Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identifying the context dependencies of plant-herbivore interactions across a species’ range Loughnan, Deirdre 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_september_loughnan_deirdre.pdf [ 2.66MB ]
Metadata
JSON: 24-1.0308710.json
JSON-LD: 24-1.0308710-ld.json
RDF/XML (Pretty): 24-1.0308710-rdf.xml
RDF/JSON: 24-1.0308710-rdf.json
Turtle: 24-1.0308710-turtle.txt
N-Triples: 24-1.0308710-rdf-ntriples.txt
Original Record: 24-1.0308710-source.json
Full Text
24-1.0308710-fulltext.txt
Citation
24-1.0308710.ris

Full Text

Identifying the context dependencies of plant-herbivore interactions  across a species’ range. by  Deirdre Loughnan  Hons. B.Sc., The University of Toronto, 2014  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Geography)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2016  © Deirdre Loughnan, 2016   ii Abstract  Spatial variation in abiotic and biotic factors creates local contexts that influence the intensity and outcome of plant-herbivore interactions. Previous studies have tried to represent the complexity of these context dependencies using latitudinal clines, but this approach has proven insufficient for many systems. Despite substantial variation in herbivore damage across populations, the relative contribution of local community factors to explaining this variation is less well known.   I investigated plant-herbivore interactions across the entire range of Quercus garryana, specifically testing the relative importance of climate and population traits on herbivore communities, leaf traits, and the extent of damage caused by different insect feeding guilds. I performed similar analyses on trees grown in a common garden, allowing me to detect the relative importance of environmental and genetic contributions to leaf defense traits and susceptibility to herbivory.   Although I observed no variation in herbivore diversity among populations in the field, the abundance of herbivores declined with increasing latitude, while also responding to variation in population size, leaf traits, and climate. Leaf traits were also influenced by climate factors, but in addition varied with tree size and within the growing season. Differences in herbivore damage were best explained by long-term trends in spring precipitation, leaf traits, and population size, with no relationship to latitude. The relative importance of each of these factors depended on when the damage occurred as well as the insect feeding guild causing the damage. Conversely,   iii the extent of damage in the common garden was constant across trees of different provenance, providing further support for the importance of climate in driving variation in herbivory.   The findings of this study demonstrate the importance of climate, irrespective of latitude, on plant-herbivore interactions and in mediating the effects of other community factors. As such, they highlight the importance of conducting studies across diverse ecosystems and climate gradients. Only by understanding the underlying drivers of selection can we begin to draw generalizations and develop a predictive framework of plant-herbivore interactions to inform conservation and effective habitat management.     iv Preface  This thesis is original, unpublished work by the author, Deirdre E. E. Loughnan. Supervision and guidance for this research was provided by Dr. Jennifer L. Williams (University of British Columbia, Geography). The common garden portion of this study was made possible by the work of Colin A. Huebert and Dr. Sally Aitken (University of British Columbia, Forestry). All data were collected and analyzed by the author.  Portions of this work were presented in a talk at the Ecological Society of America Conference (Fort Lauderdale, Florida, August 2016).   v Table of Contents  Abstract.......................................................................................................................................... ii!Preface........................................................................................................................................... iv!Table of Contents ...........................................................................................................................v!List of Tables ................................................................................................................................ vi!List of Figures.............................................................................................................................. vii!Acknowledgements .................................................................................................................... viii!Chapter 1: Introduction ................................................................................................................1!Chapter 2: Materials and Methods ..............................................................................................6!2.1! Study species................................................................................................................... 6!2.2! Tree sampling and trait measurements ........................................................................... 7 2.3! Statistical analysis......................................................................................................... 10 Chapter 3: Results........................................................................................................................14 Chapter 4: Discussion..................................................................................................................18 Chapter 5: Conclusion.................................................................................................................24!References.....................................................................................................................................40!Appendix A...................................................................................................................................49!  vi List of Tables  Table 1 Model ranking & parameters for explaining insect abundance........................................26 Table 2 Statistical tests for differences in herbivory and traits between sampling periods...........27 Table 3 Statistical tests for differences in herbivory & traits among populations in situ..............28 Table 4 Model ranking & parameters for explaining leaf traits.....................................................29 Table 5 Model ranking & parameters for explaining variation in total herbivory........................ 30 Table 6 Model ranking & parameters for explaining herbivory by feeding guilds.......................31!Table 7 Statistical tests for differences in herbivory among populations for common garden samples.......................................................................................................................................... 33 !   vii List of Figures  Figure 1 Population locations and Quercus garryana var. garryana distribution........................ 34 Figure 2 Variation in insect abundance with latitude and community traits ................................ 35 Figure 3 Leaf trait variation with precipitation and tree size........................................................ 36 Figure 4 Herbivory damage with latitude for both sampling periods........................................... 37 Figure 5 Variation in total damage with community traits ........................................................... 38 Figure 6 Herbivore damage of summer samples across population size classes.......................... 39    viii Acknowledgements  This project was made possible by the support and guidance of many people. Foremost, I would like to acknowledge my supervisor, Dr. Jennifer L. Williams, for her tremendous guidance and support throughout this project. The experience I have gained, both through her teaching and mentorship, are invaluable. I am particularly grateful for having had the opportunity to develop my own research questions and Jenn’s support in seeing them come to fruition. It would not have been possible had it not been for the time she dedicated to this project and to my personal development as a scientist. I am grateful to my supervisory committee, Dr. Sally Aitken and Dr. Greg Henry for their insights in the review of my research proposal and feedback throughout this project and Dr. Sally Aitken’s constructive comments on my written work. Dr. Gregory Crutsinger is thanked for his guidance in the development of this research project. Lucas Jarron and Riley Finn are thanked for their hard work as my research assistants in the field and in the lab. Thanks also go to Charlotte Trowbridge and Cameron Hunter, both for their time volunteering to collect data in the field and for making my time at UBC so enjoyable. Special thanks to my peers and community members in the Florum discussion group for providing advice on research methods and analysis. Finally I would like to thank my loved ones and friends for their support and encouragement to pursue my passions throughout years of study.    Thanks also to the Centre for Forest Conservation Genetics for their help in the establishment and maintenance of the common garden used in this experiment, with funding provided by the British Columbia Ministry of Forests, Lands and Natural Resources Operations and the Forest Genetics Council of British Columbia’s Genetic Conservation Subprogram.   ix Further thanks to the numerous land managers, owners, and community groups who allowed me to conduct research on their properties (arranged alphabetically): Townsend Angell (Facilities Operations, Reed College, Portland, Oregon), Irvin Banman (Nature Conservancy Canada, Duncan, British Columbia), Jerymy Brownridge (Operations and Management Services, Government House, Victoria, British Columbia), Chris Chyde-lay (Park Services, Oak Bay Municipality, British Columbia), Sarah Deumling (Zena Forest Products LLC, Rickreall, Oregon), Tim Ennis (Nature Conservancy Canada, Duncan, British Columbia), Gary Groth (Parks and Land Departments, Douglas County, Oregon), Joelle Hammerstad (Seattle Parks and Recreation, Seattle, Washington), Angela Holmes (School District 71, Comox Valley, British Columbia),  Jackson County Parks (Jackson County, Oregon), Dr. Dalton M. Johnson (Parks Division, Thurston County, Washington), Michele LaFontaine (Earth & Space Sciences Department, Pierce College, Lakewood, Washington), Ernie Mansueti (Parks and Recreation, Municipality of North Cowichan, British Columbia), Michael Perkins (Parks and Recreation, City of West Linn, Oregon), Anne Schuster (Wolf Haven International, Tenino, Washington), Laura Singer (Sequim Prairie Garden Club, Sequim, Washington), Nancy Swain (Benton County Fairgrounds, Corvallis, Oregon).  Financial support was provided by the UBC Department of Geography, UBC Faculty of Arts. Funding was also provided by the UBC Biodiversity Research Centre through a Biodiversity Research: Integrative Training and Education Fellowship and a Natural Sciences and Engineering Research Council (NSERC): Canada Graduate Scholarship (Masters) awarded to the author, as well as a NSERC Discovery Grant awarded to Dr. Jennifer Williams.  1 Chapter 1: Introduction  The abundance and distribution of species across large spatial scales are strongly affected by biotic interactions. However, the community conditions, or context, that meditate the outcome of these interactions, and therefore shape entire communities, are often poorly understood. These context dependencies are the result of variation in specific abiotic or biotic factors, altering the effects of species interactions on an individual’s fitness (Schemske et al. 2009, Maron et al. 2014). By altering the strength of interactions between plants and herbivores for example, these context-dependencies can affect the extent of plant growth and reproduction (Maron et al. 2014). Frequently this variation is regarded as noise (Maron et al. 2014), however, there is increasing recognition that knowing what conditions are most pertinent to individual fitness is integral to understanding plant community dynamics.   To account for the additive effects of the highly context-dependent nature of plant-herbivore interactions (Maron et al. 2014), many studies have taken the indirect approach of studying variation in interactions across latitude. Latitude is generally thought to correlate with clines in abiotic factors, such as temperature (Pearse and Hipp 2012, Pellissier et al. 2014) and precipitation (Rivas-Ubach et al. 2014, Moreira et al. 2015, Abdala-Roberts et al. 2016), allowing the combined effects of those factors to be accounted for. Furthermore, it may account for variation in species diversity, and the diversity of trophic levels such as herbivores, predators, or parasitoids (Dobzhansky 1950, Willig et al. 2003). The latitudinal gradient in plant defense hypothesis is based on the premise that interactions between species intensify closer to the equator (Dobzhansky 1950, Schemske et al. 2009), with greater herbivore damage at low   2 latitudes (Coley and Aide 1991, Pennings et al. 2009, Schemske et al. 2009), and selection for the expression of defense traits in these communities (Pennings et al. 2009, Marquis, Ricklefs and Abdala-Roberts 2012, Pearse and Hipp 2012). Support for this hypothesis has been found in some ecosystems, particularly from species in Atlantic Coast salt marshes (e.g. Pennings and Silliman 2005, Pennings et al. 2007, Pennings et al. 2009); however, a recent meta-analysis by Moles et al. (2011) found most studies have either observed alternate trends, such as greater damage at higher latitudes (e.g. Adams and Zhang 2009), or at the centre of a species’ ranges (e.g. Woods et al. 2012), while others found no trend at all (e.g. Andrew and Hughes 2005). Although latitude may explain some spatial variation in defense traits and susceptibility of some plant species to herbivory, these trends are more likely to be the result of underlying gradients in climate or community factors (Moreira et al. 2015), the identification of which would provide a more mechanistic understanding of what factors most affect species interactions.   Abiotic factors, such as temperature, precipitation, and nutrient availability, play a major role in determining the presence and abundance of both plant and herbivore species in a given population (Maron et al. 2014). By imposing physiological constraints and determining the availability of resources to plants, local climates impose direct selection on plant growth and phenotype (Wright et al. 2004). The resource availability hypothesis posits that in communities in which resources, such as water or nutrients, are limited there should be selection for slow growing plants that invest relatively more resources in defense, reducing the loss of tissue that is costly to produce or restore (Coley et al. 1985). Conversely, plants growing in high resource populations are predicted to have higher levels of damage, as they can tolerate herbivory and replace lost tissue at a lower cost, selecting against the investment of resources to defense (Coley   3 et al. 1985, Maron et al. 2014). Alternatively, the plant stress hypothesis predicts the opposite:  plants growing in populations with limited resources are unable to invest resources into both growth and defense, making them susceptible to herbivore damage (White 1969, Maron et al. 2014). Support for both of these hypotheses has been found for specific feeding guilds in a variety of ecosystems (eg. Koricheva et al. 1998, Endara and Coley 2011), making generalizations on how climate selects for various leaf defense traits difficult.   In addition to potentially selecting for leaf traits adapted to local abiotic conditions, climate also determines the abundance of herbivores that can be sustained by the plant community and the extent of damage they can impose by limiting plant community productivity (Maron et al. 2014). Greater herbivore damage is expected in communities with more benign or favorable climates (Hawkes and Sullivan 2001, Dostal et al. 2013), as insects living in communities with shorter growing seasons and lower winter temperatures have fewer generations per year and experience greater density-independent mortality, which may result in less damage to plants (Hairston et al. 1960, Bale et al. 2002). Recent studies have also shown that the type of herbivore damage can be influenced by climate. For example, Leckey et al. (2014) found leaf damage by leaf miners and galling insects varied in response to precipitation, with wetter communities experiencing more damage. By altering the distribution and identity of herbivores within a community, climate can have an indirect effect, influencing top-down selection on plants by the herbivore community.   The complexity of determining what community conditions are driving trait expression is confounded by the fact that many traits provide multiple advantages for a plant. In addition to providing defense from herbivores, traits may also mitigate stressful climate conditions. For   4 example, specific leaf area (SLA), which is the quotient of a leaf’s surface area over its dry mass, reflects a plants investment in leaf structure and defense, with tougher leaves having lower SLA values (Cornelissen et al. 2003). In warmer, water limited populations, increasing SLA by reducing leaf surface area, reduces the transpiration rate (Poorter et al. 2009). Similarly, high trichome densities deter herbivores by making leaves less palatable to larger herbivores and difficult for small herbivores to move across and oviposit onto (Price et al. 2011), while also reducing water loss through transpiration by creating a boundary of still air at the leaf surface (Guerfel et al. 2009).  These secondary functions limit the ability to identify the drivers of trait selection, but still reflect the additive selective pressures of the biotic interactions and climate.   Finally, population size may also influence the outcome of plant-herbivore interactions. Given the patchy nature of many ecosystems due to natural variation in substrate, climate, and human-driven habitat fragmentation, population size can vary substantially within a species. Classic island biogeography theory predicts that herbivore pressure will be greater in larger patches, both as a result of the greater herbivore population sizes that can be supported, as well as predicted higher diversity in herbivore sizes and feeding guilds (Steffan-Dewenter and Tscharntke 2002). Larger populations are also more likely to be colonized by more predator or parasitoid species, which through trophic cascades, may limit herbivore populations size and result in reduced herbivore pressure (Steffan-Dewenter and Tscharntke 2002, Elzinga et al. 2005).   The purpose of this study was to identify climate and community factors that underlie the context-dependencies of plant-herbivore interactions across the latitudinal range of a single species. To achieve this, I tested how environmental factors and plant traits influence the suite of   5 insect herbivore interactions with Quercus garryana (Garry oak), and the extent to which local insect communities and selection on leaf traits underlies these trends. Specifically, I asked: (1) What are the relative contributions of abiotic and biotic factors on defense traits and the extent of leaf damage? (2) How does insect community diversity and abundance vary among populations across the range and what factors explain these differences? (3) Do leaf traits and herbivore damage of trees grown in a common garden vary among populations, given the absence of environmental variation?  To address these questions, I quantified the amount and type of leaf damage, specific leaf area (SLA), trichome density, trunk volume, and six climate and weather conditions in situ for 18 populations of Q. garryana across its entire latitudinal range in western North America.  In addition, I sampled the insect community to determine whether the diversity or abundance of insects varies among populations. If the latitudinal gradient in plant defense hypothesis is supported, I should find negative correlations between latitude and community factors, with herbivore damage decreasing as latitude increases. However, if insect communities or trait adaptations determine the outcome of plant-herbivore interactions, communities with the least favorable weather conditions for insect populations and higher defense traits should experience the least damage. Finally to disentangle the effects of local environments and genotype, I took similar measurements in an established common garden, located near the northern limit of Q. garryana’s distribution.     6 Chapter 2: Materials and Methods  2.1  Study species Quercus garryana var. garryana Douglas ex. Hook (Fagaceae), commonly known as Garry oak or Oregon white oak, is endemic to the Garry oak ecosystem (GOE), with a distribution that spans from northern California, USA to southwestern British Columbia, Canada (Jepsen 1909, Fuchs 2001). Environmental conditions in the GOE are near-Mediterranean, with annual temperatures ranging from 8°C to 18°C and annual precipitation ranging from 170-2630 mm (Stein 1990). Northern populations in Washington State, USA and Vancouver Island, experience slightly drier conditions as a result of rain shadows created by the Olympic Peninsula in Washington State, and the Vancouver Island central spine (Vellend et al. 2008). Due to their postglacial origins, soils in this habitat are moderately infertile (Fuchs 2001) and can span a gradient of soil depths, ranging from less than 5 cm to 100 cm (MacDougall et al. 2011).  Deer are widely recognized as an important herbivore in the northern communities of the GOE. Relatively little is know of the insect herbivore communities on Q. garryana, however surveys suggest approximately 140 species of phytophagus insects can be found on this species (Fuchs 2001). Several of these species are considered exotic pests, such as Operophtera brumata (winter moth), while others, like Erynnis propertius (propertius duskywing), are threatened and of conservation concern (Fuchs 2001, Prior and Hellmann 2010).   Furthermore, the GOE has experienced significant losses in habitat area. Following European settlement, this ecosystem has largely been developed or converted to agricultural land, with   7 only about 1% of habitat remaining intact (Fuchs 2001). As such, most populations of Q. garryana are relatively small and extremely isolated, potentially making it less likely for them to sustain diverse insect communities or species of higher tropic levels.   2.2 Tree sampling and trait measurements To assess variation in herbivory and infer the relative investment of a tree to defense, I sampled leaves from 18 populations of Q. garryana dispersed across the extent of its range, a gradient of approximately 7° in latitude (Figure 1). I sampled leaves and quantified traits at the beginning and end of the 2015 growing season, with the early spring sampling (May 5 – 21, 2015) timed to correspond with full leaf emergence, but prior to leaf maturation. I expected leaves during this period to be most vulnerable and herbivore diversity greatest. I timed the sampling of individual populations to ensure they were at approximately the same phenological stage when sampled. In the spring, samples were collected from 12 trees in all populations except for the eleventh population (Appendix A), at which I sampled all 8 available trees. At the end of the growing season (August 17 - 30, 2015), I revisited all 18 populations to estimate the total annual damage from insect herbivores. I increased the sampling effort in the late summer sampling, and collected data from 15 trees in all populations, except for the eleventh (Appendix A).  To detect the relative importance of local environmental and genetic contributions to insect herbivory on Q. garryana, I also took trait measurements on trees growing in a common garden. The garden was established in 2006 at the University of British Columbia, Canada (49° 13’N, 123° 6’W), when acorns were planted from 9 natural populations of Q. garryana var. garryana (Huebert 2009). The latitudinal range of the common garden populations corresponds to that of   8 my field samples. Similar to the field sampling, I collected leaves during two sampling periods in 2015: spring (May 24 - June 5) and late summer (September 11 – 25). I sampled up to ten trees for each of the nine populations, excluding trees less than 1.3 m tall.  For each sampled tree in the field and common garden, I measured tree size, four leaf traits, and at the time of sampling, subsampled the insect community. To account for daily temporal variation in insect activity, collections were made from 9:00 and 16:00h on warm, sunny days. I standardized sampling of the branches by removing a 50 cm segment from a terminal, fourth-order or higher branch in the lower canopy. To sample the insect communities, branches were beaten and insects immediately preserved and stored in vials with 50 ml of 70% ethanol. Upon returning to the lab, I sorted and identified the insects to morphospecies. Predatory or parasitoid species were not included in this study. Furthermore, insect species of conservation concern were recorded but not collected if found. I found few insects on the trees during the late summer field sampling and in the common garden, and therefore I only conducted analyses on the insect communities collected during the early (May) field sampling.  From the 50 cm segment of branch, I removed all leaves and randomized them by blindly mixing the leaves for 20 seconds. During the spring sampling, I selected 20 leaves per tree, which I increased during the late summer sampling to 30 leaves per tree. For each leaf, I classified the leaf damage as chewed, skeletonized, or mined, visually estimating the percent damage of each type, and summed those to estimate the total percent damage per leaf. I then randomly selected five leaves to bring back to the lab to quantify specific leaf area (SLA), and with three of these five leaves I quantified the density of trichomes on the abaxial side of the leaf. I quantified   9 specific leaf area using the field methods outlined by Perez-Harguindeguy et al. (2013). I punched a 15 mm diameter hole in each leaf, creating a sample of known area. After drying the leaves at 65° Celsius for 48 hours, I assessed the mass of each sample, and calculated the SLA in units of mm3/mg. I quantified trichome density by counting the number of trichomes in a 3.5 mm2 field of view, at five times magnification, converting the counts to densities. In the field, I also measured the diameter at breast height and estimated the height of each tree, and used these measures to estimate volume, based on the assumption that a perfect cylinder accurately represents tree trunks.  To test for chemical and nutritional differences across populations, I estimated the carbon to nitrogen ratio and relative phenolic content of leaves for a subset of populations: two populations at the southern and two at the northern end of the range, with samples from both sampling periods. I also included two populations from both sampling periods from the common garden. I selected five trees from each population and randomly selected five leaves from each tree. The leaf tissue was then ground in a ball mill grinder, and 3 !g of tissue used to estimate the ratio of carbon to nitrogen via combustion analysis and an Elemental Analyzer. Since populations did not significantly differ in the mean ratio of carbon to nitrogen, no further samples were quantified.   Finally I assessed the relative percentage of total phenolics using Folin Ciocaulteau assays. To extract the phenolics, I combined approximately 25mg of ground leaf tissue with 10 ml of methanol and placed the solution on a shaker overnight. I combined 20 !l of this solution with 100 !l of Folin-Ciocaulteau reagent, 500 !l of 20% Na2Co3, and 380 !l of distilled water, vortexing the solution after each addition. After 20 minutes, I read the absorbance of 200 !l of   10 the assay and weighted the values to the exact mass of dry leaf sample used to create the solution. Again, I did not find significant differences between populations or sampling periods and therefore further assays were not conducted.  To estimate the size of each population in the field, I drew polygons around each population in Google EarthPro version 7.1,  (https://www.google.com/earth/explore/products/desktop.html) and estimated the approximate area. I divided populations into four size classes: small (<1 ha), medium (1 ha – 10 ha), large (10 ha – 100 ha) and extremely large populations (> 100 ha).  2.3 Statistical analysis I began my analyses by determining whether the populations I sampled in the field and those in the common garden differed in both herbivore damage and traits. I constructed linear models with population as a fixed effect and either total herbivore damage, SLA, or trichome density as the response variable. Using a likelihood ratio test, I compared each model to an intercept model to determine whether populations differed. I used the same approach to determine whether measurements from the two sampling periods differed. Given that I observed significant differences between populations and between the two sampling periods in the field samples, I included population as a random effect in my models and performed analyses on the two sampling periods individually. I ran similar analyses for the common garden samples, but failed to find differences between the sampling periods and between maternal populations for the extent of damage or for SLA.     11 To evaluate whether different climate variables influence herbivory, I obtained climate data for each site from 1960 to 2013 using ClimateWNA (Wang et al. 2012). Given the importance of winter (December to February) and spring (March to May) conditions on insect communities and leaf phenotypic plasticity, I only included the long-term mean spring and winter temperatures and precipitation, long-term mean annual temperature and precipitation, and the number of growing degree days with temperature exceeding 18 °C in the model selection. To detect the potential effects of recent weather, I obtained similar data for the winter (December 2014 to February 2015) and spring (March to May of 2015) prior to my sample collection, using data from Environment Canada for populations in Canada and data from the National Oceanic and Atmosphere Association for those in the USA (http://climate.weather.gc.ca; https://www.ncdc.noaa.gov/cdo-web/). Among the populations sampled, mean annual precipitation varied 2.5 fold, ranging from 404 to 1586 mm per year, while minimum winter temperature among populations also differed, ranging from -2.04 °C to 2.52 °C. Site elevation was also included as a potential predictor variable for each response variable.   I tested the effects of all community variables on SLA and trichome density by constructing linear mixed effects models, in which each leaf trait was a response variable and community factors the predictor variables. Models were constructed for the field and common garden samples separately, with latitude, elevation, tree volume, annual weather, and long-term climate variables as fixed effects, and population as a random effect. Using a hierarchical approach, I began by modeling each predictor variable alone, and determined whether these models were different from an intercept only model and to each other using Akaike’s Information Criterion, with delta AIC values greater than two indicating that the predictor variable significantly   12 improved the fit of the model (Burnham and Anderson 2002). Based on this, I selected the best precipitation, temperature, and weather response variables and included them in more complex models with the other factors, manually performing a forward-stepwise approach, ranking models based on their AIC. Of the highest ranking models with delta AIC less than 2, the simplest model was selected as the best. These analyses were performed using the lmer function in the package lme4 (Bates et al. 2014).   Variation in local insect communities for the spring field samples was assessed using estimates of Shannon’s diversity (H’) and the mean abundance of individual insects per sample per population. To account for sampling effort, I excluded the eleventh population, which had only 8 trees. Furthermore, I conducted these analyses at the population level. I constructed linear mixed effects models using each of these metrics as the response variable to determine the extent to which latitude, population size, tree volume, SLA, trichome density, long-term mean climate and annual weather explained the observed variation. I used the same stepwise regression approach described above to choose the best model.  I explored how leaf damage was affected by climate, leaf traits, or population size by developing models in which insect damage was the response variable and community conditions the predictor variables. I modeled leaf damage using generalized linear mixed effects models (GLMM) with a binomial family and the default logit function. As I was interested in the mean leaf damage per tree within a population, I averaged individual leaf measurements for each tree, for a total sample size of 212 trees in the spring field sample, 263 trees in the summer field sample, and 94 trees from both sampling periods in the common garden. I converted herbivore   13 damage per leaf into binomial data, using 100, the percent of the entire leaf area, as the binomial denominator (as discussed in Crawley 2005). The proportion of removed tissue was considered the frequency of successes, while the proportion of the remaining tissue the frequency failures. In these models, the predictor variables were mean SLA, trichome density, volume, long-term climate variables, annual weather variables, and the interaction between the two focal leaf traits, and were treated as fixed effects. Population was included as a random factor, accounting for dependences between trees within a population. To find the most parsimonious model, I applied the same method of model selection as above. In addition to modeling total herbivore damage, I also built models for chewing, skeletonizing, and mining damage individually. The frequency of mining damage observed during the spring sampling period was negligible and therefore the analyses of what factors explained variation in this type of damage at this stage in the growing season could not be explored further. In the analysis of herbivory in field samples, a total of 55 models were created, while 47 models were used to analyze herbivory in the common garden. All models of herbivore damage were built using the glmer function in the R package lme4. Finally I tested whether the relative ranking of traits in the field populations corresponded to that of populations in the common garden, using a Mann-Whitney U-test. All analyses were performed using R version 3.3.0 (R core Development Team 2016).   14 Chapter 3: Results  The abundance of insects on Quercus garryana varied significantly geographically (F1,15 = 4.88, P = 0.043), but I found no significant differences in insect diversity among populations (F1,15 = 0.45, P = 0.51). In the spring, differences in insect abundance were best explained by latitude, specific leaf area (SLA), and current spring precipitation (Table 1), with the highest abundances occurring in low latitude populations with dry spring conditions and on trees with low SLA (Figure 2).  Population differences in leaf traits were also explained by abiotic factors, specifically the long-term mean spring precipitation. Both SLA and trichome density differed between the two sampling periods and among populations (Table 2 and 3); however, the factors that explained the within-season variation only differed for trichome density (Table 4). Specifically, in spring, prior to full leaf maturation, the minimum spring temperature of the current year best explained trichome density, with higher densities occurring in cooler populations. The trichome density of mature leaves sampled later in the growing season was explained by long-term mean spring precipitation and volume (Table 4), with higher densities occurring in populations with dry springs and trees with larger volumes (Figure 3B). The SLA of leaves collected in both sampling periods was also best explained by long-term trends in mean spring precipitation and volume (Figure 3A). SLA was greater in populations with wetter springs and lower tree volumes (Figure 3A). Interestingly, both traits decreased over the course of the growing season, with the SLA of summer samples decreasing on average by 31.1 ± 1.8% and trichome densities decreasing on average by 27.7 ± 4.1%.   15 Latitude alone was a poor predictor of the extent of herbivore damage experienced by trees among populations (Spring: !21 = 1.19, P = 0.27, Summer: !21 = 0.017, P = 0.90), suggesting that the population differences were due to other abiotic or biotic factors (Table 3). Similar to leaf traits, the extent of damage differed between the two sampling periods (Table 2). Although not a significant increase, leaf damage of the spring samples in the field was greater on average in high latitude populations compared to low latitude populations (Figure 4A), with samples taken late in the season having a 21.2 ± 8.3% relative increase in damage on average.  The amount of herbivore damage was best explained by variation in climate and population size, with the relative importance of each of these factors differing between the two sampling periods (Table 2). Leaf damage in the spring was best predicted by the interaction between SLA and trichome density, and tree volume (Table 5). Trees that were small and produced leaves with lower SLA and low densities of trichomes experienced the most damage (Table 5). In contrast, the total herbivore damage incurred by the end of summer was best explained by the interaction between SLA and trichome density, long-term mean spring precipitation, and population size class (Table 5). Populations with drier average spring conditions experienced greater insect damage in large populations (Figure 5A), while damage was greatest for leaves with low SLA and low trichome densities (Figure 5B). Across all models in which population size was significant, trees growing in small populations less than 1 ha in area consistently experienced less damage than those between 10 and 100 ha in area (Figure 6).  In light of the importance of specific abiotic factors in explaining variation in insect community diversity, it is not surprising that these factors also affected the distribution of damage by   16 different feeding guilds. The amount of chewing and skeletonizing damage remained relatively constant across the growing season (Table 2), but both showed significant variation between populations (Table 3). Chewing damage in the spring was best explained by long-term spring precipitation, the interaction between SLA and trichome density, and population size class (Table 6). The most chewing damage occurred in larger populations with drier springs, and for leaves with low SLA and low trichome densities (Table 6). Later in the season, chewing damage was best explained by the interaction between SLA and trichome density and tree volume, with the most damage again occurring on leaves with low SLA and trichome densities and on trees with low volumes (Table 6). In turn, skeletonizing damage in the spring was best explained by long-term spring precipitation and leaf SLA, with the most damage occurring on leaves with high SLA and under drier spring conditions (Table 6). As the growing season progressed, this changed, with the simplest model for skeletonizing damage including only SLA (Table 6). Damage by mining insects differed between the two sampling periods, but only the summer samples were suitable for analysis (Table 2). For these samples, long-term mean spring precipitation and population size class best explained variation in mining damage late in the summer, with the most damage occurring in larger populations with drier springs (Table 6).   Finally, no variation was observed in the extent of herbivore damage among trees of different maternal populations (Table 7). Of the two leaf traits considered, only trichome density of leaves sampled late in the growing season showed significant differences in relation to provenance (Table 7). It should be noted, however, that leaf damage in the common garden was low relative to the field populations (Figure 4B), with an average damage of 4.52 ± 0.40 % for the spring samples and 4.70 ± 0.42% for summer samples. The patterns of trees’ resource allocation to   17 mechanical defenses observed in the field were not preserved in the common garden, as the ranking of trichome densities in populations in the field does not match that of trees in the common garden (W=39747.5 P=0.80).      18 Chapter 4: Discussion  Variation in herbivory across the entire latitudinal range of Quercus garryana is best explained by spring precipitation, leaf traits, and population size (Table 5). The relative importance of each of these factors, however, varied depending on when the damage occurred and by insect feeding guild (Table 5 and 6). Conversely, trees of different maternal populations in a common garden did not differ in the extent of damage they received (Table 7). This corroborates my findings in the field, that environmental variation, more than genetic differences, drive the observed variation in herbivore damage, since in its absence, variation in damage between populations disappeared.   The results of this study broadly reject the latitudinal gradient in plant defense hypothesis, as only insect abundance was partially explained by latitude (Table 1). This finding is in disagreement with several recent studies, including Abdala-Roberts et al. (2016), who found that trichome density varied with latitude, with greater densities of trichomes on leaves of Ruellia nudiflora at higher latitudes. While Andrew and Hughes (2005), also failed to find a latitudinal gradient in the percent leaf damage on leaves of Acacia falcate, they did, however, find that SLA decreased with increasing latitudes, with lower values at high latitudes reflecting greater investment in leaf mass and structure relative to photosynthetic area. The inconsistencies from the plant defense hypothesis that I observed in this study may arise at least in part from the difference in scale between this study and the seminal studies on which this hypothesis is based. Early studies that support the existence of latitudinal gradients were conducted across large spatial scales, comparing species from tropic to temperate regions (Coley and Aide 1991). In   19 contrast, the latitudinal range over which this study was conducted (7° latitude) is much narrower and thus includes less variation in climate factors. Regardless, intraspecific studies that encompass the range of a single species are important for determining the drivers of fine-scale variation that underlies global trends.  In the absence of a latitudinal gradient, long-term mean spring precipitation was a major determinant of herbivore abundance, leaf phenotype, and their relative effects on plant-herbivore interactions (Table 5). Insects in populations with drier springs were more abundant, suggesting that local climate conditions in these populations were more favorable and enabled greater population growth (Figure 2). Since this study was only conducted over a single growing season, it is difficult to know to what extent the insect community and leaf traits reflect current or past climate conditions. If, for example, insect abundance responds quickly to current conditions, the resulting lag in the response of leaf traits to climate or herbivore pressure may mean the trends I observed are the result of previous selection.  Longer-term research into the annual variation in insect population traits and leaf traits is therefore needed to draw stronger causal relationships of the indirect effects of climate on herbivore pressure.  The long-term mean in spring precipitation was also the only abiotic factor that influenced the extent of damage by various feeding guilds. At some point in the growing season, all types of damage were partially explained by long-term trends in spring precipitation, increasing with reduced precipitation (Table 6). The direction of these responses are in agreement with the expectations of how water-stress effects plant’s susceptibility to insect herbivores, as discussed by Huberty and Denno (2004). However, the variation in mining damage contrasts that observed   20 in other studies. A recent study by Lecky et al. (2014) found mining damage increased in wetter populations, potentially as a result of reduced larval mortality and increased potential for population growth (Yarnes and Boecklen 2005).    Although the expression of leaf traits in this study fulfills the expectations of the resource limitation hypothesis, the trends I observed in herbivore damage suggest that environmental stressors are of greater importance. As expected under the resource limitation hypothesis, leaves in drier populations invested more in leaf structure, producing leaves with lower SLA values and higher trichome densities (Figure 3 A and B). The slight increase in SLA as the amount of spring precipitation increased is expected given that plants in moister areas are able to increase photosynthetic leaf area inexpensively by producing more turgid cells (Ryser 1996), resulting in greater leaf area relative to dry mass. Trichome density is also expected to decline with increasing precipitation, due to reduced selection for efficient water use in wetter populations (Guerfel et al. 2009). Interestingly, herbivore damage in drier populations was highest, suggesting that lower SLA and trichome density did not confer greater defense against insect herbivores (Figure 5 A and B). This finding is more in line with predictions of the plant stress hypothesis, which predicts more damage in more stressful populations (White 1984, Cornelissen et al.  2009). The observed SLA and high trichome densities in this study may therefore be due to their alternate roles of reducing water loss and not to deter insect herbivores. Given the observational nature of this study, such causal mechanisms cannot be confirmed and further research is needed to determine whether defense traits not included here are capable of reducing herbivore damage in populations with wetter springs.    21 The observed variation in herbivory and traits may be also influenced by the variation in plant and herbivore phenology caused by current climate conditions. Given the number of populations I sampled and the scale of the study, leaves were only approximately at the same stage in development during the spring sampling. There may have been small differences in their degree of expansion and allocation of resources to cell structure. These phenological differences may contribute to the difference in trait values observed between the two sampling periods (Table 2), as well as the importance of the current spring temperature in explaining variation in trichome density (Table 4).  Leaves developing in populations with cooler temperatures, and presumably lower rates of maturation, may not have expanded as much at the time of sampling as those developing under warmer temperatures, potentially contributing to higher trichome densities (Table 4). Variation in the phenology of the insect community across the range may also confound the results, as our insect sampling may have missed early and late emerging species, leading me to underestimate both insect diversity and abundance.   The extent of damage was positively correlated with population size, in agreement with island biogeography theory (Figure 6). This trend has been widely observed for the consumption of different plant tissue, including seeds (Colling and Matthies 2004, Agrens et al. 2008), and flowers (Kery et al. 2001), as well as for specific herbivore guilds (Bukovinszky et al. 2010). Interestingly, large populations of Q. garryana, that were between 10 ha – 100 ha had the most damage (Figure 6). The decline in damage in populations greater than 100 ha may be the result of trophic cascades, with very large populations able to support predator populations large enough to limit the effects of herbivores on the plant community. Elzinga et al. (2005) observed such a response in herbivore damage in large populations of Silene latifolia, as higher levels of   22 parasitism reduced herbivore populations. Other studies of plant-herbivore responses to tritrophic interactions have, however, found no relationships between parasitism and fragmentation (Braschler et al. 2003).   Finally, the absence of variation in both traits and damage in the common garden suggest a negligible influence of genotype on the interactions between Q. garryana and its insect community. This is supported by the, albeit limited, genetic research on this species, which has shown genetic diversity to vary only slightly across the species range (Fuchs 2001, Degner 2014). Furthermore the inconsistency in population ranks in trichome density between field and common garden samples suggest factors other than genetic variation between populations are driving the significant differences in this trait observed in the common garden. Collectively, these findings are in agreement with the results from the field sampling, where I found that local abiotic conditions, such as precipitation and population size, contributed the most to explaining the observed variation in herbivory. In contrast to these findings, other common garden studies have found defense traits to be genetically controlled and gradients in herbivory to persist in the absence of environmental variation (Pennings et al. 2001, Salgado and Pennings 2002). Comparisons between studies are limited by differences in protocol, however, as some studies apply feeding assays to test for differences in plant palatability (Salgado and Pennings 2002), as opposed to using naturally occurring levels of herbivory, as was done in this study. The trends I observed may be confounded by the ability of insect herbivores to colonize the common garden, which may be limited as there are no proximal, natural populations of Q. garryana nearby. It is possible that this isolation, and the resulting low herbivore pressure, prevented the detection of genetic differences in tree’s susceptibility to herbivory.    23  To fully understand the importance of genetic diversity in Q. garryana to plant-herbivore interactions will require additional research to account for these effects. The continued adoption of new genetic approaches may confirm the trends in variation observed in previous studies in which the molecular markers used only provide a proxy for the overall levels of gene diversity within a species and therefore may not reflect the diversity of adaptation to specific environmental factors (Ritland et al. 2005), such as adaptations to herbivore pressure.      24 Chapter 5: Conclusion  The extent of insect herbivory on Quercus garryana, is dependent on leaf traits, spring precipitation, and population size, with individual herbivore feeding guilds varying in their relative response to each of these factors. Although latitude explains some of the variation in herbivore abundance, in this system it did not affect the outcome of oak-insect interactions. Furthermore, the expression of certain leaf traits may not be constitutive, but rather may vary in response to current environment conditions. Collectively, the results presented here demonstrate the importance of examining individual community factors and evaluating the potential for variation in the herbivore community and genotype to contribute to the outcome of plant-herbivore interactions.   Future research should continue to include both abiotic and biotic factors when testing for context-dependencies in plant-herbivore interactions. Greater consideration of the diversity and abundance of insect community and higher trophic levels may also be crucial to account for top-down controls on herbivore pressure through trophic cascades (Marczak et al. 2011). Finally, a mechanistic understanding of the influence of individual factors will require experimentation and the direct manipulation of community conditions and herbivore communities across the landscape. Although logistically challenging, manipulating community factors and performing transplant and common garden studies at multiple points within a species’ range will allow us to identify the causal mechanisms that shape plant herbivore-interactions and further disentangle the relative importance of climate, herbivore communities, and genetic diversity.    25 Herbivore damage can have considerable deleterious effects on plant fitness and play a major role in shaping plant communities and distributions (Crawley 1989, Bruelheide and Scheidel 1999, Ehrlen 2003, Maron and Crone 2006). As species distributions change, with many insect herbivores moving poleward, identifying the context-dependencies of plant herbivore interactions also has practical applications. In addition to allowing for better predictions of changes in interactions in response to climate change, appropriate management strategies can be developed, and effective biocontrols identified (Syrett et al. 2000, Shea et al. 2005, HilleRisLambers et al. 2013). Adopting a holistic approach in studying the geographic variation in plant-herbivore interactions will lay the foundation for the predictive framework needed to achieve these important goals.     26 Table 1: Model parameters of the best model explaining variation in insect abundance on Quercus garryana in situ. Models were only developed for the insect communities collected in the spring, given the rarity with which they were found during the summer sampling.   Spring Intercept 452.42 SLA -1.66 Trichome Density  SLA " Trichome Density  Mean spring precipitation  Minimum winter temperature -0.29 Mean Annual Precipitation  Volume  Latitude -7.36                     27 Table 2: Tests for differences between spring and summer leaf samples of Quercus garryana, including each type of damage and the two focal leaf traits quantified in the field and common garden. Leaf damage by individual feeding guilds in the common garden was insufficient for individual analyses.     Field   Common Garden  !2 df P-value   !2 df P-value All damage 18.85 1 <0.001   0.099 1 0.75 Chewed damage 0.81 1 0.37      Skeletonizing damage 0.32 1 0.57      Mining damage 13.87 1 <0.001      SLA 270.81 1 <0.001   3.73 1 0.053 Trichome Density 58.46 1 <0.001   4.55 1 0.033                  28 Table 3: Tests for population differences for each type of herbivore damage and leaf trait for 18 natural populations of Quercus garryana for each sampling period.    Spring   Summer   F df P   F df P Herbivore damage All damage 47.46 17,196 <0.001   41.66 17,245 <0.001  Chewed damage 43.25 17,192 <0.001   36.12 17,245 <0.001  Skeletonizing damage 2.47 15,78 0.005   25.65 17,242 <0.001  Mining damage 2.94 17,223 0.0001   7.35 17,212 <0.001           Traits SLA 23.42 17,190 <0.001   5.75 17,245 <0.001  Trichome Density 8.45 17,190 <0.001   11.82 17,245 <0.001              29 Table 4: Ranking and parameters of models explaining variation in specific leaf area (SLA), trichome density, and insect abundance, with delta AIC values less than 2. Models were developed for both the spring and late summer samples for leaf traits of Quercus garryana measured in situ. Estimates of regression coefficients are calculated using maximum likelihood.    Specific Leaf Area   Trichome Density   Spring   Summer   Spring   Summer Rank 1 2   1 2 3 4   1 2   1 2 !AIC 0.00 0.11   0.00 1.26 1.46 1.76   0.00 1.53   0.00 1.52 Model weight 0.25 0.24   0.37 0.2 0.18 0.15   0.42 0.19   0.65 0.31                  Intercept 8.02 4.73   8.02 8.84 10.21 4.73   42.78 22.54   23.79 3.87 Mean spring precipitation 0.0095 0.010   0.0095   0.010       -0.044 -0.040 Minimum winter temperature           -3.80 -3.93     Mean Annual Precipitation      0.0024           Volume -0.086 -0.086   -0.086 -0.086 -0.085 -0.086       0.50 0.49 Latitude  0.068      0.068    0.45    0.41   30 Table 5: Ranking and parameters of models for total insect herbivore damage with delta AIC values less than 2. Models were developed for both the spring and late summer samples individually for 18 natural populations. Maximum likelihood was used to estimate the regression coefficients for each model.    Spring   Summer Rank 1 2 3   1 2 !AIC 0.00 0.43 1.87   0.00 1.12 Model weight 0.35 0.28 0.14   0.46 0.26         Intercept -2.48 -0.93 -0.021   -1.33 -1.36 SLA 0.0062 0.0065 0.0072   0.012 0.012 Trichome Density 0.016 0.015 0.016   0.043 0.045 SLA " Trichome Density -0.0022 -0.0021 -0.0022   -0.0055 -0.0056 Mean spring precipitation  -0.0099 -0.011   -0.0086 -0.0085 Volume -0.0031  -0.0028    -0.0056 Population size: 1–10 ha  0.58    0.86 0.87 Population size: 10–100 ha  1.49    1.60 1.60 Population size: >100 ha  0.75    1.49 1.49 Latitude                 31 Table 6: Ranking and parameters of models of chewing, skeletonizing, and mining leaf damage with delta AIC values less than 2. Models were developed for both the spring and late summer samples for both chewing and skeletonizing damage measured in situ; however, due to the low frequency of occurrence in spring samples, models of mining damage were constructed for the summer samples only. Estimates of regression coefficients are calculated using maximum likelihood.   Chewing damage Mining damage Skeletonizing Damage  Spring   Summer   Summer   Spring Rank 1 2   1 2   1   1 2 3 4 !AIC 0.00 1.42   0.00 1.39   0.00   0.00 0.12 1.48 1.61 Model weight 0.50 0.24   0.45 0.23   0.69   0.19 0.18 0.09 0.08                 Intercept -1.28 -1.28   -1.28 -2.47   -4.48   -1.73 -1.72 -1.90 -1.89 SLA 0.20 0.020   0.020 0.020      0.0045  0.018 0.0056 Trichome Density 0.034 0.034   0.034 0.036       -0.0045 0.011  SLA " Trichome Density -0.0048 -0.0048   -0.0048 -0.0049        -0.00074  Mean spring precipitation -0.0083 -0.0083   -0.0083    -0.0072   -0.0047 -0.0045 -0.0045 -0.0051 Minimum winter temperature               0.070 Volume  -0.00038    -0.00091          Population size: 1–10 ha 0.71 0.71   0.71    2.48       Population size: 10–100 ha 1.26 1.26   1.26    2.55       Population size: >100 ha 0.89 0.89   0.89    2.34       Latitude                    32 Table 6: (continued): Ranking and parameters of models of chewing, skeletonizing, and mining leaf damage with delta AIC values less than 2. Models were developed for both the spring and late summer samples for both chewing and skeletonizing damage measured in situ; however, due to the low frequency of occurrence in spring samples, models of mining damage were constructed for the summer samples only. Estimates of regression coefficients are calculated using maximum likelihood.   Skeletonizing damage  Summer Rank 1 2 3 4 5 6 7 8 9 !AIC 0.00 0.12 0.26 1.27 1.52 1.52 1.75 1.76 1.79 Model weight 0.11 0.10 0.10 0.06 0.05 0.05 0.05 0.05 0.05           Intercept -0.84 -1.05 -1.75 -3.35 -2.67 -2.51 -0.99 -1.04 -2.66 SLA -0.05 -0.044 -0.037 -0.044 -0.045 -0.051 -0.034 -0.050 -0.035 Trichome Density -0.0070  0.00048   -0.0068 0.0035 -0.0068 0.0041 SLA " Trichome Density   -0.00080    -0.0011  -0.0011 Mean spring precipitation -0.0073 -0.0071 -0.0073    -0.0073 -0.0077  Minimum winter temperature        0.087  Volume         -0.00019 Population size: 1–10 ha   0.66 0.42      Population size: 10–100 ha   0.97 0.94      Population size: >100 ha   1.50 1.39      Latitude               33 Table 7: Tests for population differences for total herbivore damage and leaf traits for samples collected in the common garden from 9 populations.    Spring   Summer   F df P   F df P Herbivore damage All damage 0.53 8,85 0.83   0.24 8,85 0.98           Traits SLA 51.56 8,80 0.42   15.63 8,73 0.35  Trichome Density 0.60 8,80 0.77   2.14 8,73 0.042                34  Figure 1: Map of site locations overlaid on the species distribution Quercus garryana var. garryana, shown in dark grey. The black dashed line is located at the approximate distribution edge of this variety, below which the two shrub varieties, var. semota and var.  breweri, dominate. The yellow star signifies the location of the common garden used in this study.       !!!!!!!! !!!!!!!!!!!!!!−128 −126 −124 −122 −120 −118 −1164244464850!!Common garden populationsField populations4244464850!"#$%&'()*!+,&'()*  35    Figure 2: Insect abundance across latitudes for: A) high, low and mean spring precipitation of the current growing season, B) high, low, and mean SLA. High values represent those in the top 30th percentile and low values those in the bottom 30th percentile for the factor. Each point represents the mean abundance of insects per tree for each population in situ. Estimates were calculated using linear mixed effects models.     43 44 45 46 47 48 49020406080100120LatitudeAbundanceHigh PrecipitaionLow PrecipitaionMean Precipitaion43 44 45 46 47 48 49020406080100120LatitudeAbundanceHigh SLALow SLAMean SLA!"#$%&'()*+',$("-%*&"*,'(./( 0/(43 44 45 46 47 48 49020406080100120LatitudeAbundanceHigh PrecipitaionLow PrecipitaionMean Precipitaion43 44 45 46 47 48 49020406080100120LatitudeAbundanceHigh SLALow SLAMean SLA  36  Figure 3: Variation in leaf traits for samples taken late in the growing season for natural populations. A) Variation in SLA for spring samples versus the long-term mean in spring precipitation, with lines depicting the upper and lower 20th percentile of tree volumes and the mean. B) Variation in trichome density for spring samples versus latitude, again with lines depicting the upper and lower 20th percentile of tree volume. Each point represents the mean trait value per tree. Estimates were calculated using linear mixed effects models.  !"# $"#%&'(#)*+,(-#*+&.,*,/'01(#2334#5*&.,6.#7&'8#'+&'#2339 3-:; 4#<+,.=13&#>&(),/?#233:;4#150 200 250010203040Mean Spring Precipitation (mm)Specific Leaf Area (mm2  mg!1 )Large treesSmall treesMean tree volume150 200 2500102030405060Mean Spring Precipitation (mm)Trichome Density (mm!1)Specific Leaf Area (mm2  mg!1 )lllTrichome Density (mm!1)  37  Figure 4: Mean proportion of damage per tree across latitudes, measured in situ: A) the naturally occurring populations of Quercus garryana, and B) the common garden populations, in which damage is lower overall and less variable. Each point represents the mean proportion of leaf damaged per tree.    43 44 45 46 47 48 490.00.20.40.60.81.0LatitudeProportion damage per treeMayAugust44 46 48 500.00.20.40.60.81.0LatitudeProportion damage per tree43 44 45 46 47 48 490.00.20.40.60.81.0LatitudeProportion damage per treeMayAugust44 46 48 500.00.20.40.60.81.0LatitudeProportion damage per tree!"# $"#%&'()*+#,-./.-'.0#*&1&2+#/+-#(-++#  38  Figure 5: Responses of total in situ herbivore damage to factors identified through model selection. A) Variation in herbivore damage with increasing long-term mean spring precipitation and in response to differences in population size class. B) Variation in herbivore damage with trichome density, depicting the response in relation to high SLA (top 20th percentile), low SLA (lower 20th percentile), and mean SLA. Estimates were calculated using generalized linear mixed effects models.   150 200 2500.00.20.40.60.81.0Mean Spring Precipitation (mm)Proportion leaf damage per treeSmallMediumLargeExtremely Large0 10 20 30 400.00.20.40.60.81.0Trichome Density (mm!1)Proportion leaf damage per treeHigh SLALow SLAMean SLA150 200 2500.00.20.40.60.81.0Mean Spring Precipitation (mm)Proportion leaf damage per treeSmallMediumLargeExtremely Large0 10 20 30 400.00.20.40.60.81.0Trichome Density (mm!1)Proportion leaf damage per treeHigh SLALow SLAMean SLA!"#$%&'()*(+,#-.)/''012)3(4+),5"#+6)5"($#5#-47&+)/''2)89) :9);"&5&"7&+)<(4=)*4'46()5(")-"(()  39      Figure 6: Proportion of in situ leaf damage by population size class of Q. garryana. Three populations were less than 1 ha, seven populations were 1–10 ha, four populations were 10–100 ha, and four populations were >100 ha. Red diamonds represent the mean damage per leaf per tree for each size class. Each grey dot represents the mean proportion damage per tree in a population.  !"#$%&'"()*+,-).%&**)/-&()#0"#"0'"()%-&1)2&3&4-)#-0)50--)6)7)8&) 7)9)7:)8&) 7:)9)7::)8&) ;)7::)8&)1 2 3 40.00.20.40.60.81.0Population size classMean leaf damage per population  40 References  Abdala-Roberts, L., X. Moreira, S. Rasmann, V. Parra-Tabla, and K. A. Mooney. 2016. Test of biotic and abiotic correlates of latitudinal variation in defences in the perennial herb Ruellia nudiflora. Journal of Ecology 104:580-590. Adams, J. M., and Y. Zhang. 2009. Is there more insect folivory in warmer temperate climates? A latitudinal comparison of insect folivory in eastern North America. Journal of Ecology 97:933-940. Agren, J., J. Ehrlen, and C. Solbreck. 2008. Spatio-temporal variation in fruit production and seed predation in a perennial herb influenced by habitat quality and population size. Journal of Ecology 96:334-345. Andrew, N. R., and L. Hughes. 2005. Herbivore damage along a latitudinal gradient: relative impacts of different feeding guilds. Oikos 108:176-182. Bale, J. S., G. J. Masters, I. D. Hodkinson, C. Awmack, T. M. Bezemer, V. K. Brown, J. Butterfield, A. Buse, J. C. Coulson, J. Farrar, J. E. G. Good, R. Harrington, S. Hartley, T. H. Jones, R. L. Lindroth, M. C. Press, I. Symrnioudis, A. D. Watt, J. B. Whittaker. 2002. Herbivory in global climate change research: direct effects of rising temperature on insect herbivores. Global Change Biology 8:1-16. Bates D, M. Maechler, B. Bolker and S. Walker. 2014. lme4: Linear mixed-effects              models using Eigen and S4_. R package version 1.1-7             http://CRAN.R-project.org/package=lme4. Downloaded on 9 March 2015. Braschler, B., G. Lampel, and B. Baur. 2003. Experimental small-scale grassland fragmentation alters aphid population dynamics. Oikos 100:581-591.   41 Bruelheide, H., and U. Scheidel. 1999. Slug herbivory as a limiting factor for the geographical range of Arnica montana. Journal of Ecology 87:839-848. Bukovinszky, T., R. Gols, A. Kamp, F. de Oliveira-Domingues, P. A. Hamback, Y. Jongema, T. M. Bezemer, M. Dicke, N. M. van Dam, and J. A. Harvey. 2010. Combined effects of patch size and plant nutritional quality on local densities of insect herbivores. Basic and Applied Ecology 11:396-405. Burnham, K. P., and D. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Springer, New York, New York, USA. Coley, P. D., and T. Aide. 1991. Comparison of herbivory and plant defenses in temperate and              tropical broad-leaved forests. Pages 25–49 in P. W. Price, T. M. Lewinsohn, G. W.              Fernandes, and W. W. Benson, editors. Plant–animal interaction: evolutionary ecology in              tropical and temperate regions. Wiley, New York, New York, USA. Coley, P. D., J. P. Bryant, and F. S. Chapin. 1985. Resource availability and plant antiherbivore defense. Science 230:895-899. Colling, G., and D. Matthies. 2004. The effects of plant population size on the interactions between the endangered plant Scorzonera humilis, a specialised herbivore, and a phytopathogenic fungus. Oikos 105:71-78. Cornelissen, J. H. C., S. Lavorel, E. Garnier, S. Diaz, N. Buchmann, D.E. Gurvich, P. B. Reich,        H. ter Steege, H. D. Morgan, M. G. A. van der Heijden, J. G. Pausas, H. Poorter. 2003. A         handbook of protocols for standardized and easy measurement of plant functional traits              worldwide. Australian Journal of Botany 51: 335-380. Cornelissen, T., and P. Stiling. 2009. Spatial, bottom-up, and top-down effects on the abundance of a leaf miner. Ecography 32:459-467.   42 Crawley, M. J. 1989. Insect herbivores and plant population dynamics. Annual   Review of Entomology 34: 531-564. Crawley, Michael J. 2005. Statistics: an introduction using R. John Wiley & Sons Ltd,                Chichester, West Sussex, England. Degner, J. C. 2014. Using a genotyping-by-sequencing (GBS) approach to elucidate              population structure in Garry oak (Quercus garryana). Undergraduate Thesis.              University of British Columbia, Vancouver, British Columbia, Canada. Dobzhansky, T. 1950. Evolution in the tropics. American Scientist 38:209-221. Dostal, P., E. Allan, W. Dawson, M. van Kleunen, I. Bartish, and M. Fischer. 2013. Enemy damage of exotic plant species is similar to that of natives and increases with productivity. Journal of Ecology 101:388-399. Ehrlén J. 2003. Fitness components versus total demographic effects: evaluating herbivore impacts on a perennial herb. American Naturalist 162:796–810. Elzinga, J. A., H. Turin, J. M. M. van Damme, and A. Biere. 2005. Plant population size and isolation affect herbivory of Silene latifolia by the specialist herbivore Hadena bicruris and parasitism of the herbivore by parasitoids. Oecologia 144:416-426. Endara, M. J., and P. D. Coley. 2011. The resource availability hypothesis revisited: a meta-analysis. Functional Ecology 25:389-398. Environment Canada. 2016. Historical Climate Data. http://climate.weather.gc.ca. Downloaded on 22 March 2016 Fuchs, M.A. 2001. Towards a recovery strategy for Garry oak and associated   ecosystems in Canada: Ecological assessment and literature review.   Technical Report GBEI/EC-00-030. Environment Canada, Canadian Wildlife   43 Service, Pacific and Yukon Region. Google EarthPro 7.1. 2016. https://www.google.com/earth/explore/products/desktop.html. Downloaded on 2 April 2016. Guerfel, M., O. Baccouri, D. Boujnah, W. Chaibi, and M. Zarrouk. 2009. Impacts of water stress on gas exchange, water relations, chlorophyll content and leaf structure in the two main Tunisian olive (Olea europaea L.) cultivars. Scientia Horticulturae 119:257-263. Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960. Community structure, population control, and competition. American Naturalist 94:421-425. Hawkes, C. V., and J. J. Sullivan. 2001. The impact of herbivory on plants in different resource conditions: A meta-analysis. Ecology 82:2045-2058. Huebert, C. A. 2009. The ecological and conservation genetics of Garry oak (Quercus               garryana Dougl. ex Hook). Masters Thesis. University of British Columbia,                  Vancouver, British Columbia, Canada. HilleRisLambers, J., M. A. Harsch, A. K. Ettinger, K. R. Ford, and E. J. Theobald. 2013. How will biotic interactions influence climate change-induced range shifts? Climate Change and Species Interactions: Ways Forward 1297:112-125. Huberty, A. F., and R. F. Denno. 2004. Plant water stress and its consequences for herbivorous insects: A new synthesis. Ecology 85:1383-1398. Jepson, W. 1909. The trees of California. Pages 158-161. Cunningham, Curtis, and Welch. San Francisco, California, US.  Kery, M., D. Matthies, and M. Fischer. 2001. The effect of plant population size on the interactions between the rare plant Gentiana cruciata and its specialized herbivore Maculinea rebeli. Journal of Ecology 89:418-427.   44 Koricheva, J., S. Larsson, and E. Haukioja. 1998. Insect performance on experimentally stressed woody plants: A meta-analysis. Annual Review of Entomology 43:195-216. Leckey, E. H., D. M. Smith, C. R. Nufio, and K. F. Fornash. 2014. Oak-insect herbivore interactions along a temperature and precipitation gradient. Acta Oecologica-International Journal of Ecology 61:1-8. MacDougall, A. S., M. C. Rillig, and J. N. Klironomos. 2011. Weak conspecific feedbacks and exotic dominance in a species-rich savannah. Proceedings of the Royal Society B-Biological Sciences 278:2939-2945. Marczak, L. B. C. K. Ho, K. Wieski, H. Vu, R. F. Denno, and S. C. Pennings. 2011. Latituidnal variation in top-down and bottom-up control of salt marsh food web. Ecology 92:276-281. Maron, J. L. and E. Crone. 2006. Herbivory: Effects on plant abundance, distribution, and populations growth. Proceedings of the Royal Society B-Biological Sciences 273:2575-2584. Maron, J. L., Baer, K. C., Angert, A. L. 2014. Disentangling the drivers of context-  dependent plant–animal interactions. Journal of Ecology 102:19-27. Marquis, R. J., R. E. Ricklefs, and L. Abdala-Roberts. 2012. Testing the low latitude/high defense hypothesis for broad-leaved tree species. Oecologia 169:811-820. Moles, A. T., S. P. Bonser, A. G. B. Poore, I. R. Wallis, and W. J. Foley. 2011. Assessing the evidence for latitudinal gradients in plant defence and herbivory. Functional Ecology 25:380-388.   45 Moreira, X., L. Abdala-Roberts, V. Parra-Tabla, and K. A. Mooney. 2015. Latitudinal variation in herbivory: influences of climatic drivers, herbivore identity and natural enemies. Oikos 124:1444-1452. National Oceanic and Atmospheric Administration. 2016. Climate Data Online. https://www.ncdc.noaa.gov/cdo-web/. Downloaded on 23 March 2016. Pearse, I. S., and A. L. Hipp. 2012. Global patterns of leaf defenses in oak species. Evolution 66:2272-2286. Pellissier, L., A. Roger, J. Bilat, and S. Rasmann. 2014. High elevation Plantago lanceolata plants are less resistant to herbivory than their low elevation conspecifics: is it just temperature? Ecography 37:950-959. Pennings, S. C., C.-K. Ho, C. S. Salgado, K. Wieski, N. Dave, A. E. Kunza, and E. L. Wason. 2009. Latitudinal variation in herbivore pressure in Atlantic Coast salt marshes. Ecology 90:183-195. Pennings, S. C., and B. R. Silliman. 2005. Linking biogeography and community ecology: Latitudinal variation in plant-herbivore interaction strength. Ecology 86:2310-2319. Pennings, S. C., E. L. Siska, and M. D. Bertness. 2001. Latitudinal differences in plant palatability in Atlantic coast salt marshes. Ecology 82:1344-1359. Pennings, S. C., M. Zimmer, N. Dias, M. Sprung, N. Dave, C.K. Ho, A. Kunza, C. McFarlin, M. Mews, A. Pfauder, and C. Salgado. 2007. Latitudinal variation in plant-herbivore interactions in European salt marshes. Oikos 116:543-549. Perez-Harguindeguy, N., S. Diaz, E. Garnier, S. Lavorel, H. Poorter, P. Jaureguiberry, M. S. Bret-Harte, W. K. Cornwell, J. M. Craine, D. E. Gurvich, C. Urcelay, E. J. Veneklaas, P. B. Reich, L. Poorter, I. J. Wright, P. Ray, L. Enrico, J. G. Pausas, A. C. de Vos, N.   46 Buchmann, G. Funes, F. Quetier, J. G. Hodgson, K. Thompson, H. D. Morgan, H. ter Steege, M. G. A. van der Heijden, L. Sack, B. Blonder, P. Poschlod, M. V. Vaieretti, G. Conti, A. C. Staver, S. Aquino, and J. H. C. Cornelissen. 2013. New handbook for standardised measurement of plant functional traits worldwide. Australian Journal of Botany 61:167-234. Poorter, H., U. Niinemets, L. Poorter, I. J. Wright, and R. Villar. 2009 Causes and consequences of variation in leaf mass per area (LMA): a meta-analysis. New Phytologist 182:565-588. Price, P., R. F. Denno, M. D. Eubanks, D. L. Finke, I. Kaplan. 2011. Insect Ecology: Behavior, Populations, and Communities, Part III: Species interactions. Pages 99-183. Cambridge University Press, Cambridge, UK. Prior, K. M. and J. J. Hellmann. 2010. Impact of an invasive oak gall wasp on a native butterfly: a test of plant-mediated competition. Ecology 91:3284-3293. R Core Development Team. 2013. R: A language and environment for statistical computing. R  Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/.  Downloaded on 9 March 2015. Ritland, K., D. G. W. E. Meagher, and Y.A. El-Kassaby. 2005. Isozyme variation and the                 conservation genetics of Garry oak. Canadian Journal of Botany 83:1478-1487. Rivas-Ubach, A., A. Gargallo-Garriga, J. Sardans, M. Oravec, L. Mateu-Castell, M. Perez-Trujillo, T. Parella, R. Ogaya, O. Urban, and J. Penuelas. 2014. Drought enhances folivory by shifting foliar metabolomes in Quercus ilex trees. New Phytologist 202:874-885. Ryser, P. 1996. The importance of tissue density for growth and life span of leaves and roots: A comparison of five ecologically contrasting grasses. Functional Ecology 10:717-723.   47 Schemske, D. W., G. G. Mittelbach, H. V. Cornell, J. M. Sobel, and K. Roy. 2009. Is There a Latitudinal Gradient in the Importance of Biotic Interactions? Annual Review of Ecology Evolution and Systematics 40:245-269. Shea, K., D. Kelly, A. W. Sheppard, and T. L. Woodburn. 2005. Context-dependent biological control of an invasive thistle. Ecology 86:3174-3181. Steffan-Dewenter I. and T. Tscharntke. 2002. Pollination, seed set, and seed   predation on a landscape scale. Proceedings of the Royal Society London                 268: 1685–1690. Stein, W.  1990.  Quercus  garryana  Dougl.  Ex  Hook. Oregon  White  Oak. Pages 650-657 in R. Burns and B. Honkala, editors.  Silvics of North America: 2. Hardwoods.  USDA  Handbook  654.  USDA  Forest  Serivce,  Washington,  DC, US.  Syrett, P., D. T. Briese, and J. H. Hoffmann. 2000. Success in biological control of terrestrial weeds by Arthropods. Pages 189-230 in G. Geoff and S. Wratten, editors. Biological Controls: Measures of success. Springer, Netherlands. Vellend, M., A.D. Bjorkman, A. McConchie. 2008. Environmentally biased               fragmentation of oak savanna habitat on southeastern Vancouver Island,               Canada. Biological Conservation 141:2576-2584. Wang, T., A. Hamann, D. L. Spittlehouse, and T. Q. Murdock. 2012. ClimateWNA—high-              resolution spatial climate data for western North America. Journal of Applied               Meteorology and Climatology 51: 16–29.  White, T. C. R. 1969. An index to measure weather-induced stress of trees associated with              outbreaks of Psyllids in Australia. Ecology 50:905-909. Willig, M. R., D. M. Kaufman, and R. D. Stevens. 2003. Latitudinal gradients of biodiversity:    48             Pattern, process, scale, and synthesis. Annual Review of Ecology, Evolution, and              Systematics 34: 273-309. Woods, E. C., A. P. Hastings, N. E. Turley, S. B. Heard, and A. A. Agrawal. 2012. Adaptive geographical clines in the growth and defense of a native plant. Ecological Monographs 82:149-168. Wright, I. J., P. B. Reich, M. Westoby, D. D. Ackerly, Z. Baruch, F. Bongers, J. Cavender-Bares, T. Chapin, J. H. C. Cornelissen, M. Diemer, J. Flexas, E. Garnier, P. K. Groom, J. Gulias, K. Hikosaka, B. B. Lamont, T. Lee, W. Lee, C. Lusk, J. J. Midgley, M. L. Navas, U. Niinemets, J. Oleksyn, N. Osada, H. Poorter, P. Poot, L. Prior, V. I. Pyankov, C. Roumet, S. C. Thomas, M. G. Tjoelker, E. J. Veneklaas, and R. Villar. 2004. The  worldwide leaf economics spectrum. Nature 428:821-827. Yarnes, C. T., and W. J. Boecklen. 2005. Abiotic factors promote plant heterogeneity and influence herbivore performance and mortality in Gambel's oak (Quercus gambelii). Entomologia Experimentalis Et Applicata 114:87-95.     49 Appendix A Population location  Latitude and longitude for sampled Quercus garryana populations for both field populations and the populations of origin for trees in the common garden.  Sample  Population Latitude (°N) Longitude (°W) Field Populations 1 42.09 122.36  2 43.18 123.13  3 43.29 123.13  4 44.35 123.18  5 44.55 123.09  6 45.20 122.39  7 45.28 122.37  8 46.51 123.20  9 46.54 122.50  10 47.10 122.34  11 47.32 122.15  12 48.75 123.69  13 48.76 123.67  14 48.04 123.60  15 48.48 123.39  16 48.47 123.44  17 49.15 123.15  18 49.72 125.01     Common Garden Population 1 42.46 122.61  2 45.01 123.16  3 45.28 121.34  4 46.05 122.85  5 46.82 123.01  6 46.57 121.68  7 48.79 123.69  8 48.46 123.40  9 49.72 125.01 

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.24.1-0308710/manifest

Comment

Related Items