SEARCH

SEARCH BY CITATION

Keywords:

  • bottleneck;
  • Campanulastrum americanum;
  • genetic drift;
  • inbreeding depression;
  • peripheral populations;
  • range expansion

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Many temperate taxa were confined to warmer latitudes during the last glacial maximum. As their ranges expanded when climates warmed, genetic drift and inbreeding in relatively small peripheral populations are expected to have reduced genetic diversity and the segregating genetic load. Therefore, inbreeding depression in peripheral populations might be lower than in centrally located sites. We evaluated the consequences of inbreeding for fitness traits in six central and six northern peripheral populations of the herb Campanulastrum americanum. Inbreeding reduced performance for all traits. Inbreeding depression in peripheral populations was lower than in central populations. This difference increased across the life cycle from similar levels for germination, to central populations having three times the inbreeding depression for adult traits. Geographical patterns of inbreeding depression suggest that mating system variation and potential future mating system evolution in many temperate taxa might reflect, at least in part, nonequilibrium conditions associated with historic range changes.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Contemporary distributions of animal and plant taxa are broadly the result of complex histories of range expansion and contraction over long periods of time (Hoffmann & Blows, 1994; Brown et al., 1996; Gaston, 2003; Yasuda et al., 2005; Sexton et al., 2009). Following the end of the Pleistocene in North America, many taxa migrated and expanded their ranges northwards from refugial populations in the warmer southern latitudes as land became available after the retreat of the Wisconsin Ice Sheet; similar migrations occurred in Asia, North Africa, and Europe (Brunsfeld et al., 2001; Davis & Shaw, 2001; Abbott & Brochmann, 2003; Schonswetter et al., 2005; Soltis et al., 2006; Godbout et al., 2010; Morris et al., 2010; Rebernig et al., 2010). Large-scale range expansions such as these are not without ecological and evolutionary consequences. For example, levels of genetic diversity are expected to be lower in northern relative to southern populations due to the effects of genetic drift acting via repeated bottlenecks and/or founder events (Hewitt, 1996, 2000; Ibrahim et al., 1996; Excoffier et al., 2009). Consistent with this expectation are studies that have found that peripheral populations harbour less genetic variation (Cwynar & MacDonald, 1987; Allen et al., 1996; but see Yakimowski & Eckert, 2008) and are less able to respond to selection (Pujol & Pannell, 2008).

Inbreeding depression is the reduction in fitness of inbred relative to outcrossed individuals resulting from self-fertilization or mating among relatives. Numerous forces can influence levels of inbreeding depression, including heterozygote frequency and/or interactions among loci, the number and relative proportions of deleterious alleles with large versus small effects on fitness, growth form, parity, selfing rate, life stage, and ploidy (Charlesworth & Charlesworth, 1987, 1999; Husband & Schemske, 1996; Morgan, 2001; Carr & Dudash, 2003; Scofield & Schultz, 2005; Barringer & Geber, 2008). However, studies suggest that most inbreeding depression is due to the expression of recessive deleterious alleles whose fitness effects are unmasked as genome-wide levels of homozygosity increase as a result of inbreeding (Charlesworth & Charlesworth, 1987, 1999; Carr & Dudash, 2003). Interestingly, this process facilitates the selective elimination (i.e. purging) of deleterious alleles such that inbreeding depression can decline over time (Darwin, 1876; Lande & Schemske, 1985; Husband & Schemske, 1996; Charlesworth & Charlesworth, 1990; Crnokrak & Barrett, 2002; but see Charlesworth et al., 1990; Byers & Waller, 1999). Indeed, taxa that reproduce predominantly via self-fertilization often exhibit relatively low levels of inbreeding depression (Husband & Schemske, 1996).

The relationship between inbreeding depression and range expansion is not well understood; though, it has the potential to influence both the evolution of mating systems in peripheral populations and a taxon’s ability to successfully colonize new areas. There are several reasons why we expect inbreeding depression to change with range expansion. First, the repeated bottlenecks and founder events associated with migration are expected to reduce overall levels of genetic variation due to the loss of rare alleles during the sampling process. Because the deleterious recessive alleles that underlie inbreeding depression are often rare, they are expected to be highly susceptible to loss via sampling error in founder events. At the same time, any deleterious recessive alleles that are included in colonizing populations will be at an increased frequency and therefore are less likely to be lost due to drift. In fact, the frequency of deleterious alleles harboured by small peripheral populations may increase via genetic drift despite the selection against them (Baker & Stebbins, 1965; Ibrahim et al., 1996; Excoffier et al., 2009), and such populations might actually exhibit an increased genetic load (Bataillon & Kirkpatrick, 2000). However, in these cases, levels of inbreeding depression would still decrease because the fixation of alleles due to genetic drift means there is little increase in homozygosity following inbreeding (Paland & Schmid, 2003). Indeed, a previous study found that in small populations most genetic load represents fixed alleles, whereas in large populations it is due to segregating deleterious alleles (Paland & Schmid, 2003).

A number of studies have found evidence for a reduction in levels of inbreeding depression in serially inbred populations (Crnokrak & Barrett, 2002), and population bottlenecks, which can lead to increased levels of inbreeding, can result in a decreased genetic load (Lynch & Walsh, 1998; Keller & Waller, 2002). However, few studies have explicitly addressed whether range expansion is correlated with reduced inbreeding depression among peripheral populations (but see Pujol et al., 2009), and no studies have done so for fitness traits throughout the life cycle.

We tested the hypothesis that inbreeding depression is lower in peripheral relative to central populations of the temperate North American herb Campanulastrum americanum (American bellflower). A preliminary phylogeography of this species (K. Barnard-Kubow & L.F. Galloway, unpublished data) suggests that C. americanum, like many other taxa in eastern North America, was confined to southern latitudes, perhaps including south-eastern or Appalachian refugia, before migrating northwards after glacial retreat at the end of the Pleistocene (McLachlan et al. 2005; Soltis et al., 2006; Godbout et al., 2010; Morris et al., 2010). Therefore, levels of genetic diversity and the segregating genetic load of populations presently located on or near the northern edge of the species’ range might differ substantially from those present in more centrally located populations. To address our hypothesis, we compared fitness traits of inbred and outcrossed half-siblings from 12 populations: six near the centre of the species’ natural range and six near the northern range margin. We did not quantify selfing rates; however, studies suggest that C. americanum reproduces predominantly via outcrossing (Galloway et al., 2003) due to the combined actions of protandry (Evanhoe & Galloway, 2002) and cryptic self-incompatibility (Kruszewski & Galloway, 2006). Outcrossed individuals performed better than inbred individuals for all fitness traits in the populations studied. Consistent with our hypothesis, northern peripheral populations had reduced inbreeding depression relative to more central populations, and this pattern was greater at late life stages.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Study species

Campanulastrum americanum (L.) Small (= Campanula americana L., Campanulaceae) is a monocarpic autotetraploid herb native to the eastern half of North America and associated with disturbed habitats and forest edges. Populations can be relatively small (i.e. 10–20 individuals) and ephemeral or larger (i.e. many hundreds of individuals) and stable for decades depending on habitat persistence. Seeds are gravity dispersed and germinate in the fall or spring. After germination, plants overwinter as rosettes before bolting and producing flowers the following summer. Flowers are self-compatible but protandrous, and the production of inbred seed in nature appears to be rare (multilocus outcrossing rate for a Virginia population, tm = 0.94; Galloway et al., 2003). Limited inbred seed is likely due in part to cryptic self-incompatibility, whereby a majority of outcrossed seed is produced when self and outcross pollen are placed in equal proportions on the stigma (Kruszewski & Galloway, 2006). Flowers are pollinated by insects (mostly bumblebees and halictids) with the majority of flowers occurring from mid-July through the end of August.

Experimental design

Seeds were harvested from six central and six northern populations (Fig. 1 and Appendix 1) in fall 2008. For a given site, seeds were collected from between 18 and 32 (mean = 25.9, SD = 3.8) maternal families distributed throughout the population. Thirty seeds, distributed equally among the maternal families (mean = 1.5, SD = 0.5 seeds per family) in each population, were selected, individually sown in a randomized design in plug trays filled with MetroMix, and germinated under near-optimal conditions in a growth chamber (21 °C day/14 °C night, 12 h days). Twenty-seven days after planting the temperature was decreased and seedlings were grown for 45 days at 5 °C day/5 °C night (12 h days) to simulate winter conditions and to promote vernalization. Individuals were then transplanted into Cone-tainers (Stuewe and Sons, Tangent, OR) filled with a 3 : 1 ratio of MetroMix and fritted clay and moved into a greenhouse. In the greenhouse, the plants were grown under extended daylight conditions (16 h days) and watered twice each day. Supplemental fertilizer (24N:8P:16K) was added to the water once per week.

image

Figure 1.  Range of Campanulastrum americanum in North America (shaded grey; http://plants.usda.gov/) and the 12 populations used in this study. Open circles represent the six populations in the central region of the species’ range; filled circles represent the six populations along the northern periphery. See Appendix 1 for detailed population locations.

Download figure to PowerPoint

To produce inbred and outcrossed offspring, four flowers on each individual plant were hand-pollinated: two were self-fertilized and two were outcrossed. All flowers used were emasculated prior to anthesis. Selfed flowers were pollinated using pollen obtained from a different flower on the same individual; outcrossed flowers received pollen from two randomly selected donors in the same population. For both cross types (self and outcross), anthers from pollen donor flowers were removed and applied directly to a recipient flower’s stigma (using forceps) to saturate the surface. Variation among individuals in terms of initial flower production precluded our ability to conduct all crosses simultaneously; however, all crosses within a given population were conducted within 22 (mean = 9.2, SD = 5.7) days of flowering initiation. After maturation (approximately 35 days after pollination), fruits were harvested and stored in coin envelopes until used.

Fitness traits were compared between inbred and outcrossed half-siblings. Forty seeds per cross type per population (960 total) were individually sown in a randomized design and germinated as in the previous generation. Seeds were evenly distributed among families (9–27 families/population; mean = 16, SD = 6.0) for an average of 2.3 (SD = 1.0) seeds per family per cross type. Germination was assessed after emergence was complete (approximately 21 days after seeds were sown). Plants were then vernalized for 45 days before being transplanted into Cone-tainers, moved into a greenhouse under extended daylight conditions, and watered and fertilized as in the previous generation. Rosette size, the product of the width of the largest leaf and the number of leaves present, was measured when plants were moved into the greenhouse. However, this trait was excluded from analysis because it was tightly correlated with flower production (R = 0.62). Whether or not a given plant survived to produce flowers was recorded. Flower number was assessed by recording the number of open flowers on each plant once/week, starting with the initiation of flowering and continuing for 4 weeks. These four counts were summed to produce an index of total flower number. In the greenhouse, individual flowers last approximately 6 days (Evanhoe & Galloway, 2002), suggesting that weekly flower counts reflect most of the flowers produced. Floral counts were conducted for 4 weeks because in nature, most of an individual’s flowers are produced within 4 weeks of initiating flowering (Galloway & Burgess, 2009, 2012).

Statistical analysis

Inbred and outcrossed offspring were compared for fitness components representing three distinct life stages: germination, survival to flowering and flower number. We also compared cumulative fitness, estimated as the product of proportion germination × survival to flower  × flower number. Fitness components were compared between cross types (self-fertilized vs. outcrossed), populations in the centre and northern margin of the species range (region), populations nested within each region, and families nested within each population as well as interactions between cross and region, and cross and population using a nested anova (PROC Mixed; SAS Institute., 2009). In the analysis, region, cross type and their interactions were treated as fixed effects and population nested within region, family and population by cross-type interaction were treated as random effects. The denominator degrees of freedom used to test the fixed effects are limited due to the nested nature of the design and reflect the number of populations included in the study. Proportion germination and survival were calculated for each family. Analysis was conducted on these family-level values; therefore, the effect of family was not included in the model. Log-linear analysis of individual-level traits provided similar results (not shown) but were not as robust because germination and survival approached unity in some populations. Assumptions of normality and homogeneity of variance were assessed by visual inspection of residual and normal probability plots; none of the traits required transformation to meet assumptions.

Inbreeding depression, δ, is often defined as 1 − wS/wO, where wS and wO are the mean phenotypic values of traits for inbred and outcrossed individuals, respectively. However, this method does not yield a symmetrical distribution of δ around zero when inbred individuals outperform their outcrossed relatives (i.e. outbreeding depression), as was observed for some families and traits in our study. Therefore, we compared the fitness of inbred and outcrossed individuals using a measure of relative performance (RP), defined as RP = 1 − wS/wO when wS ≤ wO, and RP = wO/wS − 1 when wS > wO (Argen & Schemske, 1993; Dudash et al., 1997; Barringer & Geber, 2008). This method yields estimates of RP that are identical to traditional estimates of inbreeding depression when outcrossed progeny outperform their inbred relatives; however, it applies equal weight to outcomes in which inbred individuals have the highest fitness (i.e. outbreeding depression). For this reason, it is a useful method with which to compute and compare unbiased estimates of the mean effects of inbreeding on fitness.

We estimated inbreeding depression for each population for fitness components and cumulative fitness. Initially, family-level values of inbreeding depression were calculated using family mean wO and wS. We then estimated inbreeding depression at the population level by averaging the inbreeding depression estimates across families for each population. Population-level inbreeding depression estimates were compared between central and northern populations for all fitness components and for cumulative fitness using a t-test (SAS Institute., 2009). Analyses of family-level estimates (not shown) produced similar patterns of statistical results, but were less precise due to limited or no replication within families (especially in later life traits).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

The difference in fitness between inbred and outcrossed progeny was generally greater for later relative to earlier life stages (Table 1a; Fig. 2). On average, inbred individuals germinated less frequently, had lower rates of survival, produced fewer flowers and exhibited reduced cumulative fitness than outcrossed individuals (Table 1a, Fig. 2). The effect of cross type on these fitness components was similar across populations (Table 1a). However, the effect of cross type differed between regions for later life-stage traits. Relative to outcrossed plants, inbred plants from central populations exhibited a greater reduction in both flower number and cumulative fitness compared to inbred plants from northern peripheral populations (Table 1a, Fig. 2).

Table 1.    (a) Analysis of variance of proportion germination, survival to flowering, flower number and cumulative fitness for inbred and outcrossed Campanulastrum americanum from six populations in the centre region and six populations near the northern limit of the species’ range. F-values are reported for fixed effects (cross type and region) and Z-values for random effects (population nested within region and family nested within population). Proportion germination and survival of each family were analysed; therefore, family was not included in the model. (b) t-test comparing population means of inbreeding depression between central and northern regions.
Sourced.f.GerminationSurvivalFlower Num.Cumulative Fitness
  1. Numerator degrees of freedom, Denominator degrees of freedom.

  2. *P < 0.05; **P < 0.01; ***P < 0.001; +P = 0.06.

(a) Comparison of cross type
Cross1, 105.57*10.66*91.27***72.97***
Region1, 100.701.9135.83***14.93**
Pop (Region)0.00.700.00.0
Family (Pop, Region)2.72**3.37***
Cross × Region1, 102.952.1024.64***4.56+
Cross × Pop (Region)0.00.00.00.14
(b) Comparison of inbreeding depression
Region10−0.523.47**6.14**2.10+
image

Figure 2.  Mean (±1 SE) (a) proportion germination (b) survival to flowering, (c) flower number, and (d) cumulative fitness of outcrossed and inbred Campanulastrum half-siblings from six populations in the centre and six populations at the northern periphery of the species’ range. See Table 1a for statistical analysis.

Download figure to PowerPoint

Northern populations exhibited less inbreeding depression than populations located near the centre of the species’ range. There was no difference in inbreeding depression for proportion germination between central and northern regions (Tables 1b; Fig. 3). However, levels of inbreeding depression for survival were significantly greater in central relative to northern regions (Table 1b, Fig. 3). Similarly, flower number exhibited greater inbreeding depression in central relative to northern populations (Table 1b; Fig. 3). Finally, for cumulative fitness central populations showed a near-significant (P < 0.06, Table 1b) greater inbreeding depression in central relative to northern populations (Fig. 3).

image

Figure 3.  Mean (±1 SE) levels of inbreeding depression for proportion germination, survival to flowering, flower number, and cumulative fitness for populations of Campanulastrum americanum growing near the centre and northern periphery of the species’ range. See Table 1b for statistical analysis.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Our results supported the prediction that northern peripheral populations of Campanulastrum americanum would exhibit less inbreeding depression than central populations. Although there was a reduction in performance of inbred individuals for all traits, the pattern of greater inbreeding depression in central populations was strongest for traits that manifest themselves late in life. Specifically, inbreeding depression for both survival and flower number was three times greater in central than peripheral populations. Cumulative fitness, perhaps expected to be the culmination of this pattern as it had the largest inbreeding depression, showed less difference between northern peripheral and central populations because it combined the distinct inbreeding depression of the later life traits with the reduced, and more variable, inbreeding depression found for proportion germination. Collectively, our data suggest that range expansion in this species facilitated an erosion of variation responsible for inbreeding depression via genetic drift, selective purging or a combination of both evolutionary forces.

Both theoretical and empirical studies suggest that deleterious alleles affecting early life stages are purged more efficiently than those affecting traits later in life (Lande & Schemske, 1985; Schemske & Lande, 1985; Charlesworth & Charlesworth, 1987, 1999; Husband & Schemske, 1996, 1997; Carr & Dudash, 2003) because most organisms experience a reduction in fecundity as they age (Williams, 1957; Charlesworth, 1993; Roach, 1993). Indeed, two previous studies of inbreeding depression in C. americanum found higher levels of inbreeding depression at later life stages (Galloway et al., 2003; Galloway & Etterson, 2007). In our study, inbreeding depression for proportion germination was relatively small and did not differ between central and northern regions. In contrast, inbreeding depression of later life traits was 2.5 times larger than of germination for central populations while comparable or slightly smaller for northern populations. The uniformly low values of inbreeding depression for proportion germination, regardless of population location, suggest that selection has acted to purge deleterious alleles for this early life trait. It is possible that this purging occurred prior to range expansion. Alternatively, the low levels of inbreeding depression for proportion germination may reflect purging in central populations but drift in the northern peripheral populations.

The reduction in levels of inbreeding depression in northern C. americanum populations for later life traits suggests a genome-wide loss of genetic diversity through drift. Previous work on a different monocarpic herb found that fixation of mildly deleterious alleles within populations reduced inbreeding depression in small populations (Paland & Schmid, 2003). If this occurred during range expansion, we would predict enhanced performance of between-population crosses, which would restore variation, relative to within population crosses. Contrary to this prediction, previous work in C. americanum found almost no heterosis in several between-population crosses (Galloway & Etterson, 2005; Barnard-Kubow & Galloway, unpublished data). However, the crosses required to test this hypothesis (i.e. between northern peripheral populations) were not conducted. A greater contribution of drift to northern than to central populations might also be expected to lead to a reduction in fitness of the northern populations due to the fixation of mildly deleterious mutations. However, there was no evidence for fitness differences between populations from different regions. Regardless, three times greater inbreeding depression in central populations than northern populations is perhaps surprising given that drift is expected to be a less potent evolutionary force in autotetraploids such as C. americanum (Moody et al. 1993).

Increased levels of self-fertilization often occur within populations located along range edges (Baker, 1967; Schoen et al., 1996; Barrett, 2002), a pattern commonly attributed to selection for reproductive assurance in regions where pollinators and/or conspecifics might be relatively rare. In studied eastern populations, C. americanum are fully self-compatible, but favour outcross pollen if it is available (i.e. cryptic self-incompatibility, Kruszewski & Galloway, 2006). This mechanism of outcrossing suggests that C. americanum’s mating system is labile, highly outcrossing when possible (Galloway et al., 2003), but with the potential to self-fertilize if no outcross pollen is available. Therefore, there may be limited selection for an altered mating system to enhance reproductive assurance. It is also possible that selection for reduced cryptic self-incompatibility has occurred at the periphery of the species’ range; however, this has not been tested. Across C. americanum’s range, floral morphology, protandry, pollinators and population sizes are comparable, providing no evidence for shifts in mating system (Galloway, personal observation). If range-wide C. americanum is predominantly outcrossing, comparable to eastern populations (Galloway et al., 2003), our results indicate that inbreeding depression can be reduced in peripheral populations, even in the absence of changes in the mating system, a result consistent with that found in European populations of Mercurialis annua (Pujol et al., 2009). Alternatively, the mating system might vary among C. americanum populations, with reduced levels of inbreeding depression in peripheral populations facilitating the invasion of alleles that positively influence selfing rates. In sum, mating system evolution in peripheral populations might be more complex than commonly believed, as any selection for reproductive assurance occurs in the context of genetic changes associated with a species’ history of range expansion.

Campanulastrum americanum is an autotetraploid (Gadella, 1964; Galloway et al., 2003); autopolyploids are commonly predicted to exhibit relatively low levels of inbreeding depression owing to the buffering effects of multiple genomes (Lande & Schemske, 1985; Husband & Schemske, 1997; Ronfort, 1999; Barringer & Geber, 2008). The relatively low levels of inbreeding depression reported here and found in three Virginia populations of C. americanum (Galloway et al., 2003) are consistent with this prediction. However, this work was conducted on plants growing in greenhouses where environmental conditions are near optimal; studies suggest that inbreeding depression is frequently greater in natural settings (Dudash, 1990; Jimenez et al., 1994; Miller, 1994; Crnokrak & Roff, 1999; Armbruster & Reed, 2005; but see Armbruster et al., 2000; Swindell & Bouzat, 2006). Indeed, substantially greater levels of inbreeding depression were found when it was measured on C. americanum plants growing in nature (Galloway & Etterson, 2007). Although our estimates of inbreeding depression might be conservative relative to those found in natural populations, there is no a priori reason to suspect that the relative differences in inbreeding depression between central and northern populations would be lower in nature, where harsher environments tend to exacerbate genetic differences.

The relationship between inbreeding depression and range expansion is complex and influenced by numerous – and sometimes conflicting – evolutionary and ecological forces, including both the need to adapt to changing biotic and abiotic environments and the ability to access mates in peripheral populations where conspecific and/or pollinator densities may be low (Hoffmann & Blows, 1994; Brown et al., 1996; Gaston, 2003; Yasuda et al., 2005; Sexton et al., 2009). Range expansion is often associated with repeated colonization events leading to a reduction in genetic variation and an increase in the genetic load due to the combined action of genetic drift and inbreeding in relatively small and isolated populations. To the extent that these processes reduce levels of inbreeding depression and facilitate the evolution of self-fertilization and/or mating among relatives, a reduction in levels of genetic variation might promote range expansion by ensuring reproductive success along range edges. However, decreased genetic variation in peripheral populations does not come without a cost. Indeed, a reduction in levels of genetic variation in northward-migrating populations of purple loosestrife (Lythrum salicaria) was found to limit local adaptation and impede further range expansion (Colautti et al., 2010). Because evolutionary and ecological forces are dynamic and not constant throughout a species’ range, the genetic architecture of populations is expected to vary. Our study highlights one important outcome of this variation and suggests that differences in mating systems and the potential for mating system evolution exhibited by many taxa in temperate areas might reflect, at least in part, nonequilibrium conditions associated with historic range changes.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

The authors thank A. Erwin, M. Geber, B. Gould, K. Kubow, D. Coltman and several anonymous reviewers for helpful comments on the previous versions of this manuscript. Thanks also to J. Alsayegh, E. Crabtree, A. Dai, L. Dierkes and F. Kilkenny for assisting with the experimental set-up and data collection and to J. Engleman, L. Norbeck, and the Franklin and Marshall College greenhouse staff for help with plant care.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Appendix

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendix

Appendix 1

Geographical locations of the 12 Campanulastrum americanum populations used in this study.

Population IDRegionStateLatitudeLongitude
1CentralMO37.1328−91.2772
2CentralIL38.6074−89.8992
3CentralIN39.3355−86.5102
4CentralTN36.0821−86.2961
5CentralKY37.8999−84.3024
6CentralWV37.9931−80.3618
7NorthernMN44.8178−93.0075
8NorthernWI43.3510−89.9499
9NorthernMI42.3010−85.3568
10NorthernOH41.1147−81.5181
11NorthernPA41.0080−80.0833
12NorthernPA39.9472−76.3676