Non-native plant species can disrupt plant–pollinator interactions by altering the foraging behavior of pollinators and by contributing to the transfer of interspecific pollen to native plant species. In this study I investigate these potential effects in fennel (Foeniculum vulgare), which has become common throughout coastal California since its introduction from Europe in the 1800s. Fieldwork for this study took place on Santa Cruz Island (Santa Barbara County, CA) where fennel has become locally abundant following the removal of non-native sheep and pigs. I conducted a controlled and replicated fennel-flower removal experiment to test three hypotheses: (i) fennel shares pollinators with native plants, (ii) fennel flower removal alters visitation by insect pollinators to flowers of native plants, and (iii) fennel flower removal decreases the transfer of fennel pollen to flowers of native plants. The native plant considered in this study was the Santa Cruz Island buckwheat (Eriogonum arborescens), which is endemic to Anacapa, Santa Cruz and Santa Rosa Islands. Overlap in the insects visiting both plant species was relatively low, but buckwheat and fennel did share certain pollinators (e.g., Augochlorella pomoniella, Colletes sp., Hylaeus sp. and V. pennsylvanica). Native green sweat bees preferentially visited buckwheat, whereas yellowjacket wasps (Vespula pensylvanica) preferentially visited fennel. After the experimental removal of fennel flowers, insect visitation to buckwheat flowers was significantly greater in the control group than in the fennel flower removal group, but fennel pollen transfer was unaffected by the removal of fennel flowers. These results indicate that fennel acts as a magnet species for certain pollinating insects and that this attraction spills over to the native Santa Cruz Island buckwheat. These results also indicate that insects readily transfer fennel pollen beyond the 13 x 13 m area of our study plots. Future research could focus on the extent to which fennel pollen deposition on native plant stigmas hinders plant reproduction.

Invader-removal experiments can lend insight into the ecological resilience of biological assemblages and the extent to which they can recover from perturbations. Argentine ant invasions on Santa Cruz Island have disrupted the diversity and functionality of ant assemblages. This thesis sought to quantify changes in native ant species following invader removal in nine experimental plots paired with nine control plots on Santa Cruz Island. Stable isotope analysis (variation in δ15N) was used to estimate relative trophic positions of and resource acquisition by native ant species. Spatiotemporal variation in δ15N was also investigated to assess its relative impact on the trophic niches of twelve native ant species. Stable isotope analysis was conducted on an assemblage of native ants in 2017 and 2019 following the removal of the Argentine ant from Santa Cruz Island conducted by The Nature Conservancy with support from the National Park Service. Soil δ15N and plant δ15N values were positively related to one another. Subsequently, this correlation was used to correct for the δ15N values of native ant species. δ15N values for native ant species did not differ between control and experimental plots. Native ant species also displayed temporal constancy in δ15N values in 2017 and 2019. Overall, there was significant variation in trophic positions among native ant species. It is hypothesized that functional traits of the ant species played a role in niche diversification, and overall competition levels were minimized as a result of these differences. Species richness was lower on experimental plots as the experimental recovery is still ongoing, and additional ant species are expected to re-colonize the plots. The similar trophic positions of ants on invaded and uninvaded plots demonstrated the ability of native ants to recover to their pre-invasion state following successful eradication of the invader.

Plant-pollinator mutualisms provide essential services to both wild and cultivated ecosystems across the globe. Yet these important interactions face many environmental threats that could impact the ability of pollinators to effectively mediate reproduction in plant hosts. Climate change is one such threat. Past studies have investigated how abiotic stress can lead to mismatches in phenology and distribution in plant-pollinator relationships, but less research has focused on the effects of altered environmental conditions on plant floral traits. This dissertation centers on the impacts that warming and drought stress have on the production of floral resources, how bee pollinators respond to those changes, and how plant reproduction is ultimately affected. For this work, we utilized the Cucurbita system, including both cultivated squash (Cucurbita pepo) and free-living gourd (C. foetidissima), which is pollinated by generalist honey bees (Apis mellifera) as well as specialist squash bees (Eucera). In Chapter 1, we examined how the combined effects of warming and drought altered the C. pepo system and found that water stress increased pollen limitation in bee-pollinated plants due to (i) decreased pollen competition caused by low-levels of stigmatic pollen deposition, (ii) reduced viability of pollen produced by plants grown under low soil moisture conditions, and (iii) a reduced capacity of self-pollinated fruits to increase seed set in response to increasing soil moisture. In Chapter 2, we investigated how generalist and specialist bees responded when given a choice between C. pepo plants grown at varying soil moistures, and discovered that only generalist honey bees increased visitation with plant soil moisture, thereby increasing deposition of pollen from well-watered plants and increasing seed set. In Chapter 3, we focused on comparing the effectiveness of honey bees and squash bees as pollinators of both C. pepo and C. foetidissima and assessed that, overall, squash bees removed more pollen, deposited more pollen, and contributed more to fruit set and seed set on both Cucurbita species. The results of this dissertation may be applied to other plant-pollinator systems and reveal that the contributions of generalist and specialist pollinators to pollination services may be impacted in an altered climate.

Differences between specialist and generalist pollinators have long been a topic of interest for biologists. Plants of the genus Cucurbita (Cucurbitaceae) are visited by generalist pollinators, such as honey bees (Apis mellifera), and by specialist pollinators, known as squash bees (e.g., Peponapis pruinosa). Previous studies have examined pollinator effectiveness between Apis and Peponapis on agricultural Cucurbita species, but few have investigated the effectiveness of these pollinators in a non-agricultural context. In the summer of 2017, I conducted single visit pollination trials on 22 buffalo gourd (Cucurbita foetidissima) plants at Los Peñasquitos Canyon Preserve, San Diego County, CA to measure pollinator effectiveness between honey bees and squash bees. The percentage of fruit set from single visits made by female P. pruinosa (57.9%) was higher than that of male P. pruinosa (23.5%). Single visits by Apis never resulted in successful fruit set. Control fruit, by comparison, set at a percentage of 85.7%. The average weight of seeds from fruit resulting from female P. pruinosa single visits was significantly higher (by 28.6%) than for single visits by males, but neither seed number nor fruit volume differed between male and female single visits. These results indicate that Apis and male P. pruinosa are less effective at pollinating buffalo gourd compared to female P. pruinosa. Differences in how these pollinators gather nectar from Cucurbita may be one explanation for differences in fruit set. These differences may be exaggerated in non-agricultural systems, where generalist pollinators like Apis have a variety of floral resources to choose from.

Functional traits are popular tool to assess functional ecology. Functional traits determine how an organism interacts with its community. The utility of functional traits has been well documented in plant and vertebrate systems but has not been as thoroughly applied to invertebrate systems. Morphology is a common functional trait used to predict ecology in vertebrate and plant models. This study tested the ability of morphology to predict trophic ecology of native ants from Santa Cruz Island. Trophic interactions represent a large portion of inter- and intraspecific interactions within a community. Analyzing these interactions is key to understanding ecosystem functioning. To assess trophic ecology, I used stable isotope analysis of δ15N and δ13C as well as predation and scavenging assays. To assess morphology, I measured 27 morphological traits from eight species of ant. The 10 most cited measurements were analyzed using linear regressions to assess the relationships between morphology and trophic ecology. Significant relationships exist between morphology and stable isotope values but not between morphology and predation or scavenging ability. The lack of relationship between morphology and predation or scavenging ability could be due to environmental conditions or due to morphology’s inability to predict predation or scavenging behavior. In future works, I would perform phylogenetic corrections to account for relatedness between ant species and apply similar observational studies to other ant community assemblages.

This thesis examines mitochondrial DNA variation exhibited by Aphaenogaster patruelis, an ant species that is confined to five oceanic islands off the coast of southern California and Baja California. The aims of this study are to assess inter-island relationships and to determine if there is cytonuclear discordance. Aphaenogaster patruelis is a putative relict known only from San Clemente Island, Santa Catalina Island, San Nicolas Island, Santa Barbara Island, and Isla Guadalupe. Samples were collected from each of the islands, and DNA extraction was performed to sequence the 12S and CO1 mitochondrial genes. Based on these sequence data, I constructed median-joining haplotype networks and a CO1 phylogenetic tree. Samples on three islands were represented by a single haplotype, but samples on two islands were represented by two, distinct haplotypes. The presence of two, distinct haplotypes suggests multiple colonization events on these islands and indicates the presence of cytonuclear discordance as the split is not evident in previous analyses based on nuclear DNA. Given the lack of an extant mainland population of Aphaenogaster patruelis, reconstructing evolutionary relationships among island populations is not possible, but our results do suggest repeated colonization on individual islands, either through movements between islands or between islands and a now extinct mainland population.

The ecological effects of species introductions can change in magnitude over time, but how and why such impacts exhibit temporal variation remains incompletely understood. In this study we used stable isotope analysis to estimate how trophic position changes as a function of the stage of invasion for the Argentine Ant (Linepithema humile), a widespread, abundant, and ecologically disruptive invader. Previous studies in southern California found that the trophic position of the Argentine ant was higher at the leading edge of invasion than at sites invaded several years earlier. To assess if a reduction in relative trophic position over time is a common feature of ant invasions, we expanded the temporal and spatial scale of sampling and estimated the relative trophic position of Argentine ants along three invasion chronosequences: Rice Canyon (San Diego Co. CA), the Sacramento River Valley (Yolo and Solano Cos., CA), and San Nicolas Island (Ventura Co., CA). Resampling Rice Canyon in 2019, 16 years after the original survey, revealed a surprising increase in Argentine ant trophic position. At the two other invasion chronosequences, the trophic position of the Argentine ant did not change with the stage of invasion. These findings suggest that changes in relative trophic position associated with invasion processes may reflect short-term responses, which are not evident in long-term sampling efforts. These findings also point to the potential value of historical data and repeated, annual sampling.

Habitat fragmentation due to human activity has led to pollinator declines worldwide, yet little is known about how this diversity loss affects pollination services in natural ecosystems. In this study, we investigate the mechanistic links between habitat fragmentation, a proxy for pollinator diversity loss, and the quantity (conspecific pollen deposition) and quality (heterospecific pollen proportion) of pollination services in coastal sage scrub habitats in the San Diego region. We documented pollinator visitation and pollen deposition across ten focal plant species in six natural reserve and six scrub fragment plots. At the level of the community as a whole, habitat fragmentation per se was not a significant driver of conspecific pollen deposition nor was it a significant driver of heterospecific pollen proportion. However, habitat type (reserves vs. fragments) formed statistically significant interactions with other variables in both conspecific and heterospecific proportion models, suggesting that fragmentation can indirectly affect pollination services. The western honeybee, Apis mellifera, was the most numerically dominant floral visitor across all study plots. Due to its high abundance, A. mellifera could influence both conspecific pollen deposition and heterospecific pollen proportion among reserve and fragment plots, potentially altering pollination services within coastal sage scrub habitats. Overall, habitat fragmentation can impart complex effects within plant-pollinator networks, as habitat type was shown to affect plant species differently with regards to changing pollinator variables. These differences in responses from plant species and pollinators could potentially result in a restructuring of plant-pollinator networks.

The Argentine ant (Linepithema humile) is a facultative scavenger that opportunistically consumes both live prey and carrion. Despite its stature as a widespread and disruptive invader, how it interacts with other arthropods remains incompletely understood; published studies report a wide range of impacts on non-ant arthropods. Here, we examine interspecific and intraspecific variation in susceptibility of different arthropods to predation by the Argentine ant. We used mesh-covered enclosures containing different arthropods to measure susceptibility to predation. Mesh screens permitted Argentine ants to move in and out of the cages, but mesh size was too small to permit escape of other arthropods. For four different arthropod taxa, the percent killed by the Argentine ant after 24 h was as follows: roaches (25%), crickets (70%), wax moth larvae (100%), and harvester ants (100%). For these same arthropods, L. humile consumed carrion and killed prey to a similar degree, indicating the importance of scavenging in general and, in particular, for arthropods that the Argentine ant can’t readily subdue as prey. In a second experiment, we found that increasing body size protected crickets from ant predation. These findings provide experimental evidence for interspecific and intraspecific variation in susceptibility to predation by the Argentine ant and provide a potential framework for understanding the contrasting impacts resulting from this invasion. Our results further illustrate the importance of scavenging for ant invasions.

Invasive species damage invaded ecosystems, displacing native species and decreasing biodiversity. In this study we focus on a specific invader, Argentine Ant (Linepithema humile), in the salt marshes of San Diego County. Argentine ants are a widespread and ecologically disruptive invader. Although ants are known to forage in intertidal environments, impacts caused by introduced ants in intertidal habitats are undescribed with no published information about ant foraging in Pacific coastal marshes. To learn more about the ecological effects of Argentine ant foraging in salt marshes we conducted an activity survey using baits to measure recruitment in marsh and adjacent upland habitat. Second, we conducted a lab experiment to test how saline conditions affect colony growth. Last, we used stable isotope analysis to determine the extent to which the Argentine ant assimilates resources from intertidal environments by comparing this invader with a native ant (Dorymyrmex insanus) and fiddler crabs (Uca sp.). Our lab experiment showed that there is no significant effect of salinity on ant colony growth. Our activity surveys revealed that the Argentine ant occurs commonly in intertidal marsh habitat with activity levels not different from those observed in adjacent upland habitat. Stable isotope analyses showed that the Argentine ant had a significantly higher δ15N value and a significantly lower δ13C compared to D. insanus. This study is the first known to measure Argentine ant foraging behavior in salt marshes in San Diego. Our results indicate that Argentine ants readily invade salt marshes in a similar manner as terrestrial habitats.