UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

How well do plant functional traits and leaf-litter traits predict rates of litter decomposition? Zukswert, Jenna Michelle 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_may_zukswert_jenna.pdf [ 67.8MB ]
Metadata
JSON: 24-1.0300015.json
JSON-LD: 24-1.0300015-ld.json
RDF/XML (Pretty): 24-1.0300015-rdf.xml
RDF/JSON: 24-1.0300015-rdf.json
Turtle: 24-1.0300015-turtle.txt
N-Triples: 24-1.0300015-rdf-ntriples.txt
Original Record: 24-1.0300015-source.json
Full Text
24-1.0300015-fulltext.txt
Citation
24-1.0300015.ris

Full Text

HOW WELL DO PLANT FUNCTIONAL TRAITS AND LEAF-LITTER TRAITS PREDICT RATES OF LITTER DECOMPOSITION? by  Jenna Michelle Zukswert  B.A., Smith College, 2013  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2016  © Jenna Michelle Zukswert, 2016 ii  Abstract Foliar functional traits have been shown to strongly co-vary with each other and with rates of litter decomposition, demonstrating an “after-life effect” of foliar traits on ecosystem processes. Leaf-litter traits are often used to indicate substrate quality in decomposition studies. Chemical traits have been studied more extensively in the context of decomposition than physical traits such as specific leaf area (SLA) and leaf toughness, which impart information on litter structure and decomposer access. I investigated relationships among foliar and litter traits, and between traits and early mass-loss, in 14 plant species native to British Columbia. Both physical and chemical traits were measured in foliage and leaf litter of each species. Foliar traits novel to this kind of study include cuticle thickness and distance to lumen (DTL); novel litter traits include leaching loss and water uptake after 2 and 24 hours. Decomposition, as net proportion of mass lost over time, was measured in litterbags installed in a temperate rain forest at the University of British Columbia Farm in Vancouver. Mass loss was divided into two phases: Phase I from 0 to 3 months, and Phase II from 3 to 12 months.  Foliar traits co-varied in ways predicted by the leaf economics spectrum hypothesis, and litter traits similarly co-varied. Trait-based relationships among species differed when using foliar traits and using litter traits, suggesting that the same traits measured in foliar and litter impart different meaning in the context of decomposition. Phase I was best predicted by leaching loss and litter traits, suggesting that leaching dominates Phase I, and Phase II was best predicted by foliar functional traits such as leaf dry matter content and nitrogen that relate to relative mesophyll abundance, suggesting that decomposer activity dominates Phase II. Physical traits predicted mass loss as well or better than chemical traits, and using both types of traits in correlative studies may provide insights into the processes that underlie litter decomposition. iii  Preface This thesis represents original, unpublished work by Jenna M. Zukswert. I was responsible for developing the research questions, collecting and analyzing the data, and writing the manuscript. I was guided by my supervisor, Dr. Cindy Prescott, in developing my research questions and analyzing my data. The fieldwork for this study took place at UBC Farm, where I coordinated with Veronik Campbell. Dr. Lacey Samuels, Dr. Miki Fujita, Garnet Martens, and Kevin Hodgson assisted with developing a protocol to measure cuticle thickness, and I borrowed solutions and equipment from the Samuels lab and UBC Bioimaging Facility. The blinding program used at first to measure cuticle thickness in an unbiased way was written by David Williams-King. Dr. Patrick Martone provided assistance in measuring toughness and let me use the Instron machine.  Ergosterol concentrations were measured by the Analytical Chemistry Laboratory of the B.C. Ministry of the Environment in Victoria, BC, led by Clive Dawson. Water-soluble extractables (WSE), non-polar extractables (NPE), acid-soluble carbohydrates (ASC), and acid-unhydrolyzable residue (AUR) were measured by the Chemical Services Laboratory of the Pacific Forestry Centre, Natural Resources Canada in Victoria, BC, led by David Dunn. Carbon and nitrogen were measured by Dr. Alice Chang of the Stable Isotope Facility in the UBC Forest Sciences Centre. Site description data and forest floor pH and C:N measurements were made by Kirsten Corrao, and Dr. Maja Krzic provided guidance in forest floor and soil sampling. Allen Larocque, Tim Philpott, and Albert Wu helped install the litterbags in the field, and Juliana Antonio helped collect the last set of litterbags.  iv  Dr. Valerie LeMay provided the R script used to perform cluster analyses and was consulted for advice on data analysis, as was Dr. Jason Barker. Drs. Cindy Prescott, Lacey Samuels, and Maja Krzic contributed to the thesis revision process.    The study design and preliminary results (up to 9 months of mass loss) have previously been presented at the following conferences: 1.  Zukswert, J. M. and Prescott, C. E. 2015. Associations between physical functional traits and leaf litter decomposition rates of 16 plant species native to British Columbia. Poster presented at the 79th Soil Science Society of America’s Annual Meeting, Minneapolis, MN.  2. Zukswert, J. M. and Prescott, C. E. 2015.  Relationships between plant functional traits and short-term leaf litter decomposition rates in 16 plant species native to British Columbia. Poster presented at the 100th Ecological Society of America’s Annual Meeting, Baltimore, MD.  3. Zukswert, J. M. and Prescott, C. E. 2015. Exploring the role of plant functional traits and anatomy in litter decomposition. Poster presented at the 6th annual CONFORWest meeting, San Juan Island, WA.   v  Table of Contents Abstract.......................................................................................................................................... ii!Preface........................................................................................................................................... iii!Table of Contents ...........................................................................................................................v!List of Tables .............................................................................................................................. viii!List of Figures............................................................................................................................... ix!List of Abbreviations ................................................................................................................... xi!Acknowledgements ..................................................................................................................... xii!Dedication ................................................................................................................................... xiii!Chapter 1: Introduction ................................................................................................................1!1.1! Project Overview ............................................................................................................... 1!1.2! Literature Review............................................................................................................... 2!1.2.1! The Process of Litter Decomposition and Transformation......................................... 2!1.2.2! Functional Traits and Leaf Litter Traits in Decomposition ........................................ 6!1.2.2.1! Chemical Traits and Decomposition.................................................................... 7!1.2.2.2! Physical Traits and Decomposition ..................................................................... 9!1.2.3! Plant Cuticles and Their Role in Decomposition...................................................... 12!1.2.4! Relationships Between Functional Traits: The Leaf Economics Spectrum.............. 18!1.2.5! Influence of Traits on Soil Organic Matter............................................................... 19!1.3! Research Objectives......................................................................................................... 21!Chapter 2: Methodology..............................................................................................................24!2.1! Study Area and Plot Description...................................................................................... 24!2.2! Sample Collection............................................................................................................ 25!vi  2.3! Trait Measurements: Foliar Functional Traits ................................................................. 28!2.4! Trait Measurements: Leaf Litter ...................................................................................... 32!2.5! Decomposition Rate Measurements ................................................................................ 35!2.6! Data Analysis ................................................................................................................... 38!Chapter 3: Results........................................................................................................................60!3.1! Relationships Among Foliar Traits and Among Litter Traits.......................................... 60!3.2! Trait-Based Relationships Among Species...................................................................... 61!3.3! Comparison of Traits Between Foliage and Litter........................................................... 62!3.4! Relationships Between Traits and Mass Loss Over Time ............................................... 64!3.4.1! Trends in Mass Loss During the First Year of Decomposition ................................ 64!3.4.2! Traits That Co-Varied with Mass Loss During the First Year of Decomposition.... 65!3.4.3! Traits That Co-Varied with Phase-I Mass Loss........................................................ 68!3.4.4! Traits That Co-Varied with Phase-II Mass Loss....................................................... 70!Chapter 4: Discussion................................................................................................................104!4.1! Relationships Between Foliar and Litter Traits Among Species ................................... 104!4.2! Relationships Between Traits and Early Mass-Loss Rates............................................ 110!Chapter 5: Conclusions .............................................................................................................128!Bibliography ...............................................................................................................................131!Appendix 1: Cuticle Images of Spruce and Subalpine Fir.....................................................146!Appendix 2: Trait Tables ..........................................................................................................147!Appendix 3: Principal Components Analysis and Cluster Analysis with Polystichum munitum ......................................................................................................................................154!Appendix 4: Regression Tree Analysis Tables ........................................................................160!vii  Appendix 5: Linear Regressions...............................................................................................169!Appendix 6: A Comparison of Litters at the Same Stage of Decomposition........................178!A6.1! Introduction................................................................................................................. 178!A6.2! Research Objective ..................................................................................................... 181!A6.3! Methodology ............................................................................................................... 182!A6.4! Results......................................................................................................................... 184!A6.5! Discussion ................................................................................................................... 185!A6.6! Tables and Figures ...................................................................................................... 192! viii  List of Tables  Table 2.1. Horizon Data from Soil Pit..........................................................................................45 Table 2.2. Site Description Data for Each Plot.............................................................................45  Table 2.3. Vegetation Composition in Each Plot..........................................................................46 Table 2.4. Forest Floor Descriptions for Each Plot.......................................................................47  Table 3.1. Correlations Between All Foliar Traits and Two Principal Components....................74  Table 3.2. Correlations Between All Litter Traits and Two Principal Components.....................76 Table 3.3. Correlations Between Select Foliar Traits and Two Principal Components................78  Table 3.4. Correlations Between Select Litter Traits and Two Principal Components................80 Table 3.5. Correlations Between Foliar and Litter Traits.............................................................84 Table 3.6. Comparison of Foliar and Litter Traits........................................................................84 Table 3.7. Important Traits in Regression Tree Analysis of Year-One Mass Loss......................92 Table 3.8. Important Physical Traits in Regression Tree Analysis of Year-One Mass Loss........94 Table 3.9. Important Chemical Traits in Regression Tree Analysis of Year-One Mass Loss......94 Table 3.10. Regression Analysis for Year-One Mass Loss...........................................................95 Table 3.11. Important Traits in Regression Tree Analysis of Phase-I Mass Loss........................96 Table 3.12. Important Physical Traits in Regression Tree Analysis of Phase-I Mass Loss..........98 Table 3.13. Important Chemical Traits in Regression Tree Analysis of Phase-I Mass Loss........98 Table 3.14. Regression Analysis for Phase-I Mass Loss..............................................................99 Table 3.15. Important Traits in Regression Tree Analysis of Phase-II Mass Loss.....................100 Table 3.16. Important Physical Traits in Regression Tree Analysis of Phase-II Mass Loss......102 Table 3.17. Important Chemical Traits in Regression Tree Analysis of Phase-II Mass Loss....102 Table 3.18. Regression Analysis for Phase-II Mass Loss...........................................................103 ix  List of Figures  Figure 2.1. Daily Rainfall Totals..................................................................................................43 Figure 2.2. Total Rainfall From 2004 Through 2015...................................................................44  Figure 2.3. Study Plots at UBC Farm...........................................................................................48 Figure 2.4. P. menziesii Control Images.......................................................................................49  Figure 2.5. P. munitum Control Images........................................................................................50 Figure 2.6. Pinus spp. Cuticle Images..........................................................................................51 Figure 2.7. G. shallon and P. munitum Cuticle Images................................................................52 Figure 2.8. P. tremuloides and P. balsamifera Cuticle Images....................................................53 Figure 2.9. A. amabilis and P. menziesii Cuticle Images..............................................................54 Figure 2.10. B. papyrifera and L. occidentalis Cuticle Images....................................................55 Figure 2.11. A. macrophyllum and A. rubra Cuticle Images........................................................56 Figure 2.12. T. plicata and T. heterophylla Cuticle Images..........................................................57 Figure 2.13. Litterbag Station.......................................................................................................58 Figure 2.14. Leaching Loss After 24 Hours and Mass Loss After 1.5 Months............................59  Figure 3.1. Ordination of All Foliar Traits....................................................................................75 Figure 3.2. Ordination of All Litter Traits....................................................................................77 Figure 3.3. Ordination of Select Foliar Traits...............................................................................79  Figure 3.4. Ordination of Select Litter Traits................................................................................81  Figure 3.5. Cluster Analysis of Woody Plants with Foliar Traits.................................................82 Figure 3.6. Cluster Analysis of Woody Plants with Litter Traits.................................................83 Figure 3.7. Carbon, Nitrogen, and C:N in Foliage and Litter.......................................................85 Figure 3.8. Chemical Proportions in Foliage and Litter...............................................................86 Figure 3.9. SLA and Thickness in Foliage and Litter...................................................................87 Figure 3.10. Toughness and pH in Foliage and Litter..................................................................88 Figure 3.11. Litter Mass Remaining of 14 Plant Species From 0 to 12 Months...........................89 Figure 3.12. Proportion of Mass Lost During Phase I, Phase II, and the First Year....................90 Figure 3.13. Comparison of Mass Lost By Leaching and Mass Lost During Phase I..................91 x  Figure 3.14. Regression Tree to Predict Year-One Mass Loss.....................................................92 Figure 3.15. Regression Trees to Predict Year-One Mass Loss Using Only Physical and Only  Chemical Traits .............................................................................................................................93 Figure 3.16. Regression Tree to Predict Phase-I Mass Loss.........................................................96 Figure 3.17. Regression Trees to Predict Phase-I Mass Loss Using Only Physical and Only   Chemical Traits..............................................................................................................................97 Figure 3.18. Regression Tree to Predict Phase-II Mass Loss.....................................................100 Figure 3.19. Regression Trees to Predict Phase-II Mass Loss Using Only Physical and Only  Chemical Traits............................................................................................................................101  Figure 4.1. Litter Samples Collected at Year One......................................................................127                               xi  List of Abbreviations ASC = acid-soluble carbohydrates (e.g., cellulose and hemicellulose) AUR = acid-unhydrolyzable residue (Klason lignin) C = carbon C:N = carbon-to-nitrogen ratio (%C to %N) CM = cuticular membrane CP = cuticle proper cp = complexity parameter (regression tree analysis) DTL = distance to lumen Fp = force-to-punch Fps = specific force-to-punch LDMC = leaf dry matter content LES = leaf economics spectrum LMA = leaf mass area N = nitrogen  NPE = non-polar extractables P = phosphorus PC = principal component PCA = principal components analysis SLA = specific leaf area SMR = soil moisture regime SNR = soil nutrient regime SOM = soil organic matter WSE = water-soluble extractables   Species Abbreviations Aa = Abies amabilis (amabilis fir) Al = Abies lasiocarpa (subalpine fir) Am = Acer macrophyllum (bigleaf maple) Ar = Alnus rubra (red alder) Bp = Betula papyrifera (paper birch) Gs = Gaultheria shallon (salal) Lo = Larix occidentalis (western larch) Pb = Populus balsamifera var. trichocarpa (black cottonwood) Pc = Pinus contorta var. latifolia (lodgepole pine) Pe = Picea engelmanii x glauca (Engelmann spruce hybrid) Pp = Pinus ponderosa (ponderosa pine) Pmt = Polystichum munitum (sword fern) Pmz = Pseudotsuga menziesii var. menziesii (coastal Douglas-fir) Pt = Populus tremuloides (trembling aspen) Th = Tsuga heterophylla (western hemlock) Tp = Thuja plicata (western redcedar)  xii  Acknowledgements I would first and foremost like to thank my supervisor, Dr. Cindy Prescott, for helping me shape this project, assisting in sample collection, and providing support along the way. I have learned so much from you. I would also like to acknowledge my committee, which consisted of Dr. Prescott, Dr. Lacey Samuels from Botany, and Dr. Maja Krzic, for providing guidance and important insights.  Several other members of the UBC faculty and staff helped with data collection and analysis. Thank you to Veronik Campbell of the UBC Farm for giving me permission to collect samples and run my litterbag study at the farm. Thank you to Miki Fujita and Dr. Samuels in Botany, Garnet Martens and Kevin Hodgson from the Bioimaging Facility (BIF), and Dr. David Bird (visiting scholar in Botany) for assisting me with cuticle thickness measurements. Thanks also to Dr. Patrick Martone in Botany for assistance with leaf toughness measurements, and to students Kyra Janot and Sam Starko for helping me troubleshoot the instrument. Thank you to Valerie LeMay as well for advice on data analysis and R scripts for cluster analysis.  Thank you to David Dunn from the Pacific Forestry Centre and Clive Dawson from the B.C. Ministry for the Environment for performing chemical analyses on my litter samples.  I would like to extend special thank you to my lab group, the Belowground Ecosystem Group (BEG) at UBC. Special thanks the Dr. Alice Chang for laboratory support, to Tim Philpott, Allen Larocque, Albert Wu, Kirsten Corrao, and Juliana Antonio for field assistance, and to Jason Barker, Tim Philpott, Relena Ribbons, Julia Amerongen Maddison, and Dylan Mendenhall for assistance in data analysis and interpretation. Thank you to all BEG members for your support.  Several Green College members also assisted me with this project. Special thanks to David and Kent Williams-King for developing a blinding program for cuticle measurements (inspired by a suggestion from Amy Smith) and to Peter Chen for assisting with fieldwork preparation.  Funding for this project was made possible by NSERC, the Cordula and Gunter Paetzold Memorial Fellowship, the R. Howard Webster Fellowship from Green College, a graduate teaching assistantship in the Biology program at UBC, and the UBC Faculty of Forestry. Faculty of Forestry funding consisted of a Strategic Recruitment Fellowship, three graduate teaching assistantships, and four internal awards (Mary and David Macaree Fellowship, George S. Allen Memorial Scholarship, VanDusen Graduate Scholarship, and Donald S. McPhee Fellowship). I am incredibly grateful to have received this support during my degree.  And special thanks to my family and friends who have been so very supportive throughout this process.  xiii  Dedication       For my family   1  Chapter 1: Introduction 1.1 Project Overview Decomposition is an ecological process that restores nutrients and organic matter to forest soils. The rate at which plant biomass decomposes has important implications for climate change mitigation in that soil organic matter (SOM) can serve as a substantial carbon sink (Dungait et al. 2012). In addition, soil nutrient availability as a result of decomposition strongly influences net ecosystem productivity and, consequently, carbon sequestration in plant biomass (Fernández-Martínez et al. 2014). Decomposition rates can also be used to improve dynamic global vegetation models, which predict how vegetation will change as the climate changes and how in turn these changes in vegetation will influence atmospheric carbon dioxide (Chave et al. 2009). Developing models to accurately predict decomposition rates and investigating the mechanisms behind decomposition, which are currently poorly understood, will allow for improvements in global carbon flux models and perhaps in carbon sequestration efforts (Prescott 2010). Doing so requires greater understanding of and insight into the process of decomposition itself.  This study is an attempt to assess the predictive ability of plant functional traits and leaf-litter traits regarding decomposition rate in early stages of decomposition with the goal of uncovering further relationships between physical traits, which have not been as thoroughly studied as chemical traits, and leaf litter decomposition rates. The correlative nature of this study means that new mechanisms cannot necessarily be discovered from these data. Rather, this study is exploratory and intended to help identify potential directions for research that could help elucidate and describe particular mechanisms in greater depth.  The first part of this study involves a comparison of plant traits among species, and between green foliage and leaf litter, in order to better understand the ways in which traits co-2  vary, possibly reflecting differences in life-history strategies. This then leads into an investigation into how these traits, or assemblages of traits, are associated with early decomposition rates, in particular how they associate with net mass loss after 3 months and from 3 to 12 months. Together, this study attempts to shed a bit more light into the process of decomposition, particularly in early stages, as well as spark more questions to be explored further.  1.2 Literature Review 1.2.1 The Process of Litter Decomposition and Transformation The rate of decomposition is typically measured in terms of net mass loss over time. Litter decomposition rate is commonly measured using the litterbag method, first described in the 1950s, in which a known amount of leaf litter is placed into a mesh bag and left to decompose, either in a laboratory or in the field, and then is collected at a later time to weigh the remaining litter (Bocock and Gilbert 1957, Pérez-Harguindeguy et al. 2013). Often, several subsets of litterbags are deployed and then collected after different time intervals in order to model how decomposition progresses over time. Litter mass loss measured in this way is often modeled using the following exponential decay function: BT+1 = BTe-kt in which BT is the initial mass of litter placed in the litter, BT+1 is the mass remaining at time T+1, e is Euler’s number (2.718), t is the number of years, and k is the rate constant (Olson 1963, Binkley and Fisher 2013). Percent or proportion of mass remaining is also often used to express mass loss over time (e.g., Gholz et al. 2000). Expressing decomposition as mass loss, however, is misleading in that litter tends to be transformed over time, rather than just lost. Therefore, these 3  mass remaining measurements include not only litter material remaining, but also the mass of material of microbial or faunal origin as well (Paul and Clark 1996, Binkley and Fisher 2013, Hobara et al. 2014). Further method refinement is needed to distinguish and quantify these forms of mass and more fully describe how litter is transformed over time. Decomposition consists of several related processes: leaching (loss of material by water), catabolism (chemical degradation of substances), and comminution (physical degradation of substances), the latter two of which are carried out by an assortment of faunal and microbial decomposers (Swift et al. 1979). Soil mesofauna between 0.1 and 2 mm, which consist mainly of arthropods, are important in decomposition in that they break down freshly senescent leaf litter and mix it within the forest floor, making it more accessible to microbial decomposers (Cotrufo et al. 1998, Binkley and Fisher 2013). Soil macrofauna, which are larger than 2 mm and include larger invertebrates as well as small mammals, mix leaf litter into the soil and influence soil structure at a scale not possible by mesofauna, which is believed to increase the rate of decomposition. They also ingest litter material and excrete faecal material that is more nitrogen-rich, i.e., has a lower carbon-to-nitrogen ratio than the plant litter (Binkley and Fisher 2013). Earthworms are perhaps the most influential of soil fauna, altering soil structure in ways that increase pore size and therefore the movement of water and nutrients through the soil, decreasing the thickness of and in some cases eliminating forest floor horizons, and stimulating nutrient cycling processes (Bohlen et al. 2004, Binkley and Fisher 2013). Many litterbag studies, however, do not include the influence of macrofauna if the mesh size of the litterbag excludes macrofauna (Prescott 2005).   Microbes, which include bacteria and fungi, do the majority of the actual decomposing. In acidic forest soils, fungi tend to dominate the process of decomposition (de Boer et al. 2005, 4  Binkley and Fisher 2013). Fungi in terrestrial ecosystems break down materials in litter, such as cellulose and lignin, by secreting enzymes outside of their bodies (exo-enzymes) that break down the complex organic matter into simpler compounds (Ekschmitt et al. 2005). Different types of fungi produce enzymes that decompose different substrates; for example, acid-unhydrolyzable residue (AUR), which was formerly referred to as lignin in many studies, can only be broken down by white rot fungi. Plants produce many types of recalcitrant materials such as lignin, polyphenolics, and cutin, that passively maintain their physical integrity, both during their life and afterwards, and fungi have co-evolved to produce enzymes specialized enough to break down these compounds (Ekschmitt et al. 2005). The complete decomposition of leaf litter therefore requires a sufficiently diverse community of microbial decomposers capable of collectively producing enzymes to break down all of the compounds in leaf litter (Ekshmitt et al. 2005). Many factors influence the likelihood of finding all of the necessary decomposers in the right place and time to decompose a given substrate.   The presence and activity of these microbes and soil fauna is influenced by the climate, the conditions of the forest floor, and by the quality of the litter substrate that they are decomposing. In general, decomposition tends to proceed faster when conditions are warmer and wetter, as enzyme activity is higher as temperature increases (up to an optimum point; Coûteaux and Bottner 1995) and moisture content, which influences soil aeration, can be limiting for microbes (Griffin 1963, Swift et al. 1979, Voroney 2007). For this reason, actual evapotranspiration (AET), an index that incorporates both temperature and moisture, is often an effective predictor of decomposition rates across biomes, with greater AET correlating with greater decomposition rates (Coûteaux and Bottner 1995, Aerts 1997, Gholz et al. 2000).  5   The conditions of the forest floor, which are influenced by climate as well as by the surrounding vegetation and soils type, can also influence decomposer communities and activity and thereby influence decomposition. Tree species can have a considerable influence on microbial communities by providing the substrate that forms the bulk of the forest floor, in particular the F horizon, as well as by influencing microclimate (Ellison et al. 2005, Prescott and Grayston 2013). Soil C:N predicted decomposition rates better than climate variables in lowland tropical forests (Dale et al. 2015), probably due to the influence on the decomposer community. Of recent research interest now is the “Home-Field Advantage”, or HFA, hypothesis, which suggests that plant litter tends to decompose more quickly in environments where the plant species is naturally found, due to the adaptation of the decomposer community to decomposing that particular type of litter (Keiser et al. 2014). Related to HFA is the substrate matrix interaction (SMI) hypothesis, which suggests that the more dissimilar the decomposing litter (substrate) is to the forest floor (matrix), the less able the decomposers will be to degrade the material (Freschet et al. 2012a). These and other related hypotheses have come about due to the recognition of the importance of context in litter decomposition, both in terms of the forest floor conditions and of the soil biota consequently found at the site (Austin et al. 2014). Beyond climate and the forest floor, the composition and quality of the litter substrate itself has a large influence on the rate of decomposition. It was previously believed that substrate quality was an important regulator of litter decomposition on a local scale, while climate had more of influence on this process on a global scale (Aerts 1997). Recent evidence suggests, however, that substrate quality can be important regulator of decomposition on a global scale as well, though climate is still considered an important regulator of decomposition rate (Cornwell et 6  al. 2008; Zhang et al. 2008). Substrate quality is expressed as the chemical or physical traits of leaf litter prior to decomposition.   1.2.2 Functional Traits and Leaf Litter Traits in Decomposition Measurements of substrate quality in decomposition studies are often of leaf litter traits, which may be chemical, such as nutrient content, or physical, such as leaf thickness. In recent years, however, more studies have incorporated functional traits into their investigations of decomposition. Functional traits are morphological, physiological, or phenological traits that influence the performance of individuals, such as their growth or survival, which in turn influences their fitness (Violle et al. 2007, Pérez-Harguindeguy et al. 2013). For the purpose of this study, the terms “functional trait” or “foliar trait” will be used to refer to traits of living plants, or foliage, as the plant functional traits would not necessarily serve the same function in a leaf once it has senesced. The term “litter trait” will be used to refer to traits of leaf litter.   Plant functional traits have been used to study an assortment of population, community, and ecosystem properties and processes. Examples include population vital rate elasticities (e.g., elasticity in survival and fecundity), response to herbivory, and decomposition rates (Vioelle et al. 2007, Freschet et al. 2012b, Schuldt et al. 2012, Adler et al. 2014). The popularity of trait-based models, for decomposition as well as for other ecological processes, is evident in the development of a global database of plant traits called TRY as well as a handbook of standardized methods for measuring plant traits (Kattge et al. 2011, Pérez-Harguindeguy et al. 2013). Trait-based models have even been proposed for other ecological systems, such as root-infecting fungi, due to the success of trait-based models for plants (Aguilar-Trigueros et al. 2014).   7   Although functional traits are measured in green leaves, not in leaf litter, and therefore may not accurately represent the material present on the forest floor, functional traits do convey information about plant life-history strategies (Cornwell et al. 2008, Freschet et al. 2010, Bakker et al. 2011). Associations between functional traits and decomposition serve to determine whether trait-based responses to evolutionary pressures that living plants face, such as herbivory and soil nutrient availability, have an “after-life” effect on decomposition rate, and therefore on ecosystem processes, after the leaves have died (Santiago 2007, Freschet et al. 2010, Jackrel and Wootton 2015). In other words, are there biogeochemical consequences to the assemblages of traits that living plants possess? And how well do these functional traits predict decomposition rates, in comparison to leaf-litter traits that reflect substrate quality?  Trait-based models of litter decomposition have shortcomings. Namely, they historically have considered chemical traits more than physical traits (Makkonen et al. 2012) and have focused primarily on changes in mass loss and not on changes in the structure of the substrate.   1.2.2.1 Chemical Traits and Decomposition A large emphasis has previously been placed on the relationship between chemical traits and mass-loss rate. Chemical traits can impart information about the nutrient content, secondary metabolic defences, chemical recalcitrance, and photosynthetic rate of the plant tissue, which all influence the plant’s performance. Examples of chemical traits include standard measurements of carbon chemistry (proportions of water-soluble extractables, non-polar extractables, acid-soluble carbohydrates, and acid-unhydrolyzable residues), nutrient content (as concentrations or proportions), carbon-to-nitrogen ratio, and pH. 8  Regarding foliar and litter carbon compounds, there are many types of organic compounds that can be produced in a leaf. Standard fractions of organic compounds often isolated to indicate substrate quality include water-soluble extractables (WSE), non-polar extractables (NPE), acid-soluble carbohydrates (ASC), and acid-unhydrolyzable residue (AUR). These fractions are often isolated from each other in a stepwise manner, first with dichloromethane to isolate NPE, then with hot water to isolate WSE, and then placed at 450 °C in a muffle furnace to isolate AUR (Trofymow et al. 1998). Water-soluble extractables include water-soluble sugars and phenolics (Trofymow et al. 1998). Many of these soluble compounds tend to be removed from litter first during decomposition through leaching, contributing to fast initial rates in the first phase of decomposition followed by slower rates in later phases (Vaieretti et al. 2005, Berg and McClaughert 2008). Non-polar extractables include soluble fats, waxes, and oils (Trofymow et al. 1998), which therefore include compounds comprising the cuticle, to be discussed in section 1.2.3. ASC, which consists of cellulose and hemicellulose, is obtained following sulphuric acid hydrolysis (Trofymow et al. 1998). AUR was formerly known as Klason lignin (Gallardo and Merino 1993, Prescott 2010). It is now known, however, that AUR includes other recalcitrant plant compounds as well, such as cutin (Gallardo and Merino 1993). AUR, and in particular the ratio of AUR:N, has often been found to predict decomposition rates well (Zhang et al. 2008, Jackson et al. 2013a). In addition to carbon chemistry of foliage and litter, nutrients have also been used to predict decomposition rates. In northern temperate forests, nitrogen tends to be a limiting nutrient, particularly for microbes (Swift et al. 1979, Kaye and Hart 1997). Both foliar and litter N are often strong predictors of decomposition rate (Cornwell et al. 2008, Zhang et al. 2008, Bakker et al. 2011). Ratios of nutrients to each other or of carbon or “lignin” (AUR) to nutrients 9  are also used to predict decomposition rates, in particular C:N (Cornelissen et al. 2006, Zhang et al. 2008). C:N has been shown to be a more powerful predictor than either C or N content alone (Pérez-Harguindeguy et al. 2000)  One chemical trait influenced by nutrient concentrations as well as by other aspects of underlying leaf chemistry is pH. Different compartments of the foliar or litter tissue, such as vacuoles, the cytoplasm, and the cell wall, may have very different pH values; the tissue pH is therefore an overall average of the pH of these components considered together (Cornelissen et al. 2011). Despite being an average of the different leaf components and compounds, there are some broad trends in leaf pH. For example, higher concentrations of base cations, such as calcium, tend to increase pH. Higher concentrations of organic acids conversely tend to decrease pH (Cornelissen et al. 2006, 2011). Leaf-litter pH has been found to explain a sizeable proportion of variation in decomposability between species, up to 25% alone, and also can help improve models with other traits such as concentrations of lignin (acid detergent lignin determined using the Van Soest method), cellulose, and N (Cornelissen et al. 2006). pH is also a reliable trait for species-specific models because it tends to vary more with species than with environmental factors, such as soil pH, and the ranking of pH among species is similar when comparing foliage and litter (Cornelissen et al. 2006, 2011).  1.2.2.2 Physical Traits and Decomposition Less emphasis has been placed on the role of physical functional and leaf litter traits in decomposition, which impart information about the structure of the leaf. While the chemical composition of the leaf influences its decomposability as well as its structure, measurements of physical traits help impart information and insight into how accessible the litter is to 10  decomposers, and in foliage, how investment in certain structures and tissue properties during the plant’s life may have consequences for nutrient cycling after the leaf senesces and abscises. In this way, physical traits help produce a more comprehensive view of how decomposition proceeds. Several physical traits that have been previously assessed to determine their relationship to decomposability include specific leaf area, leaf dry matter content, leaf toughness, and leaf thickness. Specific leaf area (SLA), the area per unit dry mass of a leaf, is one of the most effective physical functional-trait predictors of decomposition rate (Vaieretti et al. 2005, Cornelissen et al. 2006, Bakker et al. 2011). SLA reflects a trade-off between growing rapidly through substantial allocation of resources towards photosynthesis and nutrient acquisition, in that leaves with larger SLA values tend to have broader, thinner leaves that are more efficient at photosynthesis (Santiago 2007), and growing less rapidly but instead possessing long-lived leaves that better conserve nutrients (Pérez-Harguindeguy et al. 2013). Plants with larger SLA values tend to have thin leaves that have higher nutrient concentrations (Santiago 2007). They also tend to be short-lived, more palatable to herbivores, and more easily broken down by decomposers (Vaieretti et al. 2005, Cornelissen et al. 2006, Bakker et al. 2011). The reciprocal of SLA (1/SLA, also known as leaf mass area, or LMA) has also been shown to correlate with leaf mechanical properties such as leaf tensile strength (Onoda et al. 2011). In general, SLA is a good overall indicator of leaf structure. Vaieretti et al. (2005) speculated that SLA could directly influence decomposition rate in that leaves with higher SLA have a greater area of substrate available to microbes, but also acknowledged the possibility that SLA is correlated with other plant traits and strategies and therefore could indirectly affect decomposition rate.  11  Another physical trait related to decomposition is leaf dry matter content (LDMC), which is the ratio of dry weight to water-saturated weight. Foliage or litter with greater LDMC values tends to have less spongy and nutrient-rich mesophyll, having instead relatively more supportive, structural material (Kazakou et al. 2006). LDMC is also a good indicator of soil fertility; plants in more nutrient-rich environments tend to have lower LDMC values (Hodgson et al. 2011). Lower LDMC values are associated with faster decomposition rates (Freschet et al. 2012b, Schuldt et al. 2012).  Leaf toughness is a measurement that tests the resistance of a leaf to physical breakage. This toughness could be due to the presence of fibres, lignin, or possibly silica found in plant cell walls (Cornelissen and Thompson 1997), and therefore, while it is a physical trait, it is an emergent property of a suite of chemical traits. It can be measured by tensile tests, shearing tests, or punch-and-die tests, all of which can be used in trait-based decomposition studies (Pérez-Harguindeguy et al. 2013). Tougher leaves are generally more resistant to herbivores, pathogens, and decomposers (Pérez-Harguindeguy et al. 2000). Cornelissen et al. (1999) demonstrated in a study of 120 deciduous and coniferous plant species from Argentina and the UK that leaf tensile strength was negatively correlated with litter mass-loss rates, suggesting that leaves with more physical defences decompose more slowly. Pérez-Harguindeguy et al. (2000) also found that leaf tensile strength correlated strongly with decomposition rates of 52 angiosperm species in Argentina. Leaf toughness measured using punch-and-die tests has also been shown to correlate with decomposition rate; in a study of nine Mediterranean tree and shrub species, toughness measured as force-to-punch and the ratio of toughness to nutrients were the best predictors of mass loss (Gallardo and Merino 1993). Force-to-punch has not been used as frequently as other measurements of toughness because it is highly dependent on lamina thickness and can produce 12  misleading results when uncorrected (Choong et al. 1992, Aranwela et al. 1999). Leaf toughness measurements corrected for punch circumference and leaf thickness, however, have been correlated with leaf lifespan and index of sclerophylly in leaves, which is measured as the ratio of fibre-to-protein in leaves (Choong et al. 1992, Onoda et al. 2011). Therefore, it has been proposed for use as a predictor of plant-herbivore interactions and leaf litter decomposition rate.  These physical trait measurements have been shown to correlate with decomposition rate. Although chemical traits may often be more practical to measure, leaves are chemically and structurally heterogeneous, and therefore measurements of overall leaf chemistry may not fully reveal structural influences on decomposition (Choong et al. 1992). Further physical measurements that may help elucidate the role of leaf structure in litter decomposition, particularly early stages of decomposition, relate to cuticle structure and the process of leaching.  1.2.3 Plant Cuticles and Their Role in Decomposition One physical functional trait rarely measured in regards to its role in decomposition is cuticle thickness. The cuticle is the “skin of plants” (Müller and Riederer 2005), a layer of lipid material found on the exterior of epidermal cells that covers all exposed aerial surfaces of all terrestrial plants and serves as the first point of contact between plants and other organisms (Baker 1982, Müller and Riederer 2005). The primary function of the cuticle is to limit non-stomatal water loss, but additional functions include regulating movements of gases and solutes between the plant and its surroundings, shielding the plant from ultraviolet radiation, and defending the plant against pathogens and herbivores (Müller and Riederer 2005, Bargel et al. 2006).  13   Certain cuticle characteristics that help actualize these functions are common to all plants. All plant cuticles, for example, consist of a cutin matrix in which intracuticular waxes are embedded (Holloway 1982). Cutin is a complex, insoluble lipid polymer composed of saturated C16 ω -hydroxy and unsaturated C18 hydroxy-epoxy fatty acid monomers (Bargel et al. 2006). The three-dimensional structure of cutin is still uncertain, but it consists of pores averaging 0.3 to 0.5 nm in diameter into which aliphatic and aromatic cuticular waxes are embedded (Bargel et al. 2006). Regarding waxes, all plant cuticles have been found to contain very long chain fatty acids (VLCFAs), primary alcohols, aldehydes, alkanes, and alkyl esters, but the exact types of these compounds and their distributions in the cuticle vary widely among species; a single species could have more than 50 different types of these chemical structures, not including other types of compounds aside from these five that can also be embedded in cuticles (Jetter et al. 2006). Other chemicals that can be found in plant cuticles include secondary metabolites such as triterpenoids, phenylpropanoids, and flavonoids (Jetter et al. 2006).   Just as the chemical composition can vary widely with species, so too can the structure of the plant cuticle, among species and also during different stages of development (Holloway 1982). The outermost layer of the cuticle is epicuticular wax. This layer can vary greatly in structure and composition, containing structures defined as plates, tubes, ribbons, rodlets, filaments, and dendrites (Baker 1982). These diverse structures can be found on different parts of the needle of a single species, as observed in scanning electron microscopy (SEM) images of Tsuga heterophylla (western hemlock) needles (O’Reilly et al. 1989). Recent evidence suggests that these diverse structures are the result of the accumulation of one type of chemical compound, suggesting a strong link between chemical composition and physical structure (Jetter et al. 2006). The surface of the epicuticular wax layer strongly influences the wettability of the 14  cuticle as well as the ease with which fungal pathogens can adhere to the leaf surface and subsequently germinate and penetrate the leaf (Bargel et al. 2006, Carver and Gurr 2006, Nguyen et al. 2014). The roughness of the cuticle surface strongly influences the hydrophobicity and self-cleaning properties of the cuticle; the greater the water-contact angles of the epicuticular wax structures, the more readily water droplets can roll off the surface of the plant, removing particles such as fungal spores in the process and reducing the amount of water that enters the cuticle (Bargel et al. 2006, Nguyen et al. 2014). Environmental factors such as temperature, humidity, and sunlight can influence the morphology of epicuticular waxes, but cuticular waxes tend to be more strongly dictated by genetics than by environmental factors (Baker 1982, Jetter et al. 2006).   Below the epicuticular wax is a layer, composed entirely of lipids, called the cuticle proper (CP, Müller and Riederer 2005). The thickness of the CP varies widely among species, but is generally less than 200 nm (Bargel et al. 2006). The CP is a continuous layer of cutin with embedded intracuticular waxes that in some cases contain non-ester bonds that help protect against fungal invaders; these non-ester bonds make the cuticle more resistant to esterase enzymes produced by fungi (Müller and Riederer 2005).    The epicuticular wax, CP, and the cuticular layers below the CP collectively form the cuticular membrane (CM). The structure of the cuticle below the CP can vary considerably between species. For some species, the only component of the CM below the epicuticular wax film is the CP (Holloway 1982). For others, there is a more distinct cuticular layer underneath. Layers beneath the CP consist of polysaccharides as well as lipids because they are formed from depositing cutin and wax through and on top of the cell wall material (Holloway 1982). The morphology of these layers may differ widely between different species. For some, there is a 15  distinct cuticular layer above the secondary cell wall. For others, the secondary cell wall is cutinized through to the primary cell wall. In still others, the cell wall surrounding the entire epidermal cell, not just the surface, may be cutinized. In some cases, cuticular pegs also extend between neighboring epidermal cells (Holloway 1982). Holloway (1982) cautioned against making large generalizations about cuticle morphology, strongly suggesting that the cuticle morphology of each species be considered independently.  Cuticle morphology has also found to differ within species - even within the same individual. Cuticle thickness of Pseudotsuga menziesii (Douglas-fir) needles was found to increase with tree height and negatively co-vary with osmotic potential at full turgor, which suggests a role of cuticle thickness in regulating transpiration (Woodruff et al. 2010). Cuticle thickness was also found to differ in different positions in the crown of Betula papyrifera (paper birch) trees, with the thickest cuticles found in the upper and outermost parts of the crown (Ashton et al. 1998). And on average, abaxial (underside of the leaf) cuticles tend to be thinner and weaker than adaxial (top of the leaf) cuticles (Hunt and Baker 1982, Price 1982, Onoda et al. 2012). Serving functions in regulating water loss, solute loss, and defence against microbes and herbivores, it would be expected that the plant cuticle will also have a strong influence on the process of litter decomposition as well. This idea was suggested by Onoda et al. (2012), who found that leaves of Australian evergreen woody plants that had thicker cuticles had greater tensile strength and a higher modulus of elasticity, which means that those leaves were more stiff. This result suggests that cuticle thickness is directly related to mechanical leaf properties, which, they speculated, suggests that cuticle thickness may therefore influence decomposition rate. This was also suggested by Tian et al. (1997), who produced micrographs of forest-floor 16  thin-sections depicting fungal penetration of plant material following the destruction or separation of the cuticle from the rest of the plant, in their case Abies spp. (fir) needles and Betula spp. (birch) leaves. In their images, the cuticle seemed to act as a barrier against decomposers. They suggested, therefore, that the ease of hyphae penetrating the cuticle, and consequently accessibility of litter substrate to decomposers, will influence the decomposition rate. Their observations also suggested that the cuticle and epidermis are among the most difficult parts of the leaf to decompose, more so than the mesophyll and vascular bundles. In aquatic decomposition of Eucalyptus globulus (blue gum), the cuticle was also found to be a barrier to decomposition; it was found to stay intact for five weeks (Canhoto and Graça 1999). Cutin is notoriously difficult to break down and is one of the most recalcitrant plant compounds (Swift et al. 1979), consequently often well preserved in soil (Riederer et al. 1993). Ratios of cutin to phosphorus and nitrogen were significantly related to post-leaching decomposition of tree and shrub species in southwestern Spain (Gallardo and Merino 1993). Because many of these leaves were still intact, not broken, such that the cuticle remained an effective barrier to decomposers, the amount of cutin in leaf litter relative to nutrients was assumed to strongly influence decomposition rate (Gallardo and Merino 1993). The role of the cuticle in regulating water loss and solute retention may also influence leaching that occurs in early stages of decomposition. Solutes pass through the cuticle by passive diffusion, a process that is in part influenced by the thickness of the cuticle (Price 1982, Müller and Riederer 2005). Lipophilic and polar pathways exist through the heterogeneous cuticular membrane, allowing for the regulated passage of select non-polar and polar molecules across the cuticle (Müller and Riederer 2005). In this way, water and solutes would diffuse in and out of the cuticular membrane slowly, which would hypothetically slow the rate of mass loss by leaching. 17  This and the cuticle’s water repellency would both help to reduce or slow leaching (Bargel et al. 2006). Taylor and Parkinson (1988a) suggested that the rate of water uptake of leaf litter after soaking for two hours in distilled water relates to the strength of the cuticle. Water uptake and leaching loss, therefore, are two measures of leaf litter that could in part reflect and relate to the role of the cuticle in early stages of decomposition.  Leaching loss rate, the rate at which water-soluble material is removed from the litter, is measured by calculating the change in mass of oven-dried, senesced leaves that have soaked in distilled water (Taylor and Parkinson 1988b). Although leaching loss rate is primarily a physical trait, like toughness, it is the emergent property of a suite of chemical traits, relating to the composition of the leaf and its cuticle and how accessible water is to the leaf (Swift et al. 1979). Plant materials typically leached in early stages of decomposition include inorganic elements such as calcium, magnesium, and potassium, as well as simple, water-soluble organic compounds, including sugars, proteins, and low-weight phenolics (Taylor and Parkinson 1988b, Berg and McClaughert 2008, Ibrahima et al. 2008). The inconsistent ability of C:N to predict leaching loss, however, suggests that other factors, such as structural properties, may influence leaching as well (Soong et al. 2014).  In these ways, cuticle thickness and leaching loss may be related to mass loss during decomposition, in particular during early stages of decomposition. Taylor and Bärlocher (1996) hoped to find a correlation between adaxial cuticle thickness and leaching loss, but did not, suggesting changes to leaf structure during air-drying or differences in chemical composition as possible reasons for this unexpected result. To my knowledge, no trait-based studies of in situ decomposition have considered measurements of cuticle thickness or leaching loss as possible 18  predictors. Doing so might shed more light onto the process of decomposition itself and encourage further investigation into the role of the cuticle in the process of decomposition.  1.2.4 Relationships Between Functional Traits: The Leaf Economics Spectrum Trait-based studies have revealed that certain plant traits tend to co-vary. For example, leaves with shorter life spans and greater SLA values tend to have higher nutrient concentrations (N and P) and higher photosynthetic rates. These observations led to a new hypothesis articulated by Wright et al. (2004) stating that certain physical and chemical traits that reflect the life history strategy of plants co-vary along a “leaf economics spectrum” (LES). They used leaf trait data from 2,548 species to identify six traits that composed this spectrum: leaf mass area (LMA, the reciprocal of SLA), photosynthetic assimilation rate, leaf N, leaf P, dark respiration rate, and leaf lifespan. On one end of the LES are plants having high values of traits, such as leaf N and SLA, for which a high value indicates that plants with them tend to grow quickly and have short-lived, nutrient-rich leaves that are relatively easy for microbes to decompose. These plants allocate more energy and resources towards acquiring nutrients and fewer towards conserving them. On the other end of the spectrum are plants having high values of traits such as leaf lifespan for which high values are often associated with plants that allocate more energy and resources towards conserving nutrients and producing small, thick, tough leaves that live longer and are better protected from decomposers. These relationships can be expressed collectively as single variables determined by principal component analysis (PCA). Wright et al. (2004) determined that one principal component (PC) variable incorporating photosynthetic capacity, LMA, and N concentrations accounted for 82% of the variation in these traits among the species studied. Subsequent PCAs using different species, more traits, and both foliage and leaf-litter have 19  further confirmed that traits generally tend to co-vary in a way predicted by the LES hypothesis (Kazakou et al. 2006, Bakker et al. 2011, Birouste et al. 2012, Makkonen et al. 2012, Jackson et al. 2013a, Pietsch et al. 2014).  Research following the publication of the LES by Wright et al. (2004) produced evidence of possible effects of this spectrum on ecological processes, such as herbivory and decomposition. On a local scale (Santiago 2007, Bakker et al. 2011, Freschet et al. 2012b) as well as on a global scale (Cornwell et al. 2008, Pietsch et al. 2014), evidence exists to support the hypothesis that plants’ nutrient acquisition strategies reflected by this spectrum correlate with decomposability. In particular, plants with thin, nutrient-rich leaves and high photosynthetic rates tend to decompose more readily than plants with more recalcitrant leaves and lower photosynthetic rates. In this way, these suites of related, co-varying traits influence belowground ecosystem processes and soil organic matter formation.  1.2.5 Influence of Traits on Soil Organic Matter  By the ways in which they influence microbial activity, chemical and physical leaf-litter traits can influence the process of soil organic matter (SOM) formation. SOM is composed primarily of carbon compounds deriving from plant and microbial biomass (Kögel-Knabner, 2002). A recently articulated framework of SOM formation acknowledges that microbes are less efficient at metabolizing recalcitrant litter compounds, such as lignin, and the more labile compounds that they do metabolize are therefore more often incorporated into SOM (Cotrufo et al. 2013). This framework suggests that the likelihood of litter compounds becoming incorporated into SOM depends on how efficiently the compounds can be used by microbes (“Substrate Use Efficiency”), which reflects chemical traits and depends on the accessibility of 20  the compounds to microbes, which may be influenced by physical traits among other factors (Cotrufo et al. 2013). Factors that influence the likelihood of microbial decomposition of organic matter in soil include the chemical composition of the litter substrate (e.g., Cotrufo et al. 2013), formation of chemical complexes with soil minerals (Dungait et al. 2012; Cotrufo et al. 2013, Mueller et al. 2015), relative locations of microbes and SOM in soil (Ekschmitt et al. 2005, 2008; Dungait et al. 2012), and environmental conditions (Grandy and Neff 2008).  A new hypothesis put forth by Lehmann and Kleber (2015), the soil continuum model, summarizes our current understanding of these many factors that influence SOM formation, suggesting that SOM exists not as large “humic substances” in the soil, but rather as a continuum of transforming organic compounds. Factors that influence the degree of decomposition of these organic matter particles include presence of soil biota and microbial enzymes, soil mineralogy, and incorporation of SOM into aggregates.   Much research has been done on the relationships between the chemical composition of litter and decomposition by microbes, but not as much has been done on relationships with physical accessibility to substrates or with physical functional traits (Papa et al. 2014). Physical traits associated with defence, such as water uptake rate after two hours as an indicator of cuticle strength, leaf toughness, cell wall microporosity, and leaf dry matter content, and traits that affect the microclimate experienced by the microbes, probably influence rates of microbial decomposition as well, though these relationships have yet to be explored in depth.   Short-term decomposition studies may also be important in determining the influence of leaf litter traits on soil chemistry and SOM. Mambelli et al. (2011), for example, observed that C and N from initial litter chemistry of Pinus ponderosa (ponderosa pine) needles and roots in the organic and mineral soil within 1.5 years of decomposition in microcosms. They suggest that 21  initial litter chemistry and also litter anatomy at least indirectly influence SOM formation in early stages of decomposition, indicating structural differences in leave and roots to be a possible reason for differences in C and N contributions to SOM.   Our understanding of leaf-litter decomposition has progressed substantially in recent years through the application of functional traits to study ecological processes as well as the recognition of decomposition’s role in SOM formation, the carbon cycle, and in mitigating climate change: the number of studies relating to decomposition increased sharply after this connection between decomposition and climate change was made clear in the 1990s and has remained high since (Prescott 2010). Despite our progress in this area, there remains more to discover, in particular regarding the relationships between plant traits and decomposition rate. This study was intended to begin addressing these gaps in our current understanding.  1.3 Research Objectives The objective of this study was to explore relationships between leaf traits and mass-loss rate in early stages of decomposition. By doing so, I hoped to help provide a more clear idea of the physical process of litter decomposition and to help improve predictions of mass loss during early stages of decomposition. This study addressed two research questions related to this topic.  1. What relationships exist between foliar functional traits and leaf-litter traits, and among species?  Relationships between select physical and chemical traits in foliage and litter of 14 different plant species native to British Columbia were assessed using PCA to determine how these traits co-varied with each others, using cluster analysis to determine how species grouped 22  based on their trait similarity (rather than phylogenetic similarity), and using non-parametric t-tests to determine how traits measured in both foliage and leaf litter differed from each other. I expected to find that traits would co-vary in a way that reflects the leaf economics spectrum (Wright et al. 2004); specifically, that traits with high values that are thought to correspond to greater nutrient acquisition as opposed to conservation, such as SLA and N content, would be positively correlated and in opposition to traits with high values thought to correspond with greater nutrient conservation, such as thickness, C:N, and LDMC. Traits measured in this study that are not often measured in functional trait or decomposition studies include cuticle thickness and distance-to-lumen (DTL, which includes both the secondary cell wall and cuticle thickness of the epidermal cells) as well as leaching loss and water uptake. I expected to find that cluster analysis would group traits in a way that also reflected relationships hypothesized by the LES hypothesis, grouped according to similar life history strategies; for example, I expected the broadleaf, deciduous tree species to be more similar to each other than the coniferous, evergreen tree species. Additionally, given the findings that foliar and litter traits could be equally used to predict decomposition rate (Cornwell et al. 2008), I expected values of foliar traits to be similar to those of leaf litter traits, or at least that the ranking of the traits within the species I studied would be the same regardless of whether the trait was measured in foliage or leaf litter.  2. What relationships exist between traits and early mass-loss rates during decomposition? Relationships between leaf traits and early mass loss rates were assessed using PCA, classification and regression tree analysis, and linear regression to see which traits, or combination or traits, were more strongly correlated with the proportion of mass loss after 3 months (Phase I), mass loss from 3 to 12 months (Phase II), and mass loss over the entire study 23  period, from 0 to 12 months. I expected, because early stages of decomposition are often dominated by leaching (Swift et al. 1979), that leaching loss and the proportion of water-soluble extractables (WSE) would be more important predictors of decomposition in early stages, as well as structural traits that would influence decomposer access to tissue (e.g., cuticle thickness and leaf toughness). In later stages, I expected that nutrient quality (which I measured as C:N and  proportion of N) would have a greater influence on decomposition rates in Phase II of this study, as I predicted that by this point more substrate would have been made accessible to microbes and therefore the actual nutrient content of the litter itself would more directly influence decomposition rate. 24  Chapter 2: Methodology 2.1 Study Area and Plot Description This study was conducted in southern British Columbia, in the Coastal Western Hemlock, Eastern Very Dry Maritime (CWHxm1) biogeoclimatic zone. The litterbag study was conducted at University of British Columbia’s farm, hereafter UBC Farm, in Vancouver (49°15’N, 123°14’W). The mean daily temperature in Vancouver over the study period (2 December 2014 to 2 December 2015) was 11.4 °C, and overall temperature ranged from -5.6 °C to 28.3 °C (Environment Canada 2015). A total of 1090.4 mm of rain fell in Vancouver during this period (Figure 2.1), which is 11.4 mm less than fell during the same period the year before (1101.8 mm, Environment Canada 2015) and just less than the mean rainfall during the same period from 2004 through 2014 (mean = 1115.3 mm, Figure 2.2). There was a drought in the summer of 2015, during the middle of this study period, which led to citywide water restrictions. The forest at UBC Farm is a temperate rain forest dominated by Thuja plicata (western redcedar), P. menziesii var. menziesii (coastal Douglas-fir), and Tsuga heterophylla (western hemlock), with Acer macrophyllum (bigleaf maple) trees also present in the canopy and Alnus rubra (red alder) trees present along the perimeter. Soils at this site can be classified as Bose Duric Humo-Ferric Podzol (Luttmerding 1984). These soils have LFH, Ahe, and Bf horizons, which have a loamy sand texture, are well drained, and are moderately acidic (Table 2.1, Luttmerding 1984).   I established six plots for the litterbag study, to be described in Section 2.5, along the perimeter of the UBC Farm forest. These six plots, labeled A through F, were installed away from the Agroforestry Trail that runs through the forest, at least 10 m away from the trail. Site characteristics were described and vegetation data were collected in 10 × 10 m plots; shrubs were 25  identified and shrub cover was measured in 1 × 1-m plots within each 10 × 10-m plot. Canopy cover ranges from 50 to 90% in these plots, and the soil nutrient regime for plots B through F ranges from medium to rich (Table 2.2, BC Ministry of Forests and Range and BC Ministry of Environment 2010). Plot A is in the stem exclusion successional stage, while plots B through F are in the understory reinitiation stage (Table 2.2). All plots except for plot F have T. plicata trees in the canopy (Table 2.3). The pH of the forest floor in these plots ranged from 4.15 to 4.68 and the carbon-to-nitrogen ratio (C:N) ranged from 23.4 to 27.8 (Table 2.4). Plot F was later removed from the study due to dissimilarities in vegetation and probable dissimilarities in light levels, as the plot was located near a large canopy gap (Figure 2.3).  2.2 Sample Collection I chose 16 plant species native to British Columbia for this study and collected data for all of them, but I only analyzed data for 14 species in this thesis (excluding A. lasiocarpa and P. engelmannii x glauca). Of the 16 species, nine were coniferous tree species, five were broadleaf tree species, and two were understory species. The coniferous tree species were Picea engelmannii x glauca (spruce), Abies lasiocarpa (subalpine fir), Abies amabilis (amabilis fir), Tsuga heterophylla (western hemlock), Pinus ponderosa (ponderosa pine), Pinus contorta var. latifolia (lodgepole pine), Thuja plicata (western redcedar), Pseudotsuga menziesii var. menziesii (coastal Douglas-fir), and Larix occidentalis (western larch). The five broadleaf tree species were Populus balsamifera var. trichocarpa (black cottonwood), Populus tremuloides (trembling aspen), Alnus rubra (red alder), Acer macrophyllum (bigleaf maple), and Betula papyrifera (paper birch). Gaultheria shallon (salal) and Polystichum munitum (sword fern) were also 26  studied. I collected foliage and litter samples of each species, which were used to measure physical and chemical traits. I collected foliage samples from each of the 16 species between 27 May and 31 July 2014 from southern and southwestern British Columbia. Samples collected at UBC Farm were P. menziesii, T. plicata, T. heterophylla, P. balsamifera, A. rubra, A. macrophyllum, G. shallon, and P. munitum. Samples collected near the Skimikin tree-species trial (50°48’N, 119°26’W, Thomas and Prescott 2000) near Salmon Arm, BC were L. occidentalis, B. papyrifera, and P. tremuloides. This forest is located within the moist-warm Interior Cedar-Hemlock biogeoclimatic zone (ICHmw). A. lasiocarpa, P. engelmannii x glauca, P. ponderosa, and P. contorta were collected en route to Vancouver from Salmon Arm along the Coquihalla Highway. A. amabilis were collected at Cypress Bowl Provincial Park. I collected sun leaves as opposed to shade leaves and avoided juvenile leaves as well as the current-year growth in evergreen species (distinguished by their lighter green color and location on the tips of branches). Samples were returned to lab in plastic Ziploc™ bags into which a single moist paper towel was placed and were stored in a cold storage room at 4ºC. Physical functional traits measured on foliage include specific leaf area (SLA), leaf dry matter content (LDMC), leaf toughness as force-to-punch (Fp) and specific force-to-punch (Fps), leaf thickness, cuticle thickness, and distance to lumen (DTL, to be defined in Section 2.3). Chemical traits measured include C:N, pH, and proportions of water-soluble extractables (WSE; e.g., simple sugars, water-soluble phenolics), non-polar extractables (NPE; e.g., soluble fats, waxes, and oils), acid-soluble carbohydrates (ASC; e.g., cellulose and hemicellulose), and acid-unhydrolyzable residue (AUR, formerly known as Klason lignin; Trofymow et al. 1998). 27  I collected freshly senesced leaf litter from the same 16 species between 25 August and 2 November 2014. I collected T. heterophylla on 25 August 2014 from a cement pathway running through the Pacific Spirit Regional Park forest in Vancouver, BC. I collected P. menziesii samples on 16 October 2014 from plastic trays lined with fiberglass mesh with 1 × 1-mm holes that had been placed in a small stand of such trees at UBC Farm earlier in the season, and collected T. plicata from trees adjacent to the P. menziesii trees on the same day. I collected T. plicata litter from standing trees; this litter could be easily brushed from the branch. Species from which I collected litter on 18 October 2014 from the Agroforestry Trail portion of the UBC Farm include A. macrophyllum, P. munitum, and G. shallon. I collected P. ponderosa litter on 26 October 2014 from beneath a tree near the Totem Park residences at the University of British Columbia, and A. rubra on 27 October 2014 from a small stand of trees near the Agroforestry Trail portion of UBC Farm. I collected A. amabilis litter on 2 November 2014 from trees at Cypress Bowl Provincial Park, near the start of the Yew Lake Barrier Free Interpretive Trail. Species from which colleagues collected litter outside of Vancouver include P. contorta, P. tremuloides. L. occidentalis, P. engelmannii x glauca, A. lasiocarpa, and P. balsamifera, all of which were collected on 22 or 23 September in the Kananaskis region of Alberta, and B. papyrifera, which was collected 14 October 2014 along the Coquihalla Highway in British Columbia. Immediately upon collection, I refrigerated the litter samples at 4°C. After several days, I spread all of the litter out in the laboratory and allowed it to air dry for seven days. The air-dried litter was then stored at room temperature in brown paper bags.  28  2.3 Trait Measurements: Foliar Functional Traits I calculated SLA by dividing the area of one side of a leaf by its oven-dry weight (Pérez-Harguindeguy et al. 2013). The area was measured using a LI-3100 area meter (LI-COR, Lincoln, NE, USA), with either 0.1-mm2 or 1.0-mm2 resolution depending on the size of the leaf. I calibrated this leaf scanner using a 10-cm2 metal disk when I changed the resolution to 0.1 mm2 and a 50-cm2 metal disk when I changed the resolution to 1.0 mm2; I calculated a conversion factor by measuring the area of the disk ten times, determining the mean measured area, and dividing this mean by either 10 or 50 cm2. Leaf samples were then oven-dried at 70°C for more than 72 hours, transported in a desiccator from the drying oven to a balance, and weighed immediately. I removed the petioles from broadleaf samples before measuring area.   I measured LDMC by dividing the oven-dry weight of the leaf by the water-saturated weight of the leaf. I rehydrated the leaf samples upon collection by storing them in humidified plastic bags (a sealed plastic Ziploc™ bag with a moist paper towel in it) at 4ºC; samples stored in this way were considered water-saturated. The leaf samples were weighed and then placed in a drying oven at 70°C. I determined LDMC by dividing the oven-dry weight of the leaf in milligrams by the fresh weight in grams (Pérez-Harguindeguy et al. 2013).  I measured leaf toughness using a punch-and-die mounted to an Instron machine. The punch was flat-edged, round, 1.04 mm in diameter, and had a clearance of 0.1 mm, as recommended in a study by Aranwela et al. (1999) and in the most recent plant functional trait handbook (Pérez-Harguindeguy et al. 2013). Toughness was measured on laminar tissue, avoiding veins if possible, for ten leaves per species. Veins could not be avoided for coniferous needles, so the punch necessarily passed through them. Force to punch (Fp) was calculated by dividing the force it takes to punch the leaf (in Newtons) by the circumference of the punch, 29  which is 3.27 mm. Force is divided by circumference rather than punch area because the force of the punch is concentrated at the edges of the punch (Onoda et al. 2011, Pérez-Harguindeguy et al. 2013). Force-to-punch was also divided by the thickness of the leaf, which was measured using a digital micrometer at the time toughness was determined, in order to obtain measurements of specific force-to-punch (Fps, Onoda et al. 2011). For broadleaf samples, I measured thickness at the spot symmetrical to where force-to-punch was measured, on the opposite side of the central vein. For conifer needles, I measured thickness on either side of the spot where Fps was measured and calculated the mean of the two thickness measurements. I took care to avoid measuring thickness where Fps would be measured in order to minimize any mechanical disruption of structure that could influence the calculations. I measured cuticle thickness on freehand sections of green leaves that had been stained with nile red, a fluorescent stain that brightly stains lipids but is insoluble in water (Fowler et al. 1985). I stained sections with 0.001% aqueous nile red solution for 5 to 10 minutes and then mounted in distilled water on glass microscope slides. Three leaves or needles were studied for each species, and three free-hand sections of each leaf or needle were prepared. Sections were viewed using a Axioplan light and fluorescent microscope (Zeiss, Jena, Germany) under a 40× objective lens (400× magnification in total) using a Rhodamine B filter cube using excitation wavelengths between 540 and 552 nm (BP 546/12) and capturing emission wavelengths of 590 nm or longer (LP590). Four images of each section were obtained using a QImaging Retiga 1300 camera and QCapture software; two images were taken of the adaxial cuticle and two were taken of the abaxial cuticle. The adaxial cuticle was defined as the cuticle that receives the most sunlight on the plant, or that is typically found on the “top” of the leaf. For most species, this side of the cross-section was characterized by columnar, palisade mesophyll cells. The abaxial cuticle 30  was defined as the cuticle on the “back” of the leaf, which often had more spongy mesophyll and faced away from the sunlight. In the case of P. engelmannii x glauca, the needle faces inwards towards the stem, so the back of the needle, which was found on the thicker side of the needle, was considered the adaxial side, while the top of the needle, facing in towards the stem, was considered the abaxial side. Although this is the reverse of the others in a “top” and “back” specificity, this designation corresponds in function. To ensure that nile red stained the cuticle and to identify autofluorescence present in unstained sections, I produced sets of control images in December 2015 with a representative broadleaf species (P. munitum) and representative conifer species (P. menziesii), taking images of the abaxial and adaxial cuties of stained and unstained sections from the same tissue using the same exposure time and settings (Figures 2.4 and 2.5).   Twelve images were taken per leaf, which resulted in 36 images per species. Each image was edited for optimal brightness and contrast without oversaturation upon analysis. I numbered the cells consecutively from left to right in each image and chose three cells using a random number generator. I measured the cuticular membrane, or cuticle, which includes the epicuticular wax, cuticle proper, and cuticularized layers, of the three selected cells in each image using the “Measure” tool in ImageJ (labeled “C” in Figures 2.4 through 2.10). I also measured the distance to lumen (DTL), which included the secondary cell wall in species if it was clearly visible and distinct from both the cuticle and the lumen (labeled “D” in Figures 2.6 through 2.12). In cases where the secondary wall was distinct from the cuticle and the lumen, it was a lighter shade than the lumen, but did not appear to have been stained as brightly as the cuticle; I tested the similarity in shade objectively by viewing the image in a colour filter other than “grayscale”, such as “16 colors”, to help better distinguish the shades and determine whether the secondary 31  cell wall was more similar to the lumen or the cuticle. Therefore, the cuticle thickness and DTL of 12 cells were measured for each cross-section, of 36 cells for each leaf or needle, and of 108 cells for each species.  A blinding program was written that renamed and shuffled the images after all had been obtained so that the species identity was obscured during measurements; the first 250 images were measured blindly. Later, however, I performed the rest of these measurements and repeated measurements for the first 250 species knowing the species identity of the photograph, since doing so would help inform me of the anatomy of the leaf and help me more accurately identify the cuticle proper and DTL. The mean abaxial cuticle thickness, adaxial cuticle thickness, abaxial DTL, and adaxial DTL were measured for each species. The chemical functional traits considered were C:N, pH, and proportions of WSE, NPE, ASC, and AUR. Samples were oven-dried at 70°C and then homogenized and ground with a mortar and pestle, using liquid nitrogen to aid in this process, until 2.0 g of ground material was obtained. Exceptions include P. munitum, P. tremuloides, T. heterophylla, A. lasiocarpa, P. engelmannii x glauca, and P. menziesii, where only between 0.5 and 1.75 g of material could be obtained. Ground samples included pieces of more than one leaf in the case of A. macrophyllum, for which individual leaves weighed more than 2.0 g. Petioles were removed before the samples were ground. Four replicates of each species were used for C:N analysis. Proportions of N and C were measured on 0.002 g replicate samples using a vario EL cube elemental analyzer (Elementar, Hanau, Germany) in August 2014. Proportion of C was divided by proportion of N in order to calculate C:N.  I measured leaf pH following the method by Cornelissen et al. (2006), which is also the method recommended in Pérez-Harguindeguy et al. (2013). For each species, I added ground 32  foliage to a 1.5-mL Eppendorf tube until the tube was filled with sample to the 0.1 mL mark. I then added 0.8 mL of distilled water to each tube to create an 8:1 volume ratio of water to sample. I inverted and tapped the tubes in order to mix the contents, placed the tubes in a tube holder, and placed the tube holder on a laboratory shaker at 1,000 rpm for one hour. I vortexed the samples for approximately one second to fully mix the sample, and then centrifuged them at 5,000 rpm for 5 minutes and again for 2 minutes using a MIKRO 120 microliter centrifuge (Hettich Lab Technology, Beverly, MA, USA). I measured the pH of the supernatant using a semi-micro pH probe (Thermo Scientific #RK-05712-34) and an Oakton pH meter (Oakton Instruments, Vernon Hills, IL, USA) that had been calibrated using 4 and 7 pH buffers. I measured pH of four replicates for each species. Proportions of WSE, NPE, ASC, and AUR were measured in oven-dried, ground samples that had been sent to the Analytical Chemistry Laboratory at the B.C. Ministry for the Environment. These measurements were made using the methods described in McClaugherty et al. (1985) and Ryan et al. (1990) and summarized by Trofymow et al. (1998). Non-polar extractables were removed from the oven-dried ground leaf litter with dichloromethane (TAPPI 1976) and water-soluble extractables were removed with hot water (TAPPI 1981). After the acid-soluble carbohydrates (ASC, e.g., cellulose and hemicellulose) were removed by sulphuric acid hydrolysis, AUR was determined by placing the remaining mass of the litter, which at this point contained acid-unhydrolyzable residue and ash, in a muffle furnace at 450 °C to remove the ash.   2.4 Trait Measurements: Leaf Litter Traits measured on litter included SLA, relative leaching loss and water uptake after 2 and 24 hours, leaf toughness, leaf thickness, pH, C:N, and proportions of WSE, NPE, ASC, and 33  AUR. These measurements were made on litter samples that had been collected between late August and early November 2014 and which were subsequently air-dried.  I measured SLA for samples collected in 2014 for all species except A. macrophyllum; measurements were made on A. macrophyllum samples collected in 2015 instead because by the time these measurements were made, no intact 2014 samples of A. macrophyllum were left. Before measuring area, I rehydrated leaves by placing them in plastic Ziploc™ bags between wet paper towels, which were then placed in a cold storage room at 4°C for 24 hours. Area and oven-dried weight were measured the same way for litter as for foliage, which is described in Section 2.3   Leaf thickness and toughness were also measured on litter samples that had been rehydrated in Ziploc™ bags between wet paper towels for 24 hours. I measured thickness for each leaf sample for which Fp and Fps were measured. Toughness was measured as described in Section 2.3, using a 1.04-mm punch-and-die attached to an Instron system. Ten samples were measured for each species.   Leaching loss rate and water uptake rate were measured after 2 and 24 hours using a modified version of the method developed by Taylor and Parkinson (1988a). Air-dried litter was oven-dried at 55°C for 48 hours prior to measurement. At time zero, I submerged 1.0 g of oven-dried litter into 300 ml of distilled water in a glass pint-sized Mason jar. Three replicates were used for each time point (i.e., three replicates for 2-hour measurements and three for 24-hour measurements) for each species. After either 2 or 24 hours, I removed the litter from the jar, blotted it between paper towels to remove excess water, and determined the wet weight. I then placed this litter in a drying oven at 55°C for 48 hours to measure the final oven-dried weight. I measured water uptake by subtracting the final wet weight from the final oven-dried weight, and 34  determined the proportion of the final wet weight contributed by absorbed water by dividing the water absorbed by the final wet weight. I calculated relative leaching loss by dividing the final oven-dry weight by the initial oven-dry weight (to obtain proportion of mass remaining after leaching) and then subtracted this value from 1.0. Fiberglass screen mesh was used to submerge samples of P. balsamifera, P. tremuloides, P. munitum, B. papyrifera, G. shallon, A. macrophyllum, and A. rubra, which otherwise floated above the surface of the water. Samples of A. rubra had some holes in the leaves; A. rubra leaves are notorious for having high levels of insect herbivory, with 45 to 100% of A. rubra leaves reported to exhibit visible signs of insect herbivory (Jackrel and Wootton 2015). Therefore, I chose the leaves with the least amount of visible damage for this measurement. I used 0.30 grams of litter and 90 mL of water for A. amabilis due to a shortage of sample remaining, and used 0.34 to 0.37 g of litter in 100 to 110 mL of water in 100 mL graduated cylinders to accommodate the length of the P. ponderosa needles used in these measurements, ensuring that all of the litter was submerged during the measurement. In both cases, the ratio of litter to water was the same as it would have been using 1.0 g of litter and 300 mL of water. The litter-to-water ratio was smaller for A. macrophyllum, as more water needed to be added to completely submerge the samples, which ranged from 0.71 to 1.66 g (one or two complete, intact leaves, minus the petiole). Leaching loss and water uptake were measured for A. macrophyllum samples collected in 2015 rather than 2014 because by the time these measurements were being made, no intact A. macrophyllum samples from 2014 were left.  I also measured pH and C:N of litter samples, following the same procedures described in Section 2.3 for green foliage. I measured litter pH in April 2015 and packed tin capsules for C:N 35  in April, which were then analyzed in June. Proportions of WSE, NPE, ASC, and AUR in leaf litter were measured as described for green foliage in Section 2.3.  2.5 Decomposition Rate Measurements I measured net mass loss during decomposition of the 16 plant species using leaf litter samples collected in autumn 2014, when samples were collected for litter trait measurements, following the popular litterbag method (Bocock 1957). Litterbags were constructed of fiberglass mesh with 1 × 1-mm holes; the dimensions of the bags varied according to the size of the litter, as recommended by Pérez-Harguindeguy et al. (2013). The smallest bags were 10 × 10 cm, which were used for litter from A. amabilis, A. lasiocarpa, L. occidentalis, P. engelmannii x glauca, P. menziesii, and T. heterophylla. Bags approximately 15 ×15 cm in size were used for litter from A. rubra, P. tremuloides, B. papyrifera, P. balsamifera, P. contorta, G. shallon, T. plicata, and P. munitum. Litterbags that were 10 × 25 cm were made for P. ponderosa, and 20 × 25-cm bags were made for A. macrophyllum.   I made 42 litterbags for each species, constructing six bags per species to be collected at each of seven time points. I made 15 extra litterbags for each A. rubra and P. menziesii for the part of the study described in Appendix 6. I filled most litterbags with approximately 1.0 g of air-dried litter per species, although A. amabilis, A. lasiocarpa, P. munitum, and L. occidentalis were filled with less (0.25, 0.5, 0.75, and 0.5 g, respectively) due to either a shortage of litter or, in the case of L. occidentalis, because the litterbags were too small to hold 1.0 g. I lined litterbags for L. occidentalis with a second layer of mesh to prevent spillage. To determine the moisture content of air-dried litter for each species, I weighed a subsample of each species before and after oven-drying it at 70°C until constant weight. The moisture content was used to estimate the 36  initial oven-dried weight from the air-dried litter used to fill the litterbags. Upon filling each litterbag, I folded the open end of each litterbag and stapled it closed with at least six staples. I attached two tags to each litterbag; one metal tag had a number and one metal tag had a species and plot identifier. I placed litterbags in Ziploc™ re-sealable plastic bags when transporting them to and from the field.   Litterbags for all species except L. occidentalis were installed in the Agroforestry Trail forest at UBC Farm on 2 December 2014, after all tree species had shed their leaves that autumn; litterbags of L. occidentalis were installed 15 December 2014, after a second layer of mesh was added to the litterbag. I placed litterbags in six plots, which were labeled A through F. All plots had T. plicata and P. menziesii trees, and most had A. macrophyllum trees either in the plot or near the plot. Seven subplots were chosen haphazardly at each site and one litterbag from each species was placed in each subplot (Figure 2.13). Subplots were less than 1 × 1 m in area. I clearly delineated the perimeter of each subplot with flags and orange tape.  Before placing the bags on the forest floor, I brushed aside the loose, uppermost layer of litter on the forest floor. Snow was present on the forest floor at the sites on 2 December, but not on 15 December 2014. I did not remove understory vegetation and avoided placing litterbags on top of moss. I placed the litterbags on the forest floor such that the folded end of the bag was facing up, and I pinned each litterbag into place using two paperclips that had been unfolded and refolded into a V-shaped piece of wire. Any litter that had fallen out of the litterbags during transport to the field (in Ziploc™ re-sealable plastic bags), particularly of the litterbags for P. menziesii, P. engelmannii x glauca, and L. occidentalis, was brought back to lab, weighed, and subtracted from the initial air-dried weight (Suffling and Smith 1974). I calculated the initial oven-dried weight of litter in each 37  litterbag by multiplying the corrected air-dried weight by the ratio of the oven-dried weight to air-dried weight of the sample previously used to determine the moisture content of that species.  After six weeks (on 13 January 2015 for all species except L. occidentalis, which was collected on 26 January 2015), I collected one subset of litterbags from each of the seven plots and immediately transported it back to the lab, where I removed debris from the outside of the bags and placed the litterbags in paper bags, which I then placed in an oven at 70°C until they reached constant weight, which was at least 72 hours. I placed litterbags that appeared to have possible mineral-soil contamination in separate paper bags so that they would not contaminate others. I then removed the contents of the litterbags, placed them on brown paper, and brushed them with a paintbrush to remove excess organic soil and debris. I weighed the cleaned litter and recorded its final oven-dried weight. This process was repeated after 3 months (5 March 2015, with L. occidentalis on 19 March), 6 months (2 June 2015, with L. occidentalis on 15 June), 9 months (2 September 2015, with L. occidentalis on 15 September), and 12 months (2 December 2015, L. occidentalis samples were collected at this time as well). To estimate how much mass would be remaining after 365 days for L. occidentalis, which I collected after 352 instead of 365, I calculated the mean daily percent mass loss from the 9-month samples to the 352-day samples at each plot and then subtracted 13 times that amount to the values for mass remaining values after 352 days. Litterbags for the two remaining time points (24 months and 36 months) are not used in this thesis.      I calculated the proportion of mass remaining for each litterbag by dividing the final oven-dried weight by the initial oven-dried weight. After observing that mass loss seemed to occur rapidly at first and then level off in many species, I decided to divide mass loss into two phases, one lasting from 0 to 3 months (Phase I) and one from 3 to 12 months (Phase II). To 38  calculate Phase-I mass loss, I subtracted the proportion of mass remaining from each litterbag sample collected after 3 months from 1. To calculate Phase-II mass loss, I subtracted the proportion of mass remaining from each litterbag sample collected after 12 months from the proportion of mass remaining from each litterbag of the same species at the same plot collected after 3 months. This distinction between Phase I and Phase II was intended to separate the phase of early decomposition dominated by leaching from the phase of early decomposition dominated by microbial catabolism (Swift et al. 1979). A. lasiocarpa and P. engelmannii x glauca were found to have lost much more mass after 6 weeks than expected due to leaching after 24 hours (Figure 2.14) and were also small enough to pass through the 1 × 1-mm mesh; I concluded that some of the needles fell out of the litterbags during the study and therefore removed these species from the analysis.  2.6 Data Analysis To determine and visualize associations among foliar traits and among leaf-litter traits, I performed a principal component analysis (PCA) on the mean values of each trait for each species, excluding A. lasiocarpa and P. engelmannii x glauca. I performed separate PCAs for foliar traits and for litter traits. The foliar traits included in the first PCA were SLA, LDMC, leaf thickness, force-to-punch (Fp), specific force-to-punch (Fps), abaxial cuticle thickness, adaxial cuticle thickness, abaxial DTL, adaxial DTL, pH, C:N, proportions of N and C, and proportions of WSE, NPE, ASC, and AUR. The litter traits included in the PCA were SLA, leaf thickness, Fp, Fps, leaching loss after 2 and 24 hours, water uptake after 2 and 24 hours, pH, C:N, proportions of N and C, and proportions of WSE, NPE, ASC, and AUR. These PCA analyses were then repeated without P. munitum to quantify these relationships with only woody plants. 39  PCA analyses were performed in R version 3.0.2 (2013) using the FactoMineR package (Lê et al. 2008).  To see how similar species were to each other based on the traits that were measured in both foliage and litter of the 13 woody plant species, I performed a cluster analysis using Euclidean distance and standardized variables. These variables include SLA, leaf thickness, Fps, pH, C:N, proportions of N and C, and proportions of WSE, NPE, ASC, and AUR. I used the complete linkage method, which means that during the hierarchical clustering process, the smallest maximum distance between the forming cluster and individual points to be clustered (in this case, individual species) is used to determine which species should be added to the tree diagram next (Manly 2004). I performed one cluster analysis using green foliar traits and one using leaf-litter traits. Cluster analyses were performed in R version 3.0.2 (2013).  To see how relationships among species based on traits changed from foliage to litter and thereby better understand how traits change from foliage to litter, PCA analyses with only woody plants were done using only the traits that were measured in both foliage and litter, which are the same traits used in the two cluster analyses. I then graphed the relationship between values of each trait in foliage and litter, graphing the mean and, when applicable, creating double boxplots using the boxplotdbl package in R to see how the variation changed in foliage and litter (Tomizono 2013). I also calculated the correlation between the species mean values of each of these traits in foliage and litter, using a natural logarithm transformation for leaf thickness, proportion of N, and proportion of NPE to meet assumptions of normality, and tested for the significance of these correlations using the “rcorr” function in the R package Hmisc (Harrell 2015).  I then used pair-wise t-tests to determine how traits changed from foliage to litter, whether or not they changed in one direction. To correct for the number of traits tested, I applied 40  a Bonferroni correction to the probability level (∝, 0.05) for both the correlation tests and the paired t-tests; because 11 traits were tested, the probability level was divided by 11, resulting in adjusted probability level (∝’) value of 0.0045. If the p-value of the paired t-test was below 0.0045, the trait was found to differ significantly in one direction when measured in foliage and litter. Tests in which p-values were below 0.05 but greater than 0.0045 were considered marginally significant.  To quantify relationships between traits and mass loss for each phase, and for the entire study period, I used a combination of PCA, regression tree analysis, and linear regression. Prior to analyses, I removed the A. macrophyllum sample from plot A because the proportion of mass lost after 12 months was a large outlier due to the large amount of soil on the litter that could not be removed without damaging the sample. To the PCA analyses described above, I added proportion of mass lost during Phase I, Phase II, and over one year as supplementary variables, meaning that I would be able to see how they correlated with the trait variables without changing the ordination and the relationships between the trait variables. In doing so, I used the function “imputePCA” function in the missMDA R package (Husson and Josse 2016) to impute the missing value of mass loss for A. macrophyllum at plot A in order to maintain a balanced dataset for these analyses. These analyses helped me to see which traits were more strongly correlated, both positively and negatively, with mass loss in each phase. The first principal component variable (PC1) produced in all analyses (which explained most of the variation between species) reflected strong relationships between traits that would be expected under the leaf economic spectrum hypothesis (Wright et al. 2004) and were used in subsequent linear regression models to predict Phase I, Phase II, and year-one mass loss. 41   I also used regression tree analysis to identify influential trait variables and build models to predict mass loss, comparing these results to the results of the PCA analyses described earlier. Regression tree analysis, often referred to as CART (classification and regression tree analysis), is a nonparametric method that is gaining popularity in ecological studies because it is based on the same principles as least squares regression, but does not require the data to be normally distributed, can accommodate interactions in the predictor variables, and can account for missing values (De’ath and Fabricius 2000).  Regression tree analysis can both predict values of a response variable as well as convey the relative importance of the predictor variables; the “importance value” of a variable relates to the number of splits in the resulting dichotomous trees that the variable appears in when the tree-making algorithm runs (Breiman et al. 1984; Zuur et al. 2007).  Untransformed mean measurements of each trait variable, both foliar and litter traits, were used to predict the proportion of mass lost during each phase and over the study period. The experimental units were each species at each plot (e.g., A. macrophyllum at plot B; n = 64). Tests were performed in R version 3.0.2 (2013) using the rpart package (Therneau et al. 2013), with the complexity parameter (cp, or essentially the minimum amount by which the split can increase the R2 of the model) set to 0.05, instead of the default 0.01, to reduce the sensitivity of the model to small amounts of variation explained. In addition to performing regression tree analyses with all of the variables, I performed two additional regression tree analyses for each of the three response variables – one using only physical traits and the other using only chemical traits.  Lastly, I ran simple linear regressions with all trait variables to predict the proportions of mass lost in Phase I, Phase II, and over the study period, transforming the proportions of mass lost by multiplying by 100 (thereby calculating percentage) and then taking the natural logarithm of this percentage to more closely meet assumptions of normality and equal variance, and 42  transforming trait variables when appropriate. The experimental units were each species at each plot (e.g., A. macrophyllum at plot B; n = 64). Analyses were performed in R using the “lm” function. To correct for the number of F-tests performed, I applied a Bonferroni correction to the probability level (∝, 0.05); because 36 tests were run for each response variable (all traits and the four principal component variables, one PC1 variable for each PCA performed using only woody plants), the probability level used to indicate significance was 0.0014. The coefficient, R2, and p-value for all tests are reported in Appendix 5.    43    Figure 2.1. Daily Rainfall Totals During the Study Period. Daily rainfall (mm) from 2 December 2014 through 2 December 2015. The dashed line at 5 March 2015 represents the end of Phase I and beginning of Phase II for the litterbag study described in section 2.5.  Data were obtained from Environment Canada (2015) and were collected at the Vancouver International Airport.  44   Figure 2.2. Total Rainfall from 2004 through 2015. Total rainfall (mm) from 2 December to 2 December of the following year in Vancouver, BC, beginning in 2004. The red bar on the far right indicates rainfall during the study period (2 December 2014 through 2 December 2015). The mean rainfall from 2 December through the following 2 December from 2004 through 2015 is 1115.3 mm. Data were obtained from Environment Canada (2015) and were collected at the Vancouver International Airport. 45  Table 2.1. Horizon Data from Soil Pit. Select morphological and chemical properties of a shallow soil pit dug in the Agroforestry Trail forest1 at UBC Farm on 10 August 2015.   Horizons Depth (cm) Color Texture Consistence Grade Bulk Density (g cm-3) pH LFH 5-0  -- -- -- -- -- -- Ahe 0-5 7.5 YR 5/3 Loamy sand Very friable Weak 0.52 4.4 Bf 5-35 7.5 YR 5/6 Loamy sand Very friable Weak 0.90 5.2 1Soil pit was dug in plot E      Table 2.2 Site Description Data for Each Plot. Site position and description1 of each of the six plots established in the Agroforestry Trail forest at UBC Farm.  Site Elevation (m) Aspect Slope Position Slope  Successional Status Canopy Cover Site Series SMR2 SNR3 Plot A 82 220 (SW) Upper slope 5 (9%) Stem exclusion 90%    Plot B 96 275 (W) Upper slope 6 (10%) Understory reinitiation 50% 01 (Hwd - Kindbergia) Slightly dry and fresh Rich Plot C 79 230 (SW) Upper slope 7 (11%) Understory reinitiation 70% 07  (Cw - Foamflower) Very moist Rich Plot D 78 187 (S) Midslope 8 (13%) Understory reinitiation 80%  Unknown Rich Plot E 69 328 (NW) Midslope 4 (8%) Understory reinitiation 70% 06 (HwCw - Deer Fern) Moist Medium Plot F 70 176 (S) Flat 4 (7%) Understory reinitiation 60% 07  (Cw - Foamflower) Moist Rich 1 Kirsten Corrao collected most of this data and described the plots 2 Soil moisture regime 3 Soil nutrient regime  46  Table 2.3. Vegetation Composition in Each Plot. Vegetation data1 for each of the six plots established in the Agroforestry Trail forest at UBC Farm.  Site Dominant  Codominant  Subcanopy  Shrubs H2 S3 Notes Plot A T. heterophylla T. plicata T. plicata -- 0 0 Large P. menziesii tree just outside of plot Plot B T. plicata -- -- D. expansa (50%) P. munitum (80%) H. helix (1%) I. aquifolium (1%) 0.75 0.49 Close to road with many A. rubra trees Plot C T. plicata  T. plicata P. menziesii -- P. munitum (80%) R. parviflorus (5%) R. spectabilis (10%) D. expansa (5%) T. latifolia (1%) I. aquifolium (1%) 0.80 0.37 Possibly influenced by proximity to farm and road, as inferred by presence of oak seedling  Plot D -- -- T. plicata T. plicata (40%) P. munitum (40%) G. shallon (50%) R. spectabilis (5%) 1.21 0.69 two dominant P. menziesii trees just outside of established plot Plot E T. plicata  T. plicata  -- D. expansa (70%) G. shallon (20%) P. munitum (5%) R. spectabilis (30%) 1.09 0.60  Plot F P. menziesii  -- A. macrophyllum T. plicata  P. munitum (40%) G. shallon (40%) D. expansa (20%) R. spectabilis (10%) 1.26 0.69 Numerous T. plicata saplings outside of plot  1 Data collected by Kirsten Corrao 2 Shannon Diversity Index of shrubs 3 Simpson’s Diversity Index of shrubs   47  Table 2.4. Forest Floor Descriptions for Each Plot. Description of the forest floor (LFH horizons) in each of the six plots established in the Agroforestry Trail forest at UBC Farm.   Site Humus form Depth (cm) Forest Floor pH1 Forest Floor C:N2 Comments Plot A Mor or Moder 8-0 4.19 ± 0.05 24.09 ± 0.22 charcoal Plot B Mor or Moder 8-0 4.68 ± 0.01 27.76 ± 0.90  Plot C Mor or Moder 10-0 4.58 ± 0.04 24.76 ± 0.65  Plot D Mor or Moder 3-0 4.49 ± 0.01 24.06 ± 0.16  Plot E Mor or Moder 9-0 4.15 ± 0.08 24.05 ± 0.16 charcoal Plot F Mor or Moder 12-0 4.53 ± 0.03 23.39 ± 0.29 charcoal, many invertebrates (spiders, insects) 1 Measured from composite of five samples in laboratory by Kirsten Corrao 2 Measured from composite of five samples with elemental analyzer             48      Figure 2.3. Study Plots at UBC Farm. The six plots near the Agroforestry Trail at UBC Farm in which litterbags were installed on 2 December 2014. Images were taken in December 2014. Top row from left to right: plots A, B, and C. Bottom row from left to right: plots D, E, and F. The large canopy gap in plot F near the large A. macrophyllum tree is evident in the background of this image. 49      Figure 2.4. P. menziesii Control Images. Unedited images of unstained (left) and nile red stained (right) sections of P. menziesii needle cross-sections; the adaxial cuticle (top row) and abaxial cuticle (bottom row) were examined. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and an exposure time of 24.8 ms.  50      Figure 2.5. P. munitum Control Images. Unedited images of unstained (left) and nile red stained (right) sections of P. munitum leaf cross-sections; the adaxial cuticle (top row) and abaxial cuticle (bottom row) were examined. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and an exposure time of 24.8 ms for the adaxial cuticle and 89.5 ms for the abaxial cuticle.  51       Figure 2.6. Pinus spp. Cuticle Images. Adaxial (left) and abaxial (right) sides of P. ponderosa needle (top) and P. contorta var. latifolia needle (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 52      Figure 2.7. G. shallon and P. munitum Cuticle Images. Adaxial (left) and abaxial (right) sides of G. shallon leaf (top) and P. munitum leaflet (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 53      Figure 2.8. P. tremuloides and P. balsamifera Cuticle Images. Adaxial (left) and abaxial (right) sides of P. tremuloides leaf (top) and P. balsamifera var. trichocarpa leaf (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 54      Figure 2.9. A. amabilis and P. menziesii Cuticle Images. Adaxial (left) and abaxial (right) sides of A. amabilis needle (top) and P. menziesii needle (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscopy with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 55      Figure 2.10. B. papyrifera and L. occidentalis Cuticle Images. Adaxial (left) and abaxial (right) sides of B. papyrifera leaf (top) and L. occidentalis needle (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 56      Figure 2.11. A. macrophyllum and A. rubra Cuticle Images. Adaxial (left) and abaxial (right) sides of A. macrophyllum leaf (top) and A. rubra leaf (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 57       Figure 2.12. T. plicata and T. heterophylla Cuticle Images. Adaxial (left) and abaxial (right) sides of T. plicata scale (top) and T. heterophylla needle (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscope with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 58   Figure 2.13. Litterbag Station. One of 42 litterbag stations installed near the Agroforestry Trail at UBC Farm on 2 December 2014. Each litterbag station had 16 litterbags: one bag of litter from each species.  59   Figure 2.14. Leaching Loss After 24 Hours and Mass Loss After 1.5 Months. Relationship between proportion of mass lost after soaking in distilled water for 24 hours and proportion of mass lost after 1.5 months in the field, with a least-squares linear regression line fit through the 16 points. A. lasiocarpa (circled in blue) and P. engelmannii x. glauca (circled in red) were excluded from further analyses due to the much greater proportion of mass lost in the field relative to the proportion of mass lost due to leaching.  60 Chapter 3: Results 3.1  Relationships Among Foliar Traits and Among Litter Traits I examined the relationships among the measured functional traits of green foliage (hereafter called “foliar traits”) and among the measured traits of leaf litter (hereafter called “litter traits”). For foliage, traits related to the first principal component variable (PC1) produced in the principal components analysis (PCA) performed with all of the measured foliar traits were mostly physical functional traits. PC1 was strongly, positively correlated with leaf dry matter content (LDMC), toughness expressed as force-to-punch (Fp), C:N, leaf thickness, adaxial and abaxial cuticle thickness, and adaxial and abaxial distance to lumen (DTL), and negatively correlated with specific leaf area (SLA) and proportion of N (Table 3.1, Figure 3.1). PC1 explained 44% of the variance in these foliar trait measurements (Table 3.1). The second principal component variable (PC2), which explained an addition 18% of the variance, was strongly correlated with chemical traits. PC2 was strongly positively correlated with pH and proportion of acid-soluble carbohydrates (ASC) such as hemicellulose and cellulose, and was negatively correlated with proportion of C (Table 3.1, Figure 3.1).  For litter, the traits strongly correlated with PC1 of the PCA performed with all of the litter traits measured in this study, which explained 45% of variance in these litter traits, included a mix of physical and chemical traits. Traits strongly and positively correlated with PC1 include SLA, proportion of water-soluble extractables (WSE), leaching loss after 2 and 24 hours, and water uptake after 2 and 24 hours (Table 3.2, Figure 3.2). Toughness expressed as both Fp and specific force-to-punch (Fps), leaf thickness, and ASC were negatively correlated with PC1 (Table 3.2, Figure 3.2). PC2, explaining 17% of the variation in litter traits, was strongly  61 correlated with chemical traits, positively correlated with pH, proportion of acid-unhydrolyzable residue (AUR), and ASC, and negatively correlated with C:N and WSE (Table 3.2, Figure 3.2).    When performing PCA using only the 12 traits that were measured in both foliage and litter, PC1 and PC2 differed between functional traits and litter traits in terms of the traits with which they were more strongly correlated. In the PCA using only foliar traits, PC1 explained 42% of the variance in foliar traits among species and was strongly correlated with both physical and chemical traits, positively correlated with Fp, leaf thickness, C, and C:N, and negatively correlated with SLA, N, and pH (Table 3.3, Figure 3.3). In comparison, PC1 for litter traits explained 36% of the variance, strongly and positively correlated with Fp, Fps, leaf thickness, and ASC, and strongly and negatively correlated with SLA and WSE (Table 3.4, Figure 3.4). PC2 for foliar traits explained 22% of the variance and was strongly and positively correlated with Fps, ASC, and pH, negatively correlated with C and proportion of non-polar extractables (NPE; Table 3.3, Figure 3.3). PC2 for litter traits explained 23% of the variance and was strongly and positively correlated with pH, ASC, and AUR, and negatively correlated with C:N and WSE (Table 3.4., Figure 3.4).    3.2 Trait-Based Relationships Among Species A hierarchical tree diagram produced from a cluster analysis using the 11 foliar traits that were also measured in litter (Fp was excluded) divided the 13 woody plant species into two main groups: one group included species that produce needles and the other included all other species (Figure 3.5). Of the needle-leaf species, T. heterophylla diverged most from the other five species. Of the non-needle-leaf species, G. shallon and T. plicata were more similar to each other  62 than they were to the broadleaf species. In general, groupings largely reflected similarity in leaf form.  The trait-based relationships among species differed when litter traits were used instead of foliar traits, as shown in the tree produced by cluster analysis of corresponding litter traits (Figure 3.6). Unlike in the tree produced using foliar traits, there is no clear distinction between the needle-leaf species and the other species. A. rubra was least similar to any of the other species. B. papyrifera and L. occidentalis, a broadleaf species and a needle-leaf species, were grouped together, as were needle-leaf P. menziesii and broadleaf G. shallon. The other groupings in this tree were of needle-leaf species (with coniferous T. plicata) and of broadleaf species, as in the tree produced using functional traits, but this tree produced more groupings of species with diverse leaf forms.     3.3 Comparison of Traits Between Foliage and Litter Many of the 11 traits that were measured in both foliage and litter showed a positive correlation between mean measurement in foliage and mean measurement in litter for the 13 woody plant species used in this study. Only one of these correlations, however, was statistically significant (leaf thickness, ln-transformed, p < 0.001, Table 3.5). Many of the others were strong (r < 0.5) and marginally significant (0.05 > p > 0.0045, Table 3.5), with the exception of proportion of N (ln-transformed), C:N, WSE, and NPE (ln-transformed). Of these latter traits, N, C:N, and WSE changed uni-directionally from foliage to litter; specifically, N and WSE decreased and C:N increased (Table 3.6). Regardless of the proportion of N found in foliage, the proportion of N in litter was always at a threshold value between 0.005 to 0.015, excluding the N-fixing species A. rubra, which had the highest proportion of N in both foliage and litter  63 (Figure 3.7). Due to this, and an apparent lack of relationship between proportion of C in foliage and in litter, C:N did not seem to exhibit a positive relationship between values in foliage and litter. All C:N values, except for those of T. plicata, were greater in litter than in foliage; proportions of C and N as well as C:N in T. plicata did not differ between foliage and litter (Figure 3.7). Similar to N, WSE decreased from foliage to litter, which was accompanied by increases in ASC and AUR (Figure 3.8). NPE increased in many species, but decreased in two species, including T. heterophylla, which had a very large proportion of NPE in foliage compared to that of other species (proportion of NPE = 0.28 in foliage; Figure 3.8). Despite positive associations, if only marginally significant, between foliar and litter values for the remaining traits, individual species varied in the direction of change when the relationship between foliage and litter traits strayed from a 1:1 ratio. For example, SLA was greater in foliage than in litter for P. tremuloides, B. papyrifera, P. munitum, and T. plicata, but greater in litter than foliage for L. occidentalis and A. macrophyllum; no significant difference was detected in the remaining species, which tended to have lower mean SLA values (Figure 3.9). Similarly, thickness measurements were mostly comparable in foliage and litter, except for P. contorta, T. plicata, and P. ponderosa, for which litter was on average thicker than foliage (Figure 3.9). Fps and pH exhibited similar patterns in that some species had greater values in litter and some in foliage, but the values mostly did not greatly differ between foliage and litter (Figure 3.10). When considering the amount of variation in individual measurements of Fps, only P. ponderosa and P. menziesii had tougher litter than foliage; the distributions of values did not differ greatly otherwise, and there was quite a bit of overlap between species as well (Figure 3.10). pH was measured in composite samples, meaning that distributions do not represent differences in individual samples but instead variability in the measurement of the same  64 composite sample; therefore, it is not possible to compare the distribution of pH values for individual samples in this thesis. The two species that differed most in pH between foliage and litter, however, were P. menziesii, which had a higher pH in litter than in foliage, and T. plicata, which had a higher pH in foliage than in litter (Figure 3.10).  3.4 Relationships Between Traits and Mass Loss Over Time Foliar traits and litter traits were used to predict the proportion of mass lost during decomposition during the first year and also during two distinct phases of decomposition during the first year. Following an examination of trends in mass loss, three statistical techniques – PCA, regression tree analysis, and simple linear regression – were used to identify suites of traits that best explained variation in mass loss over this time.   3.4.1 Trends in Mass Loss During the First Year of Decomposition  Mass loss occurred quickly for many species within the first three months, and then slowed over time, with P. munitum decomposing most slowly (mean proportion lost = 0.046 ± 0.021 standard error after 12 months) and B. papyrifera most quickly (mean proportion lost = 0.73 ± 0.084 standard error after 12 months; Figure 3.11). Mass loss during this initial period, which I will refer to as Phase I (Figure 3.11), was also strongly correlated with the proportion of litter mass lost by leaching after soaking for 24 hours in distilled water (r = 0.74, n = 64). I will refer to the proportion of mass lost from 3 to 12 months as Phase II of decomposition (Figure 3.11).  The 14 species studied varied in the proportion of mass lost in Phase I and in Phase II (Figure 3.12). Eight species lost a greater proportion of initial mass loss on average during Phase  65 I of decomposition than during Phase II, even though Phase I was one-third as long as Phase II. These species include P. munitum, A. macrophyllum, A. amabilis, T. heterophylla, T. plicata, P. balsamifera, L. occidentalis, and A. rubra. All others lost more mass from 3 to 12 months after litterbag installation than during the first 3 months of decomposition.    While those species that lost more mass by leaching tended to lose more mass in Phase I (e.g., P. tremuloides, B. papyrifera, L. occidentalis, P. balsamifera, and A. rubra), there were several exceptions; A. macrophyllum, for example, lost less mass in Phase I than might be expected given how much had leached after 24 hours in a laboratory setting, while P. contorta and T. plicata lost more mass than expected from their leaching loss (Figure 3.13). Therefore, mass was probably lost during Phase I as a result of other processes as well, such as catabolism and comminution, rather than solely leaching. Although mass loss during the time period from 0 to 1.5 months was also strongly correlated with leaching loss after 24 hours (r = 0.70, n = 64), there were some species for which the proportion of mass lost during the first 1.5 months of the study was less than the proportion that had leached over 24 hours in a laboratory (Figure 2.14). Therefore, for those species, leaching was probably not complete after 1.5 months.   3.4.2 Traits That Co-Varied with Mass Loss During the First Year of Decomposition  During the first year, the foliar traits most strongly correlated with the proportion of mass lost, as determined by PCA, were SLA and leaf thickness; SLA was positively correlated with year-one mass loss, and thickness was negatively correlated. These relationships were visible when using all foliar traits and when using the subset of foliar traits that were measured in both foliage and litter (Figures 3.1 and 3.3). Proportion of N was also positively correlated with SLA and year-one mass loss, though not as strongly, and Fp, C:N, and LDMC were all also negatively  66 correlated with mass loss, and positively correlated with thickness (Figures 3.1 and 3.3). Litter traits that positively correlated with year-one mass loss include leaching loss after 2 hours and after 24 hours; Fps and C were negatively correlated with year-one mass loss (Figure 3.2). When referring to the PCA performed using the subset of litter traits measured in foliage and litter, year-one mass loss is most strongly correlated with SLA, and most negatively correlated with Fps, C, and ASC (Figure 3.4).  PC1 in all PCAs performed with the 13 woody plant species in this study explained 36 to 45% of the variation in traits among these species. PC1 was found in all cases to correlate with year-one mass loss better than PC2, and this correlation was slightly higher when using foliar traits than when using litter traits (r =  -0.68 using foliar traits and -0.57 when using litter traits, Tables 3.3 and 3.4).  Consistent with the PCA performed using all measured litter traits, leaching loss after 24 hours was the strongest predictor of year-one mass loss in regression tree analysis performed using all traits (Figure 3.14). Once split into two groups, foliar adaxial and abaxial DTL were used to further predict year-one mass loss, though their contribution to explaining the variation in year-one mass loss was much smaller in comparison, as reflected in the much shorter branch lengths (Figure 3.14). In particular, of the species that lost less than a proportion of 0.097 of mass due to leaching after 24 hours in distilled water, those with a foliar adaxial DTL less than 7.3 µm lost less mass over one year (mean proportion of initial mass = 0.27) than those with a greater DTL (mean proportion of initial mass = 0.38, Figure 3.14). Conversely, of the species that lost more than a proportion of 0.097 of mass due to leaching, those with a larger abaxial DTL lost less mass (mean proportion of initial mass = 0.50) than those with a smaller abaxial DTL (mean proportion of initial mass = 0.64, Figure 3.14). Variables other than leaching loss  67 and abaxial DTL that were important, as indicated by the number of splits they contributed to in the regression tree algorithm, include SLA, C:N, and N of foliage, as well as litter water uptake after 24 hours (Table 3.7). Adaxial DTL, toughness (Fps) and litter NPE were less important (Table 3.7). The physical foliar traits and litter traits were able to produce a regression tree that explained slightly more variation in year-one mass loss than the chemical foliar traits and litter traits measured in this study, but the difference in variation explained between them was small (68% as opposed to 64%; Figure 3.15). The regression tree produced using only physical traits was identical to the one produced using all traits, with SLA, LDMC, and leaf thickness in foliage as well as litter water uptake after 24 hours indicated as important (Table 3.8). The tree produced using only chemical traits used foliar C:N to primarily divide the samples into groups based on year-one mass loss; litter C was used to further divide the samples with high C:N ratios in foliage (Figure 3.15). Foliar C:N and N were the most important chemical traits explaining year-one mass loss, followed by foliar AUR and litter C:N, pH, and NPE (Table 3.9).  In simple linear regression, foliar LDMC explained most of the variation in ln-transformed percent mass lost in year one (R2 = 53%, Table 3.10). Foliar SLA explained nearly as much variation (R2 = 48%) as did litter SLA, but only when A. macrophyllum was omitted; otherwise, litter SLA accounted for 13.16% of variation (Table 3.10). Other trait variables that produced significant linear regressions were leaching loss after 24 hours (R2 = 42%) and water uptake after 24 hours (R2 = 21%), foliar N (ln-transformed proportion, R2 = 38% including A. rubra), foliar C:N (R2 = 34%), litter toughness (Fps, ln-transformed, R2 = 34%), litter AUR (R2 = 29%), and litter pH (R2 = 20%, Table 3.10).    68  The important trait predictors of year-one mass loss that were consistently found in all three statistical methods were a) foliar SLA, LDMC, C:N, and N, and b) litter Fps and leaching loss after 24 hours, with foliar LDMC being strongest in linear regression and litter leaching loss after 24 hours and foliar C:N being strongest in regression tree analyses. Foliar leaf thickness and Fps, as well as litter pH, AUR, C, and water uptake after 24 hours were important using two of the three methods, as well as litter SLA when A. macrophyllum samples were removed in linear regression. Foliar abaxial DTL, adaxial DTL, abaxial cuticle thickness, and pH were only important in one method, as were litter ASC, NPE, and leaching loss after 2 hours.  3.4.3 Traits That Co-Varied with Phase-I Mass Loss  The traits that best correlated with the proportion of mass lost during Phase I slightly differed from those correlating with year-one mass loss. In PCA using foliar traits, Phase-I mass loss was most strongly, positively correlated with N and negatively correlated with C:N (Figures 3.1 and 3.3). Leaf thickness, Fp, and abaxial DTL were also negatively correlated, but not as strongly, and WSE was more strongly, positively correlated with Phase-I mass loss when only the subset of traits measured in both foliage and litter was used in the PCA (Figures 3.1 and 3.3). Phase-I mass loss was also slightly positively correlated with SLA in foliage (Figure 3.1). Like year-one loss, Phase-I mass loss was also strongly correlated with leaching loss after 2 and 24 hours, but it was more strongly correlated with litter WSE (Figures 3.2 and 3.4). Phase-I mass loss was also negatively correlated with litter ASC (Figures 3.2 and 3.4).  As with year-one loss, PC1 variables explained Phase-I mass loss better than PC2 variables. However, Phase-I mass loss was more strongly correlated with the PC1 variables  69 created from PCAs using litter traits than foliar traits (r =  -0.62 using foliar traits and -0.71 using litter traits; Tables 3.3 and 3.4).  Accordingly, the regression tree that explained the most variation in Phase-I mass loss, when considering all foliar traits and litter traits, used only litter WSE, explaining 71% of variation in Phase-I mass loss (Figure 3.16). Other variables of importance include litter traits, such as proportion of AUR, Fps, and leaching loss after 24 hours, as well as abaxial and adaxial cuticle thickness in foliage (Table 3.11).  As expected from the previous regression tree analysis, the best tree produced using only chemical traits used only litter WSE to explain Phase-I mass loss (Figure 3.17). When using only physical traits to predict Phase-I mass loss, leaching loss after 24 hours best explained variation in Phase-I mass loss, with foliar adaxial cuticle thickness further explaining variation among species that lost less than a proportion of 0.097 of mass due to leaching in distilled water for 24 hours (Figure 3.17). Specifically, species with smaller adaxial cuticles lost a smaller proportion of initial mass than those with thicker cuticles (Figure 3.17). Both trees explained a similar proportion of variation in Phase-I mass loss; the physical tree explained 70% of variation, and the chemical tree explained 71%. Abaxial DTL, LDMC, SLA, and thickness in foliage, as well as water uptake in litter after 24 hours, were other important traits in the regression tree analysis of Phase-I mass loss using only physical traits, while additional important traits in the regression tree analysis explaining Phase-I mass loss using only chemical traits include litter AUR, ASC, and C as well as foliar C and N to a lesser extent (Tables 3.12 and 3.13).  Of the 32 trait variables included in the study, the natural logarithm of Phase-I mass loss (which was multiplied by 100) was predicted well in simple linear regressions using 15 traits, including 9 litter traits and 6 foliar traits (Table 3.14). All litter leaching loss and water uptake  70 traits produced significant linear regressions, with litter leaching loss after 24 hours accounting for the most variation in Phase-I loss (R2 = 51%). Chemical litter traits, specifically WSE, AUR, and ASC, produced significant regressions as well, with AUR accounting for the most variation out of the three trait variables (R2 = 43%), followed closely by litter WSE (42%). Litter SLA produced a significant regression, but only if A. macrophyllum was omitted from the analysis (R2 = 43%, Table 3.14). Foliar LDMC, SLA, and N (ln-transformed proportion) also accounted for significant variation in Phase-I mass loss (R2 = 39%, 33%, and 33% respectively; Table 3.14). Other traits that produced significant linear regressions to explain Phase-I mass loss include AUR in foliage (R2 = 29%), Fps in foliage (ln-transformed, R2 = 25%), litter C (R2 = 21%), and foliar pH (R2 = 16%, Table 3.14).   The important predictors of Phase-I mass loss indicated by all three methods include litter WSE, ASC, and leaching loss after 24 hours, as well as the proportion of N in foliage, with litter leaching loss after 24 hours being the strongest predictor in linear regression and with litter WSE and leaching loss after 24 hours being the strongest predictor in regression tree analysis. Foliar leaf thickness, SLA, pH, AUR, LDMC, and abaxial DTL were important in two out of the three methods, as well as litter leaching loss after 2 hours and water uptake after 24 hours. Abaxial and adaxial cuticle thickness, adaxial DTL, ASC, C, C:N, and Fps in foliage were only important in one of the three methods, as well as litter C and water uptake after 2 hours.   3.4.4 Traits That Co-Varied with Phase-II Mass Loss   The foliar traits positively correlated with the proportion of mass lost from 3 to 12 months of decomposition, or Phase II, include SLA and pH, although these relationships are not particularly strong (Figures 3.1 and 3.3). Foliar N was positively correlated with Phase-II mass  71 loss, but less strongly correlated than with Phase-I loss and year-one loss (Figures 3.1 and 3.3). Foliar leaf thickness and LDMC were most strongly, negatively correlated with Phase-II mass loss (Figures 3.1 and 3.3). The litter traits most strongly, positively correlated with Phase-II mass loss include SLA, N, water uptake after 2 and 24 hours, and to a lesser extent leaching loss after 24 hours; litter traits negatively correlated with Phase-II mass loss include toughness (as Fp and Fps) and thickness (Figures 3.2 and 3.4). Litter C:N was also negatively correlated with Phase-II mass loss, though not as strongly (Figures 3.2 and 3.4).   As with year-one and Phase-I mass loss, PC1 variables explained Phase-II mass loss better than PC2 variables. This correlation between Phase-II mass loss and PC1 was slightly higher when using foliar traits than when using litter traits (r =  -0.50 using foliar traits and -0.29 when using litter traits, Tables 3.3 and 3.4).  Regression tree analyses to explain Phase-II mass loss explained less variation than those for year-one and Phase-I mass loss, explaining between 35% and 44% of variation in mass loss. The best tree produced in regression tree analysis using all foliar traits and litter traits explained 44% of variation in Phase-II mass loss and divided the samples first by foliar LDMC, with those having a mean LDMC of 363 mg g-1 or greater losing less mass (mean proportion of initial mass = 0.17) than those with a smaller LDMC (mean proportion of initial mass = 0.36; Figure 3.18). Of those samples with greater LDMC values, those with a lower initial litter pH lost less mass (mean proportion of initial mass = 0.14) than those with higher pH values (mean proportion of initial mass = 0.22; Figure 3.18). The most important variables were foliar traits, including LDMC, SLA, Fps, N, and leaf thickness. Litter thickness was important as well, with pH, C:N, and N in leaf litter important to a lesser extent (Table 3.15).    72  The tree explaining Phase-II mass loss produced in regression tree analysis using only chemical traits explained slightly more variation than the tree produced using only physical traits, but both trees explained less than half of the variation in Phase-II mass loss (Figure 3.19). The tree produced using only physical traits, splitting the samples into two groups based on foliar LDMC values, accounted for 35% of variation in Phase-II mass loss. The tree produced using only chemical traits accounted for 40% of variation in Phase-II mass loss, first splitting the samples by foliar N, and then further splitting the samples with low proportion of N (< 0.026) into two groups based on litter pH (Figure 3.19). In particular, samples with greater proportions of N in foliage lost more mass during Phase II (mean proportion of initial mass = 0.33), and of those with lower proportions, those with higher litter pH values (> 5.5) lost more mass (mean proportion of initial mass = 0.22) than those with lower values (mean proportion of initial mass = 0.14; Figure 3.19). The most important physical traits were SLA, LDMC, leaf thickness, and Fps in foliage, as well as water uptake after 2 hours and thickness in litter (Table 3.16). The most important chemical traits were C:N and proportion of N in foliage, followed by litter NPE, C:N, and N, and then pH in both foliage and litter (Table 3.17).  As in regression tree analysis, the trait that best predicted Phase-II mass loss (ln-transformed) in simple linear regression was foliar LDMC (R2 = 24%, Table 3.18). Several variables identified as important in PCA and regression tree analysis also produced significant regressions to predict Phase-II loss, including foliar SLA (R2 = 23%), pH of litter (R2 = 21%) and foliage (R2 = 19%), and foliar N (ln-transformed, R2 = 16% when including A. rubra). Linear regressions using PC1 of foliage were also significant, accounting for 16% and 19% of variation in Phase-II mass loss; linear regressions using PC1 of litter were not significant (Table 3.18, Table A5-3).   73  The traits variables predicting Phase-II loss in all three methods include N, SLA, pH, and LDMC of foliage, with foliar LDMC being the strongest predictor in linear regression and foliar LDMC and N being the strongest in regression tree analyses. Foliar leaf thickness as well as litter SLA, C:N, N, pH, water uptake after 2 hours, and leaching loss after 24 hours were important in two out of the three methods. Several traits were only important in one method; these include litter Fps and water uptake after 24 hours, which were only important in PCA, and foliar Fps, C, C:N, and abaxial and adaxial cuticle thickness, as well as litter NPE, WSE, and ASC, which were only important in regression tree analysis.                 74 Table 3.1. Correlations Between All Foliar Traits and Two Principal Components. Correlations (r, ranging from -1 to 1) between each foliar trait and the first two principal component dimensions (PC) produced in a PCA using all woody plant species except for A. lasiocarpa and P. engelmannii x glauca. Correlations greater than 0.5 or less than -0.5 are marked in bold. Supplementary variables include proportion of litter mass lost during Phase I, Phase II, and during the first year. PC1 explained 44% of the variation in foliar traits among species, and PC2 explained 18%.  Foliar Trait PC1 PC 2 SLA -0.88 0.29 LDMC 0.61 -0.44 Fp 0.82 -0.05 Fps 0.32 0.36 Leaf Thickness 0.85 -0.18 N -0.93 -0.01 C 0.43 -0.67 C:N 0.95 0.05 pH -0.41 0.74 Abaxial Cuticle Thickness 0.84 0.30 Adaxial Cuticle Thickness 0.76 0.48 Abaxial DTL 0.84 0.21 Adaxial DTL 0.67 0.42 WSE -0.20 -0.38 NPE 0.18 -0.28 ASC -0.16 0.80 AUR 0.33 0.44 Phase-I Loss -0.58 -0.01 Phase-II Loss -0.47 0.27 Year-One Loss -0.63 0.20      75  Figure 3.1. Ordination of All Foliar Traits. Ordination plot of a PCA run on 13 woody plant species with all foliar traits. Proportion of mass lost during Phase I, Phase II, and during the first year were added as supplementary variables. Dim1 is PC1 and Dim2 is PC2. 76 Table 3.2. Correlations Between All Litter Traits and Two Principal Components. Correlations (r, ranging from -1 to 1) between each litter trait and the first two principal component dimensions produced in a PCA using all woody plant species except for A. lasiocarpa and P. engelmannii x glauca. Correlations greater than 0.5 or less than -0.5 are marked in bold. Supplementary variables include proportion of litter mass lost during Phase I, Phase II, and during the first year. PC1 explains 45% of the variation in litter traits among species and PC2 explains 17%.   Litter Trait PC 1 PC 2 SLA 0.81 0.10 Fp -0.88 -0.26 Fps -0.89 0.06 Leaf Thickness -0.78 -0.41 N 0.40 0.36 C -0.37 0.06 C:N -0.39 -0.59 pH 0.24 0.70 WSE 0.59 -0.67 NPE 0.09 -0.37 ASC -0.70 0.54 AUR -0.31 0.75 Leaching Loss 2 Hrs  0.76 -0.24 Leaching Loss 24 Hrs 0.86 -0.06 Water Uptake 2 Hrs 0.84 0.05 Water Uptake 24 Hrs 0.95 0.14 Phase-I Loss 0.67 -0.29 Phase-II Loss 0.28 0.07 Year-One Loss 0.54 -0.08      77  Figure 3.2. Ordination of All Litter Traits. Ordination plot of a PCA run on 13 woody plant species with all litter traits. Proportion of mass lost during Phase I, Phase II, and during the first year were added as supplementary variables. Dim1 is PC1 and Dim2 is PC2. 78 Table 3.3. Correlations Between Select Foliar Traits and Two Principal Components. Correlations between foliar traits that were measured in both foliage and litter and the first two principal component dimensions (PC) produced in a PCA using all woody plant species except for A. lasiocarpa and P. engelmannii x glauca. Correlations greater than 0.5 or less than -0.5 are marked in bold. Supplementary variables include proportion of litter mass lost during Phase I, Phase II, and during the first year. PC1 explained 42% of the variation in foliar traits among species, and PC2 explained 22%.  Foliar Trait PC1 PC 2 SLA -0.93 0.04 Fp 0.88 0.32 Fps 0.29 0.63 Leaf Thickness 0.87 0.00 N -0.92 -0.22 C 0.50 -0.68 C:N 0.91 0.22 pH -0.62 0.51 WSE -0.14 -0.21 NPE 0.20 -0.52 ASC -0.26 0.86 AUR 0.28 0.49 Phase-I Loss -0.62 -0.28 Phase-II Loss -0.50 0.14 Year-One Loss -0.68 -0.03      79  Figure 3.3. Ordination of Select Foliar Traits. Ordination plot of a PCA run on 13 woody plant species with foliar traits that were measured in both green foliage and leaf litter. Proportion of mass lost during Phase I, Phase II, and over a year were added as supplementary variables. Dim1 is PC1 and Dim2 is PC2. 80 Table 3.4 Correlations Between Select Litter Traits and Two Principal Components. Correlations between leaf litter traits that were measured in both foliage and litter and the first two principal component dimensions (PC) produced in a PCA using all woody plant species except for A. lasiocarpa and P. engelmannii x glauca. Correlations greater than 0.5 or less than -0.5 are marked in bold. Supplementary variables include proportion of litter mass lost during Phase I, Phase II, and during the first year. PC1 explained 36% of the variation in litter traits among species, and PC2 explained 23%.  Litter Trait PC1 PC 2 SLA -0.83 0.03 Fp 0.89 -0.23 Fps 0.90 0.08 Leaf Thickness 0.74 -0.37 N -0.40 0.48 C 0.46 0.18 C:N 0.44 -0.65 pH -0.29 0.66 WSE -0.60 -0.69 NPE -0.04 -0.30 ASC 0.64 0.51 AUR 0.34 0.76 Phase-I Loss -0.71 -0.28 Phase-II Loss -0.29 0.07 Year-One Loss -0.57 -0.08      81  Figure 3.4. Ordination of Select Litter Traits. Ordination plot of a PCA run on 13 woody plant species with litter traits that were measured in both foliage and litter. Proportion of mass lost during Phase I, Phase II, and over a year were added as supplementary variables. Dim1 is PC1 and Dim2 is PC2.  82  Figure 3.5. Cluster Analysis of Woody Plants with Foliar Traits. Cluster analysis using 11 foliar traits to group 13 woody plant species native to British Columbia. These traits, which were measured in both foliage and litter, include SLA, thickness, Fps, N, C, C:N, pH, AUR, ASC, WSE, and NPE.   83  Figure 3.6. Cluster Analysis of Woody Plants with Litter Traits. Cluster analysis using 11 litter traits to group 13 woody plant species native to British Columbia. These traits, which were measured in both foliage and litter, include SLA, thickness, Fps, N, C, C:N, pH, AUR, ASC, WSE, and NPE. 84 Table 3.5. Correlations Between Foliar and Litter Traits. Pearson’s correlation coefficients (r) and significance values of relationships between mean foliar and litter traits used in cluster analysis (n = 13; P. munitum, A. lasiocarpa, and P. engelmannii x glauca were not included). Correlations in bold are marginally significant (p < 0.05), and those underlined are significant (p < 0.0045) Trait r p SLA  0.71 0.006 Fps 0.59 0.035 Leaf Thickness (ln transformation) 0.93 <0.001 C 0.60 0.030 N (ln transformation) 0.47 (0.25*) 0.11 (0.42*) C:N 0.39  (0.26*) 0.18 (0.42*) pH 0.67 0.012 WSE 0.32 0.29 NPE (ln transformation) 0.16 0.61 ASC 0.50 0.082 AUR 0.63 0.021 * without A. rubra   Table 3.6 Comparisons of Foliar and Litter Traits. Paired t-test results of comparison between foliar and litter traits used in cluster analysis (n = 13; P. munitum, A. lasiocarpa, and P. engelmannii x glauca were not included). Tests in bold are marginally significant (p < 0.05), and those underlined are significant (p < 0.0045). Trait Foliar*  Litter* t p SLA  10.4 10.0 0.267 0.794 Fps 3.51 3.25 0.663 0.520 Leaf Thickness (ln transformation) 0.37 0.38 0.74 0.475 C 0.495 0.501 -1.33 0.208 N (ln transformation) 0.018 0.010 4.88 <0.001 C:N 33.55 60.07 -4.76 <0.001 pH 4.59 4.69 -0.572 0.577 WSE 0.35 0.25 4.1 0.0015 NPE (ln transformation) 0.071 0.059 -0.26 0.80 ASC 0.356 0.386 -2.22 0.046 AUR 0.207 0.293 -6.78 <0.001 *Means are reported in original units (not transformed when applicable)  85   Figure 3.7. Carbon, Nitrogen, and C:N in Foliage and Litter. Comparison of the mean proportion of N (a), proportion of C (b), and C:N (c) in foliage and litter of 13 woody plant species used in this study. A 1:1 line bisects each graph, with species having higher values in foliage than litter falling on the right side of the line and those with lower values in foliage than litter falling on the left side.   86   Figure 3.8. Chemical Proportions in Foliage and Litter. Comparison of mean proportion of water-soluble extractables (WSE, a), non-polar extractables (NPE, b), acid-soluble carbohydrates (ASC, c), and acid-unhydrolyzable residue (AUR, d) in foliage and litter of 13 woody plant species used in this study. A 1:1 line bisects each graph, with species having higher values in foliage than litter falling on the right side of the line and those with lower values in foliage than litter falling on the left side.  87    Figure 3.9. SLA and Thickness in Foliage and Litter. Comparison of specific leaf area (SLA, a and b) and leaf thickness (c and d) between foliage and litter of 13 woody plant species. Panels a and c show relationships between means, and panels b and d show relationships between distributions of values. A 1:1 line bisects each graph, with species having higher values in foliage than litter falling on the right side of the line and those with lower values in foliage than litter falling on the left side.  88   Figure 3.10. Toughness and pH in Foliage and Litter. Comparison of specific force-to-punch (Fps, a and b) and pH (c and d) between foliage and litter of 13 woody plant species. Panels a and c show relationships between means, and panels b and d show relationships in distributions of values. A 1:1 line bisects each graph, with species having higher values in foliage than litter falling on the right side of the line and those with lower values in foliage than litter falling on the left side.  89  Figure 3.11. Litter Mass Remaining of 14 Plant Species From 0 to 12 Months. Mean proportion of mass remaining (with standard error) of leaf-litter samples of 14 species collected after decomposing for 42, 93, 182, 276, and 365 days. Mass loss after 93 days (encompassing the first three data points) was considered Phase I, and mass loss from 93 to 365 days was considered Phase II.  See List of Abbreviations for species’ identities.  90   Figure 3.12. Proportion of Mass Lost During Phase I, Phase II, and the First Year. Mean proportion of mass lost (with standard error) during Phase I (blue), Phase II (red), and over one year (violet) in 14 plant species decomposing in litterbags at the UBC Farm in Vancouver, BC (n = 5 for each, except for Am where n = 4). See List of Abbreviations for species’ identities.  91  Figure 3.13. Comparison of Mass Lost By Leaching and Mass Lost During Phase I. Mean proportion of mass lost in Phase I (with standard error, turquoise, n = 5 for all species except Am where n = 4) for each of the 14 species used in the litterbag study, compared to the proportion of mass lost by leaching in distilled water for 24 hours (with standard error, navy blue, n = 3 for all species). See List of Abbreviations for species’ identities. 92  Figure 3.14. Regression Tree to Predict Year-One Mass Loss. Tree diagram produced from regression tree analysis of 13 woody plant species (n = 64) using both foliar traits and litter traits to predict proportion of mass lost during the first year. This model explains 68% of variation in year-one mass loss (relative error = 0.32).  Table 3.7. Important Traits in Regression Tree Analysis of Year-One Mass Loss. Variable importance for the 10 most important traits used to predict proportion of mass lost from 0 to 12 months in regression tree analysis. Variable Importance Value Litter Leaching Loss 24 Hrs 15 Foliar Abaxial DTL 15 Foliar SLA 14 Foliar C:N 13 Foliar N 13 Litter Water Uptake 24 Hrs 13 Foliar Adaxial DTL 4 Foliar Fps 2 Litter Fps 2 Litter NPE 2  93  Figure 3.15. Regression Trees to Predict Year-One Mass Loss Using Only Physical and Only Chemical Traits. Tree diagrams produced from regression tree analysis of woody plant species (n = 64) using only physical traits (left) and only chemical traits (right) to predict proportion of mass lost from 0 to 12 months. The model using physical traits explains 68% of variation in year-one mass loss (relative error = 0.32) and the model using chemical traits explains 64% of variation in year-one mass loss (relative error = 0.36).       94 Table 3.8. Important Physical Traits in Regression Tree Analysis of Year-One Mass Loss. Variable importance for the 10 most important physical traits used to predict proportion of mass lost from 0 to 12 months in regression tree analysis. Variable Importance Value Foliar Abaxial DTL 17 Litter Leaching Loss 24 Hrs 16 Foliar LDMC 15 Foliar SLA 15 Litter Water Uptake 24 Hrs 14 Foliar Leaf Thickness 10 Foliar Adaxial DTL 4 Foliar Fps 2 Litter Fps 2 Foliar Abaxial Cuticle  1  Table 3.9. Important Chemical Traits in Regression Tree Analysis of Year-One Mass Loss. Variable importance for the 10 most important chemical traits used to predict proportion of mass lost from 0 to 12 months in regression tree analysis. Variable Importance Value Foliar C:N 21 Foliar N 21 Foliar AUR 10 Litter C:N 10 Litter NPE 10 Litter pH 10 Litter C 4 Foliar C 3 Foliar pH 3 Litter AUR 3  95 Table 3.10. Regression Analysis for Year-One Mass Loss. Traits, coefficient values, R2, p-values, and root mean squared error (SEE) of significant (p < 0.0014) linear regressions run using traits and PC1 variables to predict the proportion of mass lost (ln-transformed) after one year. The cutoff p-value was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). The regressions listed below met the linear regression assumptions of normality and equal variance of residuals. Additional regression results are listed in Tables A5-1 and A5-2 of Appendix 5.  Trait Coefficient (b) R2 p SEE Foliar LDMC -0.0033 53% 1.2 × 10-11 0.25 Foliar SLA 0.035 48% 2.9 × 10-10 0.26 Litter SLA   0.019 (0.03971*) 13% (47%*) 0.0032 (2.0 × 10-9*) 0.34 (0.28) Litter Leaching Loss 24 Hrs 3.3 42% 5.7 × 10-9 0.28 Foliar N (ln) 0.50 (0.49**) 38% (33%) 6.1 × 10-8 (2.3 × 10-6**) 0.29 (0.29) Litter Fps (ln) -0.39 34% 4.2 × 10-7 0.29 Foliar C:N -0.015 34% 4.6 × 10-7 0.29 Litter AUR -3.5 29% 5.0 × 10-6 0.31 Litter Water Uptake 24 Hrs 1.3 21% 1.6 × 10-4 0.32 Litter pH 0.22 20% 2.4 × 10-4 0.32 PC1 Foliage All Traits*** -0.080 37% 1.2 × 10-7 0.29 PC 1 Litter All Traits 0.071 29% 5.4 × 10-6 0.31 PC1 Foliage Same Traits*** -0.11 44% 2.9 × 10-9 0.27 PC1 Litter Same Traits -0.094 29% 3.6 × 10-6 0.30 * without A. macrophyllum ** without A. rubra *** assumptions not met    96  Figure 3.16. Regression Tree to Predict Phase-I Mass Loss. Tree produced from regression tree analysis of woody plant species (n = 64, A. macrophyllum at plot A omitted) using both foliar traits and litter traits to predict proportion of mass lost after 3 months (Phase I). This model explains 71% of variation in Phase I mass loss (relative error = 0.29).  Table 3.11. Important Traits in Regression Tree Analysis of Phase-I Mass Loss. Variable importance for the 10 most important traits used to predict proportion of mass lost after 3 months (Phase I) in regression tree analysis.  Variable Importance Value Litter WSE 20 Litter AUR 18 Litter Fps 17 Litter Leaching Loss 24 Hrs 14 Foliar Abaxial Cuticle Thickness 12 Foliar Adaxial Cuticle Thickness 12 Litter Leaching Loss 2 Hrs 2 Litter ASC 2 Foliar ASC 2 Foliar pH 2  97               Figure 3.17. Regression Trees to Predict Phase-I Mass Loss Using Only Physical and Only Chemical Traits. Tree diagrams produced from regression tree analysis of woody plant species (n = 64) using only physical traits (left) and using only chemical traits (right) to predict proportion of mass lost after 3 months (Phase I). The model using physical traits explains 70% of variation in Phase I mass loss (relative error = 0.30) and the model using chemical traits explains 71% of variation in Phase I mass loss (relative error = 0.29).        98 Table 3.12. Important Physical Traits in Regression Tree Analysis of Phase-I Mass Loss. Variable importance for the ten most important physical traits used to predict proportion of mass lost after 3 months (Phase I) in regression tree analysis. Variable Importance Value Litter Leaching Loss 24 Hrs 18 Foliar Abaxial DTL 16 Foliar LDMC 16 Foliar SLA 15 Litter Water Uptake 24 Hrs 14 Foliar Leaf Thickness 11 Litter Fps 3 Foliar Adaxial Cuticle 2 Foliar Fps 2 Foliar Adaxial DTL 2   Table 3.13. Important Chemical Traits in Regression Tree Analysis of Phase-I Mass Loss. Variable importance for the ten most important chemical traits used to predict proportion of mass lost after 3 months (Phase I) in regression tree analysis. Variable Importance Value Litter WSE 24 Litter AUR 21 Litter ASC 15 Litter C 13 Foliar C 9 Foliar N  9 Litter pH 2 Foliar pH 2 Foliar ASC 2 Foliar AUR 2  99 Table 3.14. Regression Analysis for Phase-I Mass Loss.  Traits, coefficient values, R2, p-values, and root mean squared error (SEE) of significant (p < 0.0014) linear regressions run using traits and PC1 variables to predict the proportion of mass lost (ln-transformed) after decomposing for three months  (Phase I). The cutoff p-value was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Additional regression results are listed in Tables A5-3 and A5-4 of Appendix 5.  Trait Coefficient (b) R2 p SEE Litter Leaching Loss 24 Hrs 3.9 51% 3.0 × 10-11 0.27 Litter AUR -4.6 43% 3.5 × 10-9 0.29 Litter SLA  0.022 (0.042*) 15% (43%*) 0.0016 (1.1 × 10-8*) 0.36 (0.29) Litter WSE 3.2 42% 6.1 × 10-9 0.30 Foliar LDMC -0.0030 39% 3.4 × 10-8 0.30 Foliar N (ln) 0.50 (0.4663**) 33% (26%**) 6.4 × 10-7 (3.3 × 10-5**) 0.32 (0.33) Foliar SLA 0.031 33% 8.2 × 10-7 0.32 Litter Water Uptake 24 Hrs 1.6 30% 3.2 × 10-6 0.33 Foliar AUR -7.5 29% 4.4 × 10-6 0.33 Litter Leaching Loss 2 Hrs 11.5 27% 1.0 × 10-5 0.33 Foliar Fps (ln) -0.53 25% 2.2 × 10-5 0.34 Litter ASC -4.0 24% 4.1 × 10-5 0.34 Litter C -9.6 21% 1.6 × 10-4 0.35 Litter Water Uptake 2 Hrs 1.1 16% 9.4 × 10-4 0.36 Foliar pH 0.23 16% 0.0011 0.36 PC1 Foliage All Traits -0.077 29% 4.9 × 10-6 0.33 PC 1 Litter All Traits*** 0.095 43% 3.7 × 10-9 0.29 PC1 Foliage Same Traits -0.10 34% 3.6 × 10-7 0.32 PC1 Litter Same Traits*** -0.12 43% 3.1 × 10-9 0.29 *without A. macrophyllum **without A. rubra *** assumptions not met   100  Figure 3.18. Regression Tree to Predict Phase-II Mass Loss. Tree diagram produced from regression tree analysis of woody plant species (n = 64, A. macrophyllum at plot A omitted) using both foliar traits and litter traits to predict proportion of mass lost from 3 to 12 months (Phase II). This model explains 44% of variation in Phase II mass loss (relative error = 0.56).   Table 3.15. Important Traits in Regression Tree Analysis of Phase-II Mass Loss. Variable importance for the 10 most important traits used to predict proportion of mass lost from 3 to 12 months (Phase II) in regression tree analysis. Variable Importance Value Foliar LDMC 19 Foliar SLA 19 Foliar Fps 13 Foliar N 12 Foliar Leaf Thickness 11 Litter Thickness 11 Litter pH 5 Litter C:N 4 Litter N 4 Foliar C:N 2  101  Figure 3.19. Regression Trees to Predict Phase-II Mass Loss Using Only Physical and Only Chemical Traits. Tree diagrams produced from regression tree analysis of woody plant species (n = 64) using only physical traits (left) and using only chemical traits (right) to predict proportion of mass lost from 3 to 12 months (Phase II). The model using physical traits explains 35% of variation in Phase II mass loss (relative error = 0.65) and the model using chemical traits explains 40% of variation in Phase II mass loss (relative error = 0.60).        102 Table 3.16. Important Physical Traits in Regression Tree Analysis of Phase-II Mass Loss. Variable importance for physical foliar traits and litter traits used to predict proportion of mass lost from 3 to 12 months (Phase II) in regression tree analysis. Variable Importance Value Foliar SLA 22 Foliar LDMC 18 Foliar Leaf Thickness 12 Litter Water Uptake 2 Hrs 12 Foliar Fps 9 Litter Thickness  9 Foliar Abaxial Cuticle 5 Litter SLA 5 Foliar Adaxial Cuticle 4 Litter Leaching Loss 24 Hrs 3   Table 3.17. Important Chemical Traits in Regression Tree Analysis of Phase-II Mass Loss. Variable importance for chemical foliar traits and litter traits used to predict proportion of mass lost from 3 to 12 months (Phase II) in regression tree analysis. Variable Importance Value Foliar C:N 22 Foliar N  22 Litter NPE 14 Litter C:N 10 Litter N 10 Foliar pH 7 Litter pH 4 Litter ASC 3 Litter WSE 3 Foliar C 2  103 Table 3.18. Regression Analysis for Phase-II Mass Loss. Traits, coefficient values, R2, p-values, and root mean squared error (SEE)  of significant (p < 0.0014) linear regressions run using traits and PC1 variables to predict the proportion of mass lost (ln-transformed) after decomposing from three to 12 months (Phase II). The cutoff p-value was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Additional regression results are listed in Tables A5-5 and A5-6 of Appendix 5.  Trait Coefficient (b) R2 p  SEE Foliar LDMC -0.0034 24% 4.7 × 10-5 0.50 Foliar SLA 0.039 23% 5.8 × 10-5 0.50 Litter pH 0.36 21% 1.1 × 10-4 0.50 Foliar pH 0.37 19% 2.8 × 10-4  0.51 Foliar N (ln) 0.51 (0.50*) 16% (14%) 0.0011 (0.0041) 0.52 (0.54) PC1 Foliage All Traits -0.084 16% 9.2 × 10-4 0.52 PC 1 Litter All Traits 0.043 4.2% 0.10 0.56 PC1 Foliage Same Traits -0.11 19% 2.7 × 10-4 0.51 PC1 Litter Same Traits -0.058 4.6% 0.089 0.55 * without A. rubra    104 Chapter 4: Discussion 4.1 Relationships Between Foliar and Litter Traits Among Species   According to the leaf economics spectrum (LES) hypothesis, traits co-vary such that species with high rates of photosynthesis, high specific leaf area (SLA), high nutrient concentrations, and low dry matter content (LDMC), which therefore are expected to allocate more resources in nutrient acquisition than nutrient conservation, occupy one end of the LES, while species that have lower rates of photosynthesis, lower SLA and nutrient concentrations, and greater LDMC, occupy the other end (Wright et al. 2004). I hypothesized that foliar traits with high values suggesting greater allocation of resources towards nutrient acquisition include SLA, N, WSE, and pH (Wright et al. 2004, Vaieretti et al. 2005, Cornelissen et al. 2006). Conversely, foliar traits with high values suggesting a greater allocation of resources towards nutrient conservation include LDMC, leaf thickness, toughness (Fps), C:N, AUR, cuticle thickness, and DTL (Gallardo and Merino 1993, Pérez-Harguindeguy et al. 2000, Freschet et al. 2012b).  Consistent with these predictions informed by the LES, foliar traits did co-vary along the first principal component (PC) axis such that high SLA, N, WSE, and pH occupied one end of the LES and high toughness, thickness, C, C:N, NPE, AUR, cuticle thickness, and DTL occupied the other (Figure 3.1). SLA, LDMC, Fp, leaf thickness, N, C:N, cuticle thickness, and DTL mostly strongly contributed to PC1, and thereby to this spectrum (Table 3.1). The directionality of these relationships reflects relationships between SLA, N, pH, C, C:N, leaf thickness, lignin, and toughness observed by Wright et al. (2004), Bakker et al. (2011), Jackson et al. (2013a), and Pietsch et al. (2014). The opposition between LDMC and water-soluble compounds was also observed by Birouste et al. (2012). These results provide further evidence that relationships  105 between plant traits may reflect an evolutionary trade-off that plants face between growing rapidly through greater allocation of resources in photosynthesis and nutrient acquisition, and growing less rapidly but instead possessing long-lived leaves that better conserve nutrients.  Novel in my study is the inclusion of cuticle thickness and distance to lumen (DTL) in these trait comparisons. Both abaxial and adaxial cuticle thickness, as well as DTL, co-varied positively with C:N, leaf thickness, toughness, and LDMC - the traits for which high values are probably associated with nutrient conservation, though my study does not explicitly test this association. This is consistent with the hypothesis that leaves at the “nutrient conservation” end of the LES are harder to break down. Cutin, a defining component of the cuticle and a component of acid-unhydrolyzable residue (AUR), is notoriously difficult to break down, which suggests that thicker cuticles would be more difficult to break down too (Swift et al. 1979, Gallardo and Merino 1993). The association between toughness and cuticle thickness has been demonstrated before (Gallardo and Merino 1993, Onoda et al. 2012), but through correlations of cutin concentration and force-to-punch, as or thickness of chemically isolated cuticles and tensile strength, not through correlations of cross-sectional cuticle thickness and Fps. The positive association between LDMC and cuticle thickness further supports the assertion by Onoda et al. (2012) that although the cuticle is a thin fraction of the leaf cross-section, it may account for a large proportion of leaf dry matter. DTL was measured in this study to accommodate variations in secondary cell wall thickness. Due to the strong correlation between cuticle thickness and DTL, however, the extent to which DTL simply reflects cuticle thickness or additionally imparts new information related to the secondary cell wall remains to be determined. To do so, it would be necessary to measure secondary cell wall thickness and determine how this measurement relates to cuticle thickness and DTL.  106 Although this spectrum only applies to foliar traits, I hypothesized that relationships among litter traits would mirror those among foliar traits. I expected that higher values of the additional traits measured in leaf litter, including leaching loss after 2 and 24 hours and water uptake after 2 and 24 hours, would be associated with the end of the spectrum often correlated with nutrient acquisition rather than conservation. Because leaves from plants that occupy the end of the LES associated with nutrient conservation tend to be tougher, have a higher dry matter content, and may have thicker cuticles, I hypothesized that these leaves would lose less mass due to leaching and take up less water. The relative ranking of trait values among species were similar between foliage and litter for many traits, but not all. For example, proportion of N was between 0.005 and 0.015 in leaf litters of all non-N-fixing species but was much more variable among species in foliage (Figure 3.7). This is consistent with the observation that trees tend to produce litter with similar N content as a result of N resorption during senescence (Killingbeck 1996). WSE also decreased in all species from foliage to leaf litter (Figure 3.8), which could be due to the reduction of photosynthesis, and reduced production of water-soluble sugars, due to the breakdown of Rubisco and chlorophyll during senescence  (Collier and Thibodeau 1995, Guiboileau et al. 2010).  Despite these differences attributable to biochemical changes during senescence, many of the litter traits exhibited relationships that would be consistent with a litter trait spectrum. SLA, leaching loss after 2 and 24 hours, water uptake after 2 and 24 hours, and to a lesser extent N, WSE, and pH occupied one end, while toughness, leaf thickness, C:N, ASC, AUR, and C occupied the other (Figure 3.2). The traits most strongly correlated with this spectrum include SLA, toughness (as Fp and Fps), leaf thickness, WSE, ASC, leaching loss after 2 and 24 hours,  107 and water uptake after 2 and 24 hours (Table 3.2). This existence of a litter trait spectrum has also been observed by Bakker et al. (2011), Jackson et al. (2013a), and Makkonen et al. (2012), who noticed positive correlations among litter SLA, WSE, and to a lesser extent N, which were all negatively correlated with toughness, lignin, and C. Makkonen et al. (2012) also measured water saturation capacity, a measurement similar to water uptake, which was also positively correlated with SLA. They also measured cellulose and hemicellulose separately, finding that cellulose was positively correlated with SLA while hemicellulose was not. ASC in my study does not differentiate between these carbohydrates and may obscure different associations among them and other traits. Also novel in my study is the inclusion of leaching loss and water uptake as leaf-litter traits. Leaching loss and water uptake are physical traits that reflect the overall performance of a suite of traits; they are affected by the chemical composition of leaf litter as well as by the physical structures that would prevent leaching loss, i.e., the cuticle (Taylor and Parkinson 1988a). In particular, leaching loss is thought to reflect the loss of sugars and proteins from litter, and water uptake after 2 hours is thought to relate to cuticle thickness (Taylor and Parkinson 1988a, b). Water absorption measured in this way has also been shown to positively correlate with litter thickness, litter area, and water-soluble sugar concentrations, suggesting that these traits may be associated with water acquisition and retention as well (Ibrahima et al. 2008). Makkonen et al. (2012) suggested that water saturation capacity reflected water acquisition and retention, consequently relating to the microclimate experienced by decomposers; perhaps this is true of water uptake as well. If so, the association of water uptake with other traits that strongly predicted mass loss in my study, such as SLA and leaching loss, suggest a possible mechanism for greater decomposability: favorable moisture conditions for decomposers associated with  108 greater water uptake. The results of my study, however, do not provide direct evidence for this mechanism.  As a result of changes in select traits between foliage and litter, trait-based relationships among species differed when using foliage and using litter. In particular, the clear distinction between broadleaf and needle-leaf species in cluster analyses using foliar traits disappeared when using litter traits, probably due to the reduction in variation of N and WSE as a result of senescence (Figures 3.5 and 3.6). Changes in other traits may have influenced these differences in relationships as well. For example, while SLA and thickness did not change in species with lower mean values of these traits, they did change for species with greater means, increasing in litter thickness and both increasing and decreasing in litter SLA (Figure 3.9). It is possible that the deviation in SLA at higher values is due to a difference in resolution when measuring large and small leaves using the LI-COR leaf scanner; perhaps the 1-mm2 resolution for larger leaves exaggerated differences in SLA values compared to the 0.1-mm2 resolution used for smaller leaves. It would be best to perform all measurements using the same resolution, though this was not possible in my study due to the inability of the scanner to detect needle area with 1-mm2 resolution and to measure the entire area of large leaves with 0.1-mm2 resolution. It is also important to acknowledge for all traits that for many of these species, the litter did not come from the same source as the foliage due to logistical issues. It is possible that differences in trait values between foliage and litter in these species may represent differences between populations, not just between foliage and leaf litter. Four of the species did come from the same source populations, however; they include P. menziesii, P. munitum, G. shallon, and A. rubra, all collected at UBC Farm from the same population of plants. When focusing on these species, WSE, AUR, C, N, and C:N changed consistently in all species (Tables A2-3 and A2-4),  109 which is in agreement with the trends observed across all species (Tables 3.5 and 3.6). SLA decreased in all species except for G. shallon, for which the value did not significantly change, and thickness decreased for all species except for A. rubra, for which the value did not significantly change (Tables A2-1 and A2-2). These observations are consistent with the observation of decreasing leaf area during senescence, possibly due to constriction of cells when senescing leaves dry out (Lin and Wang 2001), which would produce a lower SLA if area decreased more relative to mass lost during senescence. Toughness and pH also did not change in one direction for these four species (Tables A2-1 through A2-4). These results from species obtained from the same populations more convincingly suggest that while some traits change consistently in one way across all species, others may not. The only way to accurately determine if these traits change from foliage to leaf litter is to measure the trait in the same leaf specimen as foliage and as litter, which, due to the destructive nature of these protocols, is currently impossible. In future studies, it would be best for foliage and litter to come from the same population in order to minimize extraneous variables and approach the best estimate of the extent to which traits change following senescence. To quantify the extent to which these trait values vary among populations, samples of the same species could be collected from several locations and populations.  The changes in relationships among species based on foliar traits and based on leaf-litter traits suggest that foliar and litter traits may not be interchangeable predictors of decomposition rate. Past studies have claimed that these traits do not change drastically during senescence as a result of incomplete resorption (Killingbeck 1996, Cornwell et al. 2008, Bakker et al. 2011), and while this may be true for some traits, this is not necessarily true for all. It is important to remember that one trait measured in foliage imparts different meaning from the same trait  110 measured in litter, especially in the context of decomposition. In foliage, the trait may represent a functional trait, imparting information on the growth and evolutionary strategy of the plant (Bakker et al. 2011). In litter, which does not face the same evolutionary pressures as foliage, this trait may not serve the same function and may also not be the same or a comparable value at all. Therefore, though it may be the same trait in terms of how it is measured, its function for the leaf, and interpretation in the context of decomposition, may differ between foliage and leaf litter.    4.2 Relationships Between Traits and Early Mass-Loss Rates During the first year of decomposition, mean proportion of mass lost ranged from 0.73 in B. papyrifera to 0.046 in P. munitum. With the exception of A. macrophyllum, more mass loss occurred in deciduous species than in evergreen species. Additionally excluding L. occidentalis, and P. munitum, broadleaf species decomposed faster than needle-leaf species (Figure 3.11). These results are consistent with past observations of deciduous leaves decomposing faster than evergreen leaves in early stages (Prescott et al. 2000, Cornwell et al. 2008, Rahman and Tsukamoto 2013). Faster decomposition of deciduous species with short-lived foliage is consistent with observations of correlations between the LES and decomposability, suggesting that the species which allocate more resources towards conserving nutrients and living longer decompose more slowly compared to those that allocate more resources towards nutrient acquisition (Cornwell et al. 2008, Freschet et al. 2012b).  Also consistent with past observations are the distinct ways in which broadleaf and needle-leaf species lost mass. The fastest decomposing species B. papyrifera lost epidermal and mesophyll tissue between the vascular veins first, leaving a “skeletal” leaf behind (e.g., Figure  111 4.1a), which is consistent with the “skeletonization” pattern of mass loss in Betula spp. observed by Tian et al. (1997). Many other deciduous leaves decomposed in a similar way, though not always to the same extent, even within the same litterbag (e.g., P. tremuloides in plot B, Figure 4.1b). These differences may reflect the heterogeneous distribution of decomposers in the forest floor (Ekschmitt et al. 2005). In contrast, many of the needles appeared outwardly intact throughout the decomposition process (e.g., P. contorta, Figure 4.1c), which is consistent with observations by Ponge (1991) and Tian et al. (1997) suggesting that needle decomposition begins with the penetration of the needle by fungi, which hollow out the mesophyll tissue inside the needle. The presence of black spots on many of the needles (e.g. Figure 4.1c), evidence of fungal colonization, is consistent with similar observations in Pinus sylvestris needles in the L1 layer of the forest floor, which are in the earliest phases of decomposition (Ponge 1991).   Despite being broadleaf, P. munitum lost the least mass of all species in this study, which is consistent with past observations of slower decomposition of fern litter relative to that of angiosperm species (Figure 4.1d; Enrich and Ogden 1987, Preston et al. 2000, Allison and Vitousek 2004, Cornwell et al. 2008). One possible explanation could be the relatively lower foliar concentrations of calcium in ferns (Funk and Amatangelo 2013). Calcium, which was not measured in this study, is necessary for and associated with greater diversity and abundance of faunal decomposers such as earthworms (Reich et al. 2005); therefore, a reduction in calcium supply could lead to slower decomposition. Another compound not measured in this study that could possibly contribute to slower decomposition is silicon; Polystichum ferns have among the highest concentrations of silicon among plants (Hodson et al. 2005), a compound that is associated with increased toughness (Cornelissen and Thompson 1997) and therefore possibly slower decomposition. Ferns may also decompose more slowly due to higher concentrations of  112 toxic defence compounds in their foliage, as suggested by Preston et al. (2000), though this explanation is most often used to explain slow decomposition of non-polypod ferns, specifically Pteridium spp. (Preston et al. 2000, Rasmussen et al. 2003, Amatangelo and Vitousek 2009), not more recently derived polypod ferns like P. munitum. Another explanation could be the relatively high concentrations of lignin and cellulose fibres in fern foliage (Wardle et al. 2002), an explanation supported by the data in that P. munitum litter had the highest proportion of ASC and second-highest proportion of AUR (Table A2-4). Enrich and Ogden (1987), who also observed slower decomposition in ferns over 12 months, consider ferns to be “sclereophyllous”, further suggesting that the relative proportion of lignified sclerenchyma tissue could explain this observation.  Perhaps surprising is the low mass loss measured in A. macrophyllum. With a high SLA, thin cuticle, and low C:N in litter relative to other species, one might expect, based on observed correlations between species on the “nutrient acquisition” end of the LES and decomposability, that these traits would cause A. macrophyllum to lose proportionally more mass, not less. Accordingly, removing the high litter SLA values of A. macrophyllum from the linear regression produced a significant and much stronger regression, suggesting that A. macrophyllum may be an exception to this pattern (Table 3.10), though it is also important to remember that the A. macrophyllum samples on which SLA were measured, while collected from the same population of trees, were collected one year after litterbags were installed. One reason for this deviation in expected decomposability and actual mass loss could be the high proportion of recalcitrant AUR found in A. macrophyllum litter. Of all species, A. macrophyllum had the highest proportion of AUR in litter (Table A2-4). AUR, or Klason lignin, has often strongly predicted mass-loss rates, with litter having high AUR concentrations decomposing more slowly (Cornwell et al. 2008,  113 Freschet et al. 2012b, Koehler and Tranvik 2015), and with de-lignified cell walls breaking down faster in the presence of fungal cellulases than those with lignin (Ding et al. 2012). Although A. macrophyllum leaves have been shown to have faster estimated turnover rates compared to P. menziesii in terms of biomass and nutrients N, P, K, Ca, S, and Mg due to higher concentrations of these nutrients in soil (Fried et al. 1990, Turk et al. 2008), total soil C content did not differ between forests dominated by A. macrophyllum and dominated by coniferous T. heterophylla and P. menziesii, as would be expected if A macrophyllum were to decompose more rapidly (Turk et al. 2008). This suggests that, while A. macrophyllum may be losing nutrients more rapidly, it may not be fully decomposing; instead, part of this litter may be transforming into recalcitrant SOM due to decomposer activity (Turk et al. 2008). My observations support this hypothesis.     Of all foliar and litter traits measured, the traits that in all three statistical methods (PCA, regression tree analysis, and linear regression) correlated with mass loss during the first year of decomposition are a) foliar SLA, LDMC, C:N, and N, and b) litter toughness (as Fps) and leaching loss after 24 hours. Each of these traits, except for leaching loss, has correlated with mass loss. Foliar SLA has strongly correlated with mass loss after the first few months and up to two years, as well as or better than leaf-litter traits (Vaieretti et al. 2005, Cornelissen et al. 2006, Santiago 2007, Bakker et al. 2011) and correlates with foliar N (Wright et al. 2004, Bakker et al. 2011, Kang et al. 2014), another predictor of year-one mass loss. pH has predicted year-one and year-two mass, as well as further enhanced predictions made using other chemical traits and foliar SLA (Cornelissen et al. 2006). Foliar LDMC has also correlated with mass loss (Cornelissen et al. 2006, Freschet et al. 2012b), as has toughness (Gallardo and Merino 1993), which had also correlated with C:N, another strong predictor of decomposition rate (Pérez- 114 Harguindeguy et al. 2000). And while leaching loss as measured in this study has not to my knowledge been shown to predict year-one mass loss, part of mass loss within the first year is often attributed to leaching (Swift et al. 1979, Gallardo and Merino 1993, Vaieretti et al. 2005). Therefore, these predictors of year-one mass loss are consistent with previous observations. Although leaching loss was an important predictor of year-one mass loss, it is possible that leaching loss is most relevant for the first phase of decomposition, and not as important during the rest of the first year. Dividing the first year into two distinct phases of decomposition based on the shape of the mass loss curves provides further insights into the role of these and other traits in early mass loss, suggesting more about the mechanisms that may drive decomposition during the first year.  Accordingly, different traits were important in predicting Phase I and Phase II. The important traits predicting Phase-I mass loss in all three statistical methods used were a) litter WSE, ASC, and leaching loss after 24 hours and b) foliar N. Therefore, mass loss during the first three months of decomposition is most strongly predicted by litter traits as well as foliar N. In particular, WSE, ASC, and leaching loss all relate to the process of leaching; water-soluble components are lost during this process, which would accordingly increase the proportion of ASC in litter. This observation is consistent with others suggesting that early mass loss is dominated by leaching loss (Swift et al. 1979, Gallardo and Merino 1993, Trofymow et al. 2002). Conversely, Phase-II mass loss was most strongly related to foliar N, SLA, pH, and LDMC, suggesting that it is best predicted by foliar traits, not litter traits, and not traits that necessarily indicate leaching loss. Instead, these foliar traits, which have been shown to be part of various leaf and plant economics spectra, reflect plant functional strategies (Wright et al. 2004, Freschet et al. 2012b, Diaz et al. 2016).      115  With foliar traits, such as foliar N shown to predict Phase I and the four foliar traits shown to predict Phase II, it is important to consider that these foliar traits may convey different meanings in the context of decomposition than their litter counterparts and to accordingly consider what they represent instead. For example, foliar N is not interchangeable with litter N. The stronger relationship of foliar N with decomposition, especially early mass-loss rates, than litter N has also been observed by Bakker et al. (2011), who suggest that it relates to other traits that may better predict decomposition. But what are those traits, if they do not include measured leaf-litter traits?   One possible explanation for the association between foliar N and decomposition, as well as between foliar N and SLA, and in other studies photosynthetic rate (Wright et al. 2004), could be the relative abundance of mesophyll cells within the leaf. Within leaves, the largest source of N is Rubisco, an important CO2-fixing enzyme involved in photosynthesis (Thomas 2012), and an additional source of N is chlorophyll (Taiz and Zeiger 2010). Both of these sources of N are located in chloroplasts, which are found in the mesophyll and contain up to 70% of N-containing leaf proteins (Evert and Eichhorn 2006, Guiboileau et al. 2010). Coincidentally, it is this mesophyll tissue that is the first to disappear as a result of decomposer activity in needles (Ponge 1991, Tian et al. 1997) and in broadleaves (Tian et al. 1997). Mesophyll is largely comprised of thin-walled parenchyma cells, specifically chloroplast-containing chlorenchyma, characterized by many plasmodesmata that facilitate movement of molecules between cells and relatively large amounts of intercellular space allowing for gas exchange (Evert and Eichhorn 2006). These characteristics would make this tissue accessible to both microbial decomposers and soil fauna - more accessible than the densely packed fibres and vascular bundles. Although they did not quantify this distinction in their study, Enright and Ogden (1987) determined that the leaf litters  116 they considered to be most “mesophyllous” decomposed faster than those they considered “sclerophyllous”, which is consistent with this hypothesis. Their interpretation is also consistent with the strong, negative relationship observed between year-one and Phase-II mass loss and LDMC, as perhaps leaves with greater LDMC have a greater proportion of sclerenchyma, which is found in supportive, structural tissue within the leaf; the lignified cell walls of sclerenchyma are less hydrated than those of collenchyma and parenchyma, which form the bulk of mesophyll tissue (Evert and Eichhorn 2006). My study does not provide direct evidence of the role of mesophyll in the mechanism of decomposition. Developing a metric that would relay the relative amount of mesophyll to other tissues in the leaf, in particular to structural tissues that contain higher proportion of sclerenchyma, would perhaps more clearly show a relationship between plant anatomy and decomposition and serve as a better mass-loss predictor. Perhaps the sclerophylly index (ratio of crude fibre to crude protein in leaves, with crude fibre representing structural tissue and crude proteins representing mesophyll), observed by Choong et al. (1992) to be strongly related to toughness, would also imply this relationship and could also be used to predict mass loss.  Although litter N was found to be not important for Phase-II mass loss in all three statistical methods, it is interesting to note that litter N was more closely related to mass loss in Phase II than Phase I in PCA (Figures 3.2 and 3.4), implying that perhaps litter nutrient content, of importance to decomposers, is more important in later phases of early mass loss. This is consistent with the finding that foliar traits, perhaps related to the relative abundance of mesophyll tissue that may be more easily broken down by decomposers, more strongly predicted mass loss during later phases of decomposition, after leaching had mostly ceased. In this way, the results of my study suggest that, while both processes may be occurring simultaneously  117 throughout the first year of decomposition, Phase I is dominated by leaching and Phase II by decomposer activity.  Further support for the hypothesis that Phase I is dominated by leaching comes from higher rainfall in Phase I than in Phase II: the average daily rainfall during Phase I was 4.7 mm, and during Phase II was 2.4 mm (Figure 2.1). Greater rainfall during Phase I, which coincides with fresh litter being present on the ground, could partly explain the strong relationship between mass loss in the field during Phase I and leaching loss.  This connection highlights the important of interpreting these results in the context of the study system; in arid ecosystems, for example, the extent to which leaching predicts early mass loss may depend on differences in regional climate. Rainfall during this period was within the range of annual rainfall from 2004 through 2015 (Figure 2.2). Therefore, it is reasonable to assume that patterns of mass detected in this study are representative of mass loss occurring at this site.   During year-one mass loss, and during both phases, physical traits, specifically LDMC, leaching loss, and SLA, were among the strongest predictors of mass loss. Physical structures are composed of chemicals, and thereby related to chemical composition, but beyond chemical traits, they convey more information about how the arrangement of these chemicals in litter in turn may influence decomposer access and effects of water (the cause of leaching) on leaf litter. The results of my study therefore suggest that physical traits, though not as routinely used in trait-based decomposition studies (Makkonen et al. 2012), can be effective in predicting mass loss and provide further possibilities for interpretations based on decomposer access to substrate. Considering both physical and chemical traits together helps to create a better conceptual framework for understanding the mechanisms of decomposition.     118  New traits included in this study include leaching loss, water uptake, cuticle thickness, and distance to lumen, or DTL. Leaching loss was a strong predictor of year-one mass loss and especially Phase-I mass loss, explaining the more variation in Phase-I mass loss (ln-transformed) than all other traits in linear regression and having the greatest importance in regression tree analysis with only physical traits (Tables 3.7, 3.10, 3.12, and 3.14). These results are consistent with past observations of leaching dominating early mass-loss rates (e.g., Gallardo and Merino 1993, Trofymow et al. 2002) and suggest that leaching loss could serve as an effective and cost-effective predictor of early mass-loss, reflecting an emergent property produced from both chemical composition and structural characteristics, in particular surface properties. Significant and positive correlations between water uptake after 2 hours and cuticle thickness, particularly with abaxial cuticle thickness (r = -0.79, p < 0.001 for abaxial cuticle thickness; r = -0.65, p < 0.001 for adaxial cuticle thickness), confirm similar claims by Taylor and Parkinson (1988a) and suggest a cost-effective metric that incorporates cuticle properties. I did not measure cuticle thickness of leaf litter in this study, but given the recalcitrance of cutin (Swift et al. 1979), I expect that cuticle thickness might not significantly differ between foliage and leaf-litter. DTL measurements of P. contorta litter, abaxial and adaxial combined, were comparable to those measured in leaf litter (mean = 4.02 ± 0.11 SE µm in litter, n = 18, compared to 3.97 ± 0.18 SE µm in foliage, n = 36; Zukswert, unpublished data), which suggests that this may be true, though DTL includes the secondary cell wall and therefore may obscure relative differences in cuticle and cell wall thickness. DTL in P. ponderosa litter, however, was slightly thicker (mean = 4.65 ± 0.12 SE µm in litter, n = 18, compared to 3.84 ± 0.16 µm in foliage, n = 36; Zukswert, unpublished data). Measuring cuticle thickness in litter of different genera, including broadleaf  119 and deciduous species, will help confirm whether cuticle thickness measurements in foliage can accurately approximate those in litter.      Cuticle thickness and DTL correlated positively with foliar C:N and negatively with foliar SLA and N (Figure 3.1), which were good predictors of year-one mass loss, but these measurements alone did not strongly predict mass loss. These measurements did contribute in regression tree analysis, but not necessarily in expected ways. For example, in the regression tree predicting year-one loss, species that had lost a proportion of more than 0.097 of initial mass and had abaxial DTL measurements greater than 2.8 µm lost more mass over 12 months than the species with DTL measurements less than 2.8 µm (Figure 3.14). However, of the species that lost a proportion less than 0.097 of initial mass, in regards to adaxial DTL this pattern was reversed (Figure 3.14). For several of the species that leached less and had smaller adaxial DTL values and lower predicted mass loss, the adaxial DTL measurements did not greatly differ from those of the cuticle (e.g., P. contorta, P. ponderosa; Table A2-6, Table A4-2), and for some of the species that decomposed faster (e.g., P. menziesii), the cuticle was noticeably distinct from and smaller than the secondary cell wall. Perhaps differences in the cuticle composition, e.g., the extent to which they are lignified, or differences in the proportion of cuticle to cell wall, could explain these results. The microscopy technique used in this study to image cuticle thickness and DTL did not clearly distinguish between cuticle and secondary cell wall, a distinction that can be difficult to make in some species due to the gradual transition from cuticle to cell wall (Holloway 1982), but other microscopy techniques using light microscope exist to more clearly visualize this distinction. Phloroglucinol, for example, can be used to visualize the distribution of lignin in plant tissue and Sudan IV stain (0.09% in 1:1 EtOH:glycerol) can also be used to visualize cuticles (Ruzin 1999). Lignin also auto-fluoresces and can be visualized using confocal  120 microscopy (e.g., Donaldson et al. 2001). Foliar adaxial cuticle thickness also helped predict mass loss in a counterintuitive way in litter from species that lost a proportion less than 0.097 of initial mass in the regression tree analysis that was performed using only physical traits (Figure 3.17). It is also possible that, due to the small amount of variation that these traits explain within the resulting regression tree analyses, evident by the much smaller branch lengths relative to those for leaching loss, that the proportions of variation explained by cuticle thickness and DTL in these cases are minimal and not biologically significant.  Cuticle thickness can vary significantly within a single tree; for example, it can vary with height and with leaf vapor pressure deficit at maximum stomatal conductance (Woodruff et al. 2010). Additionally, three-dimensional imaging of cuticle structure using confocal microscopy revealed that cuticle morphology and consequently thickness can vary widely even within a single leaf (Buda et al. 2009). Despite these shortcomings, cuticle thickness measurements obtained from fluorescent microscopy images of free-hand sections stained with nile red did vary as would be expected according to the LES, suggesting that this method, which can in a relatively short time produce many measurements, does produce a reliable, though general estimate of cuticle thickness.  The lack of strong relationships between measures of cuticle thickness and mass loss and the counterintuitive directions of influence may indicate that characteristics of the cuticle, other than thickness, more strongly influence decomposition. Given that the surface of epicuticular wax is the first point of contact between decomposers and litter (Müller and Riederer 2005) and that surface roughness influences the ability of fungi colonize foliage (Carver and Gurr 2006), perhaps surficial properties have a stronger influence. Kearns and Bärlocher (2008) used scanning electron microscopy (SEM) images to estimate leaf surface roughness and observed  121 greater colonization of rougher leaves in microcosms to simulate decomposition in streams. Perhaps surface roughness influences terrestrial decomposition in these early mass-loss phases as well.   While it is informative to consider which traits are related to each phase of mass loss, it is also important to consider how much variation in mass loss these traits explain. This comparison is possible by comparing the root mean-squared error, also called the standard error of the estimate (SEE), between linear regression models using traits to predict mass loss in Phase I and Phase II. Regressions with smaller SEE values predict decomposition better than those with larger values, and SEE was smaller for all significant linear regressions predicting Phase-I mass loss than those predicting Phase-II mass loss (Tables 3.14 and 3.18). Therefore, traits better predicted the first mass-loss phase. This observation is further supported by comparisons of correlations between PC1 of foliage and leaf-litter and the two phases of decomposition (Tables 3.1 through 3.4), and comparisons of proportion of variation explained in regression tree analyses (70 to 71% of variation in Phase-I loss, compared to 35 to 40% of variation in Phase-II loss). Similarly, Vaieretti et al. (2005) observed that physical and chemical traits of foliage and leaf-litter predicted the first phase of year-one mass loss (0 to 70 days) better than the second (70 to 365 days). One reason could be that initial foliar and litter traits more accurately describe the state of the litter complex at the start of Phase I than at the start of Phase II. After three months, the litter has already undergone chemical and physical changes and therefore does not exactly resemble the initial litter. Perhaps using chemical and physical traits measured on leaf litter that has soaked in water for 24 hours, and therefore from which many of the water-soluble components have been removed, would better predict Phase II, in that characteristics of this leached litter might more accurately reflect the characteristics of litter that has been decomposing for three months  122 and has probably already lost most of its soluble components. This type of study would also better reveal how characteristics of the components of leaf litter not lost in leaching relate to later stages of mass loss and thereby relate to decomposer activity. Another reason for the greater SEE and lower proportion of variation explained by trait-based models predicting Phase-II mass loss than Phase-I mass loss could be differences in time. Effects of environmental factors, such as climate variables and forest floor conditions, at the micro-site level as well as of differences in the spatial aggregation of decomposers may accumulate over time and contribute to further variation in decomposition and consequently mass loss (Fogel and Cromack 1977, Ekschmitt et al. 2005). Phase I is only one-third as long as Phase II, which occurs later in the decomposition process; this may translate to greater variation in Phase-II mass loss due to accumulating effects of environmental factors and decomposer activity. Although I measured several environmental variables at the plot level in my study, and did not find any significant relationships between them and mass loss (Tables A5-7 through A5-9), this study was not designed to test the effects of environmental variables on decomposition. Decomposer abundance, diversity, and activity was not measured in this study, but is known to be heterogeneous on the forest floor (Ekschmitt et al. 2005) and would probably differ among plots, as well as within plots. In order to better account for these differences, Bradford et al. (2016) recommend measuring environmental factors at the level of the litterbag. Lastly, this discrepancy in the ability of traits to predict Phase-I and Phase-II mass loss may be due to greater extrapolation required for Phase II calculations. Phase I represents a difference between the initial mass and mass after 3 months of the same litterbag, but Phase-II calculations involve assuming that the proportion of mass lost after 3 months in the litterbag collected after 12 months is the same as that of the litterbag actually collected after 3 months, an assumption that may not  123 be accurate based on the heterogeneity of decomposer distribution and conditions in the forest floor. Although this currently may be the best way to estimate Phase-II mass loss, measuring environmental variables and decomposer characteristics at the level of the litterbag would at least help assess the extent to which this assumption is reliable.   The consistency of all three statistical methods in suggesting that trait-based models predict Phase-I mass loss better than Phase II highlight congruence among these methods. In terms of the best models of Phase I and Phase II predicting using each method, the conclusions differ slightly. In linear regression, traits predicted Phase-I mass loss better than Phase-II, with litter leaching loss after 24 hours predicting mass loss best for Phase I and foliar LDMC for Phase II, these models having the highest R2 and lowest SEE. In regression tree analysis, the model using litter WSE had the lowest relative error in predicting Phase-I mass loss (relative error = 0.29, 71% of variation explained) and the model using foliar LDMC and leaf-litter pH had the lowest in predicting Phase-II mass loss (relative error = 0.56, 44% of variation explained). And in the ordination produced by PCA, Phase-I mass loss was most strongly positively correlated with foliar N, foliar SLA, and litter leaching loss, and negatively correlated with foliar C:N, foliar Fp, litter ASC, and litter Fps (Figures 3.1 and 3.2). Phase-II mass loss was most strongly, positively correlated with foliar N, foliar and litter SLA, and litter water uptake, and negatively correlated with foliar thickness, foliar LDMC, litter toughness (Fp and Fps), and litter C (Figures 3.1 and 3.2).  It is difficult to say which of the methods best predicted mass loss, as they each predict mass loss in different ways and are inconsistent in how mass loss and the variables are expressed: PCA used standardized trait values, regression tree analysis used un-transformed data, and regression predicted ln-transformed mass loss using both transformed and un-transformed  124 trait values. These methods also have their own unique strengths and weaknesses, each imparting something slightly different about mass loss predictions. PCA is valuable in that it shows how all traits relate to each other, supporting the recent observation that only a relatively small set of possible trait combinations may be evolutionarily successful for vascular plants (Diaz et al. 2016). PCA also created individual variables from linear combinations of all of the trait variables (principal components) that explained a fair amount of variation when used in linear regression. Linear regression is a common way to predict decomposition and produces standard errors of the estimate that allow for reliable comparisons across models with different response variables (such as Phase-I and Phase-II loss), but doing regression using many traits can be problematic, and regression itself does not show how these traits are correlated with one another. Performing many regressions can also be time-consuming and require modifications to set probability values (e.g., Bonferroni corrections). In addition, regressions can include interactions and multiple variables, which while tools like stepwise regression exist to aid in variable selection, can be difficult for trait-based studies with skewed distributions and highly correlated predictor variables. Regression tree analysis, which is advantageous in its ability to handle non-normal distributions and highly correlated variables that interact with each other (De’ath and Fabricius 2000), has not to my knowledge been used in trait-based studies of mass loss. Its agreement with linear regression in which traits most strongly predict Phase-I and Phase-II mass loss suggest that it is effective at identifying strong predictors. It provides a further advantage over PCA and regression in that, by constructing hierarchical trees, it may help identity possible threshold values, above or below which mass loss changes. Regression tree analysis can also further identify traits that influence decomposition only in a subset of species, and helps identify traits that are repeatedly influential through ranking them by “importance”. One criticism of regression  125 tree analysis is its strong sensitivity to small differences in values (Prasad et al. 2006), but pruning the tree by adjusting the complexity parameter can help reduce this sensitivity, at least by retaining only the splits that best explain the data and eliminating splits that only explain a small proportion of variation (Therneau et al. 2013). Other tree-based methods, namely bagging and random forest, have also recently been shown to make slightly better predictions than regression tree analysis (Prasad et al. 2006); perhaps these methods can be used in trait-based studies as well. My study has shown that these methods are largely consistent in their conclusions, but do convey slightly different information, particularly by identifying relationships between traits and identifying traits that are important for some species but not others. Together, these insights help provide a broader perspective of trait relationships with early mass loss.  One common problem encountered in all three statistical methods, as well as in many other past decomposition studies, is variability, particularly in trait measurements. Intraspecific variation in traits can sometimes be as high as species-level variation (Messier et al. 2010), or account for a significant proportion of variation if not as much (Jackson et al. 2013b). Past studies on trait relationships (Wright et al. 2004, Diaz et al. 2016) and relationships between traits and decomposition (e.g., Bakker et al. 2011, Makkonen et al. 2012) have historically used mean trait values, ignoring intraspecific trait variation. Jackson et al. (2013b) observed, however, that intraspecific variation in leaf and litter traits did not predict decomposability, but species-level trait variation did. In addition, intraspecific variation is in part explained by environmental variables, such as soil chemistry (e.g., Jackson et al. 2013b), which were not found to differ greatly between my plots or significantly influence mass loss in my study (Tables A5-7 through A5-9). Therefore, the influence of my decision to use mean trait values and exclude intraspecific  126 variation on my interpretation is probably small. Including intraspecific variation in future studies, though, may provide a more comprehensive perspective and additional insights (Messier et al. 2010).  As these interpretations and discussions of methodology have demonstrated, it is difficult, if not impossible, to determine exactly what drives decomposition by using correlative studies such as this one. Due to the high degree of covariance among these traits, consistent with the LES, many traits change as one changes. This and other studies are informative in that by considering new traits, specifically physical traits, they help shed new light on possible decomposition, but these correlative methods alone will not be able to explain causation. Considering what the traits that relate to mass loss mean in relationship to decomposition and devising a method to observe that mechanism, as well as accounting for environmental variables at the level of the litterbag, decomposer characteristics, and intraspecific variation in traits, will reveal more about the process of decomposition, as well as developing innovative methods that deviate from traditional correlative studies.    127   Figure 4.1. Litter Samples Collected at Year One. Samples of B. papyrifera (a), P. tremuloides (b), Pinus contorta (c), and Polystichum munitum (d) obtained from plot B at UBC Farm on 2 December 2015, 12 months after the installation of the litterbags, and oven-dried at 70ºC.  (b) (a) (c) (d)  128 Chapter 5: Conclusions My study provides further evidence that foliar traits do co-vary in such a way that would reflect the trade-off between nutrient acquisition and nutrient conservation articulated in the leaf economics spectrum hypothesis, with cuticle thickness and distance to lumen co-varying positively with traits for which high values are often associated with nutrient conservation, and with leaching loss and water uptake co-varying positively with traits for which high values are often associated with nutrient acquisition. A similar spectrum exists using leaf-litter traits, but because some traits change from foliage and leaf-litter, trait-based relationships among species change when using leaf-litter traits instead of foliar traits. Therefore, while there is some consistency between foliar traits and litter traits, we cannot assume that they are interchangeable, particularly in the context of decomposition studies. These traits, though the same measurement, impart different meaning when measured in foliage and in leaf litter.   The first phase of decomposition, lasting 3 months, was best predicted by leaf-litter traits, particularly by leaching loss after 24 hours and litter chemistry, suggesting that Phase I of mass loss is dominated by leaching. The second phase of decomposition, from 3 to 12 months, was best predicted by foliar functional traits: negatively correlated with LDMC and highly correlated with N and SLA. These traits are associated with photosynthetic rate, so these relationships suggest that the relative abundance of mesophyll tissue in foliage could be a good predictor of mass loss of Phase II. Therefore, because decomposers preferentially break down mesophyll, this phase may be dominated by decomposer activity. The better ability of both foliar and litter traits to predict Phase-I rather than Phase-II mass loss, as indicated by the lower standard error of the estimate in linear regressions, could reflect a greater influence of environmental factors on  129 Phase-II mass loss or reflect the greater uncertainty inherent in Phase-II calculations of mass loss due to assumptions required about mass remaining after 3 months.   During both Phase I and Phase II, physical traits such as LDMC, SLA, and leaching loss predicted mass loss as well or better than chemical traits. These findings suggest that physical traits can be effective predictors of mass loss, and they additionally facilitate interpretations of the data that relate to litter structure and decomposer access to litter. In particular, physical traits provide further insight into what processes are occurring, and more so than chemical traits can provide, where in the leaf litter they are occurring. Beyond determining the influence of substrate composition on decomposition, which chemical traits accomplish, physical traits additionally help to determine whether the combination of these chemicals into specific structures influences decomposition. Using both chemical and physical traits in predicting mass loss in each phase of decomposition will help to create a better, more comprehensive framework for describing the mechanism of decomposition.  The intention of this work was to help identify potential directions for research to help elucidate mechanisms in greater depth. Several possible directions for future research that have emerged from this study include 1) investigating relationships between leaf anatomy or litter architecture and mass loss, 2) better understanding relationships between foliar and litter traits, and 3) measuring traits of leached litter for the purpose of identifying characteristics that best predict later stages of decomposition. The hypothesis that relative mesophyll abundance could help explain decomposition, as well as the recognition that physical traits provide additional insights into mass loss, suggest that considering leaf anatomy and litter architecture, in particular as they relate to mesophyll, may be a fruitful direction for research to better understand the decomposition process. Observing and quantifying the distribution of structural material in  130 leaves and litter, such as venation patterns, may be fruitful as well. The finding that foliar traits are not necessarily equivalent to litter traits suggests that an improved understanding of relationships between foliar and litter traits would help us better explain what the correlations we observe between foliar functional traits and mass loss might actually mean in terms of the mechanism of decomposition, in that it is litter, not foliage, that serves as the substrate for decomposition, not foliage. The finding of a strong relationship between leaching loss and early mass loss, at least in this temperate rainforest ecosystem, raises the question: would traits of leached litter better predict mass loss in later stages of decomposition? Perhaps characteristics of litter from which water-soluble extractables have already been leached would better predict the decomposition of material in this ecosystem after 3 months, and might help identify the characteristics of leaf litter that are important for decomposers in these later stages. Research in these three directions would help further elucidate and clarify the mechanisms of decomposition and therefore possibly improve our predictive abilities in the context of decomposition.          131 Bibliography Adler, P. B., Salguero-Gómez, R., Compagnoni, A., Hsu, J. S., Ray-Mukherjee, J., Mbeau-Ache,  C., and Franco, M. 2014. Functional traits explain variation in plant life history strategies. P. Natl. Acad. Sci. USA 111(2): 740–745. doi: 10.1073/pnas.1315179111  Aerts, R. 1997. Climate, leaf litter chemistry, and leaf litter decomposition in terrestrial ecosystems: A triangular relationship. Oikos 79(3): 439–449. Available from   http://www.jstor.org/stable/3546886[accessed 22 September 2013].  Aguilar-Trigueros, C. A., Powell, J. R., Anderson, I. C., Antonovics, J., and Rillig, M. C. 2014.   Ecological understanding of root-infecting fungi using trait-based approaches. Trends Plant Sci 19(7): 432–438. doi: 10.1016/j.tplants.2014.02.006  Allison, S. D. and Vitousek, P. M. 2004. Rapid nutrient cycling in leaf litter from invasive plants  in Hawai’i. Oecologica 141: 612-619. doi: 10.1007/s00442-004-1679-z  Amatangelo, K. L. and Vitousek, P. M. 2009. Contrasting predictors of fern versus angiosperm  decomposition in a common garden. Biotropica 41(2): 154-161. doi:10.1111/j.1744-7429.2008.00470.x  Aranwela, N, Sanson, G., and Read, J. 1999. Methods of assessing leaf-fracture properties. New  Phytol 144(2): 369–383. Available from http://www.jstor.org/stable/2588781 [accessed 6 March 2014].  Ashton, P. M. S., Olander, L. P., Berlyn, G. P., Thadani, R., and Cameron, I. R. 1998. Changes  in leaf structure in relation to crown position and tree size of Betula papyrifera within fire-origin stands of interior cedar-hemlock. Can. J. Bot. 76: 1180–1187. doi: 10.1139/b98-128  Austin, A. T., Vivanco, L., González-Arzac, A., and Pérez, L. 2014. There’s no place like home?  An exploration of the mechanisms behind plant litter-decomposer affinity in terrestrial ecosystems. New Phytol. 204: 307–314. doi: 10.1111/nph.12959  Baker, E. A. 1982. Chemistry and morphology of plant epicuticular waxes. In The Plant Cuticle.  Edited by D.F. Cutler, K. L. Alvin, and C. E. Price. Academic Press, London, UK. pp. 33-44.  Bakker, M. A., Carreño-Rocabado, G., and Poorter, L. 2011. Leaf economics traits predict litter  decomposition of tropical plants and differ among land use types. Funct. Ecol. 25: 473–483. doi: 10.1111/j.1365-2435.2010.01802.x  Baldrian, P., Vetrovsky, T., Cajthaml, T., Dobiasova, P., Petrankova, M., Snajdr, J., and  Eichlerova, I. 2013. Estimation of fungal biomass in forest litter and soil. Fung. Ecol. 6: 1–11. doi:10.1016/j.funeco.2012.10.002   132 Bardgett, R. D., and Wardle, D. A. 2010. Aboveground-Belowground Linkages: Biotic  Interactions, Ecosystem Processes, and Global Change. Oxford University Press, Oxford, UK.   Bargel, H., Koch, K., Cerman, Z., and Neinhuis, C. 2006. Structure-function relationships of the  plant cuticle and cuticular waxes – a smart material? Funct. Plant Biol. 33: 893–910. doi: 10.1071/FP06139  BC Ministry of Forests and Range and BC Ministry of Environment. 2010. Field Manual For  Describing Terrestrial Ecosystems. 2nd ed. Province of British Columbia, Victoria, BC.  Berg, B. and McClaughert, C. 2008. Decomposition as a process. In Plant Litter. Springer,  Berlin, DE. pp. 11-33.  Binkley, D. and Fisher, R. F. 2013. Ecology and Management of Forest Soils. 4th ed. John Wiley  & Sons, Ltd, West Sussex, UK.  Birouste, M., Kazakou, E., Blanchard, A., and Roumet, C. 2012. Plant traits and decomposition:  Are the relationships for roots comparable to those for leaves? Ann. Bot. 109: 463–472. doi: 10.1093/aob/mcr297  Bocock, K. L. and Gilbert, O. J. W. 1957. The disappearance of leaf litter under different woodland conditions. Plant Soil 9(2): 179-185.  de Boer, W., Folman, L. B., Summerbell, R. C., and Boddy, L. 2005. Living in a fungal world:  Impact of fungi on soil bacterial niche development. FEMS Microbiol. Rev. 29: 795–811. doi: 10.1016/j.femsre.2004.11.005  Bohlen, P. J., Groffman, P. M., Fahey, T. J., Fisk, M C., Suárez, E., Pelletier, D. M., and Fahey,  R. T. 2004. Ecosystem consequences of exotic earthworm invasion of north temperate forests. Ecosystems 7: 1–12. doi: 10.1007/s10021-003-0126-z  Bradford, M. A, Berg, B., Maynard, D. S., Wieder, W. R., and Wood, S. A. 2016. Understanding  the dominant controls on litter decomposition. J. Ecol. 104: 229–238. doi: 10.1111/1365-2745.12507  Breiman, L., Friedman, J. H., Olshen, R. A., and Stone, C. J. 1984. Classification and Regression  Trees. Wadsworth International Group, Belmont, Calif.  Buda, G. J., Isaacson, T., Matas, A. J., Paolillo, D. J., and Rose, J. K. C. 2009. Three- dimensional imaging of plant cuticle architecture using confocal scanning laser microscopy. Plant J. 60: 378–385. doi: 10.1111/j.1365-313X.2009.03960.x  Bunnell, F. L., Tait, D. E. N., Flanagan, P. W., and Van Cleve, K. 1977. Microbial respiration  and substrate weight loss – I: A general model of the influences of abiotic variables. Soil Biol. Biochem. 9: 33–40. doi: 10.1016/0038-0717(77)90058-X   133 Canhoto, C., and Graça, M. A. S. 1999. Leaf barriers to fungal colonization and shredders  (Tipula lateralis) consumption of decomposing Eucalyptus globulus. Microbial Ecol. 37: 163-172. Available from http://www.jstor.org/stable/4251637 [accessed 25 July 2014].  Carver, T. L. W., and Gurr, S. J. 2006. Filamentous fungi on plant surfaces. In  Biology of the  Plant Cuticle. Edited by M. Riederer and C. Müller. Blackwell Publishing Ltd, Oxford, UK. pp. 368-397.  Chave, J., Coomes, D., Jansen, S. Lewis, S. L., Swenson, N. G., and Zanne, A. E. 2009. Towards  a worldwide wood economics spectrum. Ecol. Lett. 12: 351–366. doi: 10.1111/j.1461-0248.2009.01285.x  Choong, M. F., Lucas, P. W., Ong, J. S. Y., Pereira, B., Tan, H. T. W., and Turner, I. M. 1992.  Leaf fracture toughness and sclerophylly: Their correlations and ecological implications. New Phytol. 121(4): 597–610. Available from http://www.jstor.org/stable/2557614 [accessed 8 October 2013].  Collier, D. E., and Thibodeau, B. A. 1995. Changes in respiration and chemical content during  autumnal senescence of Populus tremuloides and Quercus rubra leaves. Tree Physiol. 15: 759-764. Available from http://treephys.oxfordjournals.org/content/15/11/759.full.pdf [accessed 30 March 2014].  Cornelissen, J. H. C, and Thompson, K. 1997. Functional leaf attributes predict litter  decomposition rates in herbaceous plants. New Phytol. 135: 109–114.  Cornelissen, J. H. C., Pérez-Harguindeguy, N., Díaz, S., Grime, J. P., Marzano, B., Cabido, M.,  Vendramini, F., and Cerabolini, B. 1999. Leaf structure and defence control litter decomposition rate across species and life forms in regional floras on two continents. New Phytol. 143(1): 191–200. Available from http://www.jstor.org/stable/2588526 [accessed 8 September 2013].  Cornelissen, J. H. C., Quested, H. M., van Logtestijn, R. S. P., Pérez-Harguindeguy, N., Gwynn- Jones, D., Díaz, S., Callaghan, T. V., Press, M. C., and Aerts, R. 2006. Foliar pH as a new plant trait: Can it explain variation in foliar and carbon cycling processes among subarctic plant species and types? Oecologia 147(2): 315–326. Available from http://www.jstor.org/stable/20445828 [accessed 7 October 2013].  Cornelissen, J. H. C, Sibma, F., Van Logtestijn, R. S. P., Broekman, R. A., and Thompson, K.  2011. Leaf pH as a plant trait: Species-driven rather than soil-driven variation. Funct. Ecol. 25: 449–455. doi: 10.1111/j.1365-2435.2010.01765.x  Cornwell, W. K., Cornelissen, J. H. C., Amatangelo, K., Dorrepaal, E., Eviner, V. T., Godoy, O.  Hobbie, S. E., Hoorens, B., Kurokawa, H., Pérez-Harguindeguy, N., Quested, H. M., Santiago, L. S., Wardle, D. A., Wright, I. J., Aerts, R., Allison, S. D., van Bodegom, P., Brovkin, V., Chatain, A., Callaghan, T. V., Díaz, S., Garnier, E., Gurvich, D. E., Kazakou, E., Klein, J. A., Read, J., Reich, P. B., Soudzilovskaia, N. A., Vaieretti, M. V.,  134 and Westoby, M. 2008. Plant species traits are the predominant control on litter decomposition rates within biomes worldwide. Ecol. Lett. 11:1065–1071. doi: 10.1111/j.1461-0248.2008.01219.x  Cotrufo, M. F., Briones, M. J. I., and Nelson, P. 1998. Elevated CO2 affects field decomposition  rate and palatability of tree leaf litter: Importance of changes in substrate quality. Soil Biol. Biochem. 30(12): 1565–1571. doi: 10.1016/S0038-0717(98)00032-7  Cotrufo, M. F., Wallenstein, M. D., Boot, C. M., Denef, K., and Paul, E. 2013. The Microbial  Efficiency-Matrix Stabilization framework integrates plant litter decomposition with soil organic matter stabilization: Do labile plant inputs form stable soil organic matter? Global Change Biol. 19: 988–995. doi: 10.1111/gcb.12113  Coûteaux, M., Bottner, P., and Berg, B. 1995. Litter decomposition, climate, and litter quality.  Trends Ecol. Evol. 10(2): 63–66. doi: 10.1016/S0169-5347(00)88978-8  Dale, S. E., Turner, B. L., and Bardgett, R. D. 2015. Isolating the effects of precipitation, soil  conditions, and litter quality on leaf litter decomposition in lowland tropical forests. Plant Soil 349(1): 225–238. doi: 19.1007/s11104-015-2511-8.  De’ath, G., and Fabricius, K. E., 2000. Classification and regression trees: A powerful yet simple   technique for ecological data analysis. Ecology 81(11), 3178–3192. doi: 10.1890/0012-9658(2000)081[3178:CARTAP]2.0.CO;2  Díaz, S., Kattge, J., Cornelissen, J. H. C., Wright, I. J., Lavorel, S., Dray, S., Reu, B., Kleyer, M.,  Wirth, C., Prentice, I. C., Garnier, E., Bönisch, G., Westoby, M., Poorter, H., Reich, P.  B., Moles, A. T., Dickie, J., Gillison, A. N., Zanne, A. E., Chave, J., Wright, S. J., Sheremet’ev, S. N., Jactel, H., Baraloto, C., Cerabolini, Pierce, S., Shipley, B., Kirkup, D., Casanoves, F., Joswig, J. S., Günther, A., Falczuk, V., Rüger, N., Mahecha, M. , and Gorné, L. D. 2016. The global spectrum of plant form and function. Nature 529: 167–171. doi: 10.1038/nature16489  Donaldson, L., Hague, J., and Snell, R. 2001. Lignin distribution in coppice poplar, linseed and  wheat straw. Holzforschung 55: 379–385. doi: 10.1515/HF.2001.063 Dungait, J. A. J., Hopkins, D. W., Gregory, A. S., and Whitmore, A. P. 2012. Soil organic  matter turnover is governed by accessibility not recalcitrance. Global Change Biol. 18: 1781–1796. doi: 10.1111/j.1365-2486.2012.02665.x  Ekblad, A., Wallander, H., and Näsholm, T. 1998. Chitin and ergosterol combined to measure total and living fungal biomass in ectomycorrhizas. New Phytol. 138: 143–149. doi: 10.1046/j.1469-8137.1998.00891.x    Ekschmitt, K., Kandeler, E., Poll, C., Brune, A., Buscot, F., Friedrich, M., Gleixner, G.,  135 Hartmann, A., Kästner, M., Marhan, S., Miltner, A., Scheu, S., and Wolters, V. 2008. Soil-carbon preservation through habitat constraints and biological limitations on decomposer activity. J. Plant Nutr. Soil Sci. 171: 27–35. doi: 10.1002/jpln.200700051  Ekschmitt, K., Liu, M., Vetter, S., Fox, O., and Wolters, V. 2005. Strategies used by soil biota  to overcome soil organic matter stability – why is dead organic matter left over in the soil? Geoderma 128: 167–176. doi: 10.1016/j.geoderma.2004.12.024  Ellison, A. M., Bank, M. S., Clinton, B. D., Colburn, E. A., Elliott, K., Ford, C. R., Foster, D. R.,  Kloeppel, B. D., Knoepp, J. D., Lovett, G. M., Mohan, J., Orwig, D. A., Rodenhouse, N. L., Sobczak, W. V., Stinson K. A., Stone, J. K., Swan, C. M., Thompson, J., Von Holle, B., and Webster, J. R. 2005. Loss of foundation species: Consequences for the structure and dynamics of forested ecosystems. Front. Ecol. Environ. 3(9): 479–486. doi: 10.1890/1540-9295(2005)003[0479:LOFSCF]2.0.CO;2  Enright, N., and Ogden, J. 1987. Decomposition of litter from common woody species of kauri  (Agathis australis Salisb.) forest in northern New Zealand. Aust. J. Ecol. 12: 109–124. doi: 10.1111/j.1442-9993.1987.tb00933.x  Environment Canada. 2015. Historical Climate Data. [Online database]. Available from  http://climate.weather.gc.ca [accessed 15 December 2015].  Evert, R. F., and Eichhorn, S. E. 2006. Esau’s Plant Anatomy: Meristems, Cells, and Tissues of  the Plant Body: Their Structure, Function, and Development. 3rd ed. John Wiley & Sons, Inc., Hoboken, NJ.  Fernández-Martínez, M., Vicca, S., Janssens, I. A., Sardans, J., Luyssaert, S., Campioli, M.,  Chapin III, F. S., Ciais, P., Malhi, Y., Obersteiner, M., Papale, D., Piao, S. L., Reichstein, M., Rodà, F., and Peñuelas, J. 2014. Nutrient availability as the key regulator of global forest carbon balance [online]. Nature Clim. Change: doi: 10.1038/NCLIMATE2177  Fogel, R., and Cromack, Jr., K. 1977. Effect of habitat and substrate quality on Douglas fir litter  decomposition in western Oregon. Can. J. Bot. 55(12): 1632–1640. doi: 10.1139/b77-190 Foucreau, N., Puijalon, S., Hervant, F., and Piscart, C. 2013. Effect of leaf litter characteristics  on leaf conditioning and on consumption by Gammarus pulex. Freshwater Biol. 58: 1672–1681. doi: 10.1111/fwb.12158  Fowler, S. D., and Greenspan, P. 1985. Application of nile red, a fluorescent hydrophobic probe,  for the detection of neutral lipid deposits in tissue sections: Comparison with Oil Red O. J. Histochem. Cytochem. 33(8): 833–836. doi: 10.1177/33.8.4020099  Freschet, G. T., Cornelissen, J. H. C., van Logtestijn, R. S. P, and Aerts, R. 2010. Evidence of the ‘plant economics spectrum’ in a subarctic flora. J. Ecol. 98: 362–373. doi: 10.1111/j.1365-2745.2009.01615.x   136 Freschet, G. T., Aerts, R., and Cornelissen, J. H. C. 2012a. Multiple mechanisms for trait effects  on litter decomposition: Moving beyond home-field advantage with a new hypothesis. J. Ecol. 100: 619–630. doi: 10.1111/j.1365-2745.2011.01943.x  Freschet, G., T., Aerts, R., and Cornelissen, J. H. C. 2012b. A plant economics spectrum of litter  decomposability. Funct. Ecol. 26: 56–65. doi: 10.1111/j.1365-2435.2011.01913.x  Fried, J. S., Boyle, J. R., Tappeiner II, J. C., and Cromack Jr., K. 1990. Effects of bigleaf maple  on soils in Douglas-fir forests. Can. J. Forest. Res. 20: 259–266. doi: 10.1139/x90-038 Funk, J. L., and Amatangelo, K. L. 2013. Physiological mechanisms drive differing foliar  calcium content in ferns and angiosperms. Oecologia 173: 23–32. doi: 10.1007/s00442-013-2591-1  Gallardo, A., and Merino, J. 1993. Leaf decomposition in two Mediterranean ecosystems of southwest Spain: Influence of substrate quality. Ecology 74(1): 152–161. Available from:  http://www.jstor.org/stable/1939510 [accessed 1 October 2013].  Gholz, H. L., Wedin, D. A., Smitherman, S. M., Harmon, M. E., and Parton, W. J. 2000. Long- term dynamics of pine and hardwood litter in contrasting environments: toward a global  model of decomposition. Global Change Biol. 6: 751–765. doi: 10.1046/j.1365-2486.2000.00349.x  Grandy, A. S., and Neff, J. C. 2008. Molecular C dynamics downstream: The biochemical  decomposition sequence and its impact on soil organic matter structure and function. Sci. Total Environ. 404: 297–307. doi: 10.1016/j.scitotenv.2007.11.013  Griffin, D. M. 1963. Soil moisture and the ecology of soil fungi. Biol. Rev. 38: 141–166. doi: 10.1111/j.1469-185X.1963.tb00781.x  Guiboileau, A., Sormani, R., Meyer, C., and Masclaux-Daubresse, C. 2010. Senescence and  death of plant organs: Nutrient recycling and developmental regulation C. R. Biologies 333: 382–391. doi: 10.1016/j.crvi.2010.01.016  Harrell, F. E. 2015. Package ‘Hmisc’ [online]. Available from  https://cran.r-project.org/web/packages/Hmisc/Hmisc.pdf [accessed January 2016].  Hobara, S., Osono, T., Hirose, D., Noro, K., Hirota, M., and Benner, R. 2014. The roles of  microorganisms in litter decomposition and soil formation. Biogeochemistry 118: 471–486. doi: 10.1007/s10533-013-9912-7  Hodgson, J. G., Montserrat-Marti, G., Charles, M., Jones, G., Wilson, P., Shipley, B., Sharafi,  M., Cerabolini, B. E. L., Cornelissen, J. H. C., Band, S. R., Bogard, A., Castro-Díez, P., Guerrero-Campo, J., Palmer, C., Pérez-Rontomé, M. C., Carter, G., Hynd, A., Romo-Díez, A., de Torres Espuny, L., and Royo Pla, F. 2011. Is leaf dry matter content a better  137 predictor of soil fertility than specific leaf area? Ann. Bot. 108: 1337–1345. doi: 10.1093/aob/mcr225  Hodson, M. J., White, P. J., Mead, A., and Broadley, M. R. 2005. Phylogenetic variation in the silicon composition of plants. Ann. Bot. 96: 1027–1046. doi: 10.1093/aob/mci255  Holloway, P. J. 1982. Structure and histochemistry of plant cuticular membranes: An overview. In The Plant Cuticle. Edited by D.F. Cutler, K. L. Alvin, and C. E. Price. Academic Press, London, UK. pp. 1–32.  Hunt, G. M., and Baker, E. A. 1980. Developmental and environmental variations in plant  epicuticular waxes: Some effects on the penetration of naphthylacetic acid. In The Plant Cuticle. Edited by D.F. Cutler, K. L. Alvin, and C. E. Price. Academic Press, London, UK. pp. 279–292.  Husson, F. and Josse, J. 2016. imputePCA {missMDA} [online]. Available from  http://www.inside-r.org/packages/cran/missMDA/docs/imputePCA [accessed February 2016].  Ibrahima, A., Biyanzi, P., and Halima, M. 2008. Changes in organic compounds during leaf  litter leaching: Laboratory experiment on eight plant species of the Sudano-guinea Savannas of Ngaoundere, Cameroon. iForest 1: 27–33. doi: 10.3832/ifor0450-0010027  Jackrel, S. L., and Wootton, J. T. 2015. Cascading effects of induced terrestrial plant defences  on aquatic and terrestrial ecosystem function [online]. Proc. R. Soc. B. 282: 20142522. Available from http://dx.doi.org/10.1098/rspb.2014.2522 [accessed 16 April 2015].  Jackson, B. G., Peltzer, D. A., and Wardle, D. A. 2013a. Are functional traits and litter  decomposability coordinated across leaves, twigs, and wood?  A test using temperate rainforest tree species. Oikos 122: 1131–1142. doi: 10.1111/j.1600-0706.2012.00056.x  Jackson, B. G., Peltzer, D. A., and Wardle, D. A. 2013b. The within-species leaf economic  spectrum does not predict leaf litter decomposition at either the within-species or whole community levels. J. Ecol. 101: 1409–1419. doi: 10.1111/1365-2745.12155  Jetter, R., Kunst, L., and Samuels, A. L. 2006. Composition of plant cuticular waxes. In Biology  of the Plant Cuticle. Edited by M. Riederer and C. Müller. Blackwell Publishing Ltd, Oxford, UK. pp. 145-181.  Kandeler, E. 2007. Physiological and biochemical methods for studying soil biota and their  function. . In Soil Microbiology, Ecology, and Biochemistry. 3rd ed. Edited by E. A. Paul. Academic Press, Burlington, Mass. pp. 53–83.  Kattge, J., Díaz, S., Lavorel, S., Prentice, I. C., Leadley, P., Bonisch, G., Garnier, E., Westoby,  M., Reich, P. B., Wright, I. J., Cornelissen, J. H. C., Violle, C., Harrison, S. P., van Bodegom, P. M., Reichstein, M., Enquist, B. J., Soudzilovskaia, N. A., Ackerly, D. D.,  138 Anand, M., Atkin, O., Bahn, M., Baker, T. R., Baldocchi, D., Bekker, R., Blanco, C. C., Blonder, B., Bond, W. J., Bradstock, R., Bunker, D. E., Casa Noves, F., Cavender-Bares, J., Chambers, J. Q., Chapin III, F. S., Chave, J., Coomes, D., Cornwell, W. K., Craine, J. M., Dobrin, B. H., Duarte, L., Durka, W., Elser, J., Esser, G., Estiarte, M., Fagan, W. F., Fang, J., Fernández-Méndez, F., Fidelis, A., Finegan, B., Flores, O., Ford, H., Frank, D., Freschet, G. T., Fyllas, N. M., Gallagher, R. V., Green, W. E., Gutierrez, A. G., Hickler, T., Higgins, S. I., Hodgson, J. G., Jalil, A., Jansen, S., Joly, C. A., Kerkhoff, A. J., Kirkup, D., Kitajima, K., Kleyer, M., Klotz, S., Knops, J. M. H., Kramer, K., Kühn, I., Kurokawa, H., Laughlin, D., Lee, T. D., Leishman, M., Lens, F., Lenz, T., Lewis, S. L., Lloyd, J., Llusià, J., Louault, F., Ma, S., Mahecha, M. D., Manning, P., Massad, T., Medlyn, B. E., Messiers, J., Moles, A. T., Müller, S. C., Nadrowski, K., Naeem, S., Niinemets, Ü, Nöllert, S., Nüske, A., Ogaya, R., Oleksyn, J., Onipchenko, V. G., Onoda, Y., Ordoñez, J., Overbeck, G., Ozinga, W. A., Patiño, S., Paula, S., Pausas, J. G., Peñuelas, J., Phillips, O. L., Pillar, V., Poorter, H., Poorter, L., Poschlod, P., Prinzing, A., Proul, R., Rammig, A., Reinsch, S., Reu, B., Sack, L., Salgado-Negret, B., Sardans, J., Shiodera, S., Shipley, B., Siefert, A., Sosinski, E., Soussana, J. –F., Swaine, E., Swenson, N., Thompson, K., Thornton, P., Waldram, M., Weiher, E., White, M., White, S., Wright, S. J., Yguel, B., Zaehle, S., Zanne, A. E., and Wirth, C. 2011. TRY – a global database of plant traits. Global Change Biol. 17: 2905–2935. doi: 10.1111/j.1365-2486.2011.02451.x  Kaye, J. P. and Hart, S. C. 1997. Competition for nitrogen between plants and soil microorganisms. Trends Ecol. Evol. 12(4): 139–143. doi: 10.1016/S0169-5347(97)01001-X  Kazakou, E., Vile, D., Shipley, B., Gallet, C., and Garnier, E. 2006. Co-variations in litter  decomposition, leaf traits, and plant growth in species from a Mediterranean old-field succession. Funct. Ecol. 20(1): 21-30. doi: 10.111l/ j.1365-2435.2006.01080.x   Kearns, S. G., and Bärlocher, F. 2008. Leaf surface roughness influences colonization success  of aquatic hyphomycete conidia. Fungal Ecol. 1:13–18. doi: 10.1016/j.funeco.2007.07.001  Keiser, A. D., Keiser D. A., Strickland, M. S., and Bradford, M. A. 2014. Disentangling the  mechanisms underlying functional differences among decomposer communities. J. Ecol. 102: 603–609. doi: 10.1111/1365-2745.12220  Killingbeck, K. T. 1996. Nutrients in senesced leaves: Keys to the search for potential resorption  and resorption proficiency. Ecology 77(6): 1716–1727. Available from http://www.jstor.org/stable/2265777 [accessed 19 February 2014].  Koehler, B. and Tranvik, L. J. 2015. Reactivity continuum modeling of leaf root, and wood  decomposition across biomes. J. Geophys. Res-Biogeo. 120(7): 1196-1214. doi: 10.1002/2015/JG002908.  Kögel-Knabner, I. 2002. The macromolecular organic composition of plant and microbial   139 residues as inputs to soil organic matter. Soil Biol. Biochem. 34: 139–162. doi: 10.1016/S0038-0717(01)00158-4  Lê, S., Josse, J., and Husson, F. 2008. FactoMineR: An R package for multivariate analysis.  J. Stat. Softw. 25(1): 1–18. doi: 10.18637/jss.v025.i01 Lehmann, J., and Kleber, M. 2015. The contentious nature of soil organic matter. Nature 528:  60–68. doi: 10.1038/nature16069  Lin, P., and Wang, W. 2001. Changes in the leaf composition, leaf mass, and leaf area during  leaf senescence in three species of mangroves. Ecol. Eng. 16: 415–424. doi: 10.1016/S0925-8574(00)00126-9  Luttmerding, H.A. 1984. Soils of the Langley–Vancouver Map Area: Report 15, British  Columbia Soil Survey Volume 6. Province of British Columbia, Ministry of Environment, Assessment and Planning Division [online]. Available from http://sis.agr.gc.ca/cansis/publications/surveys/bc/bc15/bc15-v6_report.pdf [accessed 2015].  Makkonen, M., Berg, M. P., Handa, I. T., Hattenscwhiler, S., van Ruijen, J., van Bodegom, P.  M., and Aerts, R. 2012. Highly consistent effects of plant litter identity and functional traits on decomposition across a latitudinal gradient. Ecol. Lett. 15: 1033–1041. doi: 10.1111/j.1461-0248.2012.01826.x  Mambelli, S., Bird, J. A., Gleizner, G., Dawson, T. E., and Torn, M. S. 2011. Relative  contribution of foliar and fine root pine litter to the molecular composition of soil organic matter after in situ degradation. Org. Geochem. 42: 1099–1108. doi: 10.1016/j.orggeochem.2011.06.008  Manly, B. F. J. 2004. Multivariate Statistical Methods: A Primer. 3rd ed.CRC Press, Boca Raton,  Flor.    Manter, D. K., Kelsey, R. G., and Stone, J. K. 2001. Quantification of Phaeocryptopus  gaeumannii colonization in Douglas-fir needles by ergosterol analysis. Forest Pathol. 31: 229–240. doi: 10.1046/j.1439-0329.2001.00243.x  McClaugherty, C. A., Pastor, J., Aber, J. D., Melillo, J. M. 1985. Forest litter decomposition in  relation to soil nitrogen dynamics and litter quality. Ecology 66: 266–275. Available from http://www.jstor.org/stable/1941327 [accessed 30 September 2015].  Messier, J., McGill, B. J., and Lechowicz, M. J. 2010. How do traits vary across ecological scales? A case for trait-based ecology. Ecol. Lett. 13: 838–848. doi: 10.1111/j.1461-0248.2010.01476.x  Mille-Lindblom, C., von Wachenfeldt, E., and Tranvik, L. J. 2004. Ergosterol as a measure of   140 living fungal biomass: Persistence in environmental samples after fungal death. J. Microbiol. Methods 59: 253–262. doi: 10.1016/j.mimet.2004.07.010  Minderman, G. 1968. Addition, decomposition and accumulation of organic matter in forests.  J. Ecol. 56(2): 355–362. doi: 10.2307/2258238  Mori, K., Bernier, N., Kosaki, T., and Ponge, J. 2009. Tree influence on soil biological activity:  What can be inferred from the optical examination of humus profiles? Eur. J. Soil Biol. 45: 290–300. doi: 10.1016/j.ejsobi.2009.05.001  Mueller, K. E., Hobbie, S. E., Chorover, J., Reich, P. B., Eisenhauer, N., Castellano, M. J.,  Chadwick, O. A., Dobies, T., Hale, C. M., Jagodzinski, Kalucka, I., Kieliszewksa-Rokicka, B., Modrzynski, J., Rozen, A., Skorupski, M., Sobczyk, L., Stasinska, M., Trocha, L. K., Weiner, J., Wierzbicka, A.,  and Oleksyn, J. 2015. Effects of litter traits, soil biota, and soil chemistry on soil carbon stocks at a common garden with 14 tree species. Biogeochemistry 123(3): 313–327. doi: 10.1007/s10533-015-0083-6  Müller, C., and Riederer, M. 2005. Plant surface properties in chemical ecology. J. Chem. Ecol.  31(11): 2621–2651. doi: 10.1007/s10886-005-7617-7  Nguyen, S. H., Webb, H. K., Mahon, P. J., Crawford, R. J., and Ivanova, E. P. 2014. Natural  insect and plant micro-/nanostructured surfaces: An excellent selection of valuable templates with superhydrophobic and self-cleaning properties. Molecules 19: 13614–13630. doi: 10.3390/molecules190913614  Olson, J. S. 1963. Energy storage and the balance of producers and decomposers in ecological  systems. Ecology 44(2): 322–331. Available from http://www.jstor.org/stable/1932179 [accessed 21 February 2016].  Onoda, Y., Richards, L., and Westoby, M. 2012. The importance of leaf cuticle for carbon  economy and mechanical strength. New Phytol. 196: 441–447. doi: 10.1111/j.1469-8137.2012.04263.x  Onoda, Y., Westoby, M., Adler, P. B., Choong, A. M. F., Clissold, F. J., Cornelissen, J. H. C.,  Díaz, S., Dominy, N. J., Elgart, A., Enrico, L., Fine, P. V. A., Howard, J. J., Jalili, A., Kitajima, K., Kurokawa, H. McArthur, C., Lucas, P. W., Markesteijn, L., Pérez-Harguindeguy, N., Poorter, L., Richards, L., Santiago, L. S., Sosinski Jr., E. E., Van Bael, S. A., Warton, D. I., Wright, I. J., Wright, S. J., and Yamashita, N. 2011. Global patterns of leaf mechanical properties. Ecol. Lett. 14: 301–312. doi: 10.1111/j.1461-0248.2010.01582.x  O’Reilly, C., Arnott, J. T., and Owens, J. N. 1989. Effects of photoperiod and moisture  availability on shoot growth, seedling morphology, and cuticle and epicuticular wax features of container-grown western hemlock seedlings. Can. J. Forest. Res. 19: 122–131. doi: 10.1139/x89-016   141 Papa, G., Scaglia, B., Schievano, A., and Adani, F. 2014. Nanoscale structure of organic matter  could explain litter decomposition. Biogeochemistry 117: 313–324. doi: 10.1007/s10533-013-9863-z  Paul, E. A. and Clark, F. E. 1996. Soil Microbiology and Biochemistry. 2nd ed. Academic Press,  San Diego, Calif.   Pérez-Harguindeguy, N., Diaz, S., Cornelissen, J. H. C., Vendramini, F., Cabido, M., and  Castellanos, A. 2000. Chemistry and toughness predict leaf litter decomposition rates over a wide spectrum of functional types and taxa in central Argentina. Plant Soil 218: 21–30. doi: 10.1023/A:1014981715532  Pérez-Harguindeguy, N., Diaz, S., Garnier, E., Laborel, S., Poorter, H., Jaureguiberry, P., Bret- Harte, M. S., Cornwell, W. K., Craine, J. M., Gurvich, D. E., Urcelay, C., Veneklaas, E. J., Reich, P. B., Poorter, L., Wright, I. J., Ray, P., Enrico, L., Pausas, J. G., de Vos, A. C., Buchmann, N., Funes, G., Quétier, F., Hodgson, J. G., Thompson, K., Morgan, H. D., ter Steege, H., van der Heijden, M. G. A., Sack, L., Blonder, B., Poschlod, P., Vaieretti, M. V., Conti, G., Staver, A. C., Auino, S., and Cornelissen, J. H. C. 2013. New handbook for standardized measurement of plant functional traits worldwide. Aust. J. Bot. 61(3): 167–234. Available from http://dx.doi.org/10.1071/BT12225 [accessed September 2013].  Pietsch, K. A., Ogle, K., Cornelissen, J. H.C., Cornwell, W. K., Bönisch, G., Craine, J. M.,  Jackson, B. G., Kattge, J., Peltzer, D. A., Penuelas, J., Reich, P. B., Wardle, D. A., Weedon, J. T., Wright, I. J., Zanne, A. E., and Wirth, C. 2014. Global relationships of wood and leaf litter decomposability: The role of functional traits within and across organs. Global Ecol. Biogeogr. 23: 1046-1057. doi: 10.1111/geb.12172  Ponge, J. 1991. Succession of fungi and fauna during decomposition of needles in a small area of  Scots pine litter. Plant Soil 138: 99–113. doi: 10.1007/BF00011812 Prasad, A. M., Iverson, L. R., and Liaw, A. 2006. Newer classification and regression tree techniques: Bagging and random forests for ecological prediction. Ecosystems 9: 181–199. doi: 10.1007/s10021-005-0054-1  Prescott, C. E. 2005. Do rates of litter decomposition tell us anything we really need to know?  Forest Ecol. Manag. 220: 66–74. doi: 10.1016/j.foreco.2005.08.005  Prescott, C. E. 2010. Litter decomposition: What controls it and how can we alter it to sequester  more carbon? Biogeochemistry 101: 133–149. doi: 10.1007/s10533-010-9439-0  Prescott, C. E., and Grayston, S. J. 2013. Tree species influence on microbial communities in  litter and soil: Current knowledge and research needs. Forest Ecol. Manag. 309: 19–27. doi: 10.1016/j.foreco.2013.02.034  Prescott, C. E., Taylor, B. R., Parsons, W. F. J., Durall, D. M., and Parkinson, D. 1993.   142 Nutrient release from decomposing litter in Rocky Mountain coniferous forests: Influence of nutrient availability. Can. J. Forest Res. 23: 1576–1586. doi: 10.1139/x93-198   Prescott, C. E., Zabek, L. M., Staley, C. L., and Kabzems, R. 2000. Decomposition of broadleaf  and needle litter in forests of British Columbia: Influences of litter type, forest type, and litter mixtures. Can. J. Forest. Res. 30: 1742–1750. doi: 10.1139/x00-097  Preston, C. M., Trofymow, J. A., and the CIDET Working Group. 2000. Variability in litter quality and its relationship to litter decay in Canadian forests. Can. J. Bot. 78: 1269–1287. doi: 10.1139/b00-101  Price, C. E. 1982. A review of the factors influencing the penetration of pesticides through plant  leaves. In The Plant Cuticle. Edited by D.F. Cutler, K. L. Alvin, and C. E. Price. Academic Press, London, UK. pp. 237-252.  Rahman, M. M., and Tsukamoto, J. 2013. Leaf traits, litter decomposability and forest floor dynamics in an evergreen- and a deciduous-broadleaved forest in warm temperate Japan. Forestry 86: 441–451. doi: 10.1093/forestry/cpt015  Rasmussen, L. H., Jensen, L. S., and Hansen, H. C. B. 2003. Distribution of the carcinogenic  terpene ptaquiloside in bracken fronds, rhizomes (Pteridium aquilinum) and litter in Denmark. J. Chem. Ecol, 29(3): 771–778. doi: 10.1023/A:1022885006742  Reich, P. B., Oleksyn, J., Modrzynski, J., Mrozinski, P., Hobbie, S. E., Eissenstat, D. M., Chorover, J., Chadwick, O. A., Hale, C. M., and Tjoelker, M. G. 2005. Linking litter calcium, earthworms, and soil properties: A common garden test with 14 tree species. Ecol. Lett. 8: 811–818. doi: 10.1111/j.1461-0248.2005.00779.x  Riederer, M., Matzke, K., Ziegler, F., and Kögel-Knabner, I. 1993. Occurrence, distribution and  fate of the lipid plant biopolymers cutin and suberin in temperate forest soils. Org. Geochem. 20(7): 1063–1076. doi: 10.1016/0146-6380(93)90114-Q  Ruzin, S. E. 1999. Plant Microtechnique and Microscopy. Oxford University Press, New York,  NY.  Ryan, M.G., Melillo, J. M., and Ricca, A. 1990. A comparison of methods for determining  proximate carbon fractions of forest litter. Can. J. Forest. Res. 20: 166–171. doi: 10.1139/x90-023  Santiago, L. S. 2007. Extending the leaf economics spectrum to decomposition: Evidence from  a tropical forest. Ecology 88(5): 1126–1131. doi: 10.1890/06-1841 Schuldt, A., Bruelheide, H., Durka, W., Eichenberg, D., Fischer, M., Krober, W., Härdtle, W.,   143 Ma, K., Michalski, S. G., Palm, W., Schmid, B., Welk, E., Zhou, H, and Assmann, T. 2012. Plant traits affecting herbivory on tree recruits in highly diverse subtropical forests. Ecol. Lett. 15: 732–739. doi: 10.1111/j.1461-0248.2012.01792.x  Seitz, L. M., Sauer, D. B., Burroughs, R., Mohr, H. E., and Hubbard, J. D. 1979. Ergosterol as a  measure of fungal growth. Phytopathology 69(11): 1202–1203. doi: 10.1094/Phyto-69-1202  Soong, J. L., Calderón, F. J., Betzen, J., and Cotrufo, M. F. 2014. Quantification and FTIR  characterization of dissolved organic carbon and total dissolved nitrogen leached from litter: A comparison of methods across litter types. Plant Soil 385: 125–137. doi: 10.1007/s11104-014-2232-4  Suffling, R. and Smith, D. W. 1974. Litter decomposition studies using mesh bags: Spillage  inaccuracies and the effects of repeated artificial drying. Can J. Bot. 52: 2157–2163.  doi: 10.1139/b74-278  Swift, M. J., Heal, O. W., and Anderson, J. M. 1979. Decomposition in Terrestrial Ecosystems.  Blackwell Scientific Publications, Oxford, UK.   Taiz, L., and Zeiger, E. 2010. Plant Physiology. 5th ed. Sinauer Associates, Sunderland, Mass.  TAPPI. 1976. Alcholo-benzene and dichloromethane solubles in wood and pulp. TAPPI Official  Standard T204.   TAPPI. 1981. Water solubility of wood and pulp. TAPPI Official Standard T207.  Taylor, B. R., and Bärlocher, F. 1996. Variable effects of air-drying on leaching losses from tree  leaf litter. Hydrobiologia 325: 173–182. doi: 10.1007/BF00014982  Taylor, B. R., and Parkinson, D. 1988a. Annual differences in quality of leaf litter of aspen (Populus tremuloides) affecting rates of decomposition. Can. J. Bot. 66: 1940–1947. doi: 10.1139/b88-266  Taylor, B. R., and Parkinson, D. 1988b. Patterns of water absorption and leaching in pine and  aspen leaf litter. Soil Biol. Biochem. 20(2): 257–258. doi:10.1016/0038-0717(88)90047-8  Therneau, T., Atkinson, B., and Ripley, B. 2013. Package ‘rpart’ [online]. Available from  https://cran.r-project.org/web/packages/rpart/rpart.pdf [accessed November 2013].  Thomas, H. 2012. Senescence, ageing and death of the whole plant. New Phytol. 197: 696– 711. doi: 10.1111/nph.12047  Thomas, K. D. and Prescott, C. E. 2000. Nitrogen availability in forest floors of three tree species on the same site: The role of litter quality. Can. J. Forest. Res. 30: 1698–1706. doi: 10.1139/x00-101  144  Tian, X., Takeda, H., and Ando, T. 1997. Application of a rapid thin section method for observations on decomposing litter in mor humus form in a subalpine coniferous forest. Ecol. Res. 12: 289–300. doi: 10.1007/BF02529458  Tomizono, S. 2013. Draw a double box plot chart (2-axes box plot; box plot correlation diagram)  in R [online]. Available from https://tomizonor.wordpress.com/2013/03/15/double-box-plot/[accessed 1 November 2015].  Trofymow, J.A., and the CIDET working group. 1998. The Canadian Intersite Decomposition  Experiment (CIDET): Project and Site Establishment Report. Can. For. Serv. Information Report BC-X-378.  Trofymow, J. A., Moore, T. R., Titus, B., Prescott, C., Morrison, I., Siltanen, M., Smith, S.,  Fyles, J., Wein, R., Camiré, C., Duschene, L., Kozak, L., Kranabetter, M., and Visser, S. 2002. Rates of litter decomposition over 6 years in Canadian forests: Influence of litter quality and climate. Can. J. For. Res. 32: 789-804. doi: 10.1139/X01-117  Turk, T. D., Schmidt, M. G., and Roberts, N. J. 2008. The influence of bigleaf maple on forest  floor and mineral soil properties in a coniferous forest in coastal British Columbia. Forest Ecol. Manag. 255: 1874–1882. doi: 10.1016/j.foreco.2007.12.016  Vaieretti, M. V., Pérez-Harguindeguy, N., Gurvich, D. E., Cingolani, A. M., and Cabido, M.  2005. Decomposition dynamics and physico-chemical leaf quality of abundant species in a montane woodland in central Argentina. Plant Soil 278: 223–234. doi: 10.1007/s11104-005-8432-1  Veen, G. F., Freschet, G. T., Ordonez, A., and Wardle, D. A. 2015. Litter quality and environmental controls of home-field advantage effects on litter decomposition. Oikos 124(2): 187–195. doi: 10.1111/oik.01374  Violle, C., Navas, M., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., and Garnier, E. 2007. Let  the concept of trait be functional! Oikos 116: 882–892. doi: 10.1111/j.2007.0030-1299.15559.x  Voroney, R. P. 2007. The Soil Habitat. In Soil Microbiology, Ecology, and Biochemistry. 3rd ed.  Edited by E. A. Paul. Academic Press, Burlington, Mass. pp.25-49.  Wardle, D. A., Bonner, K. I., and Barker, G. M. 2002. Linkages between plant litter decomposition, litter quality, and vegetation responses to herbivores. Funct. Ecol. 16: 585–595. doi: 10.1046/j.1365-2435.2002.00659.x  Warrington, S. 2015. Methods for ergosterol in tissue. Laboratory protocol from the Analytical  Chemistry Laboratory, Pacific Forestry Centre, Natural Resource Canada, Victoria, BC.    Wickings, K., Grandy, A. S., Reed, S. C., and Cleveland, C. C. 2012. The origin of litter   145 chemical complexity during decomposition. Ecol. Lett. 15: 1180–1188. doi: 10.1111/j.1461-0248.2012.01837.x  Woodruff, D. R., Meinzer, F. C., and McCulloh, K. A. 2010. Height-related trends in stomatal  sensitivity to leaf-to-air vapour pressure deficit in a tall conifer. J. Exp. Bot. 61(1): 203–210. doi: 10.1093/jxb/erp291  Wright, I. J., Reich, P. B., Westoby, M., Ackerly, D. D., Baruch, Z., Bongers, F., Cavender- Bares, J., Chapin, T., Cornelissen, J. H. C., Diemer, M., Flexas, J., Garnier, E., Groom, P. K., Gulias, J., Hikosaka, K., Lamont, B. B., Lee, T., Lee, W., Lusk C., Midgley, J. J., Navas, M., Niinemets, Ü, Oleksyn, J., Osada, N., Poorter, H., Poot, P., Prior, L., Pyankov, V. I., Roumet, C., Thomas, S. C., Tjoelker, M. G., Veneklaas, E. J., and Villar, R. 2004. The worldwide leaf economics spectrum. Nature 428: 821–827. doi: 10.1038/nature02403  Zhang, D., Hui, D., Luo, Y., and Zhou, G. 2008. Rates of litter decomposition in terrestrial  ecosystems: Global patterns and controlling factors. J. Plant Ecol. 1(2): 85–93. doi: 10.1093/jpe/rtn002  Zuur, A. F., Ieno, E. N., and Smith, G. M. 2007. Analysing Ecological Data. Springer, New  York, NY.                        146 Appendix 1: Cuticle Images of Spruce and Subalpine Fir   Figure A1-1. P. engelmannii x glauca and Abies lasiocarpa Cuticle Images. Adaxial (left) and abaxial (right) sides of P. engelmannii x glauca needle (top) and A. lasiocarpa needle (bottom) cross-sections with the cuticle (C), distance to lumen (D), epidermis (E), mesophyll (M), and lumen (L) identified. Images were obtained using a Zeiss Axioplan2 fluorescence microscopy with a Rhodamine B filter cube (excitation: BP 546/12 nm, emission: LP 590 nm) at 400× magnification and processed in ImageJ. 147 Appendix 2: Trait Tables Table A2-1. Physical Foliar Traits. Species mean (with standard error, n = 10 per species) of five physical functional traits measured in foliage.  Species SLA  (m2 kg-1) LDMC  (mg g-1) Leaf Thickness (mm) Fp   (N mm-1) Fps  (N mm-2) A. amabilis 4.74  (± 0.23) 566.65  (± 20.21) 0.41  (± 0.02) 1.67  (± 0.06) 4.12  (± 0.22) A. lasiocarpa 2.76  (± 0.10) 596.97  (± 11.37) 0.78  (± 0.02) 1.13  (± 0.08) 1.48  (± 0.11) A. macrophyllum 12.79  (± 0.50) 487.56  (± 2.57) 0.17  (± 0.01) 0.54  (± 0.08) 3.26  (± 0.49) A. rubra 14.46  (± 1.00) 375.91  (± 11.43) 0.31  (± 0.01) 0.54  (± 0.04) 1.76  (± 0.14) B. papyrifera 24.43  (± 0.68) 284.22  (± 4.53) 0.21  (± 0.01) 0.40  (± 0.06) 1.91  (± 0.28) G. shallon 10.21  (± 1.12) 435.86  (± 19.75) 0.33  (± 0.02) 1.80  (± 0.15) 5.52  (± 0.48) L. occidentalis 12.36  (± 0.35) 411.86  (± 14.88) 0.40  (± 0.01) 1.29  (± 0.08) 3.22  (± 0.18) P. engelmannii  x glauca 2.34  (± 0.09) 514.84  (± 8.81) 0.94  (± 0.01) 1.93  (± 0.08) 2.05  (± 0.09) P. contorta   2.30  (± 0.05) 551.68  (± 4.85) 0.53  (± 0.01) 2.75  (± 0.10) 5.23  (± 0.20) P. ponderosa 1.86  (± 0.07) 533.09  (± 5.88) 0.64  (± 0.01) 2.56  (± 0.09) 4.00  (± 0.12) P. munitum 26.94  (± 1.49) 258.53  (± 7.34) 0.17  (± 0.00) 0.60  (± 0.04) 3.58  (± 0.29) P. balsamifera  10.93  (± 0.44) 491.86  (± 17.86) 0.27  (± 0.01) 0.73  (± 0.08) 2.70  (± 0.27) P. tremuloides 23.37  (± 0.43) 350.29  (± 6.66) 0.16  (± 0.00) 0.76  (± 0.08) 5.56  (± 0.66) P. menziesii  5.33  (± 0.22) 434.18  (± 9.00) 0.54  (± 0.01) 1.25  (± 0.07) 2.31  (± 0.11) T. plicata 4.61  (± 0.22)  414.81  (± 6.10) 0.58  (± 0.03) 1.76  (± 0.13) 3.10  (± 0.26) T. heterophylla 8.21  (± 0.32) 434.81  (± 11.47) 0.33  (± 0.01) 1.01  (± 0.03) 2.91  (± 0.13)  SLA = specific leaf area LDMC = leaf dry matter content Fp = force-to-punch Fps = specific force-to-punch  148 Table A2-2. Physical Litter Traits. Species mean (with standard error, n = 10 per species) of four physical traits measured in leaf litter.   Species SLA  (m2 kg-1) Leaf Thickness (mm) Fp  (N mm-1) Fps  (N mm-2) A. amabilis 4.85  (± 0.40) 0.33  (± 0.02) 1.13  (± 0.10) 4.07  (± 0.59) A. lasiocarpa 3.36  (± 0.20) 0.43  (± 0.03) 0.87  (± 0.08) 2.07  (± 0.16) A. macrophyllum 23.96  (± 3.11)  0.15  (± 0.01) 0.25  (± 0.03) 1.76  (± 0.21) A. rubra 10.14  (± 0.71) 0.33  (± 0.02) 0.36  (± 0.02) 1.12  (± 0.08) B. papyrifera 17.07  (± 1.22) 0.18  (± 0.01) 0.21  (± 0.02) 1.17  (± 0.09) G. shallon 11.27  (± 0.50) 0.23  (± 0.01) 0.95  (±0.13) 4.26  (± 0.67) L. occidentalis 20.38  (± 0.89) 0.32  (± 0.02) 0.62  (± 0.08) 1.89  (± 0.20)  P. engelmannii x glauca 1.62  (± 0.07) 0.82  (± 0.02) 2.47  (± 0.17) 3.01  (± 0.18) P. contorta  1.45  (± 0.14) 0.67  (± 0.02) 3.74  (± 0.12) 5.58  (± 0.19) P. ponderosa 1.96  (± 0.14) 0.76  (± 0.03) 5.16  (± 0.40) 6.92  (± 0.70) P. munitum 15.38  (± 0.81) 0.12  (± 0.01) 0.53  (± 0.02) 4.36  (± 0.30) P. balsamifera 11.91  (± 0.67) 0.23  (± 0.01) 0.59  (± 0.05) 2.53  (± 0.20) P. tremuloides 13.99  (± 0.70) 0.17  (± 0.00) 0.51  (± 0.05) 3.09  (± 0.31) P. menziesii  4.45  (± 0.34) 0.43  (± 0.02) 1.79  (± 0.18) 4.17  (± 0.30) T. plicata 3.27  (± 0.24) 0.87  (± 0.02) 2.20  (± 0.19)  2.54  (± 0.23) T. heterophylla 5.62  (± 0.62) 0.29  (± 0.01) 0.89  (± 0.08) 3.10  (± 0.20)  SLA = specific leaf area Fp = force-to-punch Fps = specific force-to-punch    149 Table A2-3. Chemical Foliar Traits. Species mean (with standard error when applicable) of eight chemical functional traits measured in foliage.   Species C* N*  C:N* pH* NPE** WSE** ASC** AUR** A. amabilis 0.5108 (±0.0145) 0.0105 (± 0.0002) 48.68 (± 0.99) 3.86 (± 0.00) 0.029  0.511  0.262  0.190 A. lasiocarpa 0.5040 (±0.0044) 0.0133 (± 0.0002)  37.95 (± 0.34)  4.50 (± 0.01) 0.124  0.377  0.294  0.198 A. macrophyllum 0.4732 (±0.0065) 0.0218 (± 0.0004) 21.71 (± 0.22) 4.72 (± 0.01) 0.002  0.354  0.400  0.236 A. rubra 0.5044 (±0.0073) 0.0305 (± 0.0005) 16.55 (± 0.14) 4.70 (± 0.00) 0.044  0.423  0.359  0.167 B. papyrifera 0.4783 (± 0.0021) 0.0281 (± 0.0001) 17.01 (± 0.13) 4.94 (± 0.02) 0.051  0.358  0.359  0.216 G. shallon 0.4694 (± 0.0049) 0.0111 (± 0.0002) 42.34 (± 1.33) 5.22 (± 0.01) 0.024  0.297  0.449  0.219 L. occidentalis 0.4816 (± 0.0055) 0.0194 (± 0.0003) 24.86 (± 0.25) 3.89 (± 0.02) 0.048  0.425  0.308 0.193 P. engelmannii  x glauca  0.4925 (± 0.0075)  0.0098 (± 0.0002)  50.48 (± 1.37)  4.53 (± 0.01) 0.050  0.349  0.391   0.201 P. contorta 0.5068 (± 0.0108) 0.0110 (± 0.0004) 46.40 (± 1.17) 4.05 (± 0.01) 0.052  0.324  0.376  0.237 P. ponderosa 0.5117 (± 0.0014) 0.0103 (± 0.0001) 49.64 (± 0.45) 3.97 (± 0.00) 0.089  0.343  0.353  0.199 P. munitum 0.4214 (± 0.0349) 0.0196 (± 0.0016) 21.49 (± 0.18) 6.24 (± 0.02) 0.010  0.182  0.525  0.277 P. balsamifera 0.5023 (±0.0044) 0.0236 (± 0.0003) 21.32 (± 0.22) 5.34 (± 0.01) 0.142  0.351  0.350  0.144 P. tremuloides 0.4920 (± 0.0044) 0.0294 (± 0.0003) 16.76 (± 0.04) 5.69 (± 0.01) 0.051  0.344  0.384  0.210 P. menziesii  0.5046(± 0.0029) 0.0153 (± 0.0001) 32.93 (± 0.27) 4.04 (± 0.01) 0.058  0.357  0.324  0.248 T. plicata 0.4830 (± 0.0066) 0.0086 (± 0.0002) 55.92 (± 0.77) 5.42 (± 0.01) 0.049  0.327  0.398  0.215 T. heterophylla 0.5160 (± 0.0078) 0.0123 (± 0.0002) 42.00 (± 0.37) 3.86 (± 0.01) 0.289  0.173  0.306  0.218 * = 4 technical replicates of composite samples per species ** = 1 technical replicate of composite samples per species C = carbon, proportion  N = nitrogen, proportion NPE = non-polar extractables, proportion WSE = water-soluble extractables, proportion ASC = acid-soluble fraction (proportion hydrolyzed)  AUR = acid-unhydrolyzable residue, proportion    150 Table A2-4. Chemical Litter Traits. Species mean (with standard error when applicable) of eight chemical traits measured in leaf litter.  Species C* N* C:N* pH* NPE** WSE** ASC** AUR** A. amabilis 0.5190 (± 0.0044)  0.0077 (± 0.0002)  67.44 (± 1.49)  3.64 (± 0.01) 0.023 ***  0.252 ***  0.357 *** 0.309 *** A. lasiocarpa 0.5253 (± 0.0022) 0.0065 (± 0.0004) 81.99 (± 4.32) 3.96 (± 0.02) 0.264 0.187 0.319 0.219 A. macrophyllum 0.5078 (± 0.0026) 0.011 (± 0.0001) 46.51 (± 0.48) 4.50 (± 0.02) 0.040  0.213  0.344  0.382 A. rubra 0.5197 (± 0.0025)  0.0233 (± 0.0008) 22.4 (± 0.68) 4.70 (± 0.01) 0.105  0.285  0.333 0.265 B. papyrifera 0.4887 (±0.0060) 0.0065 (± 0.0003) 74.94 (± 2.17) 4.87 (± 0.01) 0.142  0.257  0.348  0.238 G. shallon 0.4773 (± 0.0028) 0.0097 (± 0.0001) 49.19 (± 0.80) 5.87 (± 0.04) 0.021  0.132  0.495  0.348 L. occidentalis 0.4728 (±0.0027) 0.0058 (±0.0003) 82.02 (± 3.83) 4.07 (± 0.01) 0.042  0.416  0.337  0.189 P. engelmannii  x glauca 0.4927 (± 0.0037) 0.0059 ±(0.0002) 83.21 (± 1.94) 4.28 (± 0.01) 0.059 0.383 0.373  0.168 P. contorta 0.5202 (± 0.0021) 0.0050 (± 0.0001) 104.1 (± 1.52) 4.06 (± 0.02) 0.073  0.189  0.413 0.321 P. ponderosa 0.5141 (± 0.0074) 0.0081 (± 0.0004) 64.1 (± 3.44) 4.44 (± 0.03) 0.083  0.221  0.417  0.268 P. munitum 0.4859 (± 0.0034) 0.0145 (± 0.0002) 33.53(± 0.50) 5.88 (± 0.02) 0.004  0.096  0.524  0.368 P. balsamifera 0.4695 (± 0.0077) 0.0101 (± 0.0002) 46.49 (± 0.39) 5.70 (± 0.02) 0.023  0.374  0.356  0.241 P. tremuloides 0.5026 (±0.0032) 0.0107 (± 0.0003) 47.22 (± 1.57) 5.71 (± 0.02)  0.056  0.240  0.394  0.304 P. menziesii 0.5064 (± 0.0051) 0.0107 (± 0.0002) 47.22 (± 0.25) 5.30 (± 0.02) 0.046  0.127  0.445  0.371 T. plicata 0.4868 (±0.0033) 0.0088 (± 0.0003) 55.57 (± 1.70) 4.13 (± 0.02) 0.070  0.266  0.415 0.241 T. heterophylla 0.5235 (±0.0059) 0.0071 (± 0.0003) 73.74 (± 3.44) 3.91 (± 0.01) 0.051  0.239  0.364 0.338 * = 4 technical replicates of composite samples, ** = 1 technical replicate of composite samples *** = only 0.17 g of litter used due to limited supply of sample C = proportion of carbon  N = proportion of nitrogen NPE = non-polar extractables, proportion WSE = water-soluble extractables, proportion ASC = acid-soluble fraction (proportion hydrolyzed)  AUR = acid-unhydrolyzable residue, proportion  151 Table A2-5. Water Uptake and Leaching Loss. Mean proportion of water taken up and proportion of mass lost after leaching for 2 and 24 hours (with standard error, n = 3 per species). Species Water Uptake 2 Hours Water Uptake 24 Hours Leaching Loss 2 Hours Leaching Loss 24 Hours A. amabilis 0.3434  (± 0.0111) 0.5572  (± 0.0034) 0.0444  (± 0.0111) 0.0333*  (± 0.0000) A. lasiocarpa 0.2459  (± 0.0105) 0.4910  (± 0.0051)  0.0100  (± 0.0000) 0.0500  (± 0.0111) A. macrophyllum 0.5688  (± 0.0247) 0.7644  (± 0.0361) 0.0304  (± 0.0070) 0.0685  (± 0.0044) A. rubra 0.3545  (± 0.0232) 0.7044  (± 0.0122) 0.0698  (± 0.0459) 0.3545  (± 0.0059) B. papyrifera 0.5986  (± 0.0323) 0.7683  (± 0.0098) 0.0402  (± 0.0057) 0.1308  (± 0.0096) G. shallon 0.3545  (± 0.0243) 0.5939  (± 0.0052) 0.0134  (± 0.0033) 0.0465  (± 0.0087) L. occidentalis 0.5709  (± 0.0044) 0.7716  (± 0.0073) 0.0433  (± 0.0033) 0.1367  (± 0.0033) P. engelmannii  x glauca 0.2324  (± 0.0068) 0.5222  (± 0.0050) 0.0200  (± 0.0100) 0.0333  (± 0.0088) P. contorta   0.2285  (± 0.0053) 0.4202  (± 0.014) 0.0100  (± 0.0000) 0.0000  (± 0.0000) P. ponderosa 0.182  (± 0.0265) 0.3749  (± 0.0225) 0.0188  (± 0.0094) 0.0183  (± 0.0091) P. munitum 0.6369  (± 0.0153) 0.7283  (± 0.0029) 0.0034  (± 0.0168) 0.0133  (± 0.0067) P. balsamifera   0.4727  (± 0.0072) 0.7142  (± 0.0037) 0.0331  (± 0.0031) 0.1744  (± 0.0117) P. tremuloides 0.4463  (± 0.0174) 0.6498  (±0.0029) 0.0336  (± 0.0035) 0.1262  (± 0.0062) P. menziesii   0.3652  (± 0.0070) 0.6091  (± 0.0006) 0.0100  (±0.0000) 0.02667  (± 0.0120) T. plicata 0.19  (± 0.0090) 0.451  (± 0.0142) 0.0067  (± 0.0033) 0.0167  (± 0.0033) T. heterophylla 0.4804  (± 0.0198) 0.6308  (± 0.0089) 0.0367  (± 0.0033) 0.0533  (± 0.0067)  * n = 2 samples due to limited supply of litter    152 Table A2-6: Cuticle Thickness and DTL. Mean cuticle thickness and distance to lumen (DTL) of foliage from 16 species (µm, with standard error, n = 108 cells collectively from 3 leaves per species)  Species Abaxial Cuticle  Adaxial Cuticle  Abaxial DTL Adaxial DTL A. amabilis 3.22  (± 0.17) 3.71  (± 0.22) 5.36  (± 0.17) 7.04  (± 0.21) A. lasiocarpa 5.88  (± 0.35) 6.23  (± 0.39) 12.17  (± 0.61)  12.73  (± 0.63) A. macrophyllum 1.31  (± 0.06)  2.06  (± 0.12) 1.69  (± 0.03) 3.02  (± 0.17) A. rubra 1.62  (± 0.12) 2.30  (± 0.12) 1.85  (± 0.10) 2.75  (± 0.12) B. papyrifera 1.55  (± 0.08) 2.31  (± 0.17) 2.30  (± 0.13) 3.13  (± 0.27) G. shallon 2.62  (± 0.20) 3.55  (± 0.27) 4.24  (± 0.35) 7.40  (± 0.23) L. occidentalis 1.55  (± 0.09) 1.78  (± 0.10) 3.29  (± 0.16) 3.25  (± 0.11) P. engelmannii  x glauca 3.59  (± 0.22) 3.80  (± 0.27) 11.50  (± 0.68) 12.06  (± 0.58) P. contorta   3.61  (± 0.10) 3.59  (± 0.11) 4.07  (± 0.30) 3.88  (± 0.22) P. ponderosa 3.60  (± 0.15) 3.48  (± 0.18) 4.20  (± 0.25) 3.48  (± 0.18) P. munitum 1.41  (± 0.14) 1.21  (± 0.08) 2.62  (± 0.17) 3.95  (± 0.16)  P. balsamifera   2.39  (± 0.16) 2.08  (± 0.13) 3.28  (± 0.40) 3.84  (± 0.17)  P. tremuloides 3.21  (± 0.25) 3.56  (± 0.23) 4.02  (± 0.20) 5.13  (± 0.31)  P. menziesii   4.02  (± 0.18) 4.48  (± 0.17) 6.65  (± 0.24)  7.49  (± 0.15)  T. plicata 5.16  (± 0.32) 7.16  (± 0.49) 7.18  (± 0.35) 8.56  (± 0.35) T. heterophylla 2.73  (± 0.17) 4.09  (± 0.18) 5.55  (± 0.16) 7.13  (± 0.16)       153 Table A2-7: Principal Component 1 Variable Values. Values of the first principal component value (PC1) determined for each of 13 woody plant species. Principal component analyses were run with foliar traits (F) and litter traits (L) using all traits measured (AT) or just a select suite of traits measured in both foliar and litter (ST). Refer to tables 3.1 through 3.4 for more information about each PCA.  Species PC1: F, AT  PC1: L, AT  PC1: F, ST PC1: L, ST A. amabilis 2.56 -0.79 2.19 0.77 A. macrophyllum -3.03 0.09 -2.29 -2.17 A. rubra -3.29 3.34 -2.16 -2.30 B. papyrifera -4.33 2.74 -3.40 -1.78 G. shallon 0.53 -0.96 0.00 0.18 L. occidentalis -1.51 2.83 -0.27 -1.90 P. contorta  2.74 -4.43 2.91 3.76 P. ponderosa 2.94 -4.52 3.36 3.78 P. balsamifera  -1.68 2.47 -1.45 -2.13 P. tremuloides -2.52 1.38 -3.16 -1.29 P. menziesii  2.10 -1.75 1.17 1.25 T. plicata 3.77 -2.69 1.55 1.37 T. heterophylla 1.74 -0.06 1.55 0.47     154 Appendix 3: Principal Components Analysis and Cluster Analysis with Polystichum munitum Table A3-1. Correlations Between Foliar Traits and Two PCs in PCA Including P. munitum. Correlations (r, ranging from -1 to 1) between each foliar trait and the first two principal component dimensions (PC) produced in a principal components analysis (PCA) using all plant species except for A. lasiocarpa and P. engelmannii x. glauca. Correlations greater than 0.5 or less than -0.5 are marked in bold. Supplementary variables include proportion of litter mass lost during Phase I, Phase II, and the first year. PC1 explained 46% of the variation in foliar traits among species, and PC2 explained 20%.  Foliar Trait PC1 PC 2 SLA -0.92 0.19 LDMC 0.73 -0.36 Fp 0.81 0.15 Fps 0.24 0.42 Leaf Thickness 0.87 -0.01 N -0.83 -0.37 C 0.65 -0.64 C:N 0.91 0.28 pH -0.60 0.58 Abaxial Cuticle Thickness 0.84 0.30 Adaxial Cuticle Thickness 0.77 0.33 Abaxial DTL 0.82 0.32 Adaxial DTL 0.63 0.46 WSE 0.17 -0.65 NPE 0.28 -0.22 ASC -0.50 0.76 AUR -0.10 0.74 Phase-I Loss -0.14 -0.54 Phase-II Loss -0.17 -0.23 Year-One Loss -0.19 -0.41      155  Figure A3-1. Ordination of All Foliar Traits in PCA Including P. munitum. Ordination plot of a PCA run on 14 plant species with all foliar traits. Proportion of mass lost during Phase I, Phase II, and during the first year were added as supplementary variables. 156 Table A3-2. Correlations Between Litter Traits and Two PCs in PCA Including P. munitum. Correlations (r, ranging from -1 to 1) between each litter trait and the first two principal component dimensions (PC) produced in a principal components analysis (PCA) using all plant species except for A. lasiocarpa and P. engelmannii x. glauca. Correlations greater than 0.5 or less than -0.5 are marked in bold. Supplementary variables include proportion of litter mass lost during Phase I, Phase II, and the first year. PC1 explained 40% of the variation in litter traits among species, and PC2 explained 25%.  Litter Trait PC1 PC 2 SLA 0.82 -0.21 Fp -0.89 0.24 Fps -0.86 -0.25 Leaf Thickness -0.78 0.45 N 0.40 -0.24 C -0.38 0.25 C:N -0.41 0.50 pH 0.28 -0.74 WSE 0.46 0.75 NPE 0.04 0.61 ASC -0.48 -0.80 AUR -0.24 -0.74 Leaching Loss 2 Hrs 0.66 0.56 Leaching Loss 24 Hrs 0.81 0.35 Water Uptake 2 Hrs 0.78 -0.32 Water Uptake 24 Hrs 0.94 -0.20 Phase-I Loss 0.55 0.58 Phase-II Loss 0.26 0.25 Year-One Loss 0.58 0.45      157  Figure A3-2. Ordination of All Litter Traits in PCA Including P. munitum. Ordination plot of a PCA run on 14 plant species with all litter traits. Proportion of mass lost during Phase I, Phase II, and the first year were added as supplementary variables.  158  Figure A3-3. Cluster Analysis with Foliar Traits. Cluster analysis using 11 foliar traits to group 14 plant species native to British Columbia. These traits, which were measured in both foliage and litter, include SLA, thickness, Fps, N, C, C:N, pH, AUR, ASC, WSE, and NSE.   159  Figure A3-4. Cluster Analysis with Litter Traits. Cluster analysis using 11 litter traits to group 14 plant species native to British Columbia. These traits, which were measured in both foliage and leaf litter, include SLA, thickness, Fps, N, C, C:N, pH, AUR, ASC, WSE, and NSE. 160 Appendix 4: Regression Tree Analysis Tables   Table A4-1. Regression Tree Analysis to Predict Year-One Mass Loss. Nodes and splits resulting from regression tree analysis of woody plant species (n = 64) using both foliar traits and leaf-litter traits to predict proportion of mass lost from 0 to 12 months. This model explains 68% of variation in mass loss during the first year. Refer to Figure 3.14 for the tree diagram of this regression tree analysis.   Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 1.6 0.41 0.026   2 Litter Leaching Loss 24 Hrs < 0.097 30 0.20 0.31 0.0052    4* Foliar Adaxial DTL < 7.27 24 0.030 0.27 0.0011 Aa, Am, Pc, Pp, Th   5* Foliar Adaxial DTL ≥ 7.27 15 0.071 0.38 0.0047 Gs, Pmz, Tp  3  Litter Leaching Loss 24 Hrs ≥  0.097 25 0.55 0.56 0.022    6* Foliar Abaxial DTL ≥  2.79 15 0.18 0.50 0.012 Pb, Pt, Lo   7* Foliar Abaxial DTL< 2.79 10 0.25 0.64 0.025 Ar, Bp   * = terminal node ** = mean squared error      161 Table A4-2. Regression Tree Analysis to Predict Year-One Loss Using Only Physical Traits. Nodes and splits resulting from regression analysis of woody plant species (n = 64) using both physical foliar traits and physical litter traits to predict proportion of mass lost from 0 to 12 months. This model explains 68% of variation in mass loss during the first year. Refer to Figure 3.15 for the tree diagram of this regression tree analysis.    Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 1.6 0.41 0.026   2 Litter Leaching Loss 24 Hrs < 0.097 30 0.20 0.31 0.0052    4* Foliar Adaxial DTL < 7.27 24 0.030 0.27 0.0011 Aa, Am, Pc, Pp, Th   5* Foliar Adaxial DTL ≥ 7.27 15 0.071 0.38 0.0047 Gs, Pmz, Tp  3  Litter Leaching Loss 24 Hrs ≥  0.097 25 0.55 0.56 0.022    6* Foliar Abaxial DTL ≥  2.79 15 0.18 0.50 0.012 Pb, Pt, Lo   7* Foliar Abaxial DTL< 2.79 10 0.25 0.64 0.025 Ar, Bp   * = terminal node ** = mean squared error       162 Table A4-3. Regression Tree Analysis to Predict Year-One Loss Using Only Chemical Traits. Nodes and splits resulting from regression analysis of woody plant species (n = 64) using both chemical foliar traits and chemical litter traits to predict proportion of mass lost from 0 to 12 months. This model explains 64% of variation in mass loss during the first year. Refer to Figure 3.15 for the tree diagram of this regression tree analysis.    Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 1.6 0.41 0.026   2 Foliar C:N ≥  21.52 44 0.31 0.33 0.0071    4* Litter C ≥  0.5 29 0.086 0.29 0.023 Aa, Am, Pc, Pmz, Pp, Th   5* Litter C < 0.5 15 0.055 0.42 0.0030 Gs, Lo, To  3* Foliar C:N < 21.52 20 0.45 0.58 0.0037 Ar, Bp, Pb, Pt, * = terminal node ** = mean squared error                163 Table A4-4. Regression Tree Analysis to Predict Phase-I Mass Loss. Nodes and splits resulting from regression tree analysis of woody plant species (n = 64) using both foliar traits and litter traits to predict proportion of mass lost after 3 months (Phase I). This model explains 71% of variation in Phase-I mass loss. Refer to Figure 3.16 for the tree diagram of this regression tree analysis.    Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 0.38 0.21 0.0060   2 Litter WSE < 0.25 39 0.086 0.16 0.0022    4* Litter WSE < 0.23 24 0.021 0.14 0.0023 Am, Gs, Pc, Pmz, Pp,   5* Litter WSE ≥ 0.23 15 0.034 0.20 0.0009 Aa, Pt, Th  3* Litter WSE ≥ 0.25 25 0.058 0.29 0.0022 Ar, Bp, Lo, Pb, Tp *= terminal node ** = mean squared error                  164 Table A4-5. Regression Tree Analysis to Predict Phase-I Loss Using Only Physical Traits. Nodes and splits resulting from regression tree analysis of woody plant species (n = 64) using both physical foliar traits and physical litter traits to predict proportion of mass lost after 3 months. This model explains 70% of variation in mass loss over one year. Refer to Figure 3.17 for the tree diagram of this regression tree analysis.    Node Split N Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 0.38 0.21 0.0060   2 Litter Leaching Loss 24 Hrs < 0.09735 39 0.082 0.16 0.0021    4* Foliar Adaxial Cuticle < 3.65 19 0.015 0.13 0.0025 Am, Gs, Pc, Pp   5* Foliar Adaxial Cuticle ≥  3.65 20 0.037 0.19 0.0008 Aa, Pmz, Th, Tp  3* Litter Leaching Loss 24 Hrs ≥   0.097 25 0.063 0.29 0.0018 Ar, Bp, Lo, Pb, Pt * = terminal node ** = mean squared error              165 Table A4-6. Regression Tree Analysis to Predict Phase-I Loss Using Only Chemical Traits. Nodes and splits resulting from regression tree analysis of woody plant species (n = 64) using both chemical foliar traits and chemical litter traits to predict proportion of mass lost after 3 months (Phase I). This model explains 71% of variation in Phase-I mass loss. Refer to Figure 3.17 for the tree diagram of this regression tree analysis.    Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 0.38 0.21 0.0060   2 Litter WSE < 0.25 39 0.086 0.16 0.0022    4* Litter WSE < 0.23 24 0.021 0.14 0.0023 Am, Gs, Pc, Pmz, Pp,   5* Litter WSE ≥  0.23 15 0.034 0.20 0.0009 Aa, Pt, Th  3* Litter WSE ≥   0.25 25 0.059 0.29 0.0022 Ar, Bp, Lo, Pb, Tp * = terminal node ** = mean squared error                166 Table A4-7. Regression Tree Analysis to Predict Phase-II Mass Loss. Nodes and splits resulting from regression tree analysis of woody plant species (n = 64) using both foliar traits and litter traits to predict proportion of mass lost from 3 to 12 months (Phase II). This model explains 44% of variation in Phase-II mass loss. Refer to Figure 3.18 for the tree diagram of this regression tree analysis.    Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 0.9030609 0.20 0.014   2 Foliar LDMC ≥    363 54 0.3405333 0.17 0.0063    4* Litter pH < 4.6 34 0.1159882 0.14 0.025 Aa, Am, Lo, Pc, Pp, Th, Tp   5* Litter pH ≥   4.6 20 0.1447800 0.22 0.0034 Ar, Gs, Pb, Pmz,  3* Foliar LDMC < 363 10 0.2446100 0.36 0.0072 Bp, Pt * = terminal node                   167 Table A4-8. Regression Tree Analysis to Predict Phase-II Loss Using Only Physical Traits. Nodes and splits resulting from regression tree analysis of woody plant species (n = 64) using both physical foliar traits and physical litter traits to predict proportion of mass lost from 3 to 12 months (Phase II). This model explains 35% of variation in Phase-II mass loss. Refer to Figure 3.19 for the tree diagram of this regression tree analysis.    Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 0.90 0.20 0.014   2* Foliar LDMC ≥   363 54 0.34 0.17 0.0063 Aa, Am, Ar, Gs, Lo, Pb, Pc, Pmz, Pp, Th, Tp    3* Foliar LDMC < 363 10 0.24 0.36 0.025 Bp, Pt * = terminal node ** = mean squared error                 168 Table A4-9. Regression Tree Analysis to Predict Phase-II Loss Using Only Chemical Traits. Nodes and splits resulting from regression tree analysis of woody plant species (n = 64) using both chemical foliar traits and chemical litter traits to predict proportion of mass lost from 3 to 12 months (Phase II). This model explains 40% of variation in Phase-II mass loss. Refer to Figure 3.19 for the tree diagram of this regression tree analysis.    Node Split n Deviance Mean Mass Loss (Proportion) MSE** Species 1 root (no split) 64 0.90 0.20 0.014   2 Foliar N < 0.025845 49 0.27 0.16 0.0055    4* Litter pH < 5.5 39 0.17 0.14 0.021 Aa, Am, Lo, Pc, Pmz, Pp, Th, Tp   5* Litter pH ≥ 5.5 10 0.05 0.22 0.0043 Pb, Gs  3* Foliar N ≥ 0.025845   15 0.32 0.33 0.0053 Ar, Bp, Pt  * = terminal node ** = MSE       169 Appendix 5: Linear Regressions Table A5-1. Regression Analysis for Year-One Mass Loss Using Foliar Traits. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using foliar trait variables to predict the proportion of mass lost during the first year (ln-transformed). The cutoff p-value (p < 0.0014) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold. Trait Coefficient (b) R2 p SEE Assumptions Foliar Abaxial Cuticle -0.052 8.5% 0.020 0.35 Yes Foliar Abaxial DTL -0.074 11% 0.0063 0.34 Yes Foliar Adaxial Cuticle -0.060 5.3% 0.067 0.35 Yes, possible outlier Foliar Adaxial DTL -0.047 7.1% 0.033 0.35 Yes Foliar ASC 1.5 3.7% 0.13 0.36 Yes Foliar AUR -4.6 13% 0.0037 0.34 Yes Foliar C -7.9 12% 0.0059 0.34 Yes Foliar C:N -0.015 34% 4.565 × 10-7 0.29 Yes Foliar Fps Foliar Fps (ln) -0.077 -0.35 7.30% 13% 0.031 0.0033 0.35 0.34 Not equal variance Yes Foliar LDMC -0.0033 53% 1.2 × 10-11 0.25 Yes Foliar N (ln) 0.50 (0.49*) 38% (33%*) 6.1 × 10-8 (2.3 × 10-6*) 0.29 (0.29*) Yes Foliar NPE  Foliar NPE (ln, no Th) -0.77 (0.15) 2.40% (13.21%) 0.23 (0.0047) 0.36 (0.34) Outlier Outlier Foliar pH 0.29 29.55% 3.4 × 10-6 0.30 Not normal Foliar SLA 0.035 47.56% 2.9 × 10-10 0.26 Yes Foliar Thickness (ln) -0.31 15.03% 0.0016 0.33 Yes Foliar WSE 0.69 2.11% 0.25 0.36 Yes PC1 Foliage All Traits -0.080 36.64% 1.2 × 10-7 0.29 Not normal PC1 Foliage Same Traits -0.11 43.58% 2.9 × 10-9 0.27 Not normal * without A. rubra  170 Table A5-2. Regression Analysis for Year-One Mass Loss Using Litter Traits. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using litter trait variables to predict the proportion of mass lost over one year (ln-transformed). The cutoff p-value (p < 0.0014) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold. Trait Coefficient (b) R2 p SEE Assumptions Litter ASC -1.5 4.1% 0.11 0.36 Mostly equal variance Litter AUR -3.5 29% 5.0 × 10-6 0.31 Nearly normal Litter C -9.2 22% 8.0 × 10-5 0.32 Not normal Litter C:N -0.0046 6.7% 0.039 0.35 Yes Litter Fps Litter Fps (ln) -0.12 -0.39 30% 34% 3.4 × 10-6 4.2 × 10-7 0.30 0.29 Not equal variance Yes Litter Leaching 2 Hrs 7.1 12% 0.0051 0.34 Yes Litter Leaching 24 Hrs 3.3 42% 5.7 × 10-9 0.28 Yes Litter N (ln) 0.27 (0.11*) 7.3% (0.63%*) 0.31 (0.55*) 0.35 (0.36*) Yes Litter NPE (ln) 0.20 11% 0.0089 0.34 Yes Litter pH 0.22 20% 2.4 ×  10-4 0.32 Yes Litter SLA 0.019 (0.040**) 13% (47%**) 0.0032 (2.0 × 10-9**) 0.34 (0.26**) Am is outlier Litter Thickness (ln) -0.21 11% 0.0088 0.34 Yes Litter Water 2 Hrs 0.87 11% 0.0077 0.34 Yes Litter Water 24 Hrs 1.3 21% 1.6 ×  10-4 0.32 Yes Litter WSE 1.6 13% 0.0039 0.34 Yes PC 1 Litter All Traits 0.071 29% 5.39 × 10-6 0.31 Yes PC1 Litter Same Traits -0.094 29% 3.58 × 10-6 0.30 Yes * without A. rubra ** without A. macrophyllum     171 Table A5-3. Regression Analysis for Phase-I Mass Loss Using Foliar Traits. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using foliar trait variables to predict the proportion of mass lost during Phase I (ln-transformed). The cutoff p-value (p < 0.0014) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold.  Trait Coefficient (b) R2 p SEE Assumptions Foliar Abaxial Cuticle -0.10 8.6% 0.019 0.37 Yes Foliar Abaxial DTL -0.054 5.3% 0.068 0.38 Yes Foliar Adaxial Cuticle -0.051 3.3% 0.15 0.38 Yes (but possible outlier) Foliar Adaxial DTL -0.037 3.9% 0.12 0.38 Nearly normal Foliar ASC -0.93 1.3% 0.38 0.39 Mostly – variance Foliar AUR -7.5 29% 4.4 × 10-6 0.33 Yes Foliar C -5.0 3.9% 0.12 0.38 Yes  Foliar C:N -0.015 29% 4.5 × 10-6 0.33 Yes Foliar Fps Foliar Fps (ln) -0.14 -0.53 22% 25% 8.2 × 10-5 2.2 × 10-5 0.34 0.34 Yes Yes Foliar LDMC -0.0030 39% 3.4 × 10-8 0.30 Yes Foliar N (ln) 0.50 (0.47*) 33% (26%*) 6.4 × 10-7 (3.3 × 10-5*) 0.32 (0.33) Yes Foliar NPE (ln) 0.083 4.7% 0.086 0.38 Yes (but outlier) Foliar pH 0.23 16% 0.0011 0.36 Yes Foliar SLA 0.031 33% 8.2 × 10-7 0.32 Yes Foliar Thickness (ln) -0.29 11% 0.0084 0.37 Yes Foliar WSE 1.2 5.4% 0.064 0.38 Nearly normal PC1 Foliar All Traits -0.077 29% 4.9 × 10-6 0.33 Yes PC1 Foliar Same Traits -0.10 34% 3.6 × 10-7 0.32 Nearly normal * without A. rubra     172 Table A5-4. Regression Analysis for Phase-I Mass Loss Using Litter Traits. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using litter trait variables to predict the proportion of mass lost during Phase I (ln-transformed). The cutoff p-value (p < 0.0014) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold.  Trait Coefficient (b) R2 p SEE Assumptions Litter ASC -4.0 24% 4.0 × 10-5 0.34 Nearly normal Litter AUR -4.6 43% 3.5 × 10-9 0.29 Nearly normal Litter C -9.6 21% 1.6 ×  10-4 0.35 Yes Litter C:N -0.0057 9.0% 0.016 0.37 Yes Litter Fps  Litter Fps (ln) -0.16 -0.49 50% 48% 7.0 × 10-11 2.118 × 10-10 0.28 0.28 Not normal Not normal Litter Leaching 2 Hrs 11.5 27% 1.024 × 10-5 0.33 Yes Litter Leaching 24 Hrs 3.9 51% 2.969 × 10-11 0.27 Yes Litter N (ln) 0.31 (0.11*) 8.5% (0.50%*) 0.019 (0.60*) 0.37 (0.38*) Outlier (Ar) Yes Litter NPE (ln) 0.13 3.9% 0.12 0.38 Yes Litter pH 0.090 2.9% 0.18 0.38 Yes Litter SLA  0.022 (0.042**) 15% (43%**) 0.0016 (1.1 × 10-8**) 0.36 (0.29**) Yes Nearly normal Litter Thickness (ln) -0.19 7.5% 0.029 0.38 Yes Litter Water 2 Hrs 1.1 16% 0.00094 0.36 Yes Litter Water 24 Hrs 1.6 30% 3.2 × 10-6 0.33 Yes Litter WSE 3.2 42% 6.1 × 10-9 0.30 Yes PC 1 Litter All Traits 0.095 43% 3.7 × 10-9 0.29 A little skewed PC1 Litter Same Traits -0.12 43% 3.1 × 10-9 0.29 A little skewed * without A. rubra ** without A. macrophyllum    173 Table A5-5. Regression Analysis for Phase-II Mass Loss Using Foliar Traits. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using foliar trait variables to predict the proportion of mass lost during Phase II (ln-transformed). The cutoff p-value (p < 0.0014) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold.  Trait Coefficient (b) R2 p SEE Assumptions  Foliar Abaxial Cuticle -0.078 2.3% 0.23 0.56 Yes Foliar Abaxial DTL -0.10 8.5% 0.020 0.54 Yes Foliar Adaxial Cuticle -0.065 2.6% 0.21 0.56 Yes Foliar Adaxial DTL -0.064 5.4% 0.064 0.55 Yes Foliar ASC 4.3 13% 0.0040 0.53 Yes Foliar AUR -1.7 0.74% 0.50 0.57 Yes Foliar C -10.1 7.6% 0.028 0.55 Yes Foliar C:N -0.015 14% 0.0023 0.53 Yes Foliar Fps  Foliar Fps (ln) -0.012 -0.19 0.07% 1.6% 0.84 0.32 0.57 0.56 Not equal variance Yes Foliar LDMC -0.0034 24% 4.7 × 10-5 0.50 Nearly normal Foliar N (ln) 0.51 (0.50*) 16% (14%*) 0.0011 (0.0041*) 0.52 (0.54*) Yes Foliar NPE (ln) -1.7 4.7% 0.086 0.55 Yes (but outlier) Foliar pH 0.37 19% 2.8 ×  10-4 0.51 Yes Foliar SLA 0.039 23% 5.8 × 10-5 0.50 Yes Foliar Thickness (ln) -0.34 7.3% 0.031 0.55 Yes Foliar WSE 0.12 0.03% 0.90 0.57 Yes PC1 Foliage All Traits -0.084 16% 9.2 ×  10-4 0.52 Yes PC1 Foliage Same Traits -0.11 19% 2.7 ×  10-4 0.51 Yes * without A. rubra       174 Table A5-6. Regression Analysis for Phase-II Mass Loss Using Litter Traits. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using litter trait variables to predict the proportion of mass lost during Phase II (ln-transformed). The cutoff p-value (p < 0.0014) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (36). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold.  Trait Coefficient (b) R2 p SEE Assumptions Litter ASC 1.0 0.72% 0.50 0.57 Yes Litter AUR -2.2 4.6% 0.088 0.55 Yes Litter C -7.6 6.2% 0.047 0.55 Yes Litter C:N -0.0043 2.3% 0.23 0.56 Yes Litter Fps Litter Fps (ln) -0.063 -0.27 3.4% 6.7% 0.14 0.039 0.56 0.55 Not equal variance Yes Litter Leaching 2 Hrs 2.4 0.54% 0.56 0.57 Yes Litter Leaching 24 Hrs 2.7 11% 0.0062 0.53 Yes Litter N (ln) 0.28 (0.17*) 3.2% (0.54%*) 0.16 (0.58*) 0.56 (0.58*) Yes Litter NPE (ln) 0.31 10% 0.0091 0.54 Yes Litter pH 0.36 21% 1.1 × 10-4 0.50 Yes Litter SLA  0.014 (0.033**) 2.9%% (13%**) 0.18 (0.0052**) 0.56 (0.53**) Yes Yes Litter Thickness (ln) -0.21 4.2% 0.10 0.56 Yes Litter Water 2 Hrs 0.43 1.1% 0.42 0.57 Yes Litter Water 24 Hrs 0.74 2.9% 0.18 0.56 Yes Litter WSE -0.16 0.05% 0.86 0.57 Yes PC 1 Litter All Traits 0.043 4.2% 0.10 0.56 Yes PC1 Litter Same Traits -0.058 4.6% 0.089 0.55 Yes * without A. rubra ** without A. macrophyllum     175 Table A5-7. Regression Analysis for Phase-I Mass Loss Using Plot-Level Variables. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using litter trait variables to predict the proportion of mass lost during Phase I (ln-transformed). The cutoff p-value (p < 0.0063) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (8). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold.  Plot-Level Variable Coefficient (b) R2 p SEE Assumptions Elevation 0.0036 0.70% 0.51 0.39 Yes Slope 0.034 2.3% 0.23 0.39 Yes Shannon’s Diversity Index (H) -0.13 0.50% 0.58 0.39 Yes Simpson’s Diversity Index (S) -0.23 0.53% 0.57 0.39 Yes Forest Floor pH 0.18 0.97% 0.44 0.39 Yes Forest Floor C:N 0.0079 0.09% 0.82 0.39 Yes Canopy Cover 0.00060 0.04% 0.87 0.39 Yes Plot Identity (categorical) A: 0.0000 B: 0.025 C: -0.019 D: 0.10 E: -0.11 3.1% 0.76 0.39 Yes             176 Table A5-8. Regression Analysis for Phase-II Mass Loss Using Plot-Level Variables. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using litter trait variables to predict the proportion of mass lost during Phase II (ln-transformed). The cutoff p-value (p < 0.0063) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (8). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold. Plot-Level Variable Coefficient (b) R2 p SEE Assumptions Elevation 0.012 3.6% 0.13 0.56 Yes Slope -0.032 0.96% 0.44 0.57 Yes Shannon’s Diversity Index (H) -0.37 1.9% 0.28 0.56 Yes Simpson’s Diversity Index (S) -0.52 1.3% 0.37 0.56 Yes Forest Floor pH 0.32 1.4% 0.34 0.56 Yes Forest Floor C:N 0.066 2.9% 0.18 0.56 Yes Canopy Cover -0.0044 1.1% 0.42 0.57 Yes Plot Identity (categorical) A: 0.00 B: 0.028 C: 0.10 D: -0.41* E: -0.27 13% 0.09 0.54 Yes * Marginally significant: t  = -1.9, p = 0.063        177 Table A5-9. Regression Analysis for Year-One Mass Loss Using Plot-Level Variables. Linear regression coefficients (b), R2, p-values, and root mean squared error (SEE) of regressions performed using litter trait variables to predict the proportion of mass lost during the first year (ln-transformed). The cutoff p-value (p < 0.0063) was determined by applying a Bonferroni correction to the probability level (∝ = 0.05) by the number of regression analyses performed (8). Regressions were tested to see if they met assumptions of normality and equal variance of residuals; any deviations from these assumptions are indicated below. Significant regressions are indicated in bold.  Plot-Level Variable Coefficient (b) R2 p SEE Assumptions Met? Elevation 0.0065 2.6% 0.20 0.36 Yes Slope 0.0065 0.10% 0.80 0.36 Yes Shannon’s Diversity Index (H) -0.20 1.4% 0.35 0.36 Yes Simpson’s Diversity Index (S) -0.32 1.2% 0.39 0.36 Yes Forest Floor pH 0.22 1.7% 0.30 0.36 Yes Forest Floor C:N 0.030 1.5% 0.34 0.36 Yes Canopy Cover -0.0015 0.3% 0.67 0.36 Yes Plot Identity (categorical) A: 0.00 B: 0.031 C: 0.032 D: -0.096 E: -0.16 4.5% 0.60 0.36 Yes         178 Appendix 6: A Comparison of Litters at the Same Stage of Decomposition A6.1 Introduction In addition to the litterbag study described in the thesis, I conducted a short pilot study to test the effectiveness of exponential decay constants (often referred to as k) in predicting mass loss and attempt to determine whether two distinct tree species follow similar trajectories in decomposition; specifically, if at the same stage of mass loss their physical traits have changed in the same way.  Much of the research that has been done on decomposability and functional or litter traits investigates correlations between trait measurements and mass-loss rates determined in litterbag experiments. Our ability to predict rates of mass loss is so refined that we now can predict rates for just about any type of litter and accordingly model carbon and nutrient fluxes (Prescott 2010). However, mass loss studies are limited in that they only convey information about the net loss of litter material, failing to consider how litter is transformed during decomposition. In the transition from litter to soil, for example, litter is colonized by microbes, such as fungi (Ponge 1991), which may increase the overall mass of litter, producing what I will refer to as the “litter complex.” This complex includes the original plant litter itself, microbial biomass, and microbial byproducts (Paul and Clark 1996, Binkley and Fisher 2013). Hobara et al. (2014) found evidence for this, measuring net increases in concentrations of amino acids found in microbes, which greatly contributed to total N in the decomposed litter. Foucreau et al. (2013) measured ergosterol content as a proxy of fungal biomass and showed how fungal biomass increased over time, contributing to mass remaining in the leaves of three plant species that were decomposing in a stream. Ergosterol, an abundant sterol in fungi, is often used to approximate fungal biomass in that it is specific to fungi; ergosterol is found in fungal cell membranes, but is not found in  179 plants (Kandeler 2007). Unlike chitin, which persists long after the fungus has died, ergosterol tends to degrade soon after fungal death and therefore tends to be a good approximator of living fungal biomass (Ekblad et al. 1998). Compared to using qPCR to estimate fungal biomass, ergosterol content showed less variation than the qPCR product, suggesting that it may be a more accurate indicator (Baldrian et al. 2013).  Additionally, litter does not always decompose completely. In northern forest ecosystems, mass loss proceeds until about 20 to 30% of the original mass remains; this remaining material has been converted into humus, which loses mass slowly (Prescott 2010). Within mineral soil, litter may also become incorporated into soil aggregates where it is protected from microbial decomposition (Cotrufo et al. 2013). Thus, the process of litter decomposition involves more than simply mass loss due to the breakdown of the compounds that constitute litter and might better be referred to as “transformation”. To better understand how litter material is transformed as decomposition progresses, it is necessary to look beyond the analysis of mass loss and use direct observation to examine changes and determine what exactly happens at each stage of decomposition. Qualitative, observational studies by Ponge et al. (1991) and Mori et al. (2009), in which they use microscopy to examine samples of forest floor at various depths, have revealed that litter material is transformed by different soil microbes and fauna at different stages, and that decomposition differs beneath various tree species, producing dissimilar humus profiles. In a similar study, Tian et al. (1997) used thin-sections to observe morphological changes in litter at different depths within the forest floor to infer changes in mesophyll, vascular bundle, and epidermis structure over time during decomposition. Observational studies conducted with a focus on how the same functional and leaf-litter traits that we use to make initial, correlative  180 predictions of litter mass loss change during this process will lead to a more complete understanding of the process of decomposition as well as of the afterlife influence of plant functional traits on ecosystem processes.  In addition to direct observation, it is necessary to start comparing litters at the same stage of decomposition, rather than at the same time (Prescott and Grayston 2013). It is possible that the changes in litter composition that different species undergo during decomposition are more similar than what is suggested when we present changes in mass loss and nutrient content over time, as these same changes may simply occur at different rates. Comparing the changes that are happening at the same percent-mass-remaining value rather than the same time would help us better elucidate the process of decomposition itself. Few studies have expressed decomposition in this way. A study by Prescott et al. (1993) expressed changes in N concentration in Pinus contorta var. latifolia (lodgepole pine) litter during decomposition in one of three forests or a clear-cut as a function of percent mass remaining, rather than time. They found that the litter in all sites followed the same pattern of N concentration increase and then decline as they lost the same amount of mass; N concentrations in the needles in the clear-cut started to decline sooner than the needles in the forests, but this was because the needles reached that particular stage of mass loss faster in the clear-cut environment, not because the needles differed in their decay patterns. Another more recent study by Wickings et al. (2012) compared corn and grass decomposition at the same stage, looking for evidence to determine whether litter of different species chemically diverges, converges, and/or changes in a way mainly dictated by the particular decomposer community present. One reason why these types of study are much less common than studies comparing the decay of different species at the same time is because it is simpler to collect litter that has decomposed over the same time period; collecting litter at the  181 same decomposition stage requires producing and relying on accurate predictions of mass remaining, or else requires collecting at many time intervals in order to increase the probability of collecting two samples at the same stage of decomposition. This pilot study was an attempt to address several of these shortcomings in decomposition studies, namely to address morphological changes that occur during decomposition and comparisons of litter decomposition at the same stage rather than same time.  A6.2 Research Objective This pilot study was conducted to address the question: does decomposition progress in similar ways but at different rates for different species? This question was addressed by attempting to compare litter from Alnus rubra (red alder), a nitrogen-fixing, broadleaf, deciduous tree species native to British Columbia, and Pseudotsuga menziesii (Douglas-fir), an evergreen, coniferous tree species also native to British Columbia, at the same stage of decomposition (mass loss), rather than at the same time since placement. Exponential decay constant (k) values determined at the first time-point of the study conducted to address Question 2 of the thesis were used to predict the days at which litter of each species would reach 50% mass remaining. Samples from additional litterbags collected at both time points were used to measure changes in physical traits (thickness and specific force-to-punch) as well as relative fungal abundance (ergosterol concentration). I expected to find that the litters at 50% mass remaining had undergone similar changes in physical traits from the initial litter samples, and I expected to find similar concentrations of ergosterol in all samples, since they would be at the same stage of decomposition.   182 A6.3 Methodology  For a comparison of litters at the same stage of decomposition rather than at the same time, I chose Alnus rubra, a deciduous, broadleaf tree species, and Pseudotsuga menziesii, a coniferous, evergreen tree species. I placed three bags of each species, each filled with approximately 1.00 g of air-dried litter collected at the University of British Columbia Farm in Vancouver in October 2014, at each of five plots also used for the litterbag experiment described in Section 2.5 of the thesis: plots B, C, D, E, and F.  I used the mean exponential decay constant (k) values calculated after 6 weeks for A. rubra and P. menziesii, determined in the larger litterbag study described in Section 2.5, to forecast the dates when the A. rubra and P. menziesii litter would reach 50% mass lost (Figure A6-1). This was done using the following equation to solve for T (time), in which the biomass remaining at time T was set to half of the initial mass (BT) of each litterbag and the mean k values were determined from the 6-week samples collected in the study described in Section 2.5: BT+1 = BT × e-kT Accordingly, I collected A. rubra litterbags after 108 days (on 20 March 2015) and P. menziesii litterbags after 295 days (on 23 September 2015) and I immediately transported the litterbags to the lab upon collection. Of the three bags at each plot, I oven-dried one litterbag and used it to determine k and proportion of mass lost as described in Section 2.5 of the thesis. I brushed the contents of the second bag free of debris, transferred the litter into plastic bags, and shipped the samples on dry ice to the Analytical Chemistry Laboratory at the Pacific Forestry Centre of Natural Resources Canada in Victoria, BC, where three 0.19 to 0.22 g (fresh weight) subsample replicates of each of the five samples (n = 15) were analyzed for ergosterol concentration, which, when compared between the two species, would suggest the relative fungal  183 abundance (Seitz et al. 1979). The samples were first dissolved in a pre-extracting solution comprised of potassium hydroxide (KOH) and methanol, and then extracted using petroleum ether. The extracted ergosterol samples were dried and then rehydrated with methanol for ergosterol measurement, which was performed on the HPLC using UV detection at 280 nm. Ergosterol concentration was expressed in micrograms per gram (which is equivalent to parts per million, ppm) of oven-dried sample (Manter et al. 2001, Warrington 2015); subsamples were used to determine the moisture content and therefore the conversion factor to calculate the oven-dried weights of the samples.  I brought the third and remaining litterbag from each plot to the Department of Botany at UBC, where I measured leaf thickness using a caliper and toughness using a punch-and-die connected to an Instron system, as described in Section 2.3 of the thesis. There were between one and three A. rubra leaves per bag, so at least three parts of the sample (either two punches on one leaf and one on another leaf, or three punches on one leaf) were used for force-to-punch (Fp) measurements. For P menziesii, three needles were arbitrarily chosen from the litterbag and used for Fp measurements. Specific force-to-punch (Fps) was calculated by dividing Fp by leaf thickness, which was measured at the spot symmetrical to where force-to-punch was measured on A. rubra leaves, on the opposite side of the central vein. For P. menziesii needles, I measured thickness on either side of the spot where toughness was measured and calculated the mean of the two thickness measurements. As in the functional trait study, I took care to avoid measuring thickness where toughness would be measured, in order to minimize any mechanical disruption of structure that could influence the calculations. Because the samples did not end up reaching the same proportion of mass lost, I could not conduct meaningful statistical analyses on these data. I did, however, compare thickness and  184 Fps of foliar samples, senesced litter samples, and decomposed samples of each species to observe how the traits differed during these three distinct stages. I then compared proportion of mass lost and ergosterol concentrations from the samples of both species in each plot to see if these measurements in A. rubra and P. menziesii samples varied in the same way at each of the five plots.  A6.4 Results Despite setting the collection date to anticipate when the leaf litter would reach 50% mass lost, samples from neither species had yet lost 50% mass; on average, net mass loss for the A. rubra samples was 35% after 108 days, and net mass loss for the P. menziesii samples was 23% after 295 days (Figure A6-2). Therefore, the A. rubra and P. menziesii litter samples were not of the same stage of decomposition, so I could not perform statistical comparisons based on that assumption. There was quite a bit of variation between mass loss estimates in the different plots, and, with the exception of the samples in plot F, the relative differences in percent mass loss are the same for A. rubra and P. menziesii, in that samples lost mass more quickly in plot C and more slowly in plot E relative to the other plots (Figure A6-2).  The measured physical traits of the two litter species, which include thickness and toughness (as Fps), did differ between the three phases (foliage, senesced litter, decomposing litter) in distinct ways. In A. rubra samples, thickness was greatest in foliage and leaf litter and was smaller in samples that had decomposed for three months, while in P. menziesii samples, thickness was less in both fresh and decomposing leaf litter than foliage (Table A6-1). Fps was less in both fresh and decomposing leaf litter than in foliage in A. rubra samples, while Fps was  185 greatest in fresh leaf litter but was roughly equivalent in foliage and decomposing leaf litter in P. menziesii samples (Table A6-1). Ergosterol concentrations, which indicate the extent of fungal presence, were higher on average in A. rubra samples collected after 108 days (502.9 ± 41.0 SE ppm, n = 15 replicates) than in P. menziesii samples collected after 295 days (382.4 ± 9.90 SE ppm, n = 15 replicates). Ergosterol concentrations were highest in plot B for A. rubra samples, and lowest in plot F, and were higher than in P. menziesii samples in all plots except for plot F (Figure A6-3). The range of measured ergosterol concentrations was smaller in P. menziesii samples than in A. rubra samples, and there did not appear to be any parallel trends in ergosterol concentrations across plots (Figure A6-3).   A6.5 Discussion In an attempt to deviate from traditional litterbag decomposition studies, A. rubra and P. menziesii were used in a smaller pilot study to test the effectiveness of using measured exponential decay constants (k) to predict mass loss over time and, ideally, compare the characteristics of two species at the same stage of mass loss. Though the litterbag samples were collected when they were expected to have lost 50% of their mass, samples from neither species had yet reached 50% mass loss. Additionally, A. rubra samples, which had been collected 187 days prior to the P. menziesii samples, had already lost on average 12% more mass (Figure A6-2). Therefore, contrary to my predictions, the samples from each species were not at the same stage of mass loss and therefore could not be compared in this way.  There are several possible and not mutually exclusive explanations for why these mass loss measurements did not meet the predicted values, including seasonal changes in climate and  186 differences in resource quality. In regards to climate, the k value for each species was calculated for mass loss that had occurred from December 2014 through mid-January 2015 (during the first time interval in the previous study), which is a time when Vancouver, BC tends to get more rain; the average daily rainfall during that time was 5.5 mm per day, which is almost twice the average rainfall daily rainfall during the entire study period (3.0 mm per day). Decomposition tends to proceed faster in wetter conditions due to greater leaching (Swift et al. 1979, Taylor and Parkinson 1988) and greater microbial growth and mobility (Bunnell et al. 1977, Swift et al. 1979, Voroney 2007), which could in part explain why mass loss was more rapid after 1.5 months than thereafter, and why the k values calculated based on this time period predicted faster rates of decomposition. Though it is common to model litter mass loss using a single exponential equation, this model works best to model decomposition under constant conditions (Swift et al. 1979), which are not observed in situ. Perhaps one way to achieve a more accurate k value would be to use the k value calculated after the first year of mass loss, which would encompass all annual seasons. Or perhaps use the k value calculated from a time interval longer than one year; Fogel and Cromack (1977) found that k values of P. menziesii litter calculated from cumulative data after one year were consistently higher than those calculated after two years. In addition to constant conditions, the single exponential decay model is also best for modeling the decomposition substrates of uniform chemical composition, in particular pure cellulose (Swift et al. 1979), which is not the case regarding in situ studies with leaf litter samples of heterogeneous chemical composition. Different chemical fractions of leaves decompose at different rates (Minderman 1968), and A. rubra and P. menziesii are chemically heterogeneous and differ in their composition. In particular, A. rubra leaf-litter samples had on average a lower C:N (due to higher N concentrations), lower AUR, and higher WSE than P.  187 menziesii samples (Table A2-4). These collective differences in composition could also partly account for the observation that A. rubra had lost more mass than P. menziesii did, even though P. menziesii decomposed for 187 more days.  Samples from the two species also differed in the variability of mass loss and ergosterol concentrations among the five plots at which litterbags were installed. For plots B through E, which are the most similar to each other in terms of plant species composition and canopy cover, the relative mass loss between the litterbags follows the same pattern for both species. This could imply that the differences measured in mass loss could be the result of plot-level differences.  Though the plots were chosen because of their similarities, they did differ slightly in terms of vegetation composition (particularly in the understory) and canopy cover, and slightly in soil C:N and pH as well (Tables 2.2, 2.3, and 2.4). Site-level factors have been shown in previous studies to influence decomposition rates; examples of these factors include soil chemistry such as C:N (Veen et al. 2015, Dale et al. 2015) and soil moisture (Bunnell et al. 1977), both of which may be influenced by the surrounding vegetation, in particular by surrounding vegetation’s influence on soil biota communities (Reich et al. 2005, Bardgett and Wardle 2010). Differences in the site variables I measured in each plot, however, did not seem to correlate with this mass-loss trend. This could mean that the differences in these variables are not strong enough to influence mass loss or that values of these variables are heterogeneous within the plots, implying that measurements made on the litterbag scale might better reflect conditions actually experienced by the decomposing litter (Bradford et al. 2016).  Plot F breaks this pattern between A. rubra and P. menziesii, however, and again, nothing measured in my study can lend reliable evidence as to why this may be. Perhaps the low ergosterol concentrations in A. rubra at Plot F could suggest that decomposer activity was lower for the A. rubra litterbags at this site, and  188 therefore that the differences in mass loss at this site relative to others were a result of differences in decomposer abundance or community composition. This particular study, however, cannot provide enough evidence to support this hypothesis or provide further details.  In addition to variability in mass loss, I observed variability in ergosterol concentrations between the two species and between plots. The ergosterol concentrations in A. rubra were on average greater, but also encompassed a greater range of values. This could provide evidence to suggest that, even within the same forest, there can be great variability in decomposer activity on a small scale. This is likely due to the heterogeneity of the forest floor, which is produced by interactions between vegetation and soil biota communities (Bardgett and Wardle 2010). The range of ergosterol concentration for P. menziesii was not quite as large, and on average these values were lower. It is difficult to make meaningful comparisons between A. rubra and P. menziesii because the samples were not at the same stage of decomposition. This result does suggest, though, that after 295 days, P. menziesii had a lower fungal presence than A. rubra after 108 days. There are too many confounding factors to suggest a reason, but this does raise the question of how much total, both living and dead, fungal matter in contained within these tissues. Ergosterol is commonly considered a marker for living fungal biomass, though it has been detected in dead fungal biomass as well (Mille-Lindblom et al. 2004), but chitin, another fungal biomarker, would encompass total fungal biomass as opposed to just living biomass (Ekblad et al. 1998), which could in part help determine whether fungal activity was perhaps higher at one point in P. menziesii than it is now, or if fungal activity was always lower than the concentrations measured in A. rubra. One disadvantage of using chitin as a fungal biomarker, however, is the fact that chitin may be present in other tissues, such as invertebrate exoskeletons (Kandeler 2007). These sources of chitin measured in litter samples may likely to belong to decomposers,  189 though, so perhaps using chitin as a biomarker may help indicate a more diverse total decomposer presence, beyond fungi alone.           This pilot study was therefore unsuccessful in producing litter samples from different species that had lost the same proportion of mass, and consequently highlights the difficulty of using simple exponential decay functions to effectively predict mass loss over time. Few studies have been able to achieve the feat of comparing traits of diverse species at the same stage of mass loss. Wickings et al. (2012) were able to compare the litter chemistry of corn and grass at the same percent mass loss, but likely this was because they had taken samples from many time points (12), so the probability that they would be able to find samples that had the same percentage of mass loss, in their case 30% and 80%, was higher than it was in this study, which only had one time point. They also did not indicate the use of k values to predict mass loss; it is likely that they happened to have samples of both litter types with both 30% and 80% mass lost in their study. In the future, perhaps an optimal solution to achieving this objective would be to install many litterbags and collect at more frequent time intervals to increase the probability of obtaining litterbags from both species that had lost the same proportion of mass.  Despite being unable to compare A. rubra and P. menziesii in the same stage of mass loss, I did compare the thickness and toughness of these two species across the three stages from which I had measured these traits: foliage, senesced litter, and decomposed litter. I found that thickness was greatest in foliage for P. menziesii, but did not seem to change during early decomposition, while thickness did not seem to change during senescence for A. rubra but was lower after 108 days of decomposing (Table A6-1). This seems to suggest a possible difference in trajectory that these two species follow during decomposition. Observational studies of needle decomposition, suggesting that the mesophyll begins to disappear and fungi colonize the  190 hollowed needles as decomposition progresses (Ponge 1991, Tian et al. 1997), are consistent with the lack of change in thickness detected over time in P. menziesii samples. Regarding broadleaf species such as A. rubra, observations that I made during this study suggest that the soft tissue material between the veins of broadleaved species tends to decompose before the veins do, which could cause the thickness of this material to decrease; this is consistent with observations of Betula spp. litter decomposition made by Tian et al. (1997). The decrease in toughness for P. menziesii from litter to decomposed litter also fits the hypothesis that the needles are hollowing out and filling with a less tough fungi or microbial biomass. This inverse relationship between fungal biomass (expressed as ergosterol) and toughness was demonstrated by Foucreau et al. (2013), who found that as ergosterol (or fungal biomass) increased, specific force-to-punch decreased. This softening of leaf litter and increase in fungal biomass is referred to as “conditioning” (Foucreau et al. 2013). In addition, they found that in some species, thickness did not change over the study period, as I observed in P. menziesii samples. However, one of the species for which this pattern had been observed by Foucreau et al. (2013) was A. rubra, which, in my study, did not seem to follow this pattern. These results should be compared with caution, as the samples used by Foucreau et al. (2013) had been decomposing for less time, and in water.  Despite the difficulty, there is a need for studies that compare litter decomposition among different species at the same proportion of mass lost, rather than at the same time interval. There is still much about the decomposition trajectory, in particular regarding how leaf tissues change over time, that remains to be explored. My attempt to collect samples at the same proportion of mass lost was unsuccessful, but perhaps collecting from more time intervals will increase the likelihood of collecting samples at the same mass loss stage. Perhaps it would also help to use  191 decomposition models that use more parameters than only mass lost during the previous six weeks, such as traits, to make better mass loss predictions and increase the likelihood of retrieving litterbags with the predicted proportion of mass lost. To improve the accuracy of k values in making predictions, perhaps a k value calculated from a longer decomposition interval would have been able to make more accurate predictions as well. Though this study was unsuccessful, it did highlight considerations to make when pursuing these types of studies in the future and further raises the question of how decomposition trajectories differ between species.                 192 A6.6 Tables and Figures  Figure A6-1. Predicted Time to 50% Mass Loss for A. rubra and P. menziesii Litters. Exponential curves produced using the mean exponential decay constant (k) values calculated from A. rubra (red, k = 2.36) and P. menziesii (violet, k = 0.86) after decomposing for 42 days. The dotted lines indicate point on the curve when each hypothetical sample, which started at 0.92 g, reaches 50% mass remaining (0.46 g). This would take 108 days for A. rubra and 295 days for P. menziesii.    193  Figure A6-2. Proportion of Mass Lost in A. rubra and P. menziesii. Proportion of mass lost from P. menziesii (violet) and A. rubra (red) litterbags installed at UBC Farm on 2 December 2014. The solid line depicts expected mass loss (0.5), the line with short dashes represents mean mass loss of the A. rubra litter (0.35 ± 0.02, n =5), and the line with long dashes represents mean mass loss of the P. menziesii litter (0.23 ± 0.03, n =5). The A. rubra litterbags incubated for 108 days and the P. menziesii litter incubated for 295 days, which is when they were expected to have lost 50% of the starting mass based on exponential decay constant (k) values calculated after 1.5 months.  194 Table A6-1. Thickness and Toughness in A. rubra and P. menziesii Foliage, Leaf Litter, and Decomposed Samples. Mean leaf thickness (mm) and specific force-to-punch (Fps, N mm-2), with standard error, measured in foliar samples, freshly senesced leaf-litter samples, and decomposed leaf-litter samples of A. rubra and P. menziesii collected at UBC Farm in Vancouver, BC.  Species Material Type Thickness (mm) Fps (N mm-2) A. rubra Foliage 0.31 ± 0.01 1.76 ± 0.14 A. rubra Leaf Litter 0.33 ± 0.02 1.12 ± 0.08 A. rubra Decomposed 0.23 ± 0.04 1.13 ± 0.17 P. menziesii Foliage 0.54 ± 0.01 2.31 ± 0.11 P. menziesii Leaf Litter 0.43 ± 0.02 4.17 ± 0.30 P. menziesii Decomposed 0.39 ± 0.03 2.50 ± 0.34              195  Figure A6-3. Ergosterol Concentrations in Decomposing A. rubra and P. menziesii Leaf-Litter Samples. Mean ergosterol concentrations (ppm, with standard error) measured in A. rubra (red) litter samples collected after 108 days and P. menziesii (violet) litter samples collected after 295 days. One litterbag of each species from each plot was used for ergosterol measurements, and three replicates were measured from each litterbag. Litterbags were installed on 2 December 2014 at UBC Farm in Vancouver, BC.   

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0300015/manifest

Comment

Related Items