Plant–arthropod interactions of an endangered California lupine

Abstract The reintroduction of endangered plant species is an essential conservation tool. Reintroductions can fail to create resilient, self‐sustaining populations due to a poor understanding of environmental factors that limit or promote plant success. Biotic factors, specifically plant–arthropod interactions, have been shown to affect the establishment of endangered plant populations. Lupinus nipomensis (Nipomo Mesa lupine) is a state of California (California Rare Plant Rank: 1B.1) and federally (65 FR 14888) endangered endemic plant with only one extant population located along the central California coast. How arthropods positively or negatively interact with L. nipomensis is not well known and more information could aid conservation efforts. We conducted arthropod surveys of the entire L. nipomensis extant population in spring 2017. Observed arthropods present on L. nipomensis included 17 families, with a majority of individuals belonging to Thripidae. We did not detect any obvious pollinators of L. nipomensis, providing support for previous studies suggesting this lupine is capable of self‐pollinating, and observed several arthropod genera that could potentially impact the reproductive success of L. nipomensis via incidental pollination or plant predation.

& Kaye, 2007). Knowledge of herbivory and pollination has been shown to be important in maximizing reproduction and increasing establishment success of rare plant species (Archer & Pyke, 1991;Kay, 2008). Due to the often small population size of endangered plants, arthropod herbivores can have strong impacts through a reduction in plant size, growth, and recruitment (Ancheta & Heard, 2011;Myers & Sarfraz, 2017). Similarly, pollination is especially important to endangered plant species in order to maintain the population size and increase outcrossing of individuals (Horth, 2019;Reiter et al., 2017;Steffan-Dewenter et al., 2001). Maintaining genetic diversity and reducing the likelihood of inbreeding depression is essential to the conservation of rare plant species (Falk, 1990;Lee et al., 2018).  Figure 1). Historically, L. nipomensis has occurred at low densities in back dunes and inter-dune habitat. The loss of coastal back dune habitat due to land conversion, fragmentation, and competition with the invasive perennial veldt grass (Ehrharta calycina Sm., Poaceae) limits the range and potential for natural regeneration of the L. nipomensis population (Skinner & Pavlik, 1994). The entire extant population is geographically iso- The reintroduction of Nipomo Mesa lupine, and other rare plants with a limited initial population size, requires specific abiotic conditions to maximize reproductive output from reintroduction efforts.  provide an overview of abiotic microhabitat characteristics (i.e., landscape slope and aspect) and seed treatment relevant to L. nipomensis fecundity during reintroduction efforts. A foundational study of L. nipomensis by Walters and Walters (1989) primarily focuses on abiotic drivers of reproduction, specifically changes observed in flowering and fruit set as a factor of rainfall received. Walters and Walters (1989) also include a record of herbivorous arthropod interactions, but do not specifically include pollinator observations. Because other annual lupines (e.g., Lupinus bicolor Lindl.) are known to be pollinated by bees and flies, the absence of pollinator observations has led practitioners and researchers to hypothesize that L. nipomensis is capable of self-pollinating Moldenke, 1976;USFWS, 2019).
In this study, we sought to establish a baseline of plant-arthropod interactions of in situ L. nipomensis, especially targeted pollinator observations, which have not previously been studied. We surveyed arthropod use of L. nipomensis and classified plant visitors as potential pollinators when observed on or in the flowers. The primary goal of this study was to create an inventory of observed arthropod interactions with L. nipomensis to inform future research and conservation efforts.

| Study area
This study was conducted in the spring of 2017 at the Phillips 66 Oil Refinery (35.0388889, −120.5894444) in San Luis Obispo County, California, USA ( Figure 2). The region is characterized by a Mediterranean-type climate with cool, wet winters and hot, dry summers, while also receiving occasional inputs of water from coastal fog (Baguskas et al., 2016). San Luis Obispo County receives an average precipitation of 33.1 cm annually, has an average wind speed of 10.83 km/h (November-May), and average low and high temperatures ranging from 5.92 to 18.9°C during the growing season (Western Regional Climate Center, 2020). We classified cloud cover on a 4-point scale and recorded ambient air temperature and wind speed during each visit (see complete dataset Luong et al., 2021).
Phillips 66 is required to mitigate for their oil development by establishing protected areas where extant populations are regularly monitored and restricting management actions that could negatively affect L. nipomensis populations (USFWS, 2009). This protected area in which we conducted our study is a coastal back dune ecosystem and the oldest part of a dune complex. These less disturbed areas often have later successional plants with increased soil stability as well as higher plant and insect diversity (Buckler, 1979;Ferrier et al., 2012;Miller et al., 2010). The study site is dominated by non-native,

| Plant visitor observations
One monitoring plot was established per colony (n = 3). Each plot was 8m × 8m and contained 22-96 L. nipomensis individuals. We conducted arthropod visitor surveys using a standardized observationbased protocol useful for recording flower and foliage visitations (Herrera, 1990;Thompson, 2001). Observers conducted a 40-min  (Willimer, 1983). One set of early and late sampling (beginning at 9:00 and 16:00, respectively) was conducted to increase the likelihood of sampling temporal niche visitors.
During each arthropod visitor survey, L. nipomensis individuals within the monitoring plot were closely observed and any visiting arthropods (floral and otherwise) were captured using aspirations and hand or net collections (Chacoff & Aizen, 2006;Kleijn & Langevelde, 2006). It is important to note that not all flower visitors observed were pollinators, but they are considered potential or "incidental pollinators." Incidental pollinators move pollen from flower to flower while foraging for other resources (Anandhan et al., 2020;Kearns et al., 1998). Nets were only used when necessary to minimize dam- At the lab, inflorescences were dissected and any arthropods found were placed in 75% ethanol vials. Specimen collection complied with California State and Federal laws and samples were vouchered at the Invertebrate Zoology Collection at the University of California, Santa Barbara (specimen catalog numbers are included in Appendix S1).
Specimens identified to family were sorted into putative species based on morphology or morphospecies (Samways et al., 2010).
New host records were determined and recorded via the Global Biotic Interactions (GloBI) database and through literature searches (Poelen et al., 2014). Families classified as flower visitors are considered potential pollinators and were observed on or in flower parts of L. nipomensis.

| Data visualization
Interaction data between L. nipomensis and visiting arthropods was visualized in R Studio (version 1.4.1106). An interaction web was F I G U R E 2 Map of sampled Lupinus nipomensis colonies (labeled 1-3) and geographic placement of study site in San Luis Obispo County, California, USA. The study area is indicated by the red box and county lines are displayed in gray on the map of California, USA. The inset box displaying the study area is approximately 400 m × 300 m

| RE SULTS
A total of 351 arthropod individuals were observed interacting with L. nipomensis during our surveys ( Table 1) Individuals observed visiting L. nipomensis that were identified to genus and species include members of orders Diptera, Coleoptera, Hemiptera, Hymenoptera, and Lepidoptera (Appendix S1). The one Dipteran observed was Delia lupini (Coquillett, 1901) (Anthomyiidae).
Hemipterans identified to lower classifications include Orius sp.

| DISCUSS ION
Plant-arthropod interaction studies associated with restoration efforts are becoming more common due to a greater appreciation of the role arthropods play in the success of plants through positive (e.g., pollination) and antagonistic (e.g., herbivory) interactions (Bucharova et al., 2021;Cariveau et al., 2021;Sabatino et al., 2021). While our study did not detect obvious pollinators, we recorded a high number of arthropod individuals that could be incidental pollinators and inadvertently pollinate L. nipomensis while feeding on pollen or other plant resources (Gill, 1991). We observed over 200 Thysanoptera individuals present in L. nipomensis flowers (Table 1). Thysanoptera have been known to pollinate members of several angiosperm families, including fabaceous plants (Varatharajan et al., 2016;Velayudhan & Annadurai, 1986). Individuals of Thysanoptera we found within L.
nipomensis inflorescences were observed to have pollen attached to their body (Appendix S2). However, Thysanoptera are well-known flower pests that consume pollen, potentially causing withering of flowers and lowering plant reproductivity (Reitz, 2009). We also observed arthropods known to be important pollinators (i.e., Apis mellifera) visiting neighboring plants, including another lupine species, Lupinus chamissonis Eschsch. (Aslan et al., 2016;Hung et al., 2018). While Apis mellifera and other members of Apidae are known to pollinate lupine species and were detected in our survey plots, we did not observe these arthropods interacting with L. nipomensis . Our study supports previous findings that L.
nipomensis is capable of self-pollinating due to an apparent lack of visitation by arthropods known to pollinate, such as bees and flies (Cullen et al., 2021;Nye & Anderson, 1974;USFWS, 2019). If pollination is occurring among L. nipomensis individuals, this service is likely being performed by incidental pollinators, such as Thysanoptera.
Arthropod genera were observed that could affect the reproductive success of L. nipomensis due to herbivory. The Dipteran found, Delia lupini (Anthomyiidae), was collected from a gall present on a L. nipomensis individual and has been observationally implicated to reduce fecundity in a previous study of L. nipomensis (Walters & Walters, 1988). While some galls can be benign, most are detrimental to plant health and, in some cases, have been shown to threaten endangered plant species (Harris & Pitzschke, 2020;Kolesik et al., 2019). Members of Curculionidae, which are known plant pests, were observed within the stem and interacting with D. lupini galls (Johnson-Cicalese et al., 1990;Petrova et al., 2006). The impact of D. lupini gall presence on L. nipomensis reproductivity was not quantified in this study; however, it is possible there were additional individuals we did not observe that are affecting the fecundity of this lupine. Formicidae species we observed included Linepithena humile, the invasive Argentine ant (Holway, 1999). Argentine ants have been shown to impact floral visitation patterns and nesting success of other arthropods with the potential to create cascading, negative effects, and reduce pollinator visitation (Plentovich et al., 2021;Sahli et al., 2016;Underwood & Fisher, 2006). The presence of a gall-inducing and invasive arthropods may further inhibit recruitment within this singular extant population of L. nipomensis.
Plants face both biotic and abiotic barriers to reproductive success, and for rare, endangered species, these barriers can ultimately result in extirpation or extinction (Rejmánek, 2018 (Shi et al., 2005). A reduction in pollination leads to reduced outcrossing, ultimately resulting in a lower genetic diversity that further threatens already small, rare plant populations (Gray, 2019). Simultaneously, herbivorous arthropods can impact seed production and recruitment of plant species (Lucas-Barbosa, 2016).
Active intervention may be necessary to promote outcrossing via hand pollination, as well as protect L. nipomensis from herbivorous arthropods to ensure the genetic diversity and successful establishment of new individuals (Serrano et al., 2021;Walsh et al., 2019).
Our results suggest that if pollination is occurring among L.
nipomensis individuals, it is likely being performed by incidental pollinators. Additionally, we observed herbivorous arthropods that may threaten fecundity of this lupine. Further work is necessary to determine the frequency, as well as mode, of pollination and whether the potential threats this lupine faces from herbivorous arthropods will affect the establishment of novel populations during restoration efforts.

We thank the Land Conservancy of San Luis Obispo and the
Phillip-66 refinery in Arroyo Grande for their cooperation and support. We greatly appreciate the data collection assistance of Marli Restoration at UC Santa Barbara.

CO N FLI C T O F I NTE R E S T
Any findings or recommendations expressed in this material do not necessarily reflect the views of funding agencies. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Writing -review & editing (equal).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available in "Plant-arthropod interactions of an endangered California lupine"