 SDL ND 269 PB2-9 MT PD 128 M29.29 SDL ND 270 PB2-9.1 SDL PD 129 M29.36 SDL ND 271 PB2-9.13 SDL PD 130 M29.37 SDL ND 272 PB2-9.18 SDL PD 131 M29.4 SDL ND 273 PB2-9.19 SDL PD 132 M29.43 SDL ND 274 PB2-9.2 SDL PD 133 M29.55 SDL ND 275 PB2-9.20 SDL PD 134 M29.56 SDL ND 276 PB2-9.21 SDL PD 135 M29.59 SDL ND 277 PB2-9.22 SDL PD 136 M29.68 SDL ND 278 PB2-9.23 SDL PD 137 M29.71 SDL ND 279 PB2-9.24 SDL PD 138 M29.72 SDL ND 280 PB2-9.25 SDL PD 139 M29.8 SDL ND 281 PB2-9.26 SDL PD 140 M61 MT ND 282 PB2-9.4 SDL PD 141 M61.10 SDL ND 283 PB2-9.5 SDL PD 142 M61.11 SDL ND 284 PB2-9.8 SDL PD

Appendix 13. Typical trichomes of E. longifolia

Trichome

170 Appendices

Appendix 14. Images of DNA electrophoresis

Appendix 15. Images of electrophoresis of RAPD and SCoT primers

171 Appendices

Appendix 16. HPLC chromatogram of eurycomanone extract from two root samples (a, b) Ret.: Retention, N: efficiency factor, Tf: Tailing factor, k’: Retention or capacity factor, PDA: Photometric Diode Array, mAU: milli-Absorbance Units a)

172 Appendices

Appendix 17. Images of taking the root samples in the field (measuring the root diameter at 20 cm from the surface)

Appendix 18. Images of the field work a) Recording the morphological traits, geographical, soil parameters, etc.

173 Appendices b) Harvesting the leaves and seeds

c) Worms prefer the young leaves and buds

Appendix 19. The trees were illegally harvested by local people

174 Bibliography

Bibliography

Personal information Name: (Mrs.) Thi Yen Van Nationality: Vietnamese Date of birth: 2nd, June, 1983 Place of birth: Thua Thien Hue, Vietnam Email: [email protected] and [email protected] Website: http://en.huaf.edu.vn/

Education 2/2016-2021 Doctoral student at the Chair of Forest Botany, Department of Forest Sciences, Faculty of Environmental Sciences, Technical University of Dresden (TUD), Germany. 8/2009-5/2011 Degree obtained Master of Natural Resources Management, School of Environment, Resources and Development in Asian Institute of Technology (AIT), Thailand. 9/2002-6/2006 Degree obtained Bachelor of Biology Science in Hue Science of University, Hue University (HUAF), Vietnam.

Work experience 2006-Present Lecturer and Researcher in Faculty of Forestry, Hue University of Agriculture and Forestry, Hue University, Vietnam.

Address: Silviculture Department, Forestry Faculty, Hue University of Agriculture and Forestry, Hue University, Hue, Vietnam.

Note on the commencement of the doctoral procedure

1. I hereby assure that I have produced the present work without inadmissible help from third parties and without aids other than those stated; ideas taken directly or indirectly from external sources are identified as such.

2. When selecting and evaluating the material and also when producing the manuscript, I have received support from the following persons: Prof. Doris Krabel (TU Dresden), Ms. Kristin Morgenstern (TU Dresden) and Prof. Dr. Nguyen Hoang Loc (Hue University, Vietnam).

3. No further persons were involved in the intellectual production of the present work. In particular, I have not received help from a commercial doctoral adviser. No third parties have received monetary benefits from me, either directly or indirectly, for work relating to the content of the presented dissertation.

4. The work has not previously been presented in the same or a similar format to another examination body in Germany or abroad, nor has it been published.

5. I confirm that I acknowledge the doctoral regulations of the Faculty of Environmental Sciences of the Dresden University of Technology.

...... Location, date

...... Signature