Las Vegas bearpoppy (Arctomecon californica) occurrences have fluctuated during the past several decades, in part due to interannual variability in rainfall that influences recruitment and mortality events; yet, development in the Las Vegas Valley continues to threaten habitat supporting this species. Arctomecon californica was petitioned for listing under the Endangered Species Act in 2019 and is currently under review to determine whether listing is warranted (USFWS 2020). This review updates species information for A. californica and includes recent insights into the species' seed ecology, habitat requirements and suitability models, propagation and reintroduction, and pollinator biology. We include information from the past 20 years in these areas that supplement conservation and restoration actions for the species. We also identify topics with scarce information and highlight areas for future study, including the following: preservation of genetic diversity through germplasm collections, identification of mechanisms driving the species' soil endemism, maintenance of A. californica–pollinator relationships through understanding pollinator habitat, determination of the viable seed fraction and its longevity in the soil seed reserves, and prediction of population response to regional climate change based on demographic modeling.
La ocurrencia de “Las Vegas Bearpoppy” (Arctomecon californica) han fluctuado durante las últimas décadas, en parte debido a la variabilidad interanual de las precipitaciones que influyen en los eventos de reclutamiento y mortalidad. Sin embargo, el desarrollo en el valle de Las Vegas sigue amenazando el hábitat que sustenta a esta especie. Se solicitó incluir a A. californica en la lista de la Ley de Especies en Peligro de Extinción en 2019 y actualmente se está revisando para determinar si se justifica tal inclusión (Servicio de Pesca y Vida Silvestre de los Estados Unidos 2020). Esta revisión actualiza la información de la especie A. californica, e incluye conocimientos recientes sobre la ecología de las semillas de la especie, los requisitos de hábitat y los modelos de adecuación, propagación, reintroducción, así como la biología de sus polinizadores. Incluimos información de los últimos veinte años en estas áreas que complementan las acciones de conservación y restauración de la especie. También, identificamos temas con escasa información disponible y destacamos áreas para futuros estudios, incluyendo: preservación de la diversidad genética a través de colecciones de germoplasma, identificación de los mecanismos que impulsan su endemismo en el suelo, estudio del hábitat de los polinizadores para mantener la relación con A. californica, determinación de la fracción de semillas viables y su longevidad en las reservas de semillas del suelo, y generar predicciones de la respuesta de la población al cambio climático regional basada en modelos demográficos.
Arctomecon californica is a short-lived, herbaceous perennial in the Papaveraceae and grows on unique substrates in the Mojave Desert of southeastern Nevada and northwestern Arizona (Thompson and Smith 1997). Arctomecon californica has been heavily studied compared to other rare endemic plants of the Mojave Desert; however, much species and habitat information is unknown or remains equivocal. Arctomecon californica is listed as critically endangered by the state of Nevada, is covered by Clark County's Multi-Species Habitat Conservation Plan, is listed as a special-status species in Arizona (WRA Environmental Consultants 2018), and was petitioned for federal listing under the Endangered Species Act in 2019 (Cornelisse and Tyler 2019).
Previous reviews of A. californica focused some attention on practices to augment populations or reintroduce plants into unoccupied habitats that appear suitable for sustaining populations. Mistretta et al. (1996) is a comprehensive review of A. californica and concludes that a seed storage program for conserving germplasm and investigating production of commercially available stock may alleviate collection pressure for landscape purposes. In their multispecies review, The Nature Conservancy (2007) recommended salvaging surface soils containing A. californica seeds and seedlings as mitigation, in the managed Nellis Dunes OHV Recreation Area in particular, where loss and damage of rare plant habitat would require exceptional means to reclaim and restore. We build upon these reviews, provide updated information for the species, and refer to earlier material on habitat description and life history strategy to provide context for conservation actions. We direct readers to earlier literature for the historical perspective and comprehensive treatment of A. californica taxonomy, species description, distribution, status and trends, surveys and population trends, and threats. Our review draws from published and unpublished literature and focuses on recent findings to evaluate potential conservation and restoration actions for long-term protection of A. californica. Specifically, we summarize information on population dynamics, habitat, seed ecology, and pollination biology and discuss how this information relates to propagation, reintroduction, and habitat management.
Arctomecon californica occurrences are highly variable over the species' range, due in part to the influence of timing and amount of precipitation on population structure and in part to the fragmentation of habitats from urban development. The Bureau of Land Management's comprehensive survey in 1993 estimated that there were 830,000 plants on 39,500 acres across 99 populations (91 in Clark County, Nevada, and 8 in Mohave County, Arizona), although these estimates may represent a peak in population cycles influenced by above-average rainfall for the region during 1992–1993 (Mistretta et al. 1996). Arctomecon californica populations have declined since 1996 (Cornelisse and Tyler 2019), and the species' range has been reduced by half since 1994 surveys, estimated at 20,000–24,000 acres in 2000 (The Nature Conservancy 2007, Nevada Natural Heritage Program 2017). It is difficult to differentiate between disturbance effects on plant fitness or weather-driven fluctuations in recruitment and senescence as causes of population decline. For the closely related A. humilis, a major recruitment event in 1992 was the main support for population persistence over the subsequent 20 years, creating a demographic pattern that indicates population decline; long-term data sets are needed, though, to determine whether these recruitment events occur regularly at long intervals or whether they are anomalies that may not reliably support population persistence (Meyer et al. 2015). Apparent population reductions can result from observations of senescing plants; however, the seeds that persist in the soil in a dormant state awaiting appropriate conditions to emerge is another variable that characterizes plant abundance (The Nature Conservancy 2007). Indeed, these recruitment events for both A. californica and A. humilis depend upon a persistent seed bank (Meyer et al. 2015). With A. californica, many fragmented populations in the Las Vegas Valley were declared extirpated during 1996 and 1998 surveys, yet field checks conducted following above-average 2004–2005 winter rainfall found plants at 5 of the 10 populations presumed extirpated in 1998 (The Nature Conservancy 2007). However, because A. californica's seed dispersal distance is several orders of magnitude less than the distance between suitable habitat patches and because pollinator limitations in the Las Vegas Valley may reduce gene flow between populations, it is possible that these western populations may be extirpated with further loss of habitat (Meyer and Forbis 2006, The Nature Conservancy 2007). The eastern populations on protected lands are considered stable when accounting for climate pulse (above-average precipitation) and interpulse (below-average precipitation and warm seasonal temperatures) periods that drive recruitment and senescence, respectively (The Nature Conservancy 2007).
Population census may be best accomplished the year following a major recruitment event, so that maximum distribution is represented and more-complete patterns of decline and recruitment are expressed (Rominger and Meyer 2019). Population surveys using imagery gathered by UAV (unmanned aerial vehicle or drone) show promise for plant communities with sparse vegetation cover, such as those with Arctomecon humilis in Utah (Rominger and Meyer 2019, Rominger et al. 2021). This technique may be especially useful with species such as A. californica that grow on habitat susceptible to damage from vehicle and foot traffic, whether through the compaction of gypsum soils or the destruction of well-developed biological soil crusts (Rominger and Meyer 2019). For A. humilis, population density, size class distribution, and flowering proportion were determined with a single year of drone imagery by Rominger et al. (2021). Drone imagery and analysis were also combined with small-scale, on-the-ground sampling of reproductive output at each population to estimate seed production per plant and seed deposition over a certain area, which both have important ramifications for seed bank replenishment (Rominger et al. 2021). Inclusion of this type of monitoring for A. californica could be tremendously informative, especially if the monitoring is repeatedly done for the populations in the Las Vegas Valley, where habitat fragmentation and pollinator limitations may be reducing reproductive output and preventing full and/or adequate replenishment of the seed bank (Hickerson and Wolf 1998, Meyer and Forbis 2006).
Vegetation and Soil Associations
Arctomecon californica inhabits sparsely vegetated areas within communities dominated by creosote bush (Larrea tridentata) and salt-bush (Atriplex spp.), where physical crusts limit establishment of dominant vegetation (Meyer et al. 1992, Mistretta et al. 1996, Romão and Escudero 2005, Megill et al. 2011). Arctomecon californica is positively associated with shad-scale (Atriplex confertifolia) and negatively correlated with white bursage (Ambrosia dumosa), desert trumpet (Eriogonum inflatum), broom snakeweed (Gutierrezia sarothrae), and Mojave woodyaster (Xylorhiza tortifolia; Megill et al. 2011). Arctomecon californica shares its distribution with silverleaf sunray (Enceliopsis argophylla), and the 2 species are commonly surveyed together in and around Lake Mead National Recreation Area (Bangle et al. 2010).
Invasive annual species that are problematic in upland desert shrublands such as Sahara mustard (Brassica tournefortii), African mustard (Strigosella africana), red brome (Bromus madritensis ssp. rubens), and Mediterranean grass (Schismus barbatus) have been noted in A. californica habitat, usually along disturbance trails (The Nature Conservancy 2007, Bangle et al. 2010), but the direct impacts (competition) and indirect impacts (wildfire fuels) of these species on A. californica have not been quantified. Seed bank emergence trials have shown abundant seeds of invasive species in A. californica habitat, especially Mediterranean grass; although not represented in aboveground vegetation, disturbance events potentially allow these invasive plants to become established (Pereira 2019). African mustard was documented on gypsic soils at a greater frequency than on other soil types during 5 years of survey that also documented great fluctuations in annual precipitation (Abella et al. 2009), and the potential threat of competition with A. californica placed invasive plants as a high priority among research and management needs (The Nature Conservancy 2007). Invasive species may not be a widespread threat to A. californica where vegetation is typically sparse, but they may encroach where soils in fragmented populations are disturbed; thus, at a minimum, quantifying the presence/absence of invasive plants during population surveys could help trigger early intervention for removal.
Fine-scale soil mapping at the northern edge of Las Vegas Valley shows A. californica occurring on map units that are typical of spring deposits and basin floor sediments lacking a thick surface veneer of extremely gravelly alluvium, whereas map units with deep and gravelly soils do not support A. californica (Boettinger et al. 2010). Surface soils and 3-dimensional soil profiles that were sampled and analyzed indicate that A. californica inhabits finely disseminated calcium carbonate soils with petrocalcic horizons (carbonate-cemented layers), many carbonate nodules, and small (5%) to trace (<0.1%) amounts of gypsum (Boettinger et al. 2010). Arctomecon californica has a narrow distribution on the highly calcareous basin floor, as compared to the wider range of shrub vegetation on alluvial fans, the basin floor, and drainages (Boettinger et al. 2010). Geomorphic surface (e.g., spring deposit, basin floor) and fine-scale soil map unit were proposed as better indicators of A. californica habitat than surface soil chemistry (Boettinger et al. 2010) and have been included in habitat suitability models (see subsection “Habitat Suitability Models,” page 6).
Early studies considered A. californica either a gypsophile, a species restricted to soils with high gypsum content (36% to 69%; Meyer 1986, Sheldon 1994, Thompson and Smith 1997), or a gypsocline, a species primarily found on gypsum but also on other unusual or mixed soils such as soils with high boron or lithium (Meyer 1986). Thompson and Smith (1997) described soils in A. californica habitat as whitish in color, fluffy in texture, spongy and mineral-rich, forming raised crusts that are easily disturbed and having higher levels of sulfur, calcium, and soluble salts and lower levels of phosphorus than adjacent soils do. Despite the unique chemistry of gypsum soils, the effects of these elements are likely secondary to the physical structure and crust characteristics of the soil—such as biological crust cover, rock cover, and spongy texture with low bulk density—that are associated with A. californica (Meyer 1986, Thompson and Smith 1997). Indeed, true gypsophiles occur on soils only when the gypsum content is high enough to change the physical properties of the soil, emphasizing the potential importance of physical factors rather than chemistry (Meyer 1986).
More recent research confirms that A. californica can occur on a wide range of soils and parent material and is not a gypsum obligate (Childers 2004, Saxena 2005, Drohan and Merkler 2009, Boettinger et al. 2010). Gypsum soils are nutritionally impoverished due to the exchange of calcium and magnesium for other ions and are toxic to some plants because of high sulfate levels (Meyer et al. 1992, Palacio et al. 2007, Moore et al. 2014). Unlike specialist gypsophiles adapted to accumulate high foliar concentrations of sulfur, calcium, magnesium, nitrogen, phosphorus, and ash (Parsons 1976, Palacio et al. 2007), A. californica maintains moderate levels of foliar calcium and sulfur (Drohan and Merkler 2009; compared to 69 species that grow on gypsum soils—Merlo et al. 2019) and may avoid the accumulation of these elements through selective uptake (Duvigneaud and Denaeyer-de Smet 1968 cited in Moore et al. 2014, Moore et al. 2014). The calcium and sulfur levels in the leaves of A. californica suggest that it is an “indicator species” or one without a specialized physiological mechanism for coping with high levels of these select elements (Merlo et al. 2019). Further study exploring the leaf chemical compositions of A. californica growing in gypsum and non-gypsum soils, or a comparison between A. californica and known gypsophiles in the southwestern United States (e.g., Acleisanthes lanceolata, Tiquilia hispidissima, and Nama carnosa; Moore et al. 2014, Muller 2017), would be more definitive, especially as some sites with A. californica have a minimal difference in soil chemistry between occupied and neighboring unoccupied areas (Meyer 1986). In fact, soil surveys reveal that some A. californica populations have small (5%) to trace (<0.1%) amounts of gypsum in the subsoil and a total lack of gypsum in surface horizons (Boettinger et al. 2010). Furthermore, based on 2575 A. californica observations, 65.4% occur on gypsic soils, while the remaining 34.5% occur on calcid soils originating from limestone parent material with shallow calcium carbonate layers and accumulated high levels of calcium carbonate (Childers 2004).
Recent studies collectively recognize A. californica as a “gypsovag,” a species that can occur on soils with or without gypsum (Drohan and Merkler 2009, Boettinger et al. 2010). Although A. californica is not a gypsum obligate, it is possible that the species grows better or in greater numbers in habitat with a certain proportion of soil types (i.e., gypsic and calcid), edaphic conditions that may only be present in certain areas within the greater geographic extent of the species. Additionally, isolated populations of A. californica may represent ecotypes that tolerate different soil chemistries among other factors (Bangle et al. 2010). For example, 4 geographically isolated populations of A. californica in Grand Canyon exhibit unique morphology and genetics, potentially constituting an undescribed variant (Mistretta et al. 1996, Simpson 2014). Such soil chemistry–based differentiation has been found both for a widespread California aridland annual (Lasthenia californica) growing on serpentine outcrops (Rajakaruna 1998) and an edaphic endemic perennial herb (Lomatium cookii) growing in 2 isolated wetlands with different soils (Silvernail 2008).
Soils with low aggregate stability may be at risk for erosion and for surface crusting and sealing; exposed gypsum also creates hardened soil crusts when gypsum recrystallizes as water evaporates from the soil surface (Watson 1979). Communities of living organisms within the soil surface—known as cryptogamic, cryptobiotic, microbiotic, or biotic crusts—occupy some A. californica population sites in addition to physical crusts; however, it is difficult to separate the relative effects of biotic and physical crusts on plant establishment in these environments (Moore et al. 2014). Dense gypsum horizons and the penetration resistance of gypsum soils can also limit root growth and lead to sparse plant cover (Gibbens and Lenz 2001). Regionally, dominant shrubland species are typically excluded from gypsum soils when these soils are exposed on the surface, but they establish when a thin layer of alluvium covers the surface of the gypsum soils (Meyer 1986, Meyer et al. 1992). Thus, the physical nature of the surface crust of these soils limits shrub establishment and reduces plant density, leading to less resource competition for A. californica (Meyer 1986, Palacio et al. 2007, Megill et al. 2011).
Spring deposit material, petrocalcic horizons, and petrogypsic horizons are soil layers that can limit downward movement of water during the winter and spring, and these layers store moisture within the rooting zone of plants (Duniway et al. 2007). Petrocalcic horizons increase the near-surface soil water holding capacity, augmenting available soil moisture during drier periods (Herbel et al. 1972, Buxbaum and Vanderbilt 2007, Duniway et al. 2010). Winter precipitation that is absorbed by these carbonate-cemented horizons is stored deep enough that the subsoil moisture is not depleted by evaporation, but shallow enough for roots of many aridland shrub and perennial forb species to access (typically <50 cm deep; Duniway et al. 2010). Petrocalcic horizons were noted at approximately 30 cm depth for one A. californica population, with roots embedded in the petrocalcic layer (Winkel 2004). Petrogypsic horizons, cemented gypsum-derived layers that are structurally similar to petrocalcic horizons, could increase water availability on gypsum-rich soils in the same way (Casby-Horton et al. 2015). Usually these hardened horizons are described as root-limiting layers, with root mats observed on the upper surface; however, petrogypsic horizons broken apart showed presence of interior roots (Casby-Horton et al. 2015). Plants growing in gypsum-rich soils may additionally benefit from physical properties that allow water to move readily through the soil column, such as low water-holding capacity coupled with high hydraulic conductivity (Meyer and Garcia-Moya 1989, Meyer et al. 1992, Reading et al. 2012), although this dynamic has not been directly measured in A. californica habitat (Drohan and Merkler 2009). Certain plants may also be specially adapted to extract water from the crystalline structure of gypsum itself (Palacio et al. 2014). A gypsophile in Spain was able to obtain the majority (>70%) of its water during the summer from the breakdown of gypsum (i.e., calcium sulfate dihydrate) (Palacio et al. 2014). These 2 water molecules in the mineral's structure can be released through heating of the soil (initiated at 42–60 °C) or through manipulation of the soil chemistry (Palacio et al. 2014). Gypsum can rehydrate under cooler, moist conditions, allowing for continuation of the cycle (Palacio et al. 2014). Plants such as A. californica that are able to grow on gypsum soils may be specially adapted to tapping this water source during summer drought, although this adaptation has not been specifically studied for A. californica.
Diverse arbuscular mycorrhizal fungal communities are often associated with gypsum plant communities (Alguacil et al. 2009, Palacio et al. 2012, Moore et al. 2014), but species in Papaveraceae do not commonly form mycorrhizal associations, and their potential role for A. californica is unknown (Meikle et al. 2006). Arbuscular mycorrhizal fungal associations occur for both gypsophile and gypsovag species, and these associations promote tolerance of the nutrient and soil structure deficiencies of gypsum soils (Alguacil et al. 2009, Palacio et al. 2012) by increasing nutrient uptake, improving plant tolerance, and contributing to soil structure and quality (Wang 2017). The direct role that mycorrhizal colonization may play in supporting the persistence of gypsovags on gypsum soils is unclear because studies have analyzed colonization solely on gypsum soils, without comparison to non-gypsum soils (Palacio et al. 2012). Mycorrhizal vesicles, arbuscules, and internal fungal hyphae were not detected in the roots of 2 A. californica plants examined for endomycorrhizal colonization (Meikle et al. 2006), but this sample size is too small to definitively confirm or refute arbuscular mycorrhizal fungal association for the species. Establishment of A. californica seedlings does seem to be enabled by the presence of cryptogamic crusts (Bailey 2019), although the mechanism is unclear in light of preliminary data showing more experimental units of dead crusts supporting seedlings than live crusts did (6 on dead crusts vs. 1 on live crust, and 3 on no crust). Improved establishment may be largely a result of crust structure and nutrient inputs or moisture retention, but mycorrhizae can also be influenced by the presence of cryptogamic crusts, and the relationship remains unclear (Harper and Pendleton 1993, Pendleton et al. 2003).
Habitat Suitability Models
Quantitative models of species distribution and associated habitat variables can expand searches for new populations and identify which habitats could be prioritized for protection and management actions, including restoration and reintroduction. Modeling efforts for A. californica have been challenging, due in part to the coarse resolution of available geologic maps. A model for A. californica in the Gold Butte region identified suitable habitat as being close to the gypsiferous topsoil layer (66.6% variable contribution), at low elevation (20.0%), with specific surficial geology (10.8%), and having dominant bedrock material (2.6%). However, the authors note that this model potentially overpredicts suitable habitat because it did not account for habitat loss associated with human activities (Nussear et al. 2010). A model expanded to include Clark County was based on soil-based parameters (Natural Resources Conservation Service, Soil Survey Geographic Database data), an elevation constraint of 300–1120 m, and Advanced Spaceborne Thermal Emission and Reflection (ASTER) satellite imagery (Hamilton and Kokos 2011). Addition of climate variables (BioClim) to this county-wide model did not refine the soils-based model more than using elevation in the soil-based model, likely because climate is closely tied to elevation at a fine scale (Hamilton and Kokos 2011). The model with climate showed that A. californica populations tended toward warmer temperatures with a lower range of diurnal temperatures often associated with lower elevations and low isothermality, a measure of day-to-night temperature oscillations relative to annual oscillations (Hamilton and Kokos 2011, Southwest Ecology LLC 2018). There are several additional issues that these recent models highlighted. For example, the presence of gypsum is largely based on the bedrock geology of an area, and surficial geology maps are not helpful in mapping gypsiferous regions. Using different geological maps can also make refining models difficult, as geologic unit types might not be uniform or might be outdated, confusing bedrock lithology with soil type, geologic unit names, and/or landform type (Robins et al. 2014). Arctomecon californica can also occur in patches that are much smaller than the resolution of geologic units, in patches that are not detectable by satellite imagery, or on map units that are not mapped as gypsiferous but contain gypsum due to erosion from neighboring units, further complicating modeling (Hamilton and Kokos 2011).
Seed and Seedling Ecology
Pollination and Seed Set
Arctomecon californica is largely self-incompatible, requiring pollen transfer between different individuals—not between flowers within an individual—for fertilization (Sheldon 1994, Mistretta et al. 1996, Thompson and Smith 1997). However, self-compatibility can vary both within and between populations of a species (Faegri and van der Pijl 1979). For example, the closely related A. humilis was previously thought to be self-incompatible (Harper et al. 2001), but a subsequent study found that the breeding system was mixed for one population, with 60% self-compatible and 40% self-incompatible plants, although this ratio of mixed self-compatibility may not apply to all A. humilis populations (Tepedino et al. 2014). Studies focusing on self-compatibility for different populations of A. californica will elucidate the importance of pollinators to species conservation and could reveal variation in A. californica's breeding system, with possible self-compatibility in areas that are heavily fragmented or isolated. Modeling indicates that the advantage of self-compatibility is greatest when site occupancy rates are low, such as along the margins of a species' range (Pannell and Barrett 1998). For one restricted annual herb (Leavenworthia alabamica), populations at the center of its range are self-incompatible, while peripheral, isolated populations display adaptations for self-fertilization, enabling reproduction despite pollinator limitations (Busch 2005).
Reproductive failure in A. californica is primarily the result of tissue abortion of buds and capsules rather than insect herbivory (Thompson and Smith 1997). High attrition may occur when buds develop late in the flowering season and abort at the onset of environmental stress during spring (Thompson and Smith 1997). However, A. californica's reproductive attrition is low compared to other Mojave Desert perennials such as Larrea tridentata (Boyd and Brum 1983). Thus, low seed set may not play a primary role in the rare, restricted status of A. californica even though it may affect total fitness (Thompson and Smith 1997).
Upon ripening and prior to dispersal, seeds of species in the Papaveraceae family commonly display morphophysiological dormancy (Baskin and Baskin 2014), characterized by a rudimentary or linear embryo combined with physiological inhibiting mechanisms (Pereira et al. 2021). Arctomecon californica seeds fit this type of dormancy, with underdeveloped linear embryos and germination occurring only in response to specific incubation regimens (Meikle et al. 2006).
Arctomecon californica seed dispersal is facilitated by more than one vector, a mode of dispersal known as diplochory. While a small proportion of seeds fall and are trapped within the basal rosette, wind disperses A. californica seeds from the capsules to the ground within 1 m of the parent during primary dispersal, and ants subsequently collect and disperse seeds across the landscape during secondary dispersal (Megill 2007). A small proportion of seeds disperse farther from the parent plant, and seeds have been found up to 8 m away from the nearest adult plant (Megill 2007). Seed deposition in a northerly direction from the parent plant is in accordance with seasonal wind patterns after seed set (Sheldon 1994, Megill 2007).
Secondary dispersal of A. californica seeds by ants (known as myrmecochory) has been documented through seed dispersal trials and direct observations (Megill 2007). Over 99% of 23,253 intact A. californica seeds that were examined had elaiosomes, or white, fatty structures along the seed's hilum (attachment scar) that facilitate seed dispersal by ants (Fig. 1; Megill 2007). Ants transport seeds to their nests and feed the lipid-rich elaiosomes to their larvae, afterward discarding the otherwise intact seeds on the soil surface. This dispersal of seeds away from the parent plant can help seedlings avoid direct competition with adults and deter seed predators foraging close to adult plants (Giladi 2006). Intact A. californica seeds without elaiosomes that have been discarded on the surface of nests suggest that ants transport the seeds into their nests and then discard intact seeds after the elaiosomes are removed (Megill 2007). Heteromyid rodents, in contrast, generally prefer seeds larger than the average seed available in the soil (Brown et al. 1979, Price 1983), and the size of A. californica seeds (2 mm) may not allow the seed to be calorically rich enough to provide a foraging reward (Reichman 1979). Alternately, the seeds' elaiosomes may act as a rodent repellent, as in the case of Corydalis aurea (Hanzawa et al. 1985, Megill 2007), but additional studies are needed to confirm a potential role of rodent granivory or dispersal for A. californica. Seeds of the closely related A. humilis also have elaiosomes and are removed by both ants and rodents, although it is not known how each impacts seed germination (Farrall and Mull 2010).
Several studies have shown that, following dispersal, dormant seeds of Arctomecon californica accumulate over time in the soil seed reserves1 awaiting conditions for germination. Far fewer A. californica seeds have been detected in studies that use emergence methods (i.e., surface soils collected from habitat are repeatedly watered in a greenhouse until germination ceases; Abella et al. 2013, Pereira 2019) than in studies that use extraction methods (i.e., soils are sieved in a laboratory to separate seeds from soil; Sheldon 1994, Megill 2007, Megill et al. 2011, Pereira 2019). This difference occurs in part because emergence methods trigger germination of live seeds by simulating environmental cues—such as through repeated wetting and drying, or mechanical or chemical treatments—that shift seeds from a dormant state to a germinable state (Baskin and Baskin 2014). Extraction methods determine total seeds, including those that may be nonviable (i.e., unfilled, decomposing, or old seed); therefore, it is important that extracted seeds undergo a method such as staining with tetrazolium dye followed by dissection to determine viable fraction (Pereira 2019). One tetrazolium-staining protocol of A. californica seeds has produced reliable results: the radicle end of the seed is clipped off to allow dye to penetrate, the living tissue is stained, and the seed is then dissected longitudinally and analyzed. Uniform staining of the embryo signals viability, even if the staining is pink and not bright red as is typically expected (S.E. Meyer, USDA, Rocky Mountain Research Station Shrub Sciences Laboratory, personal communication). Early emergence studies for A. californica recorded minimal numbers in seed reserves (Abella et al. 2013, Pereira 2019), but a subsequent study confirmed that A. californica requires one or more seasons of chilling to stimulate germination (Pereira et al. 2021). Determining the viable seed reserves is important for understanding population viability and guiding potential seed reserve salvage or supplementation efforts. Thus, adding cold stratification to the emergence method may distinguish viable from nonviable seed fractions, but sorting the remaining soils using a combination of extraction and viability tests (e.g., poke test, dissection, or tetrazolium; Baskin and Baskin 2014) ensures that all viable seeds are accounted for.
Despite the population-wide attrition of adult plants that occurs during periods of drought (The Nature Conservancy 2007), A. californica relies on long-lived seed reserves for population persistence (Mistretta et al. 1996, Megill et al. 2011). Although primary dispersal creates a clustered distribution of seeds around parent plants, soil seed reserves may be more uniformly distributed across habitat. In one study, total seeds extracted from soils did not decline in number out to 150 cm from adult plants (Winkel 2004), although nonviable seeds were not differentiated from viable seeds, which tend to be more clustered around adult plants (Megill et al. 2011). High seed output (with some portion remaining germinable for up to 19 years based on seed burial trials; de Queiroz and Meyer 2022), combined with persistence of viable seed in the soil (Megill 2007), allows A. californica to accumulate large seed reserves in the soil (Harper and Van Buren 2004). When plant populations are sparse, broad outcrossing due to greater pollinator foraging distances among plants is known to produce genetically diverse seed in other species because distant plants are more likely to be unrelated (Ellstrand et al. 1978). In this way, A. californica, which has large interannual fluctuations in population densities, may potentially retain high genetic diversity in seed reserves despite small population sizes (Harper and Van Buren 2004). Population persistence depends on adequate seed production to replenish sizable seed reserves (Phillips and Phillips 1988 cited in The Nature Conservancy 2007). However, a population viability analysis of A. californica indicated that small, fragmented populations like those in Las Vegas Valley have low seed set due to pollinator limitations and may persist largely because of their long-developed seed reserves (Meyer and Forbis 2006). These populations would become extirpated if their seed reserves become depleted without periodic inputs during years with successful seed set.
Arctomecon californica has been found at sites after 15 years of absence (Meyer 1996 and Powell and Walker 2003, both cited in Megill 2007), which may reflect dormant seeds responding to temperature and precipitation patterns that favor seedling emergence or seeds deep in the soil being brought to the surface through bioturbation to germinate (e.g., rodent digging, ant nest construction). Megill (2007) found A. californica's seed reserves to be heterogeneously distributed, with 34% of seed reserves at 0–2 cm, 22% at 2–4 cm, 18% at 4–6 cm, and 26% at 6–15 cm depth for 2973 seeds recovered across 5 sites. Approximately 10.5% of the seeds recovered were determined to be viable in a tetrazolium stain test, with no significant difference between soil depths (Megill 2007). In addition to the presence of viable seeds in the soil profile, A. californica displays other characteristics indicating persistent seed reserves (Christoffoleti and Caetano 1998, Megill 2007), including long-lived seeds (de Queiroz and Meyer 2022) and delayed germination of fresh seeds containing immature embryos. At some sites, seeds may travel deep into the soil profile due to their small size relative to the soil particle sizes; typically seeds >7 cm below the surface are safe from most foraging granivores (Reichman 1979). Most seedlings of desert species cannot emerge from greater than a 2-cm depth (Bond et al. 1999, Cabin and Marshall 2000), so these deeply buried seeds would produce seedlings only if bioturbation brought them nearer to the surface (Megill 2007).
Emergence of A. californica seedlings has been noted during the winter or early spring months after unusually high winter precipitation (Meyer 1987 cited in Pereira 2019, Meyer and Forbis 2006). Based on these observations, Meyer (1987) surmised that winter rainfall sufficient for seedling establishment may occur as infrequently as once every 9 years (Mistretta et al. 1996), while the exact temperatures and the amount and timing of favorable rainfall events remain unknown. Seedlings tend to establish near adult plants at sites where A. californica is dense, a pattern that diminishes when disturbance limits adult plant abundance (Megill et al. 2011). High seedling mortality, typical of desert perennials, ranged between 60% and 87% for 3 sites around Lake Mead (Thompson and Smith 1997). Individual plants usually live 4–5 years and can have multiple flowering events over the plant's lifetime, beginning in their second season (Thompson and Smith 1997). Individuals develop multiple rosettes as they grow older (Fig. 2), which can increase the plant's reproductive output (Meyer 1979 cited in Mistretta et al. 1996, Thompson and Smith 1997).
One recent study suggests that cryptogamic crusts may promote A. californica seedling establishment, although the mechanism and field dynamics of this relationship have only recently come under investigation (Bailey 2019). Arctomecon humilis, a closely related species with similar growth requirements, occurs on sites where biocrust cover contributes 84% or more of the total living cover, and seeds are captured within the cryptogamic crusts (Nelson and Harper 1991, Nelson and Welsh 1993). Arctomecon californica habitat has also been described as having a heavy cryptogamic cover (Mistretta et al. 1996, Thompson and Smith 1997, The Nature Conservancy 2007). However, some population sites support minimal biocrust cover (Megill et al. 2011), and studies have not found a direct relationship between cryptogamic cover and A. californica distribution (Bangle et al. 2010, Megill et al. 2011). Arctomecon californica plants have even been observed growing in disturbed spots (hoof prints) within heavily crusted areas (Bangle et al. 2010), on topsoil piles awaiting reclamation (Cayenne Engel, Nevada Division of Forestry; Lara Kobelt, BLM, Southern Nevada District Office, personal communications), and at sites previously inundated by Lake Mead (Engel et al. 2014). Arctomecon californica seedlings also established sporadically during 2003–2006 along an unimproved road at Lake Mead National Recreation Area along sections where deep soil compaction was alleviated with heavy machinery during late fall 2002 (U.S. Geological Survey unpublished data).
Careful a priori selection of weather variables for correlating important life stages (seedling, prereproductive juvenile, adult) and phenological stages (onset of flowering, fruit maturity, plant senescence) is essential for discriminating among the environmental triggers under our current climate and for forecasting future demographic and population trends. Seasonal temperatures and precipitation, variation in timing of precipitation, and combinations of current-year and previous-year temperature and precipitation patterns influence reproduction and fruit production, phenology, and complex relationships between plant species and their pollinators for other aridland species (Meyer and Pendleton 2015, St. Clair and Hoines 2018). Although preliminary, an analysis of A. californica numbers as influenced by site-specific weather variables revealed that maximum humidity in May, minimum humidity in November, and high July rainfall were correlated with greater numbers of A. californica (Bangle et al. 2010). The ecological mechanism explaining how humidity promotes recruitment or survival is not well understood, but it may represent recent precipitation and evaporation levels, with low humidity in November playing a role in seed stratification or reducing damping off of young seedlings, for instance. Multiple stations that record daily maximum and minimum temperatures and daily rainfall across A. californica populations will discriminate between limitations driven by climate (e.g., low reproductive output, recruitment events, seedling and adult survival) and limitations driven by land use. Local long-term weather stations are also pivotal to validating the use of larger-scale climate data such as Oregon State University's PRISM Climate Group, which can justify the extrapolation of spatial weather data across the species' range.
Propagation from seed has held the most promise for developing nursery stock for habitat augmentation or reintroduction. Arctomecon californica seeds require an extended period of cold-dry or cold-wet conditions (known as stratification) to induce germination (Meikle et al. 2006, Pereira 2019, Pereira et al. 2021). This stratification requirement for A. californica germination resembles intermediate complex morphophysiological dormancy where development of the embryo is not complete at the time seeds disperse and embryo growth begins under cold stratification between 0–10 °C, regardless of when the seeds mature in nature (Baskin and Baskin 2014). One recent germination trial with A. californica seeds encompassed cold-dry storage at 4 °C for 3.5 months, alternating moderate-wet incubation at 23 °C /13 °C and cold-wet stratification at 4 °C for 5.5 months, and a final cold-wet stratification for another 4 months, resulting in 4.9% germination over the first 9 months and 18.7% germination during the final 4 months (Pereira et al. 2021). An unpublished report from Meyer (1996) used a similar protocol with comparable results (Pereira 2019). Increased germination during the final cold stratification indicates that either one long or multiple short cold stratification events are needed to produce high germination rates, but the number and duration of these chilling events is unclear (Pereira et al. 2021). Field-stratified seeds of A. californica had much lower germination rates (0.4% ± 0.8%), demonstrating that year-to-year cold periods across A. californica habitats may not always be cold enough, long enough, or frequent enough to promote germination; soil moisture may also be inadequate for seeds to imbibe water during cold events during many years (Meyer 1987 cited in Mistretta et al. 1996, Pereira et al. 2021).
For species with intermediate complex morphophysiological dormancy, gibberellic acid is used to shorten the period of cold stratification (Baskin and Baskin 2014). Arctomecon californica seeds exposed to 100 ppm and 1000 ppm gibberellic acid produced radicles but no shoots, showing that some level of cold stratification remains necessary for embryo development (Meikle et al. 2006). One trial with gibberellic acid achieved 26% germination with cold-moist stratification at 2–4 °C for 40 days, followed by treatment with 100 ppm gibberellic acid and incubation at 2–4 °C for 10 days, and then incubation at 8 °C/18 °C until germination occurred after about 10 days (Meikle et al. 2006). However, application of high levels of gibberellic acid to seeds can cause abnormal growth and low vigor of seedlings; for example, seeds of Phacelia argillacea treated with gibberellic acid produced stunted, misshapen plants with reduced seed set (Meyer 2011). For some species with this seed dormancy type, dry storage or warm stratification prior to cold stratification can also shorten the period of cold stratification needed (Baskin and Baskin 2014), but this has not been tested for A. californica.
Arctomecon californica seeds are covered by a waxy cuticle that is water permeable and does not inhibit water imbibition, although it may delay embryo development and germination by reducing levels of gas and water exchange to some degree (Meikle et al. 2006, Pereira et al. 2021). Arctomecon californica seeds treated with sulfuric acid and subjected to cold treatment germinated more rapidly than seeds exposed to cold treatment alone (Pereira et al. 2021). Arctomecon californica seeds placed in an incubator for 10 minutes at 45 °C with subsequent scraping away of the waxy cuticle had significantly higher germination compared to intact seeds, with no effect on embryo growth (Meikle et al. 2006). Elevated germination of A. californica seeds exposed to acid or physical removal of the cuticle suggests that scarification—the mechanical or chemical weakening of the seed coat to promote the processes that lead to germination—could potentially improve seedling establishment in a restoration context.
Raising plants germinated from the soil seed reserves is a potential strategy for propagating rare desert plants, but it has not been formally tested for Arctomecon californica. Although the emergence method is not suited for determining total numbers of seeds (viable + dead) in the seed reserves, this method has successfully identified the viable portion of seed reserves across a wide diversity of native and invasive plant taxa in desert shrublands (DeFalco et al. 2009, Scoles-Sciulla and DeFalco 2009, Esque et al. 2010, Jurand and Abella 2013). This method, adapted to Mojave Desert shrublands, promotes the germination of viable seeds from field-collected surface soils through repeated wetting and drying cycles in a greenhouse in combination with dormancy-breaking chemicals such as potassium nitrate and gibberellic acid (Esque 2004). By germinating seeds within the soils where A. californica grows, an emergence approach may improve seedling establishment over direct propagation from seed using nursery soil mixtures. An emergence method, including a cold stratification treatment, is currently being tested in this context for the rare Mojave Desert species Penstemon albomarginatus, Eriogonum viscidulum, and Astragalus geyeri var. triquetrus (Clark County Desert Conservation Program project #2019-USGS-190A).
The cryptogamic crusts that are often found in Arctomecon californica habitat could play a role in seedling establishment (Bailey 2019), although the mechanism may be more complex than originally thought. In a growth chamber, seeds that germinated without biocrust did not develop into seedlings, while those that germinated with either live or dead biocrust reached seedling stage; the presence of soil from the rhizosphere and the location of the seed above or below the crust did not influence initial germination (Bailey 2019). Subsequently, seeds in water-permeable pouches were buried between soil and biocrust in the field in October and retrieved the following March, resulting in seedlings that developed true leaves (Lydia Bailey, Northern Arizona University, personal communication). When these seedlings were moved from the growth chamber to the greenhouse, they died within 5 months, indicating that suitable greenhouse conditions have not yet been determined for A. californica seedlings (Lydia Bailey personal communication).
We are not aware of trials on stem or root cuttings for propagating A. californica plants. Whereas Mojave Desert perennials have been successfully propagated in a nursery setting using stem cuttings (Everett et al. 1978, Richardson et al. 1979), most species were woody shrubs. Unsurprisingly, propagation of herbaceous perennials is less common for desert species, particularly for herbs that lack organs from which they naturally reproduce asexually, such as rhizomes, stolons, or bulbils. Though challenging, vegetative propagation from cuttings of A. californica could be a useful tool in restoration because developing mature A. californica plants from seed is difficult (Lydia Bailey personal communication). For the rare Rocky Mountain medicinal herb Ligusticum porteri (Apiaceae), which arises from a single taproot, crown cuttings collected during early spring were more successful than root cuttings and provided better plant development than did propagation from seed (Panter et al. 2004). Arctomecon californica emerges as a basal rosette of leaves from which a single inflorescence arises, although typically not during its first year (Fig. 3), and multiple rosettes accumulate as plants age (Sheldon 1994). Propagation from an excised basal rosette, potentially using growth hormones that are known to establish roots in other herbaceous plants (Mladenović et al. 2016), shows some promise but has not been tried for A. californica to our knowledge. Rosette propagation could be done to limit injury and death of the donor plants, such as by collecting rosettes after above-average winter precipitation or watering donor plants prior to collection in order to augment water status, and by collecting in winter or early spring before belowground resources are mobilized for reproduction.
Micropropagation techniques that use small parts of plant tissues grown on nutrient agar to produce new plants are becoming popular for mass production of plants that are difficult to grow using conventional horticultural methods. However, most micropropagation efforts seek to produce medicinal and ornamental plants, food for humans and livestock, and lumber (Kant et al. 2010, Chokheli et al. 2020). Little research has explored micropropagation for augmenting declining populations by reintroducing rare plants.
Options for reintroducing Arctomecon californica plants into suitable habitat include outplanting nursery-raised seedling stock, salvaging seed reserves from topsoil, sowing or broadcasting seeds (either with no pretreatment or pretreatment with cold stratification or scarification), or transplanting established plants within suitable habitat (also known as plant salvage).
Outplanting of Nursery Stock
Raising Arctomecon californica for outplanting is the most resource-intensive approach to reintroduction due to preplanting costs (labor and nursery costs for a 6- to 12-month commitment to grow robust seedlings) and postplanting effort (regular supplemental irrigation during establishment and herbivore protection, although leaf alkaloids in Arctomecon species are known to deter rodents and lagomorphs—Raynie et al. 1990, 1991). This approach is also the most challenging due to the difficulties of overcoming seed dormancy, reducing seedling mortality due to fungal infection (i.e., damping off), and promoting healthy growth and survival through to outplanting (Meikle et al. 2006). We were unable to find examples where outplanting has been attempted for A. californica specifically.
Salvage of soil seed reserves by removing surface soils prior to disturbance is also resource intensive, as it requires heavy machinery and space for storing topsoil piles, but it has been carried out as a mitigation measure for large-scale activities such as urban development and power pole placement. This mitigation method has not always been successful. At the Springs Preserve in Las Vegas, 2 of 5 known A. californica populations (each about 800 ft2) had no aboveground living plants for over 3 years and were under the proposed footprint of major construction. In 2001, the top 2 inches of soil (with seed reserves) was set aside and 4 inches of the subsoil was salvaged and placed next to a third population. The topsoil was then spread on top of the subsoil. No A. californica plants were observed from the salvaged soil (Von Winkel, Southern Nevada Water Authority, personal communication). Two additional soil salvage efforts also resulted in failed A. californica emergence: (1) 40 cubic yards of gypsum soil obtained in 2005 from Nellis Air Force Base to enlarge a population at Springs Preserve and (2) a similar volume of soil obtained in 2006 from a stockpile at the Nevada Division of Forestry collected near Tule Springs (Von Winkel personal communication). In contrast, during a recent salvage effort in A. californica habitat, topsoil was removed in December 2018 and replaced after completion of a transmission line in the Rainbow Gardens region near Pabco Mine. After late rains during winter 2019–2020, roughly 80 seedlings were found growing over approximately 150 linear feet of these topsoil piles, although the majority had died by late 2020 following below-average rainfall (Cayenne Engel, Nevada Division of Forestry; and Lara Kobelt, BLM, Southern Nevada District Office, personal communications). Topsoil salvage using heavy machinery can significantly dilute Mojave Desert seed reserves when the surface soil layer is inadvertently mixed with deeper soil fractions lacking seed reserves (Scoles-Sciulla and DeFalco 2009), but when collected for nursery use, soils may also improve establishment of salvaged or greenhouse-grown plants. Alternatively, surface soils collected near existing plants by use of shovels should capture the majority of seeds (Sheldon 1994), but this method would be practical only on a small scale.
Direct seeding is the least resource-intensive approach to reintroducing Arctomecon californica, but little research exists on this method. Anecdotally, A. californica was believed to be successfully introduced into an area in Utah after seeding by a local landowner (Mistretta et al. 1996). Most recently, A. californica plants successfully established at Rainbow Gardens and Gold Butte sites seeded in spring 2018 and 2019 as part of an ongoing restoration study (Lydia Bailey personal commuication). An additional seeding trial in combination with biocrust treatments was carried out at Rainbow Gardens in fall 2020, with results yet to be reported (Lydia Bailey personal communication). Direct seeding trials that evaluate multiple treatments—including tests of seed pretreatments, granivore protection, and supplemental irrigation—and are revisited over multiple years to account for the influence of field conditions on germination and seedling establishment will strengthen our understanding of this reintroduction strategy in comparison to others.
The transplanting of salvaged Arctomecon californica plants is intermediate in resource intensity between outplanting and seeding, and this method has received the most attention as a viable strategy. The largest challenge in transplanting A. californica is preserving the single fragile taproot that can reach the top of the petrocalcic layer, with many of these taproots growing into or through this layer (Winkel 2004). The greatest depth that taproots can grow into the petrocalcic layer is unknown, but penetration may be enabled by wet conditions (Winkel 2004, Casby-Horton et al. 2015). Because the taproot grows directly under the rosettes, excavating a column of soil that is slightly larger than the aboveground portion of the plant is sufficient for salvaging A. californica, but taproot depth cannot be predicted by aboveground plant size (Winkel 2004). For example, a plant with 2 rosettes had a taproot over 33 cm in length (the root extended deeper but broke off at the petrocalcic layer), while another plant with 31 rosettes had a 26 cm taproot (Winkel 2004).
In the late 1990s and early 2000s, the Las Vegas Valley Water District and Southern Nevada Water Authority made several attempts to salvage and transplant Arctomecon californica either directly into habitat or into pots using methods meant to avoid injuring the taproot (Winkel 2004). Shallow collection down to 30 cm depth was not successful for adult plants, while some adults collected with soil to 60 cm depth survived for almost 2 years in pots, and the majority of living plants also flowered and produced seeds in spring with hand pollination (Winkel 2004). The ability of A. californica to survive in pots for many months and produce seed is useful in cases when plants cannot be immediately transplanted (Winkel 2004) and could provide a means for seed increase. Two subsequent salvage trials of seedlings were conducted with 30-cm-depth soil collection in December when the plants were dormant or in late March when plants were actively growing; seedlings were transplanted into soil salvaged from the collection site and irrigated. While 80% of the December-collected seedlings survived for 19 months after transplanting and produced seed, 73% of the plants salvaged in March died within 5 weeks of transplant (Winkel 2004). The late-March collection occurred right before A. californica plants begin to produce new rosettes and flowers, and transplantation and irrigation may have negatively impacted the physiology of the plants (Winkel 2004). Seedlings with shallow taproots were easier to salvage than the larger adult plants were, and they have the potential of living longer after transplanting, suggesting that salvage efforts should focus primarily on seedlings (Winkel 2004). Most salvaged A. californica plants that reproduced died shortly after seed production in all trials, likely because they diverted a large amount of energy to reproduction (Winkel 2004).
In December 2019, sixteen Arctomecon californica seedlings were salvaged and relocated to the Lima gypsum mine reclamation area, using a stovepipe to collect a 30-cm-deep column of soil around the taproot (Lydia Bailey personal communication). Unlike previous salvage efforts, the plants were not watered after initial transplant, but 11 of the plants flowered and set seed in spring 2020, and 7 remained alive after 10 months (Lydia Bailey personal communication). Continued monitoring of current and future salvage attempts is needed, but the rapid die-off of plants during most past efforts suggests that there may be additional factors influencing survival that have not been addressed. For example, an understanding of A. californica's mycorrhizal associations could inform potential research into mycorrhizal inoculation of disturbed or salvaged soil to help the plants weather transplant shock.
Seed production ensures that populations of A. californica are resilient to environmental changes yet ultimately depends on healthy populations of pollinators. Arctomecon californica is most commonly pollinated by the rare Mojave poppy bee (Perdita meconis), Encelia megandrena (Megandrena enceliae), and European honey bee (Apis mellifera; Hickerson 1998 cited in Portman et al. 2019, Portman et al. 2019). During surveys in 2017, A. mellifera only visited A. californica flowers at sites with 45 or more plants in bloom and commonly visited alternate floral hosts, gathering pollen from just one or a few A. californica flowers per visit (Portman et al. 2019). However, the pollinator efficiency of A. mellifera increased when aggressive interactions with male P. meconis forced A. mellifera to visit multiple A. californica plants (Portman et al. 2019). The solitary M. enceliae was historically regarded as a specialist on the widespread creosote bush (Larrea tridentata), but surveys in 1995 and 2017 showed that this pollinator is an effective cross-pollinator for A. californica due to its large size and its tendency to visit flowers on multiple plants during a single foraging trip (Hickerson 1998 cited in Portman et al. 2019, Portman et al. 2019). Perdita meconis, despite its small size, is also an effective pollinator of A. californica and was the dominant flower visitor at many recently surveyed sites (Portman et al. 2019). This species is a solitary ground-nesting bee, is a strict specialist on Arctomecon and Argemone species, and is under review for listing by the U.S. Fish and Wildlife Service (USFWS 2019).
Many desert bee species, especially foraging specialists, follow the same environmental cues as desert annual plants for emergence (Danforth 1999). The closely related Macrotera portalis (formerly Perdita portalis), a specialist on plants in the genus Sphaeralcea, follows a bet-hedging emergence pattern where only a portion of the larvae pupate under optimal conditions. Pupae can remain in diapause for multiple years, emerging upon exposure to high humidity (rainfall), which helps synchronize bee emergence and host-plant germination (Danforth 1999). It is possible that this delayed emergence and induced emergence triggered by rainfall occur in many species of specialist desert bees across a range of bee families (Danforth 1999).
Pollinator Foraging in A. californica Habitat
Plant population size and density both play a vital role in the reproductive health of Arctomecon californica. For A. humilis, greater distance between plants is correlated with lower pollination rates and seed fill, especially for plants with fewer than 10 flowers: i.e., <20% seed fill for plants ≥10 m apart, compared with 51% seed fill for plants separated by ≤3 m (Harper et al. 2001). Arctomecon californica flowers produce more seed if they are available to pollinators for the full 2 days during which stigmas are receptive (Tepedino and Hickerson 1996 cited in Bangle et al. 2010), suggesting that maximum fertilization may require multiple visits by pollinators, as is the case for A. humilis (Harper et al. 2001). Pollen deposited from multiple outcrosses raises competition in reproduction and can positively impact the vigor and genetics of the offspring (Björkman 1995, Mitchell 1997). Arctomecon humilis and A. californica are most efficiently pollinated by similar species of native solitary bees: declines in plant population density will have more impact on specialist pollinators that forage near their nests compared with those that are generalists and/or regularly travel great distances between nesting sites and plant populations (Harper et al. 2001). While the foraging range of pollinators for A. californica is not directly known, the foraging distance of bee species is generally a function of body size, with smaller bees traveling disproportionately shorter distances than larger bees (Greenleaf et al. 2007). Most solitary bees require nest and host plant sites to overlap within 100 m of each other (Zurbuchen et al. 2010), but this distance may be an overestimate for P. meconis, with its small size and strict specialization (Cornelisse 2018). The female of a similar desert specialist, P. coreopsidis, forages <5 m from its nest (Danforth 1989).
Pollinator Trends and Habitat Fragmentation
Arctomecon californica pollinator diversity and densities were relatively unchanged in 2017 compared with previous surveys conducted in 1995 on Lake Mead National Recreation Area and Bureau of Land Management lands (Hickerson 1998, Portman et al. 2019). Populations of A. californica in the western Las Vegas Valley, on the other hand, have suffered fragmentation and habitat loss, decreasing connectivity concurrent with reduced pollinator numbers and diversity (Mistretta et al. 1996, Hickerson and Wolf 1998). Fragmented A. californica populations in the Las Vegas Valley are becoming more isolated and differentiated from one another as insect pollination and corresponding levels of gene flow decline within and between populations (Hickerson and Wolf 1998). This shift reflects a decline in specialist pollinators and reliance on less effective generalist pollinators (Tepedino and Hickerson 1996 cited in The Nature Conservancy 2007, Hickerson and Wolf 1998). Therefore, plants at fragmented sites have 2 to 3 times lower seed set compared to plants in unfragmented habitat at Lake Mead National Recreation Area (Tepedino and Hickerson 1996 cited in The Nature Conservancy 2007). Follow-up pollinator surveys have not been conducted in these more fragmented Las Vegas Valley populations since 1995, yet the western portion of A. californica habitat may be comparable to A. humilis habitat in Utah around the St. George area where pollinator populations have significantly changed (Portman et al. 2019). Arctomecon humilis pollinators have become scarce, and diversity has declined over the last 20 years in southern Utah; specialized pollinators are declining in numbers and generalist pollinators are increasing (Tepedino et al. 2014). If trends continue, pollinator abundances in the region may increasingly represent the genus Lasioglossum, a prominent genus of bees that are infamously variable and unreliable in their pollinating habits (Tepedino et al. 2014).
In comparison to heavily developed land surrounding populations within the Las Vegas Valley, the less developed Lake Mead National Recreation Area and Bureau of Land Management lands surrounding the eastern portion of A. californica habitat promote pollinator habitat (Portman et al. 2019). Arctomecon californica does not produce nectar, so pollinators commonly visit other genera, including Stanleya, Enceliopsis, and Larrea (Tepedino et al. 1997, Griswold et al. 2006), making these plant associates important for pollinator persistence. Perdita meconis has been extirpated from southern Utah following Africanized honey bee invasion in 2000 (Tripodi et al. 2019), while Africanized honey bee invasion in 1998 had much less impact on populations of Perdita in southern Nevada. When Africanized honey bees arrived in Utah, alternative bee forage was low, driving Africanized honey bees to A. humilis flowers and driving away P. meconis (Portman et al. 2019). In comparison to A. humilis sites in Utah, A. californica sites have much higher floral diversity, partially due to Mojave desert tortoise (Gopherus agassizii) conservation efforts, which led to the retirement of most grazing allotments on lands at Lake Mead National Recreation Area and on lands managed by the Bureau of Land Management in Clark County (Portman et al. 2019). Portman et al. (2019) observed that many alternative flowering plants at A. humilis Utah sites (e.g., Eriogonum and Sphaeralcea) were heavily grazed, while other plants that are pollinator attractants in Nevada were reduced or entirely missing from Utah sites (e.g., Phacelia and Psorothamnus).
For management actions that target the reproductive health of A. californica populations, a better understanding of native pollinator nesting and foraging behavior is needed. Nests of solitary, ground-nesting bees are difficult to locate, and site characteristics such as exposed bare ground, litter cover, soil compaction, sloping ground, and ground cavities are often used as a proxy (Sardiñas and Kremen 2014). Perdita meconis is known to nest in open patches of gypsum soil (Clark County Department of Comprehensive Planning 2000 cited in Cornelisse 2018), while even this basic description of nesting sites is not known for less-studied pollinators such as M. enceliae. Bee emergence traps can be used to directly associate ground nesting locations with on-site nesting resources (Sardiñas and Kremen 2014), and this information could be used to monitor and protect nesting sites, which are vulnerable to surface disturbance (Hickerson and Wolf 1998).
Further Research for Species Conservation
We reviewed and presented here a body of information on Arctomecon californica that follows several notable reviews and contributions from more than a decade ago. We attempted to build on existing knowledge, with reference to relevant information previously reported that provided context to our current understanding of the species, especially with regard to its propagation potential, its reintroduction into protected habitat, and its conservation as a rare perennial desert herb existing in heavily fragmented habitats throughout its range. Our understanding of A. californica's reproductive ecology and habitat needs has developed immensely since these earlier reviews, and we highlight topics where study is still needed.
Preserve Germplasm for the Future
A conservation strategy that preserves germplasm (seeds and other tissues for propagation) in the event of extirpation of populations enables plants to be introduced into new habitats designated for protection and for repatriating previously extirpated sites throughout Arctomecon californica's range. Genetic diversity within current populations, while still high, is at risk of decreasing from inbreeding depression and genetic drift due to habitat fragmentation (Hickerson and Wolf 1998). Rigorous guidelines for germplasm collection and conservation already exist for rare species, including collecting adequate numbers of seeds and matrilines (Wall 2009, Meyer et al. 2014), and can direct future management of A. californica with little impact to intact populations.
Understand Edaphic Endemism
Managing populations of A. californica depends in part on identifying whether this species requires the chemical or physical attributes characteristic of the substrate, or simply tolerates the often physiologically stressful conditions of these soil types when other vegetation competitively excludes it from less stressful habitats (Allphin and Harper 1994, Houle and Valéry 2003). Habitat suitability models can guide site selection for carefully planned mechanistic experiments that introduce A. californica plants (by direct seeding and protection from harvester ants, by planting nursery-raised seedlings, or both) into known habitats versus perceived less-suitable habitat. Furthermore, complementary greenhouse experiments can distinguish the roles of biocrusts, soil structure, chemistry, and soil moisture and temperature dynamics to provide a more complete picture of the habitat needs of this unique species. Cryptogamic crusts and potentially mycorrhizal inoculation could improve A. californica propagation and salvage efforts (Bailey 2019), but these are largely untested; however, propagation and plant salvage experiments using cryptogamic crust are currently underway (Lydia Bailey personal communication). Potential mycorrhizal associations with A. californica cannot yet be ruled out; the limited information warrants further study. Elucidating the relationship between germination/seedling success and cryptogamic crusts in intact and disturbed habitat is also desirable to optimize reintroduction efforts.
Promote A. californica–Pollinator Relationships
Protecting habitat for key pollinators is pivotal to Arctomecon californica population persistence; such protection not only can maintain habitat patch connectivity but can also promote gene flow among populations of A. californica. Fundamentally, studies are needed to understand nesting requirements, cues for bee emergence, and foraging limits for the primary pollinators of A. californica. When integrated with A. californica population monitoring, studies of flower and fruit production and seed fill will signal inbreeding depression, genetic load, and pollinator scarcity, which can in turn portend population decline (Tepedino et al. 2014). Periodic study of A. californica's self-incompatibility across the range of its populations would track whether populations on the periphery of the species' range will develop self-compatibility and adaptations to self-fertilization over time (Busch 2005). For the most degraded habitats, reintroducing nectar plants that support A. californica's pollinators (such as Stanleya pinnata, Enceliopsis argophylla, and Larrea tridentata produced from locally sourced seeds) could be an effective ecosystem-based conservation approach (Tepedino et al. 1997) in combination with artificial pollination or active import of pollinators to restore gene flow and genetic variation for A. californica (Hickerson and Wolf 1998, Tepedino et al. 2014). Such an integrated conservation program can use rigorous implementation and monitoring to answer important outstanding research questions, including questions about the role of nectar plants (e.g., does outplanting promote native bee visitation?), the role of pollinators (e.g., does import of pollinators improve A. californica seed set?), and the desired increased fitness for A. californica (e.g., does enhanced seed bank support seedling recruitment?).
Protect Soil Seed Reserves
Seed reserves and seed dispersal dynamics can be better elucidated with broadened sampling (different methods, multiple years, wider range of sites). Combining seed reserve extraction and emergence methods, in addition to including new information about germination requirements in emergence methods, would describe the distribution of viable and nonviable seeds in the soil seed reserves, support soil salvaging efforts, and update seed dispersal knowledge (Abella et al. 2013). Methods for tracking seeds in seed dispersal trials (for example, see Vander Wall et al. 2006) and characterizing abundance of seed predators in degraded and intact habitats could better elucidate primary dispersal (distance from parent plants) and secondary dispersal (ant versus rodent removal and movement). The short-lived adults and persistent seed reserves of A. californica highlight the value of continued, standardized surveys of existing, historic, and suitable habitat, especially in years following exceptional rainfall, in order to gain a more robust understanding of demography and population dynamics.
Predict Responses to Climate Change and Land Use
The southwestern United States is a regional climate change hotspot, with projections predicting increasingly variable precipitation and potentially a shift to more spring-dominant rainfall (Seager et al. 2007, Diffenbaugh et al. 2008). This climatic trend could significantly impact A. californica's germination, which is currently believed to be dependent upon cold winter rains (Meyer 1987 cited in Mistretta et al. 1996). High reproductive attrition for A. californica at the bud and capsule stages may be linked to delayed development with onset of environmental stress in spring (Thompson and Smith 1997) or to asynchrony between flowering and abundance of pollinators, both of which can be exacerbated by changing temperatures. Better describing this reproductive attrition and knowing the temperature requirements for bud and capsule development would help researchers assess the reproductive health of A. californica populations. Less well known is how changing temperature and rainfall cues will affect synchrony between pollinators and A. californica flowering, flower and fruit production, seed viability and dormancy, and whether potential increases in invasive species in disturbed habitats will suppress A. californica recruitment. Ongoing surveys that collect information on invasive species abundance (e.g., presence/absence, frequency of plots) and on A. californica life stages (e.g., seedling, prereproductive, reproductive adult) may require little additional effort, and combined with the emerging availability of spatial climatic information, these surveys will reveal how seedling emergence cues and/or germination timing are changing through time and across habitats with varying levels of disturbance.
We are grateful for Stefanie Ferrazzano (Clark County Desert Conservation Program) and her support of our review of Arctomecon californica. We also thank Lydia Bailey (Northern Arizona University) for sharing her graduate research with us prior to publishing her results and Von Winkel (Southern Nevada Water Authority, Springs Preserve) for sharing reports and insights into propagation of desert plants. The A. californica community is a passionate one, and we are grateful for the provocative discussions during the July 2020 virtual meeting for the Rapid Species Status Assessment for Las Vegas Bear Poppy and follow-up communications with Cayenne Engel (Nevada Department of Forestry), Lara Kobelt (BLM, Southern Nevada District Office), Tiffany Pereira (Desert Research Institute), and Carrie Norman (NPS, Lake Mead National Recreation Area). We thank Tiffany Pereira and Von Winkel for reviewing this report and providing valuable feedback that improved an earlier draft. We are grateful to Susan Meyer and an anonymous reviewer for their significant comments on and edits of the final manuscript. This work was supported by the Clark County Desert Conservation Program and funded by Section 10, as project #2019-USGS-1990A, to further implement or develop the Clark County Multiple Species Habitat Conservation Plan. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
 1 The terms “seed reserves” and “seed banks” are both used in the literature to describe the natural accumulation and persistence of seeds in a dormant state within the topsoil awaiting favorable conditions for germination and seedling emergence. We use the former term in this context to avoid confusion with the latter, which is also used to reflect human-collected seeds maintained in a repository for future use in conservation or to safeguard declining populations against possible extinction (see section “Further Research for Species Conservation,” page 16).