UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Mechanisms of coexistence between native and exotic plant species Johnson, Jens 2018

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

Item Metadata


24-ubc_2018_september_johnson_jens.pdf [ 906.67kB ]
JSON: 24-1.0371237.json
JSON-LD: 24-1.0371237-ld.json
RDF/XML (Pretty): 24-1.0371237-rdf.xml
RDF/JSON: 24-1.0371237-rdf.json
Turtle: 24-1.0371237-turtle.txt
N-Triples: 24-1.0371237-rdf-ntriples.txt
Original Record: 24-1.0371237-source.json
Full Text

Full Text

MECHANISMS OF COEXISTENCE BETWEEN NATIVE AND EXOTIC PLANT SPECIES by Jens Johnson B.Sc., The University of Washington, 2015  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 2018 © Jens Johnson, 2018 ii  Committee Page The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:  Mechanisms of Coexistence Between Native and Exotic Plant Species  submitted by Jens Johnson  in partial fulfillment of the requirements for the degree of Master of Science in Geography  Examining Committee: Jennifer Williams, Geography Supervisor  Amy Angert, Botany Supervisory Committee Member   Supervisory Committee Member Rachel Germain, Botany Additional Examiner          iii  Abstract  One of the primary goals of ecology is to understand the processes that maintain biological diversity. The development of a modern coexistence theory has helped to advance this understanding by proposing a set of specific mechanisms that enable coexistence, specifically, species can coexist when niche differences between species are large enough to overcome fitness differences between species. Recent advances have used fully parameterized demographic studies to explain how traits, phenology, and evolutionary history contribute to niche and fitness differences, but there is a lack of empirical evidence of how competition and the environment interactively influence coexistence outcomes. Moreover, there is little empirical evidence of how indirect interactions between species mediate coexistence outcomes.   Using an outdoor pot experiment and observational data from natural plant communities I explored the interactions between a pair of native (Plectritis congesta) and exotic (Valerianella locusta) co-occuring, annual plant species. With this system I answered the following four questions: (1) Can Plectritis and Valerianella coexist over the long run? (2) Does environmental variation change the intrinsic interaction between these species? (3) Does environmental variation enable coexistence by providing each species with an opportunity for positive low density growth rate in certain spaces or at certain times? (4) Do indirect interactions with pollinators destabilize or promote coexistence?   The pot experiment predicted that Plectritis will exclude Valerianella over the long run. Although the coexistence outcome did not change between environmental treatments, the parameters used to calculate niche and fitness differences experienced significant changes. I did not find any evidence that Valerianella maintains abundances through variation in the environment in natural communities. Additionally, I did not find any evidence that niche differences are decreased through indirect interactions with pollinating insects.   These species are still observed to co-occur at the site level and thus Plectritis may limit, but not totally eliminate, the abundance of Valerianella. Moreover, my experiment showed how vital rates and interaction coefficients depend on the environmental context, emphasizing that abundances are driven not only by competition and environment, but also through the interaction between competition and environment.     iv  Lay Summary  Closely related, functionally similar plant species - native, Plectritis congesta, and exotic, Valerianella locusta - co-occur in the Garry oak ecosystems of BC. However, their abundances show an inverse relationship at the fine scale, suggesting that one species might competitively exclude the other. Using a combination of experimentation and field data, I quantified the differences between these species that might enable or preclude their coexistence. Due to competitive differences that outweigh niche differences, native Plectritis congesta is expected to exclude the exotic Valerianella locusta. Although this outcome did not depend on the environmental context of soil moisture, the parameters that are used to estimate niche and fitness differences do shift between treatments, highlighting an interaction between environment and competition that may influence coexistence. Additionally, I found no evidence that indirect interactions mediated by insect mutualists promote coexistence between these species.                   v  Preface This thesis is original, unpublished work by the author, Jens C. Johnson.  Supervision  and guidance for this research was provided by Dr. Jennifer L. Williams (UBC Geography).     Portions of this work were presented at the Canadian Society for Ecology and Evolution Conference (Guelph, Ontario, July 2018).                                    vi  Table of Contents   Abstract…………………………………..……………………………………………………. iii Lay Summary…………………………...…………………………………………..………… iv Preface………………………………………………………………………………………….. v Table of Contents…………………..……………………………………………….............. vi List of Tables…………………………………………………………………………………. vii List of Figures……………………………………………………………………………...... viii Acknowledgements………………………………………………………………………….. ix Introduction…………………………………………………………………………................ 1 Methods and Materials…………………………………………………………………......... 4            Study Species and Study Region…………………………………………………….. 4            Resource Competition Experiment…………………………………………….......... 5            Growing Condition Details…………………………………………………………….. 8            Field Abundance and Demography Survey…………………………………………. 9            Pollination Experiment to Test for Negative Frequency-Dependence…............. 10            Data Analysis………………………………………………………………………….. 10  Results………………………………………………………………………………………… 13            Resource Competition Experiment…………………………………………………. 13            Field Abundance and Demography Survey………………………………………... 14            Pollination Experiment to Test for Negative Frequency-Dependence …………. 18 Discussion……………………………………………………………………………………. 20 Conclusions …………………………………………………………………………………. 23 Bibliography….…………………………………………………………………………........ 24 Appendixes….……………………………………………………………………………….. 28            Appendix 1….……………………………………………………………………........ 28            Appendix 2….……………………………………………………………………........ 29       vii  List of Tables Table 1: Estimates of density-dependent population growth rate parameters…………………. 14 Table 2: Coexistence predictions from a controlled experiment with dry and  wet soil moisture conditions…………………………………………………………………………… 15                  viii  List of Figures Figure 1: Experiment Design…………………………………………………………………. 6 Figure 2: Pollination Experiment Treatment………………………………………………… 8 Figure 3: Density-dependent seed production of Plectritis in Valerianella  in dry and wet soil moisture conditions……………………………………………………... 14 Figure 4: Plectritis and Valerianella abundances in 0.1m2 and 1m2 plots in  Garry Oak plant communities………………………………………………………………... 16 Figure 5: Density-Dependent Seed Production in Natural Plant Communities………... 17 Figure 6: Relationships between predicted seed production and variation in  the community and environment composition……………………………………………... 18 Figure 7: Pollen Limitation and Density-Dependence in a Natural Plant  Community…………………………………………………………………………………….. 19           ix  Acknowledgements  This research was made possible through the incredible guidance and support of my mentors, colleagues, and friends. I would like to thank Dr. Jennifer Williams for her enthusiasm, encouragement, ingenuity and patience. Her support was critical in developing my project and, moreover, in sustaining my efforts to bring this thesis to its completion. I would also like to thank Dr. Amy Angert for providing helpful comments and discussion through both the planning and review phases of my project. Additionally, I would like to thank Rachel Germain for her interest in my work and for sharing her expertise and ideas, as well as Dr. Elizabeth Elle for providing valuable information and input. I am very grateful for the aid I received from my determined field/lab assistants Hailey Horachek and Celine Coschizza. I was also aided by insightful discussion and personal upkeep from fellow graduate students, especially including: Emily West, Stephanie Blaine, and Madelyn Ore; and from other close friends and loved ones, especially including: Charlie Ulrich, Steven Rusciano, and Sam Herle. Finally, I would like to acknowledge the inspiration given to me by past mentors, particularly Tina Harrison and Valerie Soza.   Further thanks are extended to the land managers who welcomed research on their properties: Irvin Banman (Nature Conservancy Canada) and the Capital Regional District. I also thank Meylin Zink Yi and Adam Wilkinson (UBC Botany Greenhouse) for logistical support with my experiment.   Financial support was provided by the UBC Department of Geography and the UBC Faculty of Arts, as well as a research grant to Dr. Jennifer Williams from Natural Sciences and Engineering Research Council (NSERC).        1  Introduction  One of the primary goals of ecology is to understand the processes that maintain biological diversity. In order for diversity to persist, species in biological communities must interact such that they do not drive each other toward local extinction. Consequently, understanding patterns of diversity requires an ability to predict how the abundance of one species will affect the abundance of another. The classic ecological theory of limiting similarity predicts that the most strongly competing species will exclude other species from the community when their resource requirements overlap (MacArthur et al. 1967). This theory has proved useful in predicting the coexistence of species across a range of taxa and climates (Wilson 2007), and has been used to explain non-random patterns of community assembly from the larger regional pool of species (Diamond 1975). However, limiting similarity cannot explain the instances where two seemingly similar species occur in the same community (Godoy et al. 2014b, Thompson 2010).   The development of a modern coexistence theory (Chesson 2000) has helped to resolve the gaps in limiting similarity theory by proposing a set of specific mechanisms that enable coexistence. Modern coexistence theory focuses on identifying two types of species differences that determine the stability of species coexistence: average fitness differences and niche differences (Adler et al. 2007). Average fitness differences are those differences that allow one species to access a shared resource more efficiently than another species. Here, fitness is used in an ecological rather than evolutionary sense, and populations are the unit of comparison rather than individuals (Levine and HilleRisLambers 2009). Niche differences, on the other hand, reduce the degree of overlap in resource requirements between species. Niche differences simultaneously increase intraspecific competition and reduce interspecific competition (Levine and HilleRisLambers 2009). When stabilizing niche differences are greater than average fitness differences, population growth rates exhibit negative frequency-dependence (HilleRisLambers et al. 2012). This prevents the species with highest average fitness from excluding inferior competitors by providing an advantage to growth rate when abundance is low, thereby promoting coexistence (HilleRisLambers et al. 2012). Under this theoretical framework, studies that parameterize population dynamics models with vital rates – the stages in a plant life cycle that influence population growth rates – directly quantify average fitness differences and niche differences to predict community patterns and dynamics. In doing so, these studies identify underlying mechanisms that destabilize or promote coexistence, such as particular functional traits (Adler et al. 2010, Kraft et al. 2015) or phenology patterns (Godoy et al. 2014a).  Modern coexistence studies primarily explore mechanisms of coexistence in species pairs with a history of shared interaction, meaning that the species in those systems have had time to evolve adaptive differences that function to reduce interspecific competition. In contrast, if and how exotic species, those that have been recently introduced to a system, can coexist with similar members of the native 2  community has been less explored (Germain et al. 2016). In a study of a California annual plant community, only one of 16 native and exotic species pairs were expected to coexist in the long run (Godoy et al. 2014b). In this system, the winner of competition was best predicted by relative differences in growing phenology, regardless of native versus exotic origin (Godoy et al. 2014b). In another case, a native and exotic species pair were projected to coexist through co-evolutionary processes that increase niche differences relative to average fitness differences (Huang et al. 2018).  One major challenge in studying species interactions is that changes in the environment through space and time can influence their outcome. Environmental variation can influence the intrinsic values of vital rates or alter the coefficients of species interactions (Maron et al. 2014), which are used to calculate niche and fitness differences (Godoy et al. 2014b). Consequently, recognizing the environmental conditions in which species interactions occur is critical towards forming conclusions about coexistence that transfer beyond the immediate system of study (Maron et al. 2014).   While environmental heterogeneity within and between communities poses a challenge for forming more general conclusions, environmental heterogeneity within a community may allow for patchy coexistence (Chesson 2000). The storage effect occurs when species hold trait differences that provide a competitive advantage at certain times (Angert et al. 2009, Godoy et al. 2014a) or in certain microsites (Pake and Venable 1995, Adler et al. 2006, Sears et al. 2007, Adler et al. 2009, Staples et al. 2016). The unique responses of species, due to differences in functional trade-offs such as relative growth rate versus water use efficiency, cause environmental conditions and competition to co-vary (Angert et al. 2009). As a result, the abundances of each species in a community can hold positive average growth rates when considered at a broad enough scale (Chesson 1994, Chesson 2000). Thus, through a storage effect, a species pair that is not expected to coexist at the scale of a competition experiment (e.g. Adler et al. 2010 or Godoy et al. 2014b) may still coexist through storage effects (Chesson 2000). Coexistence studies that have complimented competition experiments with data on environmental variability and the abundances of species in different sites (Staples et al. 2016) or different years (Godoy et al. 2014a) find evidence for coexistence through spatial and temporal resource partitioning despite high average fitness difference and low niche difference measured in the coupled competition experiment, showing how species may coexist through partitioning at larger scales beyond those of immediate pairwise interactions.   Indirect interactions with other species also have the potential to shift the outcome of coexistence. One relatively unexplored mechanism of coexistence is through insect-mediated interactions. Modes of pollination are considered to be a major driver of plant diversity by encouraging speciation, but only recently have studies asked if modes of pollination contribute to plant diversity by allowing overlapping species to coexist (Pauw 2012). Plant abundance is tied to insect-mediated pollination through 3  seed production (Ashman et al. 2004). When plants are pollen limited, i.e. the production of seeds is limited by the quantity or quality of pollen delivery, population growth may cease or decline (Groom 1998). Pollination and mating system evolution have been theorized to promote coexistence in plant communities if these systems cause pollen-limitation to decrease as relative abundance in the community decreases. This pattern would place negative frequency-dependence on population growth rates, thereby limiting the dominance of species that become abundant and providing an advantage to allow populations to rebound when their abundances decline (Pauw 2012). Thus, pollinator niche differentiation could act similarly to niche differentiations that reduce direct interaction for resources (Pauw 2012).   In this study I used experimental and observational approaches to examine evidence for coexistence and the underlying mechanisms between two closely related annual forbs in the Valerianaceae family, Plectritis congesta (hereafter, Plectritis) and Valerianella locusta (hereafter, Valerianella). These species, which have similar growth forms and life histories (Jacobs 2010), form a particularly interesting combination for study because they are of different geographic origins. Although they currently overlap in range in western North America where Plectritis is native, Valerianella, has been recently introduced to the region from Eurasia via escape from cultivated sources (Jacobs 2010).  Plectritis and Valerianella occur at similar rates, present in 13% and 15% of 1 m2 survey plots across seven sites in the study region in the oak savanna of southwestern British Columbia, Canada (MacDougall and Turkington 2005). These data show these species co-occur in this region, however, this does not speak to whether these species can coexist.   With this native versus invasive species pair, I tested the competitive effects of conspecific and heterospecific neighbors with an outdoor pot experiment, and corroborated these results with field data on plant abundances, demography, and environmental conditions to answer the following four questions: (1) Can Plectritis and Valerianella coexist over the long run, and at what spatial scales? I expected the exotic Valerianella would competitively exclude Plectritis because disturbance and change in this ecosystem where they co-occur has provided opportunities for invasive exotics to come to dominate plant communities (MacDougall and Turkington 2005). If this outcome were supported by the experiment, I would predict that abundances of these species would be negatively correlated in natural plant communities and, furthermore, I would predict that demographic performance of Plectritis would be more negatively impacted by Valerianella than by its conspecific neighbors in natural plant communities. (2) Does environmental variation change the intrinsic interaction between these species? Because soil moisture is a key limiting factor in this ecosystem (MacDougall and Turkington 2005) - I predicted that changes in the spatial or temporal distribution of water might allow for a greater capacity for niche differentiation - potentially enabling coexistence when this resource availability is high. (3) Does environmental variation enable coexistence by providing each species with an opportunity for positive low density growth rate in certain spaces or at certain times? Because functional trade-offs 4  may allow species to respond uniquely to environmental heterogeneity (Staples et al. 2016; Angert et al. 2009), I predicted that natural variation in key environmental factors such as soil moisture, soil depth, and/or invasive grass abundance would enable coarse scale coexistence by providing an advantage to each species in different spatial patches (4) Do indirect interactions destabilize or promote coexistence? I predicted that Plectritis pollination would show signs of negative density-dependence because high flower densities may exceed insect-pollinator availability. Since Valerianella is self-pollinating (a density-independent reproduction mode), negative density-dependence in Plectritis pollination would increase niche-differentiation and, thereby, increase the propensity for these species to coexist.  Methods and Materials  Study Species and Study Region  The Garry Oak savanna ecosystem found in southwestern British Columbia is part of a climate and vegetation complex that extends south along Pacific lowlands toward California. Seasonal variation in Pacific Ocean currents creates a sub-Mediterranean climate with wet cool winters and warm dry summers. Climate change projections for the Garry Oak ecosystems are for warmer temperatures throughout the year, with wetter winters but drier summers. Drier, warmer summers may actually increase total suitable area for Garry Oak savannas (Lea 2006). Local-scale environmental heterogeneity including soil depth, soil moisture, and shade cover, drive strong local variability in water availability from a meter to meter or finer scale (MacDougall and Turkington 2005, Vellend 2008). The vegetation of Garry Oak savanna is dominated by grasses and a diverse community of early-spring flowering herbaceous plants (Vellend 2008). Today, native flora typically constitute less than 10% of all species cover in site remnants (Trowbridge et al. 2017), and in some cases, native species have been completely extirpated.   The annual forb Plectritis congesta (sea blush) is native and common in Garry Oak communities (Erickson and Meidinger 2007). Plectritis is a winter annual that lacks dormancy and typically germinates in the fall or, alternatively, in the early spring (Young-Mathews 2012). Plectritis flowers between March and June and then sets seed and dies by the end of the summer (Young-Mathews 2012). Plectritis has floral adaptations to attract pollinators and promote outcrossing, including conspicuous flowers that are visited by pollinating insects such as bees and bee flies, protandrous flowering phenology that separates the production of pollen and the receptivity of the stigma within a flower, and a fragrant odor combined with nectar production (Ganders et al. 1977). The plants are self-compatible, and can self-pollinate in the absence of insects, but seed set of exclusively self-pollinated plants is considerably lower than seed set in populations in the field, meaning that the breeding system of the species has the potential for mixed selfing and outcrossing (Ganders et al. 1977). These plants do not reproduce vegetatively (Ganders et al. 1977).  5   Valerianella locusta (corn-salad) shares a recent evolutionary history with Plectritis as part of the Valeriana subfamily of the Valerianaceae, and exhibits a similar growth form and phenology. It was introduced to the Pacific Northwest region of North America as early as the 1880’s (Consortium of Pacific Northwest Herbaria 2017) and records indicate that it has spread throughout the Garry Oak ecosystem (MacDougall and Turkington 2005). All plants in this subfamily share traits such as sympetalous, bilaterally symmetric to strongly asymmetric flowers with an inferior, compound ovary. Other shared traits include iridoid compounds contained in the vegetative tissues that may act as anti-herbivory or anti-microbial defenses (Jacobs et al. 2010). Valerianella and several other species in the genus have relatively inconspicuous flowers and are primarily self-pollinating, although flower visitors have been observed including flies, bees, beetles and butterflies (Ernet 1977). Pollinating insects rarely, if ever, visit Valerianella in a natural Garry Oak ecosystem setting (Parachnowitsch and Elle 2005).  Resource Competition Experiment  To determine whether these species can coexist under direct competition (question 1), I set up an outdoor experiment in pots to grow Plectritis and Valerianella in density gradients of conspecific or heterospecific neighbors. From this experiment, I quantified two demographic vital rates to parameterize population dynamic models: density-independent germination rates (𝑔) and density-dependent seed production (𝐹).  Parameter estimates from the population dynamic models – specifically, intrinsic population growth rate (𝜆) and intra-/interspecific interaction coefficients (𝛼) – were then used to calculate the average fitness differences and niche differences that underlie the outcome of species interaction.  To evaluate species’ germination rates (𝑔), I sowed and monitored 432 seeds of each species (72 pots per species with 3 seeds per pot). The germination rates were determined on April 12, 2017, 32-36 days after sowing, by counting the proportion of sown seeds that had successfully germinated and survived (a small number of individuals that germinated but died before producing true leaves at this date were counted as failed germinants). Pots were kept outside on the University of British Columbia campus, Vancouver, BC, and allowed to experience natural environmental and weather conditions during the germination period.  To determine density-dependent seed production (𝐹), I also sowed seeds in 7.6 litre pots located outside on the University of British Columbia campus, Vancouver, BC. In the center of each pot I planted 1 focal individual of either species to quantify seed production at the end of the growing season. For each species, I assembled communities at four neighbor density levels: (1) 16 pots with no neighbors; (2) 12 pots with 4 heterospecific neighbors and 12 pots with 4 conspecific neighbors; (3) 8 pots with 10 heterospecific neighbors and 8 pots with 10 conspecific neighbors; and (4) 8 pots with 20 heterospecific neighbors and 8 pots with 20 conspecific neighbors (Figure 1). The background neighbors were initially sown at seeding densities above these levels of 4, 10, and 20, and then thinned following germination so as to ensure sufficient 6  densities and consistent spatial arrangements with respect to the focal plants. Higher n for low neighbor densities were selected due to higher expected variation in low-density growth. The densities across the gradient were chosen to bracket natural densities observed in the field (NCC Cowichan Garry Oak Preserve, Duncan, BC, Charlotte Trowbridge, unpublished data).    (A) (B) 7  Figure 1: Experiment Design. (A) Plectritis grown against a background of conspecific neighbors or with Valerianella. (B) Valerianella locusta grown against a background conspecific neighbors or Plectritis. These setups were conducted under either wet or dry conditions. I replicated the no competition pots 16 times, the low density pots 12 times, and the medium density and high density pots 8 times.    To explore whether the average fitness differences and niche differences between these species are context-dependent (question 2), I replicated the above density-gradient set up under a different watering regime. As natural rainfall decreased over the course of the experiment, I provided ample water to one treatment group (hereafter, the wet treatment) and decreased watering regularity and duration in the other (hereafter, the dry treatment). Additionally, I erected a 30% shade cloth over the wet treatment block to reduce moisture loss and reduce the high daytime temperatures that occurred during this last time period of the experiment.  I allowed the plants to continue growing under these watering regimes and then collected the seeds of the focal individuals at the end of the growing season. Following the flowering period, plants were observed daily for seed collection. I collected seeds non-invasively from focal plants as they matured, turning from green to brown, by gently shaking the inflorescence over a paper envelope or, when the last seeds were remaining on the last senescing inflorescences, by removing the reproductive structures. All seeds were then cleaned and counted by hand in the lab.  I also collected trait data on the focal individual in each pot. I measured the number of flowers, the number of inflorescences, the height of the flowering stem from base to top of the inflorescence and the main inflorescence size for all of the focal individuals during the peak flowering period. Immediately after collecting the last seeds from a plant, I harvested aboveground biomass to provide supporting evidence for differences in individual growth and performance.  To assess the potential for pollen limitation in the outdoor site, a parallel pollen supplementation experiment was carried out in twenty pots, each with 10 Plectritis. In each pot, I deposited pollen to the tips of all receptive stigma of two plants in each pot every other day during the duration of the flowering period (Figure 2). I collected pollen from surplus individuals grown at the experiment site using a fine brush and transferred to the stigma of each flower on the pollination treatment individuals using a toothpick. I collected and counted the seeds as the plants senesced. Seeds of two additional untreated plants from each pot were collected and used for comparison. I also measured the number of flowers, the number of inflorescences, the plant height, and the main inflorescence size for all of these plants. 8    Figure 2: Pollination Experiment Treatment. Pollen supplementation before (left) and after (right) treatment. Growing Condition Details  For both the germination rates and the density gradient, I planted seeds in Sunshine Mix #1 soil medium on March 10th - 13th, 2017. I used 7.6 L planting pots with a diameter of 22.9 𝑐𝑚, depth of 22.9 𝑐𝑚 and a 0.041 𝑚2 surface area. I obtained Plectritis seeds from Saanich Native Plants (Victoria, BC) and Valerianella seeds from Salt Spring Seeds (Salt Spring Island, BC). To ensure at least one individual of desired species identity at the center of the pot for data collection, I sowed three focal individual seeds around a marker at the center of the pot. When one of these seeds germinated, the remaining two seeds or germinants were immediately removed from the marker area to eliminate any interaction. The seeds of both species were sown directly on the surface of the soil, as recommended for Plectritis (Saanich Native Plants, personal communication; Young-Mathews 2012). Bird netting was placed over the pots to avoid seed predation.   Following the germination period, the pots were arranged on a 10 m x 8 m watering grid. Pots were placed in rows with 23 cm spacing and surrounded with a 15 cm mulch layer to moderate temperature and moisture fluctuations. Water inputs were supplied via Netafim pressure compensated spray stakes (Southern Drip Irrigation), with flow rates of 12.1 litres/hour (recommended by manufacturer for pots of 3.8 - 19 litre size) and a spray radius of ~25 cm. Pots were rotated randomly through the watering grid (within wet or dry treatment blocks) weekly to reduce any effects of small variation in spray stake water delivery. A temporary fence was constructed around the experiment area to deter herbivory from geese or other large animals.   9   All pots from both treatments were exposed to natural outdoor rainfall conditions from the start of the experiment until June 1, 2017, near when natural precipitation conditions shifted from near daily and constant to infrequent and ephemeral. Precipitation patterns generally become more infrequent and variable later in the growing season in the region’s Mediterranean dry summer climate (Lea 2006). At this date, plants were separated into wet and dry treatment groups. The semi-exposed aboveground pots dry more quickly than the soil, and thus artificial watering was necessary to maintain soil moisture conditions that would occur at ground level without any rainfall. Plants in the dry treatment received no water for 10 days, and then were maintained with 5 minutes of water every other day until the end of the experiment. Plants in the wet treatment received no water for 5 days, and then were maintained with 12 minutes of water every other day from until the end of the experiment. I added 0.25 liters of 100ppm, 20-8-20 fertilizer solution to all pots on June 5th, 2018, to replenish soil nutrients (See Appendix 1 for soil moisture conditions). Field Abundance and Demography Survey  To investigate whether spatial heterogeneity in natural communities might contribute to coexistence through a resource partitioning effect (question 3), I surveyed the abundances and fitness proxies of Plectritis and Valerianella at three locations in the southeastern area of Vancouver Island, BC: Nature Conservancy of Canada Cowichan Garry Oak Preserve, Thetis Lake Regional Park, and Stoney Hill Regional Park. The meadow in the Cowichan Preserve was managed with a prescribed fire and the plant community is in an early phase of re-establishing (NCC, personal communication). There is no known history of recent, prescribed fire-management where I surveyed at Thetis Lake or Stoney Hill Regional Parks.   Across these three sites I quantified plant abundance and environmental conditions and collected demographic data in sixty-nine 1 m2 plots. Additionally, I collected these same data in two smaller 0.1 m2 plots nested inside of the 1 m2 to investigate correlations at two different spatial scales. In both the smaller and larger plots, I counted the number of Plectritis and Valerianella. I then measured number of inflorescences, plant height, and main inflorescence size for 10 individuals of each species in each smaller 0.1 m2 plot (or as many individuals as possible if there were fewer than 10). I used these measures as proxies for fecundity.  In each plot, I assessed environmental conditions by measuring: soil depth, soil moisture, percent grass cover, percent shrub cover, and species diversity. Using a marked measuring rod, I categorically measured soil-depth in the bottom-right and top-left corners of each plot as either <7.6 cm, between 7.6 cm and 12 cm, and >12 cm. I used a soil-moisture probe to measure percent volumetric water content (%VWC) at the bottom-right and top-left corners of each plot at depths of 7.6 cm and 12 cm. I did not test soil moisture if the soil depth was less than 7.6 cm. I estimated percent grass cover and percent shrub cover by visual approximation. I estimated species diversity by counting the number of species present in each plot.  10  Pollination Experiment to Test for Negative Frequency-Dependence  To determine whether indirect interactions promote coexistence between these species (question 4), I examined the relationship between conspecific and heterospecific neighbour densities and pollen limited seed set in Plectritis. At the NCC Cowichan Garry Oak Preserve field site, I haphazardly selected 24 pairs of Plectritis individuals of relatively similar overall size and inflorescence size - and located less than 10 cm apart. Pairs were spaced at distances of 2 m or more from the next nearest pair. I measured the number of flowers, number of inflorescences, plant height, and main inflorescence size of these plants. Furthermore, I collected data on the surrounding conspecific neighbor densities, Valerianella neighbor densities and percent grass cover at both the 0.1 m2 and 1 m2 scales.   I conducted a pollen supplementation treatment on the selected plant pairs to determine if Plectritis seed production is limited by the quantity or quality of pollen delivery. Plectritis bloomed synchronously at the Cowichan site, with a peak from May 20th to May 31st, 2017. During this time, I carried out 5 rounds of hand pollination on one individual within each pair and allowed the other to be pollinated naturally. Pollen was collected from nearby neighbours using a fine brush and transferred to the stigma of each flower on the pollen supplement treatment individuals using a toothpick. At the end of the flowering period, I covered the inflorescences of all plants with fine mesh bags and supported them in an upright position with a 12” wooden stake. The mesh bag was tied so as to collect any dropped seed between the end of the pollen supplementation treatment and a return visit on June 28th, 2017, to collect seeds. I quantified seed production by counting all seeds in the bag and also those still attached to the inflorescence but matured (enlarged and brown in colour).  Data Analysis  To determine whether these species should coexist in direct competition (question 1), I used the data on the vital rates in the assembled communities to parameterize models of plant abundances. These models estimate both population growth rates in the absence of competition (λ), and the influence of neighbor density and identity on population growth rates (αconspecifc neighbors and αheterospecifc neighbors). The population dynamics of an annual plant species with no seed bank can be described as (Chesson 1990, Godoy et al. 2014a):  Equation 1:  𝑁𝑖,𝑡+1𝑁𝑖,𝑡= (1 − 𝑔𝑖)𝑠𝑖 + 𝑔𝑖𝐹𝑖   Here, 𝑁𝑖,𝑡 represents the number of ungerminated seeds for species i. The per capita growth rate for the population ((𝑁𝑖,𝑡 + 1)/(𝑁𝑖,𝑡)) is expressed as a function of the average of two vital rates: 𝑠𝑖, the survival of ungerminated seed in the soil, and 𝐹𝑖, the fecundity or the viable seeds produced per germinated individual. These terms are weighted by the average germination of the species, 𝑔𝑖 (Godoy et al. 2014a). To 11  determine how species affect the performance of one another, the per-germinant fecundity, 𝐹𝑖, can be expanded into a function that considers the effect of competing individuals in the system (Godoy et al. 2014a):  Equation 2: 𝐹𝑖 =𝜆𝑖1+ 𝑎𝑖𝑖𝑔𝑖𝑁𝑖,𝑡+𝑎𝑖𝑗𝑔𝑗𝑁𝑗,𝑡   Here, 𝜆𝑖 represents the per-germinant fecundity in the absence of competition. 𝜆𝑖 decreases when neighboring individuals of either the same species or the other species negatively impact the performance of plants. These interaction coefficients are represented by 𝑎𝑖𝑖 (an intraspecific interaction coefficient for species i), and 𝑎𝑖𝑗 (an interspecific interaction coefficient for species i). I used fecundity data from the experiment to fit equation 2 and then estimated values of 𝜆 and 𝑎 for both Plectritis and Valerianella using generalized nonlinear regression with a negative binomial distribution (“gnlr()” function) in the “gnlm” R package.   These demographic measurements were then synthesized into estimates of the stabilizing niche differences and fitness differences that determine the outcome of competition, following Godoy et al. (2014a). Niche overlap (ρ) is defined as:  Equation 3: ρ = √𝑎𝑖𝑗𝑎𝑗𝑗∗𝑎𝑗𝑖𝑎𝑖𝑖  The demographic ratios, which describes the degree to which one species produces more seeds per seed loss due to death or germination, are defined as:  Equation 4: 𝑑𝑒𝑚𝑜𝑔𝑟𝑎𝑝ℎ𝑖𝑐 𝑟𝑎𝑡𝑖𝑜 =𝜆𝑖 ∗ 𝑔𝑖1 − (1 − 𝑔𝑖)  / 𝜆𝑗 ∗ 𝑔𝑗1 − (1 − 𝑔𝑗)   The competitive response ratio, which describes the degree to which one species produces more seeds per seed loss due to death or germination, are defined as:  Equation 5:  𝑐𝑜𝑚𝑝𝑒𝑡𝑖𝑡𝑖𝑣𝑒 𝑟𝑒𝑠𝑝𝑜𝑛𝑠𝑒 𝑟𝑎𝑡𝑖𝑜 =  √𝑎𝑖𝑗𝑎𝑗𝑗∗𝑎𝑖𝑖𝑎𝑗𝑖  Fitness differences (𝜅) are defined as the product of the demographic ratio between species and the competitive response ratio between species:   Equation 6: 𝜅 =  𝑑𝑒𝑚𝑜𝑔𝑟𝑎𝑝ℎ𝑖𝑐 𝑟𝑎𝑡𝑖𝑜 ∗ 𝑐𝑜𝑚𝑝𝑒𝑡𝑖𝑡𝑖𝑣𝑒 𝑟𝑒𝑠𝑝𝑜𝑛𝑠𝑒 𝑟𝑎𝑡𝑖𝑜   Finally, coexistence conditions are considered satisfied when:   Equation 7:  ρ <  𝜅 <  1ρ   12  Given Equation 7, coexistence is possible when niche overlap (ρ) is low, that is, when ρ approaches 0, and the fitness ratio (𝜅) is relatively even between species, that is, when 𝜅 approaches 1.  To determine whether the outcome of interaction is dependent on the resource environment (question 2), I fit Equation 2 using data from the second watering treatment and then re-estimated the population parameter estimates. I used Equations 3, 4, 5, and 6 to estimate new niche differences and fitness differences and then used Equation 7 to again predict whether these species could coexist in the long run. Using this estimation approach, I could calculated confidence intervals for values of 𝜆’s and 𝑎’s.    To further address question 1, I used the data from the field surveys on the abundances and fitness proxies for Plectritis and Valerianella. I first tested whether there was a negative relationship between Plectritis and Valerianella abundances at both the 0.1 m2 scale and 1 m2 using a linear model for Plectritis abundance as a function of Valerianella abundance. To meet linear model assumptions of homoskedacity and normality, abundances were log transformed (plus 1 to handle abundances of 0 in the data set).   I fit models for predicting seed production using my seed counts alongside measurements of the number of inflorescences, plant height, and main inflorescence size. For Plectritis, I used the seed counts and measurements from the untreated plants in the field pollination experiment; for Valerianella, I used the seed counts and measurements from the focal individuals in the pot experiment, pooled across all densities, because I lacked seed counts from the field site. I selected the best models for seed count for each species using generalized linear models with negative binomial distributions (“glm.nb()” function) for fitness proxies in the “MASS” R package and comparing AIC scores of all possible nested models. The best model for Plectritis included only plant height as the predictor of seed count, while the best model for Valerianella included number of inflorescences, plant height, and main inflorescence size (Appendix 2). With these two models, I predicted the number of seeds each Plectritis and Valerianella plant in the field plots would produce given the fitness proxies I measured.    I used these predicted seed counts to see how Plectritis and Valerianella seed production was driven by conspecific and heterospecific densities in the field. I used generalized linear models with negative binomial error distribution (“glm.nb()” function) to determine how the density of each species impacted the average predicted seed production of each species. Here, I treated site as a random effect, and did not include the data from Thetis Lake where Valerianella was not found. Next, I used linear models to evaluate the impact of each plant density on the variation in seed production of each species (quantified as the absolute value of the residual). Again, I treated site as a random effect, and did not include the Thetis Lake data. Further, I looked at how the soil moisture and depth conditions, as well as the grass and shrub cover influenced average seed production of each species.  13   I evaluated whether Plectritis and Valerianella responded uniquely to environmental heterogeneity (question 3) by determining whether environmental conditions (soil moisture, soil depth, grass cover and shrub cover) differentially influenced either the abundances or predicted seed production in Plectritis and Valerianella using generalized linear models with negative binomial error distribution (“glm.nb()” function). Here, I treated site as a random effect, and excluded the Thetis Lake data.   To examine the prevalence of pollen limitation and its potential connection to coexistence (question 4), I used data from the field site to evaluate the magnitude and density-dependence of pollen-limited seed set. I tested for the presence of pollen limitation by comparing the AIC scores of nested models for seed production as a function of flowers per plant, with and without pollen treatment included as a fixed-effect (and with spatial area of the NCC Cowichan site held in both models as a random-effect). Some plants made more seeds than flowers which may be due to flowers emerging after the initial counts near the beginning of the flowering phenology window.  The pollen limitation of Plectritis would only affect coexistence if its magnitude were density-dependent. To determine if the magnitude of pollen limitation was density-dependent, I first quantified the magnitude of per plot pollen limitation as the increase in seed set of a pollen supplemented plant (relative to its flower number) in comparison to the seed set of the paired untreated plant from the same plot (relative to its flower number). Then, I evaluated the effect of conspecific densities (at 0.1 m2 and 1 m2 scales) on these magnitudes of pollen limitation using a linear model. All statistics and modeling were done in R, version 3.3.1 (R Development Core Team 2016). Results Resource Competition Experiment (1) Can Plectritis and Valerianella coexist over the long run? (2) Does environmental variation change the predicted outcome?  In the competition experiment, I found that seed production for both species was negative density-dependent under all combinations of neighbor species identity and water treatment (Figure 3). Intrinsic population growth rate (λ) of Plectritis was 17% greater in the wet treatment versus the dry treatment, while intrinsic population growth rate (λ) of Valerianella was 110% greater in the dry treatment versus the wet treatment (Table 1). The only value of 𝛼𝑥𝑥 that changed significantly (non-overlapping confidence intervals) between treatments was 𝛼𝑉𝑎𝑙𝑒𝑟𝑖𝑎𝑛𝑒𝑙𝑙𝑎 𝑥 𝑉𝑎𝑙𝑙𝑒𝑟𝑖𝑎𝑛𝑒𝑙𝑙𝑎 (Table 1). Density-independent germination rates were 67.0% for Plectritis and 89.1% for Valerianella.  14   Figure 3: Density-dependent seed production of Plectritis in Valerianella in dry (A) and wet (B) soil moisture conditions. Seed production of Plectritis (light green) and Valerianella (dark green) is plotted for all combinations of competing neighbor species and treatment. Seed production values against Plectritis neighbors are denoted with circles, while seed production values against Valerianella neighbors are denoted with triangles. Declines in seed production due to intraspecific competition are denoted with solid lines and declines in seed production due to interspecific competition are denoted with dashed lines. Dots were jittered for each neighbor density to improve legibility. Data points for 0 neighbor densities were used to model both intra- and interspecific competition and thus each data point for zero neighbor densities are plotted twice.    Table 1: Estimates of density-dependent population growth rate parameters. Standard deviations for each estimate are noted in parenthesis. See Equation 2 in methods for details on parameter estimation.  Treatment  Dry Wet λ Plectritis 364.2 (42.5) 428.1 (60.7) λ Valerianella 636.1 (89.7) 303.1 (58.2) 𝛼𝑃𝑙𝑒𝑐𝑡𝑟𝑖𝑡𝑖𝑠 𝑋 𝑉𝑎𝑙𝑒𝑟𝑖𝑎𝑛𝑒𝑙𝑙𝑎 0.30 (0.10) 0.33 (0.10) 𝛼𝑉𝑎𝑙𝑒𝑟𝑖𝑎𝑛𝑒𝑙𝑙𝑎 𝑋 𝑃𝑙𝑒𝑐𝑡𝑟𝑖𝑡𝑖𝑠 0.43 (0.11) 0.37 (0.10) 𝛼𝑃𝑙𝑒𝑐𝑡𝑟𝑖𝑡𝑖𝑠 𝑋 𝑃𝑙𝑒𝑐𝑡𝑟𝑖𝑡𝑖𝑠 0.23 (0.05) 0.40 (0.10) 15  𝛼𝑉𝑎𝑙𝑒𝑟𝑖𝑎𝑛𝑒𝑙𝑙𝑎 𝑋 𝑉𝑎𝑙𝑒𝑟𝑖𝑎𝑛𝑒𝑙𝑙𝑎 0.72 (0.15) 0.48 (0.15)   The results from the experiment predicted that these species cannot coexist, that is Plectritis will exclude Valerianella in both wet and dry conditions over the long run (Table 2), although the contribution of the underlying components differed between treatments. In the dry treatment, the demographic ratio favored Valerianella, while the competitive response ratio favored Plectritis, resulting in an overall fitness ratio that favored Plectritis (Table 2). The high niche overlap (low niche difference) in the dry treatment violates the coexistence conditions (Table 2). In the wet treatment, the demographic ratio again favored Valerianella, while the competitive response ratio again favored Plectritis (Table 2). In the wet treatment, the niche overlap decreased; however, the fitness ratio decreased by 0.18, meaning that Plectritis had an even greater fitness advantage in the wet treatment (Table 2).  Table 2: Coexistence predictions from controlled experiment with dry and wet soil moisture conditions. Coexistence is possible when niche overlap (ρ) is low, i.e. when ρ approaches 0, and the fitness ratio (𝜅) is relatively even between species, i.e. when k approaches 1. (See Equations 3, 4, 5, 6, and 7 in methods for details on parameter estimation).   Treatment  Dry Wet Niche overlap (𝛒) 0.89 0.80 Demographic ratio 2.36 0.79 Competitive response ratio 0.48 0.85 Fitness ratio (𝜿) 0.85 0.70 Outcome (coexistence predicted if: 𝛒 < 𝜿 < 1/ 𝛒) coexistence not predicted coexistence not predicted  Field Abundance and Demography Survey (1) Can Plectritis and Valerianella coexist over the long run?  In the natural plant communities I found that the abundances of Valerianella and Plectritis showed a negative relationship at both the 0.1m2 scale (F1, 105  = 37.69, P < 0.01) (Figure 4A) and at the 1m2 scale (F1, 106  = 18.84, P < 0.01) (Figure 2B). The highest abundances of Valerianella occurred when Plectritis was absent and likewise the highest abundances of Plectritis occurred when Valerianella was absent (Figure 4A and Figure 4B). I did not find Valerianella at one of the sites, Thetis Lake, and as such I did not include Plectritis abundance data in these analyses. 16   Figure 4: Plectritis and Valerianella abundances in 𝟎. 𝟏𝐦𝟐 and 𝟏𝐦𝟐 plots in Garry Oak plant communities. Using data from two plant communities where Plectritis and Valerianella co-occurred, I tested the relationship between Plectritis abundance and Valerianella abundance at the 0.1m2 scale (A) and at the 1m2 (B). Survey plots from the NCC Garry Oak Preserve are mapped in dark blue, and survey plots from Stoney Hill Regional Park are mapped in light blue.    I examined the effect of Plectritis density and Valerianella density on predicted Plectritis and Valerianella seed production and found that Plectritis density has a negative impact on predicted Valerianella reproductive output (F1, 56 = 9.74, P < 0.01), but no impact on predicted Plectritis reproductive output (F1, 56  = 0.71, P = 0.40) (Figure 5A). Valerianella density has a positive impact on predicted Valerianella reproductive output (F1, 55  = 4.50, P < 0.05) but no impact on predicted Plectritis reproductive output (F1, 55  = 0.07, P = 0.79) (Figure 5C). Although I found no relationship between seed production in Plectritis and density of either neighbor species, I found that variation in the number of seeds per Plectritis individual is influenced by Plectritis density. At low densities, predicted Plectritis fecundities range widely. This range of variation tends to 17  decrease as the density of Plectritis increases (F1, 56  = 13.2, P < 0.01) (Figure 5B). In contrast, variation in Plectritis fecundity does not change across Valerianella densities (F1, 55  = 0.01, P = 0.94) (Figure 5D). I found no evidence for variation in predicted seed production in Valerianella with respect to the density of either species (not shown).  Figure 5: Density-Dependent Seed Production in Natural Plant Communities. Using data from two natural communities where Plectritis and Valerianella co-occurred, I examined the effect of Plectritis densities (A) and Valerianella densities (C) on predicted seed production of Plectritis (light green) and Valerianella (dark green). I also tested the impact of Plectritis densities (B) and Valerianella densities (D) on the variation in predicted seed production of Plectritis   (3) Does environmental variation enable coexistence by providing each species with an opportunity for positive low density growth rate in certain spaces or at certain times?  Grass cover increased the predicted seed production in both Plectritis (F1, 49   = 6.62, P < 0.05) and in Valerianella (F1, 55 = 5.02, P < 0.05) (Figure 6A), but soil moisture at either a depth of 7.6 cm (Plectritis: F1, 39 = 2.76, P = 0.10;  Valerianella: F1, 43 = 0.18, P = 0.67) (Figure 6C) or 12 cm (Plectritis: F1, 23 = 2.60, P = 0.12; Valerianella: F1, 30 = 0.19, P = 0.67) (Figure 6D). When I compared fecundity of plants growing in soils 18  shallower versus deeper than 12 cm, I found that these species exhibited unique responses in regard to soil depth: seed production increased in deeper soils in Plectritis (F1, 49 = 7.28, P < 0.01), but showed no difference in Valerianella (F1, 51 = 0.05, P = 0.83) (Figure 6B). I did not find any relationships between percent grass cover, soil moisture, or soil depth and the abundances of either species.  Figure 6: Relationships between predicted seed production and variation in the community and environment composition. I evaluated the relationship between predicted seed production and grass cover (A), soil depth (B), soil moisture at 7.6 cm depth (C), and soil moisture at 12 cm depth (D) for Plectritis (light green) and Valerianella (dark green) growing in natural plant communities. Pollination Experiment to Test for Negative Frequency-Dependence (4) Do indirect interactions destabilize or promote coexistence?  I found that including pollination treatment (pollen supplemented versus control) significantly improves a model for seed set given the number of flowers per plant (𝜒21 = 10.94, P < 0.01) (Figure 7A). I found that pollen limitation in Plectritis was not dependent on conspecific neighbor densities in the surrounding 1 𝑚2 area (F1, 20 = 0.003, P = 0.96) 19  (Figure 7B) or in the surrounding 0.1 𝑚2 area (F1, 20 = 1.78, P = 0.20) (Figure 7C). The magnitude of Plectritis pollen limitation was not dependent on any of the other community and environmental conditions that I considered, including Valerianella densities, grass cover, shrub cover and soil moisture.  Figure 7: Pollen Limitation and Density-Dependence in a Natural Plant Community. I evaluated the relationship between seed production and individual flower count for pollen supplemented plants (dark yellow) and untreated control plants (light yellow) (A). I quantified pollen limitation of Plectritis as the difference in seeds per flower of control versus pollen supplemented plants, and then modeled its relationship with conspecific densities in the surrounding 1 m2 area (B) or in the surrounding 0.1 m2 area (C). Some plants made more seeds 20  than flowers per plant, which may be due to flowers emerging after the initial counts near the beginning of the flowering phenology window. Discussion Native Plectritis is predicted to exclude exotic Valerianella over the long run.   The mechanisms that allow a diversity of species to coexist in ecological communities, especially for closely-related and functionally-similar species to coexist in those communities, are still being explored. Theory demonstrates that species must have sufficiently high niche differences and/or sufficiently low average fitness difference to coexist at the scale of direct interaction (Chesson 2000). My experimental and observational data support the hypothesis that these species cannot coexist in direct competition due to low niche differences and/or high average fitness differences. The results from my experiment predict that Plectritis will tend to exclude Valerianella in the long run, due to a fitness difference imbalance in favor of Plectritis that overcomes the margin of niche difference (Table 2). The lack of response of Plectritis seed production to densities of either species and the negative response of Valerianella seed production to Plectritis density (Figure 5A and 5C) suggests that Plectritis regulates the abundance of Valerianella more strongly than it regulates its own abundance. This supports the idea that Plectritis should competitively exclude Valerianella abundances where and when they co-occur.   These results are surprising given the exotic origins of Valerianella. Previous work shows that between allopatrically derived species, phylogenetic distance has no effect on niche differences while driving an increase in average fitness differences (Germain et al. 2016). Overall, then, the phylogenetic distance between this species pair should drive an imbalance in average fitness that would destabilize coexistence. I would have predicted that in this pair, exotic Valerianella would be expected to exclude native Plectritis via a higher average fitness, because the competitive effect of invasive species on native species is typically stronger than the effect of native species on invasive species (Vila and Weiner 2004, Maron and Marlor 2008). In this system, it appears that native Plectritis may not resist Valerianella, but does play a role in regulating its abundance through direct interaction. This result supports previous studies that found that while ecological interactions often do not resist invasions, they can constrain the abundances of invasive species once they have established (Levine et al. 2004).    I found an inverse relationship between the abundance of Plectritis and Valerianella (at the 0.1 m2 and 1 m2 scales) across two natural communities (Figure 4), which supports the prediction that these species cannot coexist at the scale of direct interaction. However, these data also acknowledge that Plectritis and Valerianella co-occur at the site scale (and occasionally at the 0.1 m2 and 1 m2 scales), which means that other processes must limit the complete competitive exclusion of exotic Valerianella. Analyzing other elements of demographic performance that I did not 21  measure might be important to gain a more complete understanding of how patterns of these species abundances arise. In many systems, populations are more limited by space for seeds to germinate rather than by the production of seeds (Ashman et al. 2004). If this were true for Plectritis and Valerianella, testing the effect of the density of the plant community on germination rates would contribute to more accurate demographic models, especially in sites where non-native grasses dominant ground cover (MacDougall and Turkington 2005). If density-dependent germination strongly structures the dynamics in this system, it could limit the ability of Plectritis to spread locally to exclude Valerianella.  Additionally, Plectritis showed variation in seed production at low densities and this variation decreased as density increased (Figure 5C). Thus, at low densities the population dynamics of Plectritis are possibly linked to demographic stochasticity which can reduce spread rates (Palamara et al. 2016, Sullivan et al. 2017). Along with higher intraspecific competition when densities increase (Pachepsky and Levine 2011) these expansion limitations might also keep Plectritis from spreading locally to completely exclude Valerianella. Soil moisture context changes niche and fitness differences, but not enough to alter the long-term outcome of interaction.   I conducted my experiment under two different soil-moisture regimes to understand whether coexistence or exclusion was context-dependent. Although the outcome of competitive interaction remained the same in both treatments, I found that niche and fitness differences changed with water availability (Table 1). Environmental drivers such as patterns of precipitation and temperature are being increasingly recognized in their role in predicting species distributions and abundances (Ehrlen and Morris 2015). It has become clear that environmental drivers influence population dynamics through their controls on vital rates and transitions between life cycle stages (e.g. Doak and Morris 2010, Nicole et al. 2011, Dahlgren et al. 2016, Crone et al. 2018), that is, through the demographic potential of a species as quantified here. This experiment shows that environmental drivers can also influence population dynamics by altering the sensitivity of species to competition from neighbors. Changes in competitive interaction coefficients that are not proportional between species will alter the second half of the fitness difference between species, the competitive response ratio. Because a change in the fitness difference that overcomes the niche difference may change whether a species pair can coexist, understanding species abundances and distributions also requires knowledge of how the environmental context will shape aspects of intraspecific competition relative to interspecific competition. The results from this experiment, along with other context-dependent coexistence outcomes, highlight the potentially narrowed scope of studies of niche and fitness difference as measured under one set of controlled environmental circumstances (e.g. Staples et al. 2016), or niche and fitness difference as measured across environmental conditions in 22  heterogenous field-based experiments (e.g. Godoy et al. 2014a and Godoy et al. 2014b).  Although it was not enough to shift the outcome towards coexistence for these species, niche differences increased in the wet versus the dry treatment. Although the functional mechanism that increased the niche difference is unknown, increased soil moisture may have allowed these species more room to differentiate water acquisition through space or time. Phenology in invasive and native plants in another Mediterranean climate system plays a large role in coexistence, with relatively later growing species excluding earlier growing species through a fitness advantage (Godoy et al. 2014a). However, increased within-species variation in phenology or decreased overlap of between-species phenology might increase niche differences, and thus promote coexistence. Comparisons between the phenologies of the plants in my wet versus dry treatments might provide a link between the change in the environmental context and the change in the estimated niche difference.  Due to an inability to determine natural conditions in advance of the experiment, my wet treatment was in actuality wetter than any of the plots surveyed at the three field sites (Appendix 2) while the dry treatment was near the average soil moisture conditions observed. Presuming that the field measurements taken this year were representative of typical years, this means that the dry treatment represents average soil moisture conditions while the wet treatment represents a wet extreme. It would be interesting to test the outcome of competition with a third, drier extreme treatment. Given that the gap between fitness and niche differences decreased in the dry treatment here, if this trend were to continue, coexistence would be predicted under drier conditions. Plectritis and Valerianella respond similarly to variation in environmental conditions, with the exception of soil depth.  Separately from the balance between niche and average fitness differences, spatial or temporal heterogeneity may allow species to coexist (Chesson 2000). I observed co-occurrences of Valerianella with Plectritis at the site scale (and occasionally at the 0.1 m2 and 1 m2 scales) (Figure 4); this may be explained by spatial or temporal heterogeneity that provides an advantage to Valerianella in particular places or at particular times. I did not however, observe any evidence for a unique response of Valerianella that would promote stable abundances. I did not observe any influence of soil moisture on the fecundity or abundances of either species in the field. Perhaps soil moisture is too variable at the local scale and two readings from a 0.1 m2 plot are not fully indicative of the conditions that these plants are experiencing, or alternatively, the measurements from one time period may not reflect the soil moisture conditions of the entire growing season. Deeper soil depths improved the seed production of Plectritis, but not Valerianella. This unique response of Plectritis and inflexible response of Valerianella would provide an advantage to the already dominant Plectritis, furthering its ability to exclude Valerianella. Grass cover had a positive effect on seed production per plant of both species. It would be interesting to separate out the relationship between 23  grasses and the performance of these species by grass origin or functional type to test if invasive grasses are related to increased Valerianella performance but not Plectritis, since particular suites of interacting species may moderate coexistence through intransitivity (Stoufer et al. 2018). Alternatively, the diversity of co-occurring plants may be a more identifiable metric of experienced environmental difference versus abiotic measures (Staples et al. 2016).  Indirect interactions do not stabilize coexistence  I investigated whether indirect interactions via pollinators might stabilize coexistence. Although I found evidence for the presence of pollen limitation in a natural plant community (Figure 7A), its occurrence was not density-dependent (Figure 7B and 7C). Thus pollination dynamics are unlikely to affect competitive interaction coefficients in this system. However, the presence of pollen supplementation (if impo (Doak & Morris, 2010)sed through increased pollinator activity) or further pollination loss (through loss of pollinators in the community) would affect the intrinsic rate of population growth (λ) and thus might shift the outcome of coexistence via the average fitness difference.  Conclusions    My experiment predicts that native Plectritis excludes invasive Valerianella where they co-occur because of an average fitness advantage that outweighs the margin of niche difference between these species. These results are supported by the influence of Plectritis densities on Plectritis and Valerianella fecundities in natural plant communities. These species are still observed to co-occur at the site level and thus Plectritis may limit but not totally eliminate the abundance of Valerianella in natural plant communities. Some possible explanations for this pattern are that other population dynamics not considered here such as density-dependent germination rates and demographic stochasticity may limit the spread of Plectritis at the local level, limiting its ability to interact with Valerianella. I found no evidence for a storage effect that allow Valerianella to maintain stable abundances at the site scale, but more encapsulating or specific measures of abiotic and community patterns may be able to provide evidence for this mechanism of coexistence. Furthermore, I found no evidence for indirect interactions with pollinators to provide a density-dependent population response that would regulate Plectritis abundances and stabilize coexistence. Moreover, my experiment showed how vital rates and interaction coefficients depend on the environmental context, emphasizing that environmental conditions drive abundances through density-independent mechanisms (demographic potential), but also through density-dependent mechanisms (niche difference and competitive response). To make more generalizable and transferable conclusions, future coexistence studies should consider the context of interaction.     24  Bibliography   Adler, P.B., HilleRisLambers, J., Kyriakidis, P.C. Guan, Q. and Levine, J.M. 2006. Climate  variability has a stabilizing effect on the coexistence of prairie grasses. Proceedings of the National Academy of Sciences, 103: 12793-12798.  Adler, P.B., HilleRisLambers, J. and Levine, J.M. 2009. Weak effect of climate variability on coexistence in a sagebrush steppe community. Ecology, 90: 3303-3312. Adler, P.B., HilleRisLambers, J. and Levine J.M. 2007. A niche for neutrality. Ecology Letters, 10: 95-104. Benadi, G. 2015. Requirements for plant coexistence through pollination niche partitioning. Proc. R. Soc. B, 282: 150117. Chesson, P. 1994. Multispecies competition in variable environments. Theoretical Population Biology, 45: 227-276. Chesson, P. 2000. Mechanisms of maintenance of species diversity. Annual review of Ecology and Systematics, 31: 343-366. Consortium of Pacific Northwest Herbaria Specimen Database (CPNWH). 2017. Website: http://www.pnwherbaria.org (accessed April 15, 2017). Diamond, J.M. 1975. The island dilemma: lessons of modern biogeographic studies for the design of natural reserves. Biological conservation, 7: 129-146. Doak, D., and Morris, W. 2010. Demographic compensation and tipping points in climate-induced range shifts. Nature, 467: 959-962. Ehrlén, J., and Morris, W. 2015. Predicting changes in the distribution and abundance of species under environmental change. Ecology Letters, 18: 303-314. Ehrlén, J., Morris, W., von Euler, T., & Dahlgren, J. 2016. Advancing environmentally explicit structured population models of plants. Journal of Ecology, 104: 292-305. Ernet, D. 1977. Flower morphology and reproductive biology of Valerianella and Fedia (Valerianaceae). Plant Systematics and Evolution, 128: 1-22. Fuchs, M.A. 2001. Towards a recovery strategy for Garry oak and associated ecosystems in Canada: ecological assessment and literature review. Environment Canada, Pacific and Yukon Region. Ganders, F.R., Carey, K. and Griffiths, A.J. 1977. Outcrossing rates in natural populations of Plectritis brachystemon (Valerianaceae). Canadian Journal of Botany, 55: 2070-2074. Germain, R.M., Weir, J.T. and Gilbert, B. 2016. Species coexistence: macroevolutionary relationships and the contingency of historical interactions. Proc. R. Soc. B, 283: 20160047. Godoy, O. and Levine, J.M. 2014. Phenology effects on invasion success: insights from coupling field experiments to coexistence theory. Ecology, 95: 726-736. 25   Godoy, O., Kraft, N.J. and Levine, J.M. 2014. Phylogenetic relatedness and the determinants of competitive outcomes. Ecology Letters, 17: 836-844. Going, B.M., Hillerislambers, J. and Levine, J.M. 2009. Abiotic and biotic resistance to grass invasion in serpentine annual plant communities. Oecologia, 159: 839-847. HilleRisLambers, J., Adler, P.B., Harpole, W.S., Levine, J.M. and Mayfield, M.M. 2012. Rethinking community assembly through the lens of coexistence theory. Annual Review of Ecology, Evolution, and Systematics, 43: 227-248. Jacobs, B., Bell, C. and Smets, E. 2010. Fruits and seeds of the Valeriana clade (Dipsacales): Diversity and evolution. International journal of plant sciences, 171: 421-434. Lea, T. 2006. Historical Garry oak ecosystems of Vancouver Island, British Columbia, pre-European contact to the present. Davidsonia, 17: 34-50. Levine, J., Adler, P., & Yelenik, S. 2004. A meta-analysis of biotic resistance to exotic plant invasions. Ecology Letters, 7: 975-989. Macarthur, R. and Levins, R. 1967. The limiting similarity, convergence, and divergence of coexisting species. The American Naturalist, 101: 377-385. MacDougall, A.S. and Turkington, R. 2005. Are invasive species the drivers or passengers of change in degraded ecosystems? Ecology, 86: 42-55. MacDougall, A.S. and Turkington, R. 2006. Dispersal, competition, and shifting patterns of diversity in a degraded oak savanna. Ecology, 87: 1831-1843. Maron, J.L., Baer, K.C. and Angert, A.L. 2014. Disentangling the drivers of context‐dependent plant–animal interactions. Journal of Ecology, 102: 1485-1496. Maron, J., and Marler, M. 2008. Effects of native species diversity and resource additions on invader impact. The American naturalist, 172: 18-33. Matías, L., Godoy, O., Gómez-Aparicio, L., and Pérez-Ramos, I. 2018. An experimental extreme drought reduces the likelihood of species to coexist despite increasing intransitivity in competitive networks. Journal of Ecology, 106: 826-837. Nicolè, F., Dahlgren, J., Vivat, A., Till-Bottraud, I., and Ehrlén, J. 2011. Interdependent effects of habitat quality and climate on population growth of an endangered plant. Journal of Ecology, 99: 1211-1218. Pachepsky, E., and Levine, J. 2011. Density Dependence Slows Invader Spread in Fragmented Landscapes. The American Naturalist, 177: 18-28. Palamara, G., Carrara, F., Smith, M., and Petchey, O. 2016. The effects of demographic stochasticity and parameter uncertainty on predicting the establishment of introduced species. Ecology and Evolution, 6: 8440-8451. 26  Parachnowitsch, A.L. and Elle, E. 2005. Insect visitation to wildflowers in the endangered Garry Oak, Quercus garryana, ecosystem of British Columbia. The Canadian Field-Naturalist, 119: 245-253. Pake, C.E. and Venable, D.L. 1995. Is coexistence of Sonoran desert annuals mediated by temporal variability reproductive success? Ecology, 76: 246-261. Pauw, A. 2013. Can pollination niches facilitate plant coexistence? Trends in ecology and evolution, 28: 30-37. R Core Team. 2016. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. Sears, A.L. and Chesson, P. 2007. New methods for quantifying the spatial storage effect: an illustration with desert annuals. Ecology, 88: 2240-2247. Staples, T.L., Dwyer, J.M., Loy, X. and Mayfield, M.M. 2016. Potential mechanisms of coexistence in closely related forbs. Oikos, 125: 1812-1823. Stouffer, D., Wainwright, C., Flanagan, T., and Mayfield, M. 2018. Cyclic population dynamics and density-dependent intransitivity as pathways to coexistence between co-occurring annual plants. Journal of Ecology, 106: 838-851. Sullivan, L., Li, B., Miller, T., Neubert, M., and Shaw, A. 2017. Density dependence in demography and dispersal generates fluctuating invasion speeds. Proceedings of the National Academy of Sciences of the United States of America, 114: 5053-5058. Tenhumberg, B., Crone, E., Ramula, S., and Tyre, A. 2018. Time-lagged effects of weather on plant demography: drought and Astragalus scaphoides. Ecology, 99: 915-925. Thompson, K., Petchey, O.L., Askew, A.P., Dunnett, N.P., Beckerman, A.P. and Willis, A.J. 2010. Little evidence for limiting similarity in a long‐term study of a roadside plant community. Journal of Ecology, 98: 480-487. Tilman, D. 1980. Resources: a graphical-mechanistic approach to competition and predation. The American Naturalist, 116:362-393. Trowbridge, C.C., Stanley, A., Kaye, T.N., Dunwiddie, P.W. and Williams, J.L. 2016. Long‐term effects of prairie restoration on plant community structure and native population dynamics. Restoration Ecology, 25: 559-568. Vellend, M., Bjorkman, A.D. and McConchie, A., 2008. Environmentally biased fragmentation of oak savanna habitat on southeastern Vancouver Island, Canada. Biological Conservation, 141: 2576-2584. Vilà, M., and Weiner, J. 2004. Are invasive plant species better competitors than native plant species? - evidence from pair-wise experiments. Oikos, 105: 229-238. Vilà, M., Espinar, J., Hejda, M., Hulme, P., Jarošík, V., Maron, J., and Pyšek, P. 2011. Ecological impacts of invasive alien plants: a meta-analysis of their effects on species, communities and ecosystems. Ecology Letters, 14: 702-708. 27  Young-Mathews, A. 2012. Plant fact sheet for shortspur seablush (Plectritis congesta). USDA-Natural Resources Conservation Service, Plant Materials Center, Corvallis, OR                                       28  Appendices  Appendix 1: Soil moisture conditions for competition experiment. I watered the plants every other day for 5 and 12 minutes for the dry and wet treatments during June and July. I used data from the pots gathered on an afternoon in early July on a day where the plants received water and again on the following day when the plants did not receive water and then I compared the % VWC at 7.6 cm depth for the dry and wet treatment groups to confirm the effectiveness of my different watering regimes using a two-sample t-test.    Day of watering Day after watering Mean % VWC (dry treatment) 20.3 18.3 Mean % VWC (wet treatment) 34.8 33.9 Degrees of freedom 30 30 t-statistic 5.09 6.01 P-value < 0.01 < 0.01              29  Appendix 2: Model selection for seed production in Plectritis and Valerianella. I chose the best models for seed production based on the lowest relative AIC score (bolded and starred).  Species Fitness Proxies AIC Score Plectritis number of inflorescences + plant height + main inflorescence size 353.35  number of inflorescences + plant height 351.38  number of inflorescences + main inflorescence size 359.98  plant height + main inflorescence size 352.98  number of inflorescences 357.99  plant height 351*  main inflorescence size 360.06 Valerianella number of inflorescences + plant height + main inflorescence size 1892*  number of inflorescences + plant height 1894.5  number of inflorescences + main inflorescence size 1915.4  plant height + main inflorescence size 1908.5  number of inflorescences 1927.3  plant height 1911.2  main inflorescence size 2216.1  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items