Relationships Among Population Size, Environmental Factors, and Reproduction in Lupinus perennis (Fabaceae).
Human disturbance of natural populations is widely recognized as a source of habitat degradation, fragmentation, and loss (Fahrig, 2003). Historically, researchers have emphasized demographic risks to small isolated populations, including reduced population viability, shorter persistence times, and increased risk of extinction (Soule and Wilcox, 1980; Frankel and Soule, 1981; Lande, 1988; Berger, 1990; Menges, 1991). Early empirical studies showed increased inbreeding depression in small populations (Heschel and Paige, 1995) and a loss of genetic variation with reduced population size (Raijmann, 1994; Lammi et al., 1999; Cruzan, 2001). This reduced genetic variation is particularly important for threatened species by reducing their potential for adaptation to environmental change (Willi et al, 2006; Frankham, 2005; Frankham et al., 2014). Habitat loss and fragmentation may also correlate with other environmental changes that lead to ecological divergence among populations that may be further amplified by drift in small populations (Willi et al., 2007). Larger populations are also expected to have higher per capita reproductive success, a general pattern found in most (Leimu, 2006; Ouberg et al., 2006), but not all, species that have been examined (Milberg and Bertilsson, 1997; Lammi et al., 1999; Eisto et al, 2000).
Conservation and restoration planning for declining species also requires an understanding of species' responses to habitat loss and fragmentation and levels of potentially adaptive phenotypic variation. Although fitness reductions in small populations are often attributed to genetic factors (Barrett and Kohn, 1991; Byers, 1995; Agren, 1996; Groom, 1998; Fischer and Matthies, 1997; Kery et al., 2000), genetic data for species that have been subject to reductions in size and/or fragmentation are often absent or difficult to acquire. Whereas neutral molecular genetic marker studies can identify changes in mating system and loss of genetic variation, long-term adaptation and survival will be determined by whether adequate quantitative genetic variation has been retained (Kramer and Havens, 2009). However, relatively few quantitative genetic studies of rare plants have been undertaken, largely due to the need for destructive sampling and the extensive time and labor required (Oakley, 2015; Matesanz et al., 2017). Although associations among fitness components, genetic variation, and population size appear to be widespread across plant taxa (Leimu et al., 2006), reduced reproduction of small populations could also be caused by disparate nongenetic factors correlated with population size, such as population density, habitat quality, pollinator service, dispersal, and competition (Menges and Dolan, 1998; Fischer and Matthies, 1998; Lammi et al., 1999; Kery et al, 2001; Dauber et al, 2010; Lazaro-Nogal et al, 2017).
In this study we compare the reproductive success of Lupinus perennis L. (Fabaceae) populations of various sizes to determine if reproductive success was related to population size and/or associated environmental factors. High quality Midwestern oak savanna communities, which require fire disturbance to eliminate invasive woody cover (Grigore and Tramer, 1996), often feature L. perennis as an indicator species (Grigore and Windus, 1994). Lupinus perennis'numbers have greatly dwindled in the past 100 y due to clearing of the land for farming and the cutting of timber for sale (Moseley, 1928; Grigore and Tramer, 1996). Under the fire suppression that followed European settlement of Midwest savannas, regrowth of oak saplings eventually eliminates lupine from the understory (Campbell, 1933; Mayfield, 1962; Smallidge et al., 1996). Although once widely distributed on dry sandy soils from north of Florida to Maine and west to Minnesota and Indiana (Gleason and Cronquist, 1991), by the 1980s L. perennis had been listed as "potentially threatened" in Ohio (Ohio Division of Natural Areas and Preserves, 1984) and was considered extirpated or imperiled in nine states and considered vulnerable in another six states (http;//www.natureserve.org/ explorer). This herbaceous perennial became a primary target of habitat restoration efforts as it is the main larval food source for the Federally Endangered Karner Blue butterfly, Plebejus (or Lycaeides) melissa samuelis, (U.S. Fish and Wildlife Service, 1992) and a significant food source for other rare butterfly species; Erynnis peisius, the Persius Dusky Wing, and Inasalia irus, the Frosted Elfin (Shapiro, 1974; Shuey et al., 1987). These butterflies are believed to have declined in numbers because of destruction of oak savanna habitat and, more specifically, the decline of lupine populations and co- occurring nectaring plants (Grigore and Tramer, 1996). Management of this lupine for Karner Blue restoration has relied on improvements in habitat through fire, herbicides, and mechanical cutting to reduce woody cover and increase lupine growth (Boyonoski, 1992; Zaremba and Pickering, 1994; Grigore and Tramer, 1996; Smallidge et al., 1996; Grandel, 1998).
As a preliminary indicator of potential levels of adaptive phenotypic variation, we also studied seed coat color variation, an easily observable morphological trait that is polymorphic within L. perennis populations and relatively constant within plants. Plant seedling recruitment may be influenced both by the availability of "safe sites" for establishment as well as seed supply (Harper, 1977). Furthermore, because rodent seed predation significantly impacts reproductive success of other lupines (L. arboreus; Manon and Simms, 1997, 2001; Dangremond et al, 2010; Kurkjian et al, 2016), we also examined seed predation. Previously, Kappler et al. (2012) found mice (Peromyscus leucopus) in the Oak Openings actively consume L. perennis seeds, averaging 22% across sites, with predation increasing with prescribed burn frequency. Because cryptic and disruptive coloration are important in predator-prey interactions (Endler, 2006; Stevens, 2007), we conducted a field experiment to ascertain whether seed coat color influenced predation rates, which might thereby limit seedling recruitment among populations. We hypothesized seed color variation in these populations functions as camouflage for newly dispersed seeds, helping them appear cryptic against the spatially variable substrates in this habitat. Therefore, removal rates would be higher for seeds that differed in color from the background substrate, as documented for color polymorphic pine seeds and bird predators against postfire substrates (Nystrand and Granstrom, 1997) and suggested for two other wild legumes (Porter, 2013; Bishop-von Wettberg et al., 2018).
We tested the following hypotheses: (1) Populations will vary in seed production and differences are related to both population size and environmental factors; (2) Offspring emergence and growth varies with size of and environmental factors associated with the natal population; (3) Small populations harbor less seed color variation than larger populations; and 4) Seed predation rates are influenced by seed color. To test these hypotheses, we compared components of seed production (fruits per stem, seeds per plant, number of seeds per pod, and seed size) of plants in the field and relevant environmental factors (soil nutrient and moisture levels, light levels, and percent bare ground) in 16 different populations ranging in size from 25 to 8300 individuals from May-August 1996. Greenhouse experiments examined offspring vigor (seedling emergence and growth) in a common environment, while variation in seed coat color in each population was scored as an indicator of phenotypic variation. We also tested for potential effects of seed predation and environmental factors on seed coat color frequencies in a 2008 field seed removal study in one population.
ORGANISM AND STUDY SITES
Lupinus perennis is a long-lived perennial that opens blue show)' flowers around early May in Ohio, producing seed pods from mid-June to early July. The pods, which typically contain three to six seeds, dehisce by twisting, splitting the pod, and explosively propelling an average of two to three mature seeds as far as 4.8 m (Grigore and Tramer, 1996). Lupinus perennis is visited primarily by bumblebees (Bombus sp.; Bernhardt et al., 2008) and is capable of self-pollination (Shi et al., 2005). The crown grows through production of short horizontal rhizomes forming a candelabra-shaped root system (Michaels, pers. obs.), causing the plant to form clumps (Foster, 1984). Although a few isolated populations are found in northeastern counties, most Ohio lupine populations are restricted to the Oak Openings region of the northwest and extending into southeastern Michigan. This distinctive vegetation area occurs along a 35.4 km long and 9.66 km wide sand belt (Hehr, 1970; Weber and Huffman, 1989); most of which occurs in Lucas Co., extending into Henry and Fulton counties with small, scattered sand pockets in Wood Co., Ohio and Monroe Co., Michigan. Originally described in Moseley's Flora of the Oak Openings (1928) as "abundant" many of the extant northwest Ohio populations were probably once interconnected (Sears, 1926; Gordon, 1969) before major habitat fragmentation occurred as a result of both urban and agricultural development (Schetter and Root, 2011).
We studied 16 populations (across one Michigan and five Ohio counties), chosen to represent both a wide range of population sizes and the species' geographic range within the area (Table 1). The size of each population was estimated by placement of three to five parallel lm wide belt transects through each population at 5 m intervals and counting the number of individual plants, which was then multiplied by five. The numbers of individuals within the belt transects were scored early in the season to facilitate genet identification. Because L. perennis is capable of limited vegetative growth, genet limitation was assessed by careful examination of the base of each plant when ramets are first emerging from the plant crown along with shallow excavation, if needed. All stems emerging from the same localized area were scored as a genet. At the time of this study, four of the seven large populations (1800 individuals and over) were burned in the previous 1 to 2 y. Although the other three large populations had not been burned in years prior to this study, none of the small populations (under 600 individuals) had been fire managed. We do not expect that the population sizes had changed significantly owing to the longevity of this species and infrequent seedling establishment (Kelly, 1998; Plenzler and Michaels, 2015).
SEED PRODUCTION IN NATURAL POPULATIONS
To evaluate seed production, we selected 20-30 plants in each of the 16 populations by placement of a 0.6 X 0.6 m quadrat frame located by a blind-toss throw at points spaced at least 5 m apart along the established transects, choosing the plant that was closest to the center of this quadrat for sampling. To minimize potential impacts to populations of this potentially threatened species, we collected all fruits (typically 15-20) from a sample of up to three stems per plant. For each plant we estimated seed production by counting number of seeds per pod and number of seeds per plant. Although this was likely an underestimate of seeds per plant for very large individuals given no more than three stems per plant were sampled, a subsequent four year demographic study of a subset of these populations found that L.perennis plants may produce up to 68 inflorescences, with a median of two (Michaels and Mitchell, unpub. data). Seeds were individually weighed to the nearest mg. To determine population average number of fruits per stem, the number of pods was counted on 25 stems selected haphazardly from within each population.
GREENHOUSE EXPERIMENTS ON EMERGENCE AND GROWTH
We examined offspring fitness in a greenhouse emergence experiment conducted from February-March in the Bowling Green State University greenhouse. Two small populations (APN and HNR; Table 1) were excluded due to inadequate seed production, and two other populations (LKN and LCPR), similar in size to a population in a nearby State Preserve (LCPV), were not included due to space limitations. For each of the remaining 12 populations, we planted up to 48 seeds from each of 30 plants in 11 cm diameter pots containing a mixture of 25 parts sterilized potting soil, six parts milled sphagnum, 10 parts perlite, two parts vermiculite, and six parts sand. To break dormancy prior to planting, seeds were rapidly immersed and removed from distilled water that had been boiled and then cooled to 85 C (Davis, 1991). If at least 48 sibs were collected, seeds from each plant were divided into three replicates of 16 seeds per pot. When too few were available, we used only those plants producing at least 18 seeds, which were again divided to maintain three replications. Although the number of seeds planted per pot varied from 6-16 seeds, this density variation was not expected to strongly influence the results, as not all seeds were expected to emerge and this phase of the experiment was of relatively short duration. We randomly assigned one pot from each population to flats which were randomly placed on benches and rearranged every 2 d to reduce effects of spatial heterogeneity in lighting and temperature. Emergence was scored every 24 h for 7 wk under ambient lighting and greenhouse temperatures (ranging between 18-27 C during this phase of the study).
Seven weeks after initial planting, we used all seedlings from any parents producing at least 15 seedlings from the germination study in a second experiment to examine potential differences in early seedling growth in the greenhouse. From each population up to 20 seedlings from each of 15 parent plants were transplanted in early April into individual 7x7 X 14 cm tree band pots containing a growing mixture of 25 parts sterilized potting soil, two parts milled sphagnum, three parts sphagnum moss, 13 parts perlite, two parts of vermiculite, and six parts of sand. Following a 2 wk post-transplantation period to acclimate and assess transplantation mortality, the numbers of leaves, leaflets and length of the largest leaf were scored every 2 wk for 6 wk. A nondestructive synthetic measure of plant size as an indicator of leaf area, calculated as (number of leaves) x (number of leaflets) x (length of the largest leaflet), was used as an index of plant vigor.
Although plant size is sometimes used as an indirect measure of habitat quality, correlations of size can be confounded with genetically based fitness differences or age. Therefore, habitat quality was determined by direct measurements of relevant environmental factors in each population to understand how habitat factors correlated with reproduction. We evaluated habitat quality by measuring light, soil moisture and nutrient levels, and amount of bare ground in each population. We measured light levels (PAR) in each site from July 3 to July 10 every 20 m along a transect using a Quantum Sensor Photometer (LI-COR LI-89), with six readings obtained at each sample point on clear days between 10:00 and 14:00 at a height of 0.5 m to evaluate light at canopy level. We collected ten 30 cm long soil cores along each transect at two randomly selected locations per population using a standard soil probe. The ten samples were homogenized and then a 240 ml sample from each of the two locations sampled was analyzed (OSU Coop. Extension Lab) for available P, K, pH, and organic content. Using a Quickdraw Soil Moisture Probe (Soil Moisture Equipment Corp., Model #2900F), we obtained three soil moisture readings at points sampled 10-20 m apart along each transect. As an indicator of site productivity and interspecific competition, we measured percent bare ground (proportion of unvegetated cells) by ten haphazard throws (at least 10 m apart) of a square frame (0.5 m x 0.5 m, divided into a grid of 25 cells, each accounting for 4% of the total area).
SEED COLOR VARIATION
For each plant from which seeds were collected, we recorded seed coat color type. As a maternal tissue, seed coat color is relatively invariant within a pod and on a plant (Michaels, pers. obs.). The genetic basis for seed coat color in Lupinus has long been a focus for breeding studies (e.g. Lupinus pilosus, where the white seed coat color is homozygous recessive, Horovitz and Harding, 1983), while more recent work based on the sequenced genome for Lupinus angustifolius has mapped the Leuc gene responsible for white seed coats to a particular chromosome (Nelson et al, 2010). Although likely a continuously varying trait, the frequencies of six seed coat color phenotypes were initially scored across all of the populations: dark, speckled, white, white with speckles, gray, and gray with speckles.
SEED REMOVAL STUDY
Because savannas were historically subjected to fire from both natural and manmade sources and contemporary restoration often entails management by proscribed fire (Smallidge et al, 1996; Peterson and Reich, 2001), we expected fire to create heterogeneous soil surfaces that might influence post-dispersal seed predation. Intense fire can create dark substrates (consisting of burned wood chips, vegetation, and ash), whereas areas with less intense burns or unburned patches retain a lighter colored sandy substrate. Surface color heterogeneity may also develop within oak savannas when areas vary in fire history, accumulate more litter or humus, or when animal burrowing brings lighter color subsoil to the surface. Effects of seed coat and substrate colors on seed removal rates were studied in a management unit near the Wabash-Cannonball bike trail in Oak Openings Metropark (OOM, 41.556545, -83.835701), Toledo, Ohio that had been burned 2 mo previously. Surface sand and char were collected in June 2008. The char was ground using a coffee grinder and then each substrate was separately sieved with 2 mm mesh. For simplicity we used only two seed-coat color types: white, which included both solid white and white with slight speckling and dark, which included gray; gray with speckling; black; and white with more than 50% dark speckling. Seeds used in most of the experiment had been previously collected from several similar sites within OOM and stored at 4 C from previous years until we could collect, sort, and distribute new seed (from the 2008 season) into seed dishes on August 14.
During the peak of natural seed dispersal in early July, a total of 32 shallow plastic weigh boat dishes (130 mm diameter x 89 mm deep) were placed at points along a 180 m transect through the center of the lupine population. Eighteen were spaced every 10 m, with another 14 placed singly or in pairs on alternate sides 3 m from the main transect. Dishes were anchored into the ground using 18 gauge wire punched through the bottom of the dish and buried up to the lip in the substrate. One substrate (either natural sand or ground char mixed with sand) was then poured into the dish until the lip was covered. We placed twenty seeds of either white or dark seeds weekly on the surface of one of the two substrate types producing four possible substrate type/seed color treatment combinations: sand and white seed; sand and dark seed; char and white seed; char and dark seed. We haphazardly assigned eight replicates of each of the four treatments to dishes. Seed removal rates (the number of seeds taken divided by the number of seeds offered) were determined weekly for the months of June and July. To determine the number of seeds left in a dish, the contents were sieved, remaining seeds counted, and the substrate returned to the dish. Removed or chewed seeds were replaced as needed to consistently offer twenty seeds.
Dishes were checked on July 9, 16, and 23 and August 14 and 20. Predation rates were described as the ratio of number of seeds removed to the number of seeds offered per month. For analysis of spatial data, we recorded GPS coordinates of each dish with a Garmin eTrex Vista C handheld GPS unit. Normalized Difference Vegetation Index (NDYI) values were derived using Landsat ETM+ imagery (July 7 and August 8, 2008) from the USGS Global Visualization Viewer website (http://gloris.usgs.gov).
Regression analyses (Student SYST AT version 1.0, 1994) examined the relationships amongst population size and mean population number of seeds per pod, pods per stem, average mass per seed, and average number of seeds per plant. Nested ANOVA on arcsin square root transformed proportions was used to test for population size, population identity, and maternal plant effects on emergence (SAS[R], Version 9, 1996, SAS Institute Inc., Cary, North Carolina). We tested effects of population size against variation among populations, variation among populations against variation among maternal plants, and variation among plants against residual variation among replicates. Because growth data were not normally distributed, we used nonparametric analyses to test for differences in seedling size among populations and effects of population size on final plant size. Forward and backward stepwise regressions were used to determine relationships among plant performance, population size, and environmental variables. A probability of 0.10 was used as the criterion for inclusion during initial model selection.
We analyzed the relationship between population size and variation in seed coat color (number of different color types) among the populations using Spearman's correlation. To determine the effects of seed color, substrate, month, NDVI, distance to roads, and distance to water on predicting proportion of seeds removed (arcsine square-root transformed), we employed forward and backward stepwise regressions. The model entry probability was set at 0.25, and removal was set at 0.10. Interaction effects between the treatments and the covariates (dish location, NDVI, road distance, water distance) were included in preliminary analyses, but not significant and excluded from the final model, which included spatial variables, substrate, and seed color. All seed predation analyses were performed in JMP 7 (JMP[R], Version 7. SAS Institute Inc., Cary, North Carolina).
EFFECTS OF POPULATION SIZE AND ENVIRONMENTAL FACTORS ON SEED PRODUCTION
Our L. perennis populations ranged in size from 25 to 8300 individuals (Table 1). Nine of the populations were less than 1000 in size, while two were less than 100. Populations were separated by 1.61 km or more with the exception of two: LMV (located 0.8 km from LCPR) and APN (located 0.8 km from APS). Half of the populations were found in Lucas County where most of the sand deposit defining the Oak Openings occurs. Most of the populations were located in preserves or protected lands, three (LKN, LMR, and LHO) were on private land, while three others (WPR, APN, and W235) were located near or along railroad tracks and roadsides.
Regression analyses revealed several relationships of population size with measures of reproduction. Finit production (number of pods per stem) increased with population size (Fig. 1; [r.sup.2] = 0.64, P = 0.0001, n = 16), as did average number of seeds per plant (Fig. 2; [r.sup.2] = 0.44, P = 0.005, n = 16). A similar relationship held for reproduction on a per flower level as seeds produced per pod had a significant, but weaker, increase with population size (data not shown; [r.sup.2] = 0.26, P = 0.045, n=16). In contrast seed mass was unrelated to population size ([r.sup.2] = 0.07, P = 0.31, n = 16).
Because population size is likely to be influenced by habitat quality, preliminary correlation analyses tested for associations among environmental factors and population size. No environmental variables (soil moisture, proportion bare ground, light levels, pH, phosphorus, potassium, and proportion organic matter) showed significant strong correlations (P < 0.05 and r > 0.30) with population size (results not shown). Following stepwise linear regressions to determine effects of environmental factors and population size on fruit or seed production, population variation in organic matter, soil moisture, phosphorus, or potassium levels were removed from final models of pod and seed production, and there were no significant inter-correlations among those environmental variables that were retained. In the final model, higher average fruit set was associated with higher pH (Table 3; Model [r.sup.2] = 0.81, P = 0.01) and with natural log (ln) of population size (P = 0.0001). The In of population size explained 65% of the variation in fruit set, while pH explained only 17%. Higher seed mass was associated with increases in soil pH ([r.sup.2] = 0.29, F= 11.60, df= 1, P = 0.003). Finally, the number of seeds per fruit was correlated with population size (Table 3; Model r = 0.69, P = 0.003), but decreased with increased bare ground, with population size explaining 41% and bare ground 29% of this variation.
EFFECTS ON EMERGENCE AND EARLY SEEDLING GROWTH
Emergence among populations in the greenhouse experiments ranged between 20% and 48% (Fig. 3) and was not related to average population seed mass (P = 0.18, [r.sup.2] = 0.02), but did vary with population size. For populations of 600 or fewer individuals, emergence was consistently below 45% and quite variable. Inspection of the data revealed no populations had been sampled greater than 600 and smaller than 1800. However, when this natural break in the distribution was arbitrarily used to divide the populations into small and large size classes (small: 25 to 600 and large: 1800 to 8300 individuals), seedling emergence of plants from large and small population size classes did not differ significantly (nested ANOVA, Table 2). Populations also did not differ in average emergence, most likely due to the great variability in performance among plants within populations as reflected in the highly significant (P = 0.0001) effect of plants within populations. After 7 wk in the greenhouse final seedling size was significantly higher for the large population size class (Mann-Whitney U = 36.76, df= 1, P = 0.0001). However, final seedling size was not correlated with seed mass (data not shown). The environmental factors that influenced emergence and seedling size in the greenhouse differed from those that affected field measures of seed production. Emergence declined with increased bare ground (Table 3; Model [r.sup.2] = 0.515; P = 0.02), but increased with increased population size (P = 0.04). Final seedling size increased with natal population soil pH (Table 3; Model [r.sup.2] = 0.60, P = 0.01) and decreased with increased light levels (Table 3, P = 0.026), but was unrelated to population size from which seeds were collected. Organic content, potassium, phosphorus, and soil moisture levels where the seeds originated had no effect on either greenhouse emergence or growth.
SEED COLOR VARIATION AND SEED PREDATION
The speckled seed coat color was the most frequently found seed coat color across all populations, with the white seed coat color the second most common (Fig. 4). The number of seed coat colors per population was positively correlated with population size (Spearman's correlation [r.sub.s] = 0.66, n = 16, P < 0.05). Seed removal was significantly affected by month, NDVI, distances to water and roads, and seed color, but not by substrate color ([F.sub.1,56] = 0.296, P [much greater than] 0.05, model [R.sup.2] (adj). = 0.47, P < 0.0001). Seed removal rates during July and August across all treatments were substantial, averaging 65% (SE = 0.03; min = 8.3%; max = 100%), with significantly more of the seeds offered removed in August (mean = 78%) than in July (mean = 51%) ([F.sub.1,56] = 4.402, P < 0.05). This late summer increase in removal rates may reflect both seasonal increases in predator populations and conditioned visitation as dish locations were not changed between weeks. The proportion of seeds removed was greater for dark (81.1%) than for white (58.9%) seeds ([F.sub.1,56] = 8.836, P < 0.01). The distance to roads (max = 291.45 m; min = 128.64 m) and distance to bodies of water (max = 198.69 m; min = 79.13 m) were significant covariates, but their parameter estimates were low compared to that of NDV1, which had a relatively strong influence on seed removal rates in ([F.sub.1,56] = 8.611, P < 0.01). Seed removal increased as NDV1 increased (indicating denser vegetation), consistent with the expectation removal rates might increase with increased vegetative cover.
SEED PRODUCTION IN NATURAL POPULATIONS
Several components of reproductive success in Lupinus perennis were associated with population size. Higher numbers of fruits per stem, seeds per plant, and seeds per pod were all associated with increasing population size. Reduced seed production of smaller populations may arise through altered plant-pollinator interactions that limit pollen receipt and/or elevate selfing, as well as early acting inbreeding depression. In L. perennis, manually selfed flowers produce significantly fewer seeds per flower (Shi et al., 2005). Furthermore, in a study of pollinator visitation in L. perennis, pollination success significantly declined only for populations smaller than 215 plants but consistently increased with density (Bernhardt et al., 2008). In a separate study that excluded populations smaller than 120 (to avoid impacting them further), we found selfing rates were not related to population size in contrast to Shi (2004), although L. perennis performance declined with increased inbreeding (Michaels et al., 2008). These findings suggest, in this species, population size effects on pollination and mating system may only be important once population size contracts below several hundred. Evidence of such increased inbreeding in smaller populations has been observed in the congener Lupinus sulphureus ssp. Kincadii (the host for the endangered Fender's Blue butterfly), for which both population and patch size limited seed set (Severns, 2003).
Like many other oak savanna species declining from anthropogenic disruption of natural ecosystem disturbance cycles (Anderson and Bowles, 1999), reduced reproduction of L. perennis in small populations most likely developed via habitat degradation following fire-suppression (Abella et al., 2001). All small lupine populations occurred on unmanaged lands where this history of fire-suppression allowed the growth of invasive, woody, shade-producing competitors and litter accumulation. Fire has been shown to be an important factor in Midwestern oak savanna communities and for Oak Openings lupine populations by releasing nutrients accumulated in the leaf litter and directly increasing post-fire lupine seeding survival, flowering, and seed set (Grigore and Tramer, 1996). Because other management (mowing) had also been irregularly applied to selected sites and records were unavailable for some locations, a specific analysis of fire frequency was not attempted. Although some large populations had been fire-managed in the previous decade, we hypothesized small and large populations might show consistent differences in environmental factors that could influence resources for seed production. Although we found considerable variation among populations in environmental variables (bare ground, light levels, soil moisture, pH, phosphorus, potassium, and organic matter), only organic matter was significantly correlated with population size. Although organic matter declined as population size increased, this factor did not clearly influence any of the fitness components measured in the field or greenhouse.
Environmental factors explained relatively little variation in seed production. Only pH and amount of bare ground had significant relationships, increasing fruit production and decreasing seeds per pod respectively. Sites with reduced cover may be subject to increased moisture stress leading to greater fruit and seed abortion during seasonal variation in weather (H. Michaels, pers. obs.). In contrast total number of seeds set was related to population size but no other environmental variables. Seed mass, which is commonly positively correlated with emergence (Silvertown, 1984; Hendrix, 1984, Hendrix and Trapp, 1992), is often used as a measure of individual fitness in plant populations (Mitchell-Olds and Bergelson, 1990; Menges, 1991). In our study seed mass showed no relationship with population size and was responsive only to one measured source of environmental variation: seed mass increased with increased pH.
GREENHOUSE EMERGENCE AND GROWTH
Many studies have documented reduced emergence in small populations may be partially due to inbreeding depression (Dudash, 1990; Menges, 1991; Carr and Dudash, 1995; Fischer and Matthies, 1997; Kery et al., 2000). For the 12 populations selected for this experiment, variation in emergence rate was not systematically related to population size class or identity. Instead, emergence was highly variable in smaller populations with plants from the smallest populations (n = 600 or less) most likely to have variable emergence, with more consistent emergence (40-50%) associated with populations larger than 1000. Furthermore, of the original sample of 16 populations, two of the smallest populations (APN and HNR, n < 100) had insufficient seed production to include in the emergence study. When these seeds were planted for the growth experiment, no seeds germinated from APN while only two geminated from HNR. These particular populations were geographically distant from most other Ohio populations. Although these very small L. perennis populations had reduced emergence, population size itself was not consistently associated with subsequent early seedling performance. Reduced variability in emergence for large populations (>150 individuals) was first documented for Silene regia (Menges, 1991), and many papers have documented significantly reduced germination success in small populations (reviewed in Leimu et al., 2006), but relatively few have examined the reproductive components, mating system, or environmental circumstances that accompany its occurrence (Fischer and Matthies, 1997; Leimu et al, 2006).
In the common garden greenhouse environment, there was little evidence of maternal environmental effects from the natal population. Neither emergence nor growth (discussed below) were associated with variation in organic content, potassium, phosphorus, or soil moisture levels in the populations where the seed originated. Emergence in the greenhouse was only associated with the amount of bare ground, decreasing for seeds produced in natal populations with increased bare ground. Environmental factors related to maternal origins of seeds have long been known to influence seed traits and performance (Roach and Wulff, 1987; Fenner 1991). More recent physiological and genetic studies suggest a variety of maternal conditions (light quality, temperature during seed maturation) can influence dormancy and germination (reviewed in Penfield and MacGregor, 2017).
Although reduced emergence for seed associated with increased bare ground at the site of origin was evident in the greenhouse, similar responses are likely to occur outside controlled benign greenhouse environments. In field common garden restoration experiments that did not consider source population size, L. perennis seeds planted in bare sand took longer to emerge (St. Mary, 2007), as did seed planted in areas with greater canopy cover (Pavlovic and Grandel, 2008). A more recent study (Plenzler and Michaels, 2015) comparing naturally emerging seedlings in seven sites with varying management histories found increased seedling density associated with grass and lupine stem abundance but fewer with higher soil organic matter and pH. Reduced organic matter in large populations suggests management practices carried out in most of the large populations had been effective in limiting the invasion of weedy species and successional development of woody vegetation. Two small remnant populations, W235 and APN, had exceptionally high amounts of bare ground relative to the other small populations, possibly the result of occasional herbicide application by railroad companies to eliminate vegetation near the tracks. Lupines grow on sandy soils where moisture readily percolates to depths beyond reach of seedling root systems. Although vegetation removal may reduce competition and aid adult lupine persistence, management practices that create large reductions in cover may be particularly detrimental on these soils as bare ground significantly reduced seedling emergence. Plenzler and Michaels (2015) also found ground cover continues to be important into the second year, as seedling survival increased with abundance of fern and moss cover but decreased with more oak saplings and topographic wetness.
In general environmental factors were more important than population size in determining seedling growth for our populations at this relatively small landscape scale. Although populations showed significant differences in final seedling size, this was unrelated to population size. Population size effects on seedling growth in the greenhouse study may have faded over time, given resources were not limiting and conditions in the greenhouse were less stressful than the field. Greater and more persistent effects of population size may be evident in a field experiment wherein fitness differences are tested by resource limitation, competition, and more variable conditions (Ramsey, 1998). As suggested for the prairie annual, Chamaecrista fasiculata (Mannouris and Byers, 2013), fitness declines with fragmentation may have more to do with the size of the habitat fragment or its isolation than the size of the population itself.
Seedlings from populations with increased pH had greater growth, but those from populations with higher light levels had reduced growth. The pH of soil has been found to affect nutrient uptake in plants (Hinsinger, 1998) and to decrease root growth (Stalfelt, 1972), which may affect nutrients allocated to the seeds. The pH can also have a profound influence on symbiotic nitrogen-fixing bacteria and the suite of compounds that regulate symbiosis formation (Cooper, 2007). Our small populations had a wide range of pH values, but the large populations consistently had an average pH between 4.8 and 5.8. This corresponds with the optimal pH needed for root elongation in Lupinus angustifolius (Tang et al, 1992), while the pH range we observed for the large populations is the same as found among L. perennis growing in six oak savannas in Michigan's Allegan State Game Area, the nearest Midwestern natural area consistently hosting a Karner Blue butterfly population (Greenfield, 1997). Variation in field soil pH also had a significant effect on seedling growth in an experimental cohort planted within one Oak Openings population (Kelly, 1998). These pH responses are likely related to the specialized P-acquisition adaptation found in many Lupinus species (Lambers et al, 2013). Because, for many species, seed size is the main factor determining large seedling size (Atkinson, 1973; Fenner, 1978; Gross, 1984), we expected seedling growth would be correlated with seed mass, but we did not observe this pattern across our populations, although the population with the greatest seed mass, MSP (pH 7.3), produced the largest plants.
SEED COLOR VARIATION AND PREDATION
As a preliminary indicator of potential levels of adaptive phenotypic variation in L. perennis, we found more variation in seed coat color in the larger populations. This is consistent with the expectation smaller populations should have reduced genetic diversity and phenotypic variation as a result of inbreeding and genetic drift in small populations. This inference is supported by a subsequent microsatellite survey of ten Oak Openings L. perennis populations, which found a reduction in allelic diversity in small populations (Shi, 2004). However, seed coat color frequencies in these populations might also reflect selection by seed predators that preferentially removed darker seeds. Particular seed coat colors maybe selected again if they are more risible to predators, whereas seed with other colors are camouflaged in certain soil types, thereby escaping predation (Porter, 2013). Small mammals are known to be major predators of L. arboreus seeds in California dune shrub communities, where up to 86% of the seeds dispersed can be removed by rodents, leading to significant reductions in establishment (Marron and Simms, 1997, 2001). Rodent seed or fruit predation has been implicated in declines of two other rare Lupinus, Tidestrom's lupine (Dangremond et al, 2010) and Lassies lupine (Kurkjian, 2012). In our study the speckled seed coat color was predominant in the smaller L. perennis populations, which had reduced bare ground and increased organic matter. We had previously observed, in sites with increased woody vegetation and soil organic matter from fire suppression, speckled seeds are virtually hidden (Michaels, pers obs). This suggested that in areas of bare sand with low organic matter, white seeds might be less risible leading to an increased frequency of white colored seeds in the larger populations. But, these seed removal data do not support a camouflage hypothesis, as removal rates did not differ with substrate color for the two simulated backgrounds. Further study on effects of post dispersal seed predation on L. perennis offspring recruitment as a function of population size and variation in substrate types and vegetation cover are needed to further understand the ecological factors underlying this seed coat variation.
In summary our studies demonstrate that in L. perennis population size has its strongest influence on seed production: larger populations were associated with higher average seed production through fruits per stem, seeds per plant, and seeds per pod, which were all positively associated with increases in population size. While selfed L. perennis seeds have slower emergence and growth compared to naturally pollinated flowers (Shi et al., 2005), our studies of many of these same populations found levels of inbreeding depression in seedlings were similar for large and small populations (Michaels et al., 2008), indicating small populations have not purged deleterious alleles. Habitat restorations of existing small Lupine populations are likely to be ineffective, and the creation of dispersal corridors for L. perennis' specialist Lepidopteran herbivores, by augmenting existing populations or establishing new populations, will fail unless adequately sized populations are generated. Sites with vegetative cover that favors foraging by seed predators (Kappler et al., 2012) may not generate sufficient seed to be self-sustaining. Our data imply restoration designs and management practices should attempt to create and maintain Lupine populations of at least 500, preferably over 1000, individuals. Seed production and seedling growth were both influenced by soil pH, suggesting restorations that consider soil chemistry will be more successful. Smaller populations, particularly those with few seed color phenotypes, may benefit from addition of seed from nearby ecologically similar sites to infuse new variation to recover lost evolutionary potential for future responses to global change.
Acknowledgments.--This work would not have been possible without the cooperation of numerous agencies and land managers that gave access to the study sites and freely shared their accumulated years of experience in the Oak Openings and with the Lupine / Karner Blue system. Denise Gehring, Bob Jacksy, and John Jaeger of the Toledo Metroparks, The Nature Conservancy (and Terry Seidel and Gary Haase of the Kitty Todd Ohio field office in particular), Jennifer Wyndus, Ohio Division of Natural Areas and Preserves, Mitch Magditch and Peter Toison of the Toledo Zoo, and the Michigan Fish and Wildlife management staff of the Petersburg State Game Area. We would also like to thank Randy Mitchell, Karen Root, Xiujie Shi, Chris Trace)', Scott Hevner, Ryan Walsh, and Mike Tamor for their insightful comments and help on the manuscript.
ABELLA, S., J. JAEGER, D. GEHRING, R. JACKSY, K. MENARD, AND K. HIGH. 2001. Restoring Historic Plant Communities in the Oak Openings Region of Northwest Ohio. Ecolog. Restar., 19:155-160.
ANDERSON, R. C. AND M. L. BOWLES. 1999. Deep-soil savannas and barrens of the Midwestern United States. (page numbers) In: R. C. Anderson, J. S. Fralish, and J. Baskin, (eds.). Savannas, Barrens, and Rock Outcrop Plant Communities of North America, Cambridge Univ. Press, Cambridge, UK.
ATKINSON, D. 1973. Some general effects of phosphorus deficiency on growth and development. New Phytol., 72:101-111.
BARRETT, S. C. H. AND J. R. KOHN. 1991. Genetic and evolutionary consequences of small population size in plants: implications for conservation, p. 3-30. In: D. Falk and K. Holsinger (eds.), Genetics and Conservation of Rare Plants, Oxford University Press, Oxford, UK.
BEARDMORE. J. A. 1983. Extinction, survival, and genetic variation, p. 125-151. In: C. M. Schoewald-Cox, S. M. Chambers, B. MacBryde, and L. Thomas (eds.). Genetics and Conservation, Benjamin-Cummings, Menlo Park, CA.
BERGER, J. 1990. Persistence of different sized populations: an empirical assessment of rapid extinction in bighorn sheep. Conseru. Biol., 4:91-98.
BERNHARDT, C. E., R.J. Mitchell, and H.J. Michaels. 2008. Effects of population size and density on pollinator visitation, pollinator behavior, and pollen tube abundance in Lupinus perennis. Int. /. Plant Sci.. 169(7): 944-953.
BOWLES, M. L. AND S. I. APFELBAUM. 1989. Effects of land use and stochastic events on heart-leaved plantain (Plantago conluia Lam.) in an Illinois stream system. Nat. Area J.. 9:90-107.
CAMPBELL, L. W. 1933. The "Oak Openings" of northwestern Ohio. Northwest Ohio Quar. Bull., 5(2): 1-9.
CARR, D. E. AND M. R. DUDASH. 1995. Inbreeding depression under a competitive regime in Mimulus gut talus: consequences for potential male and female function. Heredity, 75:437-445.
CIDECIYAN, M. A. AND A. J. C. MALLOCH. 1982. Effects of seed size on the germination, growth and competitive ability of Rumex crispas and Rumex obtusijalius. J. Ecol, 70:227-232.
COOPER, J. E. 2007. Early interactions between legumes and rhizobia: disclosing complexity in a molecular dialogue. J Appt. Microbiol., 103:1355-1365.
CRUZAN, M. B. 2001. Population size and fragmentation thresholds for the maintenance of genetic diversity in the herbaceous endemic Scutellaria montana (Lamiaceae). Evolution, 55(8): 1569-1580.
DANGREMOND, E. M., E. A PARDINI, AND T. M. KNIGHT. 2010. Apparent competition with an invasive plant hastens the extinction of an endangered lupine. Ecology, 91(8):2261-2271.
DAUBER, J., J. C. BIESMEIJER, D. GABRIEL, W. E. KUNIN, E. LAMBORN, AND B. MEYER. (2010). Effects of patch size and density on flower visitation and seed set of wild plants: a pan-European approach. J. Ecol., 98:188-196. doi: 10.1111/j. 13652745.2009.01590.x
DAVIS, T. D., S. W. GEORGE, A. UPADHAYA, AND J. M. PARSONS. 1991. Improvement of seedling emergence of Lupinus texensis following seed scarification treatments. J. Environ. Hort., 9:17-21.
DUDASH, M. R. 1990. Relative fitness of selfed and outsrossed progenv in a self-compatible, protandrous species, Sabatia angularis L. (Gentianaceae): a comparison in three environments. Evolution, 44:1129-1139.
EISTO, A. K., M. KUITUNEN, A. LAMMI, V. SARRI. J. SCHONEN, S. SYRIASUO, AND P. M. TIKKA. 2000. Population persistence and offspring fitness in the rare bellflower Campanula cervicaria in relation to population size and habitat quality. Consent. Biol., 14 (5): 1413-1421.
ELSTRAND, N. C. AND D. R. ELAM. 1993. Population genetic consequences of small population size: implications for plant conservation. Annu. Rex' Ecol Syst., 24:217-242.
ENDLER, J. A. 2000. Disruptive and cryptic coloration. P. R. Soc. B. 273:2425-2426.
FAHRIG, L. 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev Ecol. Syst., 34:487-515.
FENNER, M. 1978. A comparison of the abilities of colonizers and closed-turf species to establish from seed in artificial swards, J. Ecol., 66:953-963.
--. 1991. The effects of the parent environment on seed germinability. Seed Sci. Rsch., 1 (2): 75-84.
FIEDLER, P. L. 1987. Life history and population dynamics of rare and common mariposa lilies (Calochortus pursh: Liliaceae). J. Ecol., 75:977-995.
FISCHER, M. AND D. MATTHIES. 1997. Mating structure and inbreeding and outbreeding depression in the rare plant Gentianella germanica (Gentianaceae). Amer. J. Botany, 84 (12): 1685-1692.
--AND--. 1998. Effects of population size on performance in the rare plant Gentianellla germanica. J. Ecol., 86:195-204.
FOSTER, C. O. 1984. Lupines. Horticulture, 5:32-35.
FRANKHAM, R. 2005. Genetics and extinction. Biol. Consent. 126:131-140.
--. C. J. A. Bradshaw, and B. W. Brook. 2014. Genetics in conservation management: Revised recommendations for the 50/500 rules, Red Lisi criteria and population viability analyses. Biol. Consent., 170:56-63
FRANKLIN, I. R. 1980. Evolutionary change in small populations, p. 135-150. In: M. E. Soule and B. A. Wilcox (eds.) Conservation Biology: An Evolutionary-Ecological Perspective, Sinauer, Sunderland, MA.
GERBER, M. A. 1985. The relationship of plant size to self- pollination in Mertensia ciliata. Ecology, 66:762-777.
GLEASON. H. A. AND A. CRONQUIST. 1991. Manual of vascular plants of the Northeastern U.S. and adjacent Canada. 2nd ed. N.Y. Bot. Garden, NY. p. 278.
GORDON, R. B. 1969. The natural vegetation of Ohio in pioneer days. Ohio Biol. Survey Bull. 3(2):1-109.
GREENFIELD, L. M. 1997. Habitat quality and utilization analysis in a spatial context: The case of Lupinus perennis L. and Lycaeides metissa samuelis Nabokov, (Lepidoptera: Lvcaenidae). M.Sc. Thesis, Michigan State University, Lansing, ML 67 p.
GRIGORE, M. T. AND E. TRAMER 1996. The short-term effect of fire on Lupinus perennis (L.). Nat. Area J, 16(1): 41-48.
--AND J. WINDUS. 1994. Decline of the Karner Blue butterfly in the Oak Openings of Northeast Ohio. p. 135-142. In: D. A. Andow, R.J. Baker and C. P. Lane, (eds.) Karner Blue Butterfly, University of Minnesota, St. Paul, MN.
GROOM, M. J. 1998. Allee effects limit population viability of an annual plant. Am. Nat., 151 (6): 487-496.
--. 2001. Consequences of subpopulation isolation for pollination, herbivory, and population growth in Clarhia continua concinna (Onagraceae). Biol Consent., 100(1): 55-63
GROSS, K. L. 1984. Effect of seed size and growth form on seedling establishment of six monocarpic perennial plants. / Ecol, 72:369-387.
HALBUR, M. M., C. M. Sloop, M. J. Zanis, and N. C. Emery. 2014. The population biology of mitigation: impacts of habitat creation on an endangered plant species. Consent Genet., 15(3): 679-695.
HARPER, J. L. 1977. Population biology of plants. London: Academic Press.
HEHR, D. W. 1970. A Comparative study of the composition of the pre-settlement vegetation and the characteristic geologic substrate of the Oak Openings and surrounding areas in northwestern Ohio. M. A. Thesis, Bowling Green State University, Bowling Green, OH.
HENDRIX, S. D. 1984. Variation on seed weight and its effects on germination in Pastinaca sativa L. (Umbelliferae). Am. J. Bol., 71:795-802.
--.AND E. J. TRAPP. 1992. Population demography of Pastinaca sativa (Apiaceae): effects of seed mass on emergence, survival and recruitment. Am. J. Bot., 79(4): 365-375.
HESCHEL, M. S. AND K. N. PAIGE. 1995. Inbreeding depression, environmental stress, and population size variation in Scarlet Cilia (Ipomopsis aggregata). Consent. Biol., 9:126-133.
HINSINGER, P. (1998). How do plant roots acquire mineral nutrients? Chemical processes involved in the rhizosphere. Adv. Agron., 64:225-265.
HOLMES, G. D., E. A. JAMES, AND A. A. HOFFMANN. 2008. Limitations to reproductive output and genetic rescue in populations of the rare shrub Grevillea repens (Proteaceae). Ann. Bot.-London, 102:1031-1041
HOLMSTROM, R. M., J. R. ETTERSON, AND D. J. SCHIMPF. 2010. Dune restoration introduces genetically distinct American beachgrass, Ammophila breviligulata, into a threatened local population. Restar. Ecol., 18:426-43.
HOROVITZ, A. AND J. HARDING. 1983. Genetics of Lupinus. XII. The mating system of Lupinus pilosus. Bol. Caz., 144(2): 276-279.
IFTNER, D. C.,J. A. SHUEY, AND J. CALHOUN. 1992. Butterflies and skippers of Ohio. Bull. Ohio Biol. Survey 9(1): 1-212
JENNERSTEN, O. 1988. Pollination in Dianthus deltoides (Caryophyllaceae): Effects of habitat fragmentation on visitation and seed set. Conserv. Biol., 2:359-366.
KAPPLER, R. H., H. J. MICHAELS, AND K. V. ROOT. 2012. Impact of mice seed predation on wild lupine in and near oak savannas. Amer. Midi. Nat., 168:18-29.
KARRON, J. D. 1989. Breeding systems and levels of inbreeding depression in geographically restricted and widespread species of Astragalus (Fabaceae). Am. J. Bot., 76(3): 331-340.
KELLY, S. D. 1998. Germination preferences and effects of micro-site variation on Lupinus perennis establishment in an oak savanna. M.Sc. Thesis, Bowling Green State University, Bowling Green, OH. 58 p.
KERY, M., D. MATTHIES, AND M. FISCHER. 2001. The effect of plant population size on the interactions between the rare plant Centiana nuciala and its specialized herbivore Maculinea rebeli. J. Ecol., 89:418-427.
--,--, AND H. H. SPILLMAN. 2000. Reduced fecundity and offspring performance in small populations of the declining grassland plants Primula vnis and Gentiana lutea. J. Ecol, 88:17-30.
KLINKHAMER, P. G. L., T.J. DE JONG, AND G.J. DE BRIAN. 1989. Plant size and pollinator visitation in Cynoglossum officinale. Oikos. 54:201-204.
KRAMER, A. T. AND K. HAVENS. 2009. Plant conservation genetics in a changing world. Trends Plant Sci., 14:599-06.
KUNIN, W. E. 1993. Sex and the single mustard--population density and pollinator behavior effects on seed-set. Ecology, 74:2145-2160.
KURKJIAN, H. M., S. K. CAROTHERS, AND E. S.JULES. 2016. Seed predation has the potential to drive a rare plant to extinction. J App. EcoL, edition and page number? doi: 10.1111/1365-2664.12808.
LACY, R. C. 1987. Loss of genetic diversity from managed populations: interacting effects of drift, mutation, immigration, selection and population subdivision. Conserv. Biol., 1(2): 143-158.
LUMBERS, H., J. C. CLEMENTS, AND M. N. NELSON. 2013. How a phosphorus-acquisition strategy based on carboxylate exudation powers the success and agronomic potential of lupines (Lupinus, Fabaceae). Am. J. Bot., 100(2): 263-288.
LAMMI, A., P. SHKAMAKI, AND K. MUSTAJARVI. 1999. Genetic diversity, population size, and fitness in central and peripheral populations of a rare plant Lychnis visearia. Conserv. Biol., 13(5): 1069-1078.
LANDE, R. 1988. Genetics and demography in biological conservation. Science, 241:1455-1460.
LEIMU, R., P. MUTIKAINEN, J. KORIGHEVA, AND M. FISCHER. 2006. How general are positive relationships between plant population size, fitness and genetic variation? J. Ecology, 4:942-952
MANNOURIS, C. AND D. E. BYERS. 2013. The impact of habitat fragmentation on fitness-related traits in a native prairie plant, Chamaecristafasciculata (Fabaceae). Biol. Jour. Linn. Soe., 108:55-67.
MARRON, J. L. AND E. L. SIMMS. 1997. Effect of seed predation on seed bank size and seedling recruitment of bush lupine (Lupinus arboreus). Oecologia, 111:76-83.
--AND--. 2001. Rodent-limited establishment of bush lupine: field experiments on the cumulative effect of granivory. J. Ecol., 89:578-588,
MATESANZ, S., M. L. RUBIO TESO, A. GARCIA-FERNANDEZ, AND A. ESCUDERO. 2017. Habitats fragmentation differentials effects genetic variation, phenotypic plasticity, and survival in populations of a gypsum endemic. Front. Plant Sci.. 8:843
MAYFIELD, H. 1962. Changes in the natural history of the Toledo region since the coming of the white man. Northwest Ohio Quarterly, 34(2): 82-104.
MENGES, E. S. 1991. Stochastic modeling of extinction in plant populations, page numbers. In: P. E. Fiedler and S. Jain, (eds.) ('.onsen-. Biol.: the theory and practice of nature conservation, population, preservation, and management. Chapman and Hall, New York.
--1991a. Seed germination percentage increases with population size in a fragmented prairie species. Consent. Biol. 5:158-104.
--AND R. W. DOLAN. 1998. Demographic viability of populations of Silene regia in midwestern prairies: relationships with fire management, genetic variation, geographic location, population size and isolation. J Ecol., 86(l):63-78.
MICHAELS, H. J., X.J. SIN, AND R. J. MITCHELL. 2008. Effects of population size on performance and inbreeding depression in Lupinus perennis. Oecologia, 154:651-661.
MILBERG, P. AND A. BERTILSSON. 1997. What determines seed set in Dracocephalum ryuschiana E. an endangered grassland plant. Flora, 192(4):361-367.
MITCHELL-OLDS, T. AND J. BERGELSON. 1990. Statistical genetics of an annual plant Impatiens capensis. II. Natural selection. Genetics, 124:417-421.
MORGAN, J. W. 1999. Effects of population size on seed production and germinability in an endangered, fragmented grassland plant. Consent. Biol.. 13(2):266-273.
MOSELEY, E. L., 1928. Flora of the Oak openings. Ohio Academy of Science Special Paper 20:79-134.
NELSON, M. N., H. T. T. PHAN, S. R. ELLWOOD, P. M. MOOLHUIJZEN, J. H.ANE, A. WILLIAMS, C. E. O'LONE, J. FOSU-NYARKO, M. SCOBIE, M. CAKIR, M. G. K.JONES, M. BELLGARD, M. KSIARKJEWICZ, B. WOLKO, S.J. BARKER, R. P. OLIVER, AND W. A. COWLING. 2006. The first gene-based map of Lupinus angustifolius L.- location of domestication genes and conserved syntenv with Medicago tmnculala. Theor. Appt. tenet., 113(2):225-225.
NYSTRAND, O. AND A. GRANSTROM. 1997. Post-dispersal predation on Pinns sylvestris seeds by Friitgilla spp: ground substrate affects selection for seed color. Oecologia. 110:353-359.
OAKLEY, C. G. 2015. The influence of natural variation in population size on ecological and quantitative genetics of the endangered endemic plant Hypericum cumulicola. Int. J. Plant Sci., 176:11-19.
OHIO DEPARTMENT OF NATURAL RESOURCES, 2006. Division of Natural Areas and Preserves, Rare Nativ e Ohio Plants. Columbus, Ohio.
OOSTERMEIJER, J. G. B., M. W. VAN EIJCK, AND J. C. M. DEN NIJS. 1994. Offspring fitness in relation to population size and genetic variation on the rare perennial plant species Gentiana pneumonanthe (Gentianaceae). Oecologia, 97:289-296.
OUBORG, N.J., P. VERGEER, AND C. MIX C. 2006. The rough edges of the conservation genetics paradigm for plants. J. Ecol., 94:1233-1248
PETERSON. D. W. AND P. B. REICH. 2001. Prescribed fire in oak savanna: fire frequency effects on stand structure and dynamics. F. col. App., 11:914-927.
PENFIELD, S. AND D. R. MACGREGOR. 2017. Effects of environmental variation during seed production on seed dormancy and germination. J. Exp Bot.. 68(4):819-825.
PLENZLER, M. A. AND H. J. MICHAELS. 2015. Seedling recruitment and establishment of Lupinus perennis in a mixed-management landscape. Nat. Area J., 35(2):224-234.
PORTER, S. S. 2013. Adaptive divergence in seed color camouflage in contrasting soil environments. New Phytol.. 197:1311-1320.
QUINN, J. F. AND A. HASTINGS. 1987. Extinction in subdivided habitats. Consent. Biol., 1:198-208.
RAMSEY, M. 1998. The effect of environment on the magnitude of inbreeding depression in seed germination in a partially self-fertile perennial herb (Blandfordia grandiflora, Eilliaceae). Int. J. Plant Sci., 159(1):98-104.
RATHCKE, B. 1983. Competition and facilitation among plants for pollination, p. 305-329. In: E. Real, (ed.). Pollination Biology, Academic Press, New York.
RAIJMANN, L. E. L. 1994. Genetic variation and outcrossing rate in relation to population size in Gentiana pneumonanthe L. Conseru. Biol., 8:1014-1026.
RICHARDS, C. M. 2000. Inbreeding depression and genetic rescue in a plant raetapopulation. Am. Natur., 155(3): 383-394.
ROACH, D. A. AND R. D. WULFF. 1987. Maternal effects in plants. Ann. Rev. Ecol. Syst., 18:209-235.
ROBINSON, G. R. AND J. F. QUINN. 1988. Extinction, turnover and species diversity in an experimentally fragmented California annual grassland. Oecologia, 76:71-82.
ROLL, J. R., R. J. MITCHELL, R. J. CABIN, AND D. L. MARSHALL. 1997. Reproductive success increases with local density of conspecifics in the desert mustard Lesquerella fendleri. Consnv. Biol., 11:738-746
SCHAAL, B. A. 1980. Reproductive capacity and seed size in Lupinus texensis. Amer. J. Bot. 67(5):703-709.
SEARS, P. 15. 1926. The natural vegetation of Ohio II: The prairies of Ohio. Ohio J. Sri., 62:128-146.
SEVERNS, P. 2003. Inbreeding and small population size reduce seed set in a threatened and fragmented plant species, Lupinus sulphureus ssp. Kincaidii (Fabaceae). Biol. Conserv., 110:221-229.
SHAPIRO, A. M. 1974. Partitioning of resources among lupine-feeding Lepidoptera. Am. Mid. Natur., 91(1):243-248.
SCHETTER, T. A. AND K. V. ROOT. 2011. Assessing an imperiled oak savanna landscape in northwestern Ohio using Landsat data. Nat. Area J.. 21:118-130
SHI, X.J., H.J. MICHAELS, AND R.J. MITCHELL. 2005. Effects of self-pollination and maternal resources on reproduction and offspring performance in the wild lupine, Lupinus perennis (Fabaceae). Sex. Plant Reprod. 18(2):55-64.
SHI, X. 2004. Inbreeding and inbreeding depression in the wild lupine. Ph.D. Dissertation, Bowling Green State University. Bowling Green, OH. 127 pp.
SUUEY, J. A., J. V. CALHOUN, AND D. C. IFTNER. 1987. Butterflies that are endangered, threatened and of special concern in Ohio. Ohio J. Sci., 87(4):98-106.
SIH, A. AND M. BALTUS. 1987. Patch size, pollinator behavior and pollinator limitation in catnip. Ecology, 68:1679-1690.
SILVERTOWN, J. 1984. Phenotypic variety in seed germination behaviour: the ontogeny and evolution of somatic polymorphism in seeds. Amer. Natur., 124:1-16.
SMALLIDGE, P. J., D. J. LEOPOLD, AND C. J. ALLEN. 1996. Community characteristics and vegetation management of Karner blue butterfly (Lyraeides melissa samuelis) habitats on rights-of-way in east-central New York, USA. J. Appt. Ecol., 33:1405-1419.
SOLLE, M. E. AND B. A. WILCOX. 1980. Conservation biology: an evolutionary-ecological perspective. Sinauer Associates, Sunderland, Massachusetts.
STEVENS, M. 2007. Predator perception and the interrelation between different forms of protective coloration. Proc. Roy. Sor. B. 274:1457-1464.
TANG, C., N. E. LONGNECKER, C. J. THOMSON, H. GREENWAY, AND A. D. ROBSON. 1992. Lupin (Lupinus angustifolius L.) and pea (Pisum sativum L.) root differ in their sensitivity to pH above 6.0. J. Plant Physiol., 140:715-719.
U. S. FISH AND WILDLIFE SERVICE. 1992. Department of the Interior. Endangered and Threatened Wildlife and Plants. Determination of endangered slatus for the Karner Blue butterfly. Final rule. Federal Register Dec. 14, 1992, 52(240):59236-44.
VANDENBERG, A. AND A. E. SLINKARD. 1990. Genetics of seed coat color in lentil. Journal of Heredity, 81(6): 484-488.
VON WETTBERG, E. B., P.L. CHANG, F. BASDEMIRE, N. CARRASQUILA-GARCIA, L. B. KORBU, S. M. MOENGA, G. BEDADA, A. GREENLON, etc. 2018. Ecology and community genomics of an important crop wild relative as a prelude to agricultural innovation. Nat. Commun., 9:649.
WEBER, A. AND M. HUFFMAN. 1989. Making a comeback: Ohio's Oak Openings. The Nature Conservancy, May: 17-21.
WILLI, V. AND M. FISCHER. 2005. Genetic rescue in interconnected populations of small and large size of the self-incompatible Ranunculus reptans. Heredity, 95:437-443.
--. J. VAN BUSKIRK J, AND A. A. HOFFMANN. 2006. Limits to the Adaptive Potential of Small Populations. Anna. Rev. Ecol,, Evol. S., 37:433-58.
--, --, B. SCHMID, AND M. FISCHER. 2007. Genetic isolation of fragmented populations is exacerbated bv drift and selection. J. Evol. Biol., 20(2):534-542
YOUNG, A. G., T. BOYLE, AND A. H. D. BROWN. 1996. The population genetic consequences of habitat fragmentation for plants. Trends. Ecol. Evol., 11:413-418.
--, A. H. D. BROWN, B. G. MURRAY. P. H. THRALL, AND C. MILLER. 2000. Genetic erosion, restricted mating and reduced viability in fragmented populations of the endangered grassland herb Rutidosis leptorrhynchiodes. p. 335-360. In: A. G. Young and G. M. Clarke, (eds.) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, U. K.
SUBMITTED 9 JANUARY 2019
ACCEPTED 14 JUNE 2019
HELEN J. MICHAELS (1), CARRIE A. CARTWRIGHT (2) and ELLEN F. WAKELEY TOMLINSON (3)
Department of Biological Sciences, Bending Green Stale University, Bowling Green, Ohio 43403
(1) Corresponding author: Telephone: (419) 372-2644; FAX: (419) 372-2024: E-mail: firstname.lastname@example.org
(2) Present address: Dept. of Investigational Theraputics, 1515 Holcombe Blvd., MD Anderson Cancer Center, Houston, Texas 77030
(3) Present address: Office of Legacy Management, LTS Department of Energy, Grand Junction, Colorado 81503
Caption: Fig. 1.--The effect of population size on mean number of pods produced per stem (fruit set) for 16 populations of Lupinus perennis in Ohio and Michigan, U.S.A. y = 1.06x - 1.62, [R.sup.2] = 0.64
Caption: Fig. 2.--The relationship between population size and mean number of seeds produced per plant for 16 populations of Lupinus perennis in Ohio and Michigan, U.S.A. Seed production samples are based on seeds collected from up to three inflorescences per plant, and therefore may underestimate seed production of any large individuals, y = 5.83x - 11.77, [R.sup.2] = 0.44
Caption: Fig. 3.--The effect of population size on seed emergence in the greenhouse. Germination proportion for the largest population may have declined because seeds from this location were sampled after peak seed dispersal due to restricted access to this site
Caption: Fig. 4.--Proportion of seed coat colors within each of 16 populations of Lupinus perennis. Seed color variation was visually divided into six color classes. Approximately 30% of seeds in populations are white or lightly speckled, although smaller populations are often missing color types with greater amounts of speckling
TABLE 1.--Population numbers, U.S.A. county and state, (IPS coordinates for population location, population abbreviation, and estimated number of plants in each population Population County State Latitude Longitude Code Size 1 Wood OH 41.23179 -83.45864 WPR 25 2 Henry OH 41.47135 -83.89253 HNR 90 3 Ashtabula OH 41.92936 -80.65155 APN 100 4 Henry OH 41.48374 -83.89696 HC2 195 5 Wood OH 41.21806 -83.47018 W235 300 6 Fulton OH 41.50952 -83.90298 FC2 400 7 Ashtabula OH 41.92822 -80.65182 APS 450 8 Lucas OH 41.37700 -83.43001 LHO 500 9 Monroe MI 41.87491 -83.69431 MSP 600 10 Lucas OH 41.53610 -83.84557 LMV 1800 11 Lucas OH 41.38761 -83.43363 LKN 3000 12 Lucas OH 41.59169 -83.77724 LCPV 3100 13 Lucas OH 41.53499 -83.83832 LCPR 3200 14 Lucas OH 41.38401 -83.45922 LMR 4500 15 Lucas OH 41.61830 -83.78563 LKT 6300 16 Lucas OH 41.58532 -83.80454 LAI5 8300 TABLE 2.--Nested ANOVA for the relationship between population source and maternal plants and proportion of seeds emerging after 7 wk for 12 populations. The analysis uses arcsin square root transformed germination proportions. Effect of population size was tested against variation among populations, and variation among populations was tested against variation among 188 maternal plants Model: emergence = population + plants (population) [r.sup.2] = Source d.f. MS F 0.69 P Population size class 1 0.607 2.57 0.140 Population 10 0.236 1.56 0.121 Plants (Population) 187 0.151 4.36 <0.001 Error 398 0.035 TABLE 3.--Regression analyses for effects of environmental factors on Lupine reproductive success. Analyses used In transformed population size and arcsin square root transformed per cent bare ground and per cent emergence. (*) P < 0.1; * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001 Model [r.sup.2] Source df Pods/Stem = Pop. Size + 0.81 Regression 9 pH Error 11 Seeds/Pod = Pop. Size + 0.69 Regression 2 Bare Ground Error 11 Emergence = Pop. Size + 0.52 Regression 2 Bare Ground Error 9 Seedling Size = pH + 0.60 Regression 2 Light Error 8 Model F P Variable P Pods/Stem = Pop. Size + 23.80 **** Pop. Size **** pH 13.95 pH ** Seeds/Pod = Pop. Size + 12.28 ** Pop. Size ** Bare Ground bare ground ** Emergence = Pop. Size + 4.78 * pop. size * Bare Ground bare ground * Seedling Size = pH + 6.04 * pH ** Light light *
|Printer friendly Cite/link Email Feedback|
|Author:||Michaels, Helen J.; Cartwright, Carrie A.; Tomlinson, Ellen F. Wakeley|
|Publication:||The American Midland Naturalist|
|Date:||Oct 1, 2019|
|Previous Article:||Habitat Preference and Phenology of Nest Seeking and Foraging Spring Bumble Bee Queens in Northeastern North America (Hymenoptera: Apidae: Bombus).|
|Next Article:||Gopher Tortoise (Gopherus polyphemus) Gut Passage Can Alter Seed Germinability.|