ABSTRACT
WALL, WADE ALAN. Population Genetics and Demography of Two Rare Plant Species Endemic to the Longleaf Pine Ecosystem. (Under the direction of William A. Hoffmann and Thomas R. Wentworth.)
Astragalus michauxii and Pyxidanthera brevifolia are two rare plant species endemic to the Fall‐line
Sandhills region of the Gulf and Atlantic Coastal Plain in the southeastern United States that are currently considered vulnerable to extirpation. The Fall‐line Sandhills, a large area of relictual dunes, are part of the longleaf pine ecosystem, a temperate savanna dominated by Pinus palustris and maintained by frequent fire. The longleaf pine ecosystem covered 37 million hectares across the southeastern United States at the time of European settlement, but land use changes led to the loss or degradation of 97% of the original area. In this dissertation, we used a variety of methods to better understand the biogeography, population genetic structure, and the effects of fire on the population dynamics of A. michauxii and P. brevifolia. Pyxidanthera brevifolia (Diapensiaceae) is an evergreen subshrub that occurs on xeric ridgetops in the Fall‐line Sandhills of North Carolina and
South Carolina (USA), with the majority of identified populations occurring on Fort Bragg Military
Installation, NC. Fort Bragg is also the only location for the other taxon within the genus, P. barbulata. We developed a germination protocol for P. brevifolia, testing for effects of light and temperature on germination success and rates. Both rates were highest under conditions of low temperature and low light. We used AFLP and cpDNA genetic markers developed for P. brevifolia to assess the taxon’s recent phylogeogeography and taxonomic relationship to P. barbulata. Results indicated there is significant morphological overlap between the two taxa very little genetic differentiation. In addition, evidence suggested that the two taxa did not exhibit any evidence of a range shift following the Pleistocene and that the northern populations of Pyxidanthera were most likely present during the Pleistocene. To investigate the population genetic variation of Astragalus michauxii, we developed eight microsatellites and genotyped 355 individuals across 22 populations.
Genetic evidence indicates that within population genetic variation accounts for 92% of the total observed genetic variation and that the species encountered a genetic bottleneck within the past.
To explore the influence of fire on the population dynamics of A. michauxii and P. brevifolia, we inventoried established demographic monitoring plots from 2007‐2010. Demographic modeling demonstrated that fire negatively affected both species in the short term by increasing mortality of smaller individuals and reducing fruit set. Results indicated that under simulated annual burning, both species would have reduced population growth rates, and that the “ideal” fire return interval may be 2‐4 years. In summary, while anecdotal evidence suggests that fire is indirectly necessary to maintain an open habitat, A. michauxii and P. brevifolia did not respond as “positively” to fire as might be expected for two species that are endemic to a fire‐dependent ecosystem. However, evidence based on climate data and pollen records indicates that the Fall‐line Sandhills were much colder and drier during the Pleistocene and that there has been a substantial shift in the dominant vegetation. If the two taxa were present in the Fall‐line Sandhills since at least the Pleistocene, they would have experienced much colder and drier conditions and, based on the inferred species composition, a much less frequent fire return interval. In conclusion, while fire may be necessary to maintain an open habitat that may have formerly been maintained climatically, it is not apparent that the two species are “fire‐adapted” in the narrow sense of this term and that a more nuanced use of the concept of fire adaptation may be appropriate. Population Genetics and Demography of Two Rare Plant Species Endemic to the Longleaf Pine Ecosystem
by Wade Alan Wall
A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy
Plant Biology
Raleigh, North Carolina
2013
APPROVED BY:
______William A. Hoffmann Thomas R. Wentworth Committee Co‐Chair Committee Co‐Chair
______Kevin Gross Ignazio Carbone BIOGRAPHY
Wade Wall was born in Asheville, NC and spent his formative years in Salt Lake City, UT, and
Marion, NC. After attending the University of North Carolina, Chapel Hill, he spent a number years doing a wide range of jobs, including building treehouses, teaching ethics to prisoners, and teaching
English to non‐native speakers. In 2005, Wade returned to school at North Carolina State University, obtaining a Master of Science degree in Botany in 2008. He currently lives in Champaign, IL with his wife, Katie, and his son, Asa.
ii ACKNOWLEDGMENTS
First and foremost, I thank Kathleen Bernadette Coyle, my wife, for her support and understanding during these past few years. Without her by my side I would not have been able to complete this process. My mother and father, Louis and Edna Wall, have always been there for me and I appreciate what they have taught me about how to live a life.
Drs. Thomas R. Wentworth and William A. Hoffmann have been sources of inspiration and great role models my years at North Carolina State University. My graduate advisory committee has also been supportive and understanding.
Many past and present students of the Plant Biology Department at NC State University have been influential and supportive. In particular, Andy Walker, Renee Marchin, and Kristen Kostelnik provided comic relief and emotional support through the whole process. I will always remember my days in the field with Andy Walker, who will remain a lifelong friend whether he likes it or not.
Norm Douglas and Jenny Xiang have been overly patient, teaching me population genetic laboratory and data analysis techniques.
The support of Sue Vitello and the Plant Biology departmental staff has been tremendous; without their assistance no graduate student would be able to complete a degree. We all owe them a debt of gratitude that can never be repaid.
I am grateful to Matthew Cleary, Brendan Dawal, Jacob Hilton, and Sherrie Emerine for excellent assistance in the field and in the laboratory. They definitely made field work more entertaining.
iii Finally, thanks to Matthew G. Hohmann, United States Army Corps of Engineers (ERDC‐CERL), and
Janet Gray, Endangered Species Branch, Fort Bragg Military Installation. Without their unwavering support and knowledge, this project would never have been completed.
iv TABLE OF CONTENTS
LIST OF TABLES .……………………………………………………………………………………………………………………………...vii LIST OF FIGURES ...... x CHAPTER 1 OVERVIEW ……………………………………………………………………………………………………………………..1 1.1 Introduction ...... 1 1.2 References ...... 6 CHAPTER 2 EFFECTS OF LIGHT AND TEMPERATURE ON GERMINATION OF PYXIDANTHERA BREVIFOLIA WELLS (DIAPENSIACEAE) ……………..………………………………………………………………………………..9 2.1 Abstract ...... 9 2.2 Introduction ...... 10 2.3 Methods ...... 13 2.4 Results ...... 14 2.5 Discussion ...... 15 2.6 References ...... 19 CHAPTER 3 EVIDENCE FOR RANGE STASIS DURING THE LATTER PLEISTOCENE FOR THE ATLANTIC COASTAL PLAIN ENDEMIC GENUS, PYXIDANTHERA MICHAUX…………………………………………………………24 3.1 Abstract ...... 24 3.2 Introduction ...... 25 3.3 Methods ...... 29 3.3.1 Sampling and Morphological Measurements ...... 29 3.3.2 Molecular methods ...... 30 3.3.3 cpDNA data analysis ...... 31 3.3.4 AFLP data analysis ...... 34 3.4 Results ...... 35 3.4.1 Morphology ...... 35 3.4.2 cpDNA ...... 36 3.4.3 AFLP ...... 38 3.5 Discussion ...... 39
v 3.5.1 Taxonomy ...... 39 3.5.2 Phylogeography of the genus Pyxidanthera ...... 41 3.6 References ...... 45 CHAPTER 4 EVIDENCE OF POPULATION BOTTLENECK IN THE ENDEMIC PLANT SPECIES ASTRAGALUS MICHAUXII (KUNTZE) F.J. HERM. …………….………………………………………………………………………………………71 4.1 Abstract ...... 71 4.2 Introduction ...... 72 4.3 Methods ...... 74 4.3.1 Population genetic methods ...... 74 4.3.2 Genetic structure and diversity ...... 75 4.3.3 Evidence of genetic bottlenecks across multiple temporal scales ...... 77 4.3.4 Estimating gene flow between populations ...... 78 4.4 Results ...... 79 4.5 Discussion ...... 81 4.6 References ...... 86 CHAPTER 5 DEMOGRAPHIC EFFECTS OF FIRE ON TWO ENDEMIC PLANT SPECIES IN THE LONGLEAF PINE‐WIREGRASS ECOSYSTEM...... 102 5.1 Abstract ...... 102 5.2 Introduction ...... 103 5.3 Methods ...... 105 5.3.1 Study area and species ...... 105 5.3.2 Field Methods ...... 106 5.3.3 Vital rates data analysis ...... 107 5.3.4 Matrix construction and analysis ...... 108 5.4 Results ...... 111 5.4.1 Effects of plant size and fire on vital rates ...... 111 5.4.2 Population growth rates, elasticities, and LTRE in relation to fire ...... 113 5.5 Discussion ...... 114 5.6 References ...... 118
vi LIST OF TABLES
Table 3.1: Chloroplast haplotype accession numbers as archived in Genbank for the atpI‐atpH intergenic spacer region (partial sequence) and the psbD‐trnT intergenic spacer region (partial sequence). Chloroplast haplotypes H1 through H12 refer to unique composite sequences from the two intergenic spacer regions...... 57 Table 3.2: Polymorphisms of the 12 cpDNA haplotypes based on the cpDNA regions atpI‐atpH and psbD‐trnT in the genus Pyxidanthera. Numbers below cpDNA regions denote position in sequence. N refers to the number of individuals identified as the corresponding haplotype, while dashes represent correspondence to consensus haplotype...... 58 Table 3.3: Analyses of Molecular Variance (AMOVA) results for Pyxidanthera barbulata using cpDNA sequences and AFLP markers. *** indicates p‐value < 0.001, ** p‐value < 0.01, * p‐value < 0.05, and NS indicates non‐significance of variation...... 59 Table 3.4: Genetic diversity indices for P. barbulata and P. brevifolia based on cpDNA sequences and AFLP markers. AFLP genetic diversity indices were only calculated for populations with more than 7 genotyped individuals (437 total specimens). %P represents the
number of polymorphic loci, DW is a measure of rare alleles per population, and He is a measure of expected heterozygosity based on the AFLP markers. π is a measure of cpDNA nucleotide diversity. N represents the number of specimens for each population for AFLP markers and cpDNA sequences (in parentheses)...... 60 Table 3.5: Summary of model statistics for the 24 IMa2 models. Included for each model (left to right) are the negative log of the probability, the number of parameters, the degrees of freedom when compared to the full model, AIC and ΔAIC, the likelihood of the model, the model probability, and the evidence ratio for each model, calculated according to Burnham and Anderson (2002)...... 62 Table 4.1: Eight polymorphic loci identified and developed for Astragalus michauxii. Column are primer pair name, sequence, repeat, microsatellite range (with fragment size in parentheses), and number of alleles observed...... 91 Table 4.2: Proportion of linkage disequilibrium (LD) tests that detected significant LD for eight polymorphic microsatellite loci...... 92 Table 4.3: Genetic variation in Astragalus michauxii populations from North Carolina and Georgia based on eight polymorphic microsatellite loci. Column headings are: N, number of
individuals; A, average number of alleles across loci; AR, average allelic richness; P,
number of private alleles; PR, private allelic richness; HO, observed heterozygosity; HE,
vii expected heterozygosity; H‐W disequilibrium, loci identified as not in Hardy‐Weinberg equilibrium...... 93 Table 4.4: Analysis of molecular variance (AMOVA) results for Astragalus michauxii populations from North Carolina and Georgia (USA). Populations were defined by Natural Heritage Program protocols. Grouping of populations into regions follows Table 2...... 94 Table 4.5: Tests for genetic bottlenecks in Astragalus michauxii using Bottleneck version 1.2 in populations with > 20 gene copies and M ratio for all populations as calculated in Arlequin 3.1. Results indicate no recent genetic bottleneck events, but evidence of a severe genetic bottleneck in the more distant past...... 95 Table 4.6: A. michauxii individuals identified by GeneClass 2 as being the result of possible interpopulation gene flow (p‐value >= 0.001). Fourteen individuals across eleven populations were identified...... 96 Table 5.1: Number of A. michauxii individuals used to estimate survivorship, transition probabilities, and reproductions for each size class. TSB refers to time since burn, with “0” indicating year that population was burned...... 122 Table 5.2: Transition matrices for A. michauxii individuals that were recently burned, 1 year post‐ fire, and 2 years post‐fire. Stages delineated based on the tallest stem: small (0.01 – 20 cm), small‐medium (>20‐40 cm), medium (>40‐80 cm), and large (> 80 cm)...... 123 Table 5.3: Number of P. brevifolia individuals used to estimate survivorship, transition probabilities, and reproductions for each size class. Populations were sampled before the burning season, so TSB = “1” indicates first measurements after population was burned...... 124 Table 5.4: Transition matrices for P. brevifolia during the 2008‐2009 and 2009‐2010 transition intervals. Seedling class is an age class, while the other 10 size classes are defined by plant area (cm2)...... 125 Table 5.5: Elasticity values for A. michauxii transition matrices. Size classes are based on tallest stem and are small (0.1 – 20 cm), small‐medium (>20‐40 cm), medium (>40‐80 cm), and large (> 80 cm)...... 127 Table 5.6: Elasticity estimates for P. brevifolia during the 2008‐2009 and 2009‐2010 transition intervals. Seedling class is an age class, while the other 10 size classes are defined by plant area (cm2)...... 128 Table 5.7: Population growth rates, stable stage distributions, and reproductive values for the three transition matrices (burned, one year post‐fire, and two or more years post‐fire) estimated for A. michauxii. Stable stage distribution values represent proportions and
viii sum to 1; reproductive values are scaled to the smallest size class, which is equal to 1...... 130 Table 5.8: Population growth rates, stable stage distributions, and reproductive values for the 6 transition matrices (2 time steps and burned, 1 year post‐fire, and 2 years post‐fire) estimated for P. brevifolia. Stable stage distribution values represent proportions and sum to 1; reproductive values are scaled to the smallest size class, which is equal to 1. Stable stage distribution does not include the seed stage and the reproductive values are scaled to seedling stage...... 131
ix LIST OF FIGURES
Fig. 2.1 Winter‐flowering Pyxidanthera brevifolia in full bloom on Fort Bragg (left) and ex situ propagated P. brevifolia seedling (right). Left photo taken 25 March 2009...... 22 Fig. 2.2: Effects of light and temperature on percentage of Pyxidanthera brevifolia seeds germinating. Error bars represent 1(+/‐) standard error, with 5 replicates (10 seeds per replicate) for each treatment. There were significant effects of both light and temperature on the percentage of germinating seeds, but also an interaction between light and temperature...... 23 Fig. 3.1: Morphological variation in leaf length, leaf width, and pubescence of P. barbulata (circles) and P. brevifolia (triangles). Solid triangles represent P. brevifolia specimens that had pubescence for half or less than half of the leaf; open triangles represent P. brevifolia specimens with pubescence greater than half of the leaf. Although there are statistically significant differences between the two varieties for leaf length, leaf width, and pubescence, there is considerable overlap between the two species...... 63 Fig. 3.2: Geographic distribution (shaded in grey) and statistical parsimony network for 12 haplotypes from 2 cpDNA regions of Pyxidanthera. State names are in bold abbreviations and numbers represent haplotypes from Table 3.2. Black dots in the haplotype network represent mutational steps; associated letters (S for South, N for North) are the most likely (>95% probability) geographic origins of mutations, inferred using Genetree 9.0. Light grey shading of haplotype network represents proportion of the associated haplotype comprised of southern individuals, and darker grey shading represents proportion comprised of northern individuals. Inset map: Sampling of Pyxidanthera populations on Fort Bragg Military Reservation. Pyxidanthera barbulata populations are represented by closed circles and P. brevifolia populations are represented by open circles...... 64 Fig. 3.3: Isolation by distance for cpDNA (left side) and AFLP (right side) markers across the entire range of Pyxidanthera barbulata. cpDNA demonstrates no isolation by geographic distance (R = 0.01, p‐value = 0.39), while AFLP markers demonstrate weak but significant (R = 0.27, p‐value = 0.02) isolation by distance at shorter distances with effects of genetic drift more evident at greater distances...... 65 Fig. 3.4: Parameter estimates for θ (southern, northern, and ancestral populations), time since divergence, and migration (gene flow) between northern and southern populations of the genus Pyxidanthera based on results from IMa2...... 66 Fig. 3.5: Non‐metric multidimensional scaling ordination of P. barbulata and P. brevifolia population genetic distances (Nei’s D) based on amplified fragment length polymorphism markers. In
x the legend, letters in parentheses represent US states. Little separation is evident among populations defined according to either taxonomic status or geographical location...... 67 Fig. 3.6: Population genetic structure for the genus Pyxidanthera as inferred from the program STRUCTURE for K 2 through 9. The individual columns represent individual genetic samples and the colors represent proportion of ancestry assigned to the different ancestral populations. Little genetic structuring is evident based on either geographic location (South vs. North) or taxonomic identity (var. barbulata vs. var. brevifolia)...... 68 Fig. 3.7: Log likelihood (Ln P(D)) and standard deviation results from program STRUCTURE for K 1 through 9. Runs included a burn‐in length of 10 000 and a post burn‐in length of 25 000 with admixture and correlated allele frequencies...... 70 Fig. 4.1: Historic range and collection sites of Astragalus michauxii. Historical range determined based on voucher specimens (UNC Herbarium Flora of the Southeast; http://www.herbarium.unc.edu/seflora). Current range is greatly restricted, with most sites in North Carolina. Survey of Georgia populations located 13 individuals and there are only two known sites in South Carolina and Alabama...... 97 Fig. 4.2: Non‐metric multidimensional scaling ordination of Astragalus michauxii population genetic distances based on eight polymorphic microsatellite loci. The Georgia populations appear separate from the North Carolina populations, while little separation appears between the North Carolina populations from Fort Bragg and Camp Mackall...... 98 Fig. 4.3: Population genetic structure for Astragalus michauxii as determined by the program STRUCTURE (K = 2‐9). Individual columns represent genetic samples and colors represent proportion of ancestry assigned to different ancestral populations. Results reflect low genetic population differentiation in A. michauxii...... 99 Fig. 4.4: Pairwise population isolation by distance for sampled Astragalus michauxii populations based on eight polymorphic microsatellite loci. Genetic distances demonstrate significant (R = 0.43, P<0.001) isolation by distance. When GA populations are removed isolation by distance is not evident (R = 0.05, P=0.36)...... 100 Fig. 4.5: M ratio values (black circles) estimated for 22 Astragalus michauxii populations across North Carolina and Georgia (USA). Horizontal lines represent 95% CIs, the vertical line is
the threshold indicative of a past genetic bottleneck, the open triangles are the critical Mc
(90% SSM), and the gray triangles are the critical Mc (80% SMM) (Garza and Williamson 2001). Population numbers are in parentheses...... 101 Fig. 5.1: time line for collection of demographic data for Astragalus michauxii (top) and Pyxidanthera brevifolia (bottom). Grey boxes represent prescribed burn season, long dashes represent data collection, and letters along x‐axis represent seasons. Data collection for A. michauxii
xi occurred during the burn season, with burned populations measured at the end of the growing season (short dashed line). Data collection for P. brevifolia occurred before the burning season; measurements were roughly 9 months post‐fire in burned populations.132 Fig. 5.2: Mortality of Astragalus michauxii (top) and Pyxidanthera brevifolia (bottom) as a function of time since last burn and size class. Error bars represent standard error of the mean .133 Fig. 5.3: Mean number of fruits produced as a function of time since last burn and size class for Astragalus michauxii (top) and Pyxidanthera brevifolia (bottom). Error bars represent standard error of the mean...... 134 Fig. 5.4: Modified box plot of post‐fire recovery rates for Astragalus michauxii (top) and Pyxidanthera brevifolia (bottom) individuals as a function of pre‐burn size. Dotted line represents recovery to pre‐burn size. Smaller individuals recover at a faster rate relative to larger individuals for both species...... 135 Fig. 5.5: Projected stochastic population growth rates under different fire‐return intervals (1‐4 years) for Astragalus michauxii (left) and Pyxidanthera brevifolia (right) using a matrix selection approach. Error bars represent bootstrapped 95% confidence intervals...... 136 Fig. 5.6: Contributions of growth, survivorship, and fecundity by size class to the difference in the population growth rate between unburned and burned Astragalus michauxii (top) and Pyxidanthera brevifolia (bottom) populations. Size classes for A. michauxii are small = 1‐20 cm, small‐medium = >20‐40 cm, medium = >40‐80 cm, and large > 80 cm. Size classes for P. brevifolia are small = 1‐50 cm2, medium = >50‐400 cm2, and large > 400 cm2...... 137
xii CHAPTER 1 OVERVIEW 1.1 Introduction
Understanding the abiotic and biotic factors that contribute to the persistence of plant populations within contemporary plant communities is important from both basic and applied science standpoints. A number of different and overlapping theories have been presented for the coexistence of multiple species within the same trophic level (Silvertown 2004), with the topic still drawing scientific attention today (Chase and Leibold 2003; Hubbell 2001). The topic is not merely an academic pursuit. Identifying the abiotic and biotic factors that influence population persistence and coexistence with other species is critical for developing active management plans for threatened and endangered species. This is becoming increasingly important as the global loss of species increases at an alarming rate due to anthropogenic causes (Pimm et al. 1995).
Numerous abiotic and biotic factors can influence population persistence at multiple spatial scales. At the landscape level, climate is one of the main drivers of population persistence within a landscape (Prentice et al. 1991). At finer scales, abiotic factors such as fire (Glitzenstein et al. 2003), soil characteristics (e.g. pH (Dodd et al. 1994; Gough et al. 2000), nutrient availability (Grime 1973), soil texture (Williams et al. 1996), and hydrology (Dimick et al. 2010) can greatly influence the population dynamics and persistence of plant species. Biotic factors related to resource competition
(e.g. for light or nutrient availability) can result in plant species failing to occupy otherwise suitable habitats (Tilman 1982). All of these factors operate across time and space to influence patterns of species persistence and, ultimately, species diversity patterns.
The species considered in this project are members of the longleaf pine ecosystem. The longleaf pine ecosystem stretched from East Texas to Florida and north to southeastern Virginia at
1 the time of European settlement, covering an estimated 37 million hectares (Frost 1993). Most of the longleaf pine ecosystem was found within the Gulf and Atlantic Coastal Plain physiographic region (GACP), and to a certain extent the two constitute a single biogeographic region. The GACP, and by extension the longleaf pine ecosystem, has been noted for its high species diversity (Estill and Cruzan 2001; Walker and Peet 1983) Changes in land use over the last few centuries have led to the loss of 97% of ecosystem (Frost et al. 1986). The dominant canopy tree over most of the area was Pinus palustris, with an often species‐rich herbaceous understory and little to no woody midstory component.
A temperate savanna physiognomy has been maintained in the longleaf pine ecosystem by a high fire frequency, with fire return intervals estimated at 1‐10 years across the region (Myers and
White 1987; Frost 1993; Huffman 2006). The ecological role of fire in the longleaf pine ecosystem has been a focal point of interest for almost a century (Andrews 1917; Chapman 1932) and has been recognized to influence the population dynamics, species diversity, and vegetation structure. Fire suppression leads to the development of a woody midstory and, ultimately, replacement of the longleaf pine canopy. Most herbaceous species in the system are relatively shade intolerant and currently rely upon frequent fire to maintain an open habitat. Shading by midstory species and accumulation of dense litter suppress many of these fire‐dependent herbaceous species.
The longleaf pine ecosystem contains the second‐highest levels of endemism in the continental United States (Takhtajan 1986), with 22 identified regions of endemism (Sorrie and
Weakley 2001). One of the endemic patterns identified occurs in the Fall‐line Sandhills region, an area of relictual dunes that occurs on the northeastern edge of the GACP, extending from North
Carolina into Georgia. The Fall‐line Sandhills consist of two main geologicformations: the Pinehurst
2 and the Middendorf. In North Carolina, the Pliocene‐aged Pinehurst formation occurs on many of the ridge tops and is generally understood to be Aeolian in origin (Bartlett 1967), while the late‐
Cretaceous Middendorf appears to be of deltaic origin (Sohl and Owens 1991). The Pinehurst formation consists mainly of sand, while the underlying Middendorf formation includes a higher percentage of clay and silt. Weathering into the Middendorf has led to differential drainage rates and results in numerous seepages that can occur on middle slopes. This creates a matrix of wetland and dry habitats across the landscape. In contrast to most the GACP, the Fall‐line Sandhills have more topography and include numerous incised streamheads.
Nine taxa have been identified as endemic to the Fall‐line Sandhills: Astragalus michauxii,
Liatris cokeri, Lilium pyrophilum, Lobelia batsonii, Lycopus cokeri, Physalis lanceolata, Pyxidanthera brevifolia, Vaccinium crassifolium var. sempervirens, and Hexastylis sorriei. Interestingly, three of these species have been described within the last decade. These taxa occur across numerous habitat types that span the hydrologic gradient. Astragalus michauxii, Liatris cokeri, P. lanceolata and P. brevifolia tend to occur in the upland, more xeric habitats. Lilium pyrophilum, V. crassifolium var. sempervirens, and H. sorriei occur in wet areas such as Sandhills seeps and wet pine savannas, while
Lycopus cokeri is generally found in Sandhills streamhead swamps.
Within a given region, the species composition of communities is generally thought of in atemporal terms, with the biogeographic history of individual plant species and past vegetation assemblages generally left to the paleo‐ecologists. However, since the Last Glacial Maximum climate has changed dramatically with ensuing range shifts for many species (Soltis et al. 2006), the disappearance of former vegetation assemblages (Jackson and Williams 2004; Overpeck et al. 1992), and the creation of new ones (Watts 1979). This is especially true in the mid and higher latitudes,
3 because climate change appears to have been more dramatic in these areas. Seen through the lens of paleo‐ecological research, communities are composed of species with differing biogeographic histories. Some species are remnants from past vegetation assemblages, while others have migrated because of climatic changes. Understanding the past biogeography of plant taxa is not simply an academic pursuit, but has wider implications for species persistence under projected climate change.
Presumably, the nine endemic taxa that were identified above have been present in the
Sandhills since the Pleistocene, although there is always the possibility that they migrated into the
Sandhills following the Last Glacial Maximum. During the Pleistocene the climate of the Fall‐line
Sandhills was colder and drier compared to the current climate, with pollen records indicating the presence of Pinus spp. (most likely Pinus banksiana) and Picea spp. with an herbaceous understory that resembles prairie assemblages (Watts 1980). Geomorphology of rivers in the Sandhills suggests that many were braided rather than meandering during the Pleistocene, becoming channelized during the wetter climatic conditions of the Holocene (Leigh 2006). Braided rivers generally occur when a sediment load threshold is reached (Leopold et al. 1964) and in some areas have been associated with exposed river banks (Tal et al. 2004), suggesting that there may have been a discontinuous vegetation layer during the Pleistocene. If we accept that these taxa maintained populations in the Sandhills during the Pleistocene, each has experienced changes in climatic conditions that in contemporary spatial terms spans thousands of km. To reiterate, the climate that these endemic taxa experienced during the Pleistocene was vastly different than that of the current period.
4 Thus, there is a certain conflict between understanding the longleaf pine ecosystem (and the species within it) as “fire‐adapted” and the fact that, at least in the northern parts of the longleaf pine ecosystem, the current assemblage of species (and the dominance of P. palustris) may only date to the middle Holocene (Goman and Leigh 2004). It is within this context that we explore the population genetics and population demography of two rare plant species that are endemic to the southeastern United States – Astragalus michauxii Sargent and Pyxidanthera brevifolia Wells.
Chapter 2 explores seed germination and seed ecology of Pyxidanthera brevifolia, a species for which a germination protocol had not been developed and little was known regarding the conditions necessary for successful germination. Chapter 3 explores the phylogeography of the genus Pyxidanthera. Chapter 4 is an analysis of the population genetic diversity within Astragalus michauxii using microsatellite markers. Finally, chapter 5 reports on the effect of fire on the short‐ term population dynamics of P. brevifolia and A. michauxii.
5 1.2 References
Andrews EF (1917) Agency of fire in propagation of longleaf pines. Botanical Gazette 64:497–508.
Bartlett CS (1967) Geology of the Southern Pines quandrangle. Master of Science thesis, University of North Carolina at Chapel Hill
Chapman HH (1932) Is the longleaf type a climax? Ecology 13:328–334.
Chase JM, Leibold MA (2003) Ecological niches: linking classical and contemporary approaches. University of Chicago Press
Dimick BP, Stucky JM, Wall W, et al. (2010) Plant‐Soil‐Hydrology Relationships in Three Carolina Bays in Bladen County, North Carolina. Castanea 75:407–420.
Dodd ME, Silvertown J, McConway K, et al. (1994) Stability in the plant communities of the Park Grass Experiment: the relationships between species richness, soil pH and biomass variability. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 346:185–193.
Estill JC, Cruzan MB (2001) Phytogeography of rare plant species endemic to the southeastern United States. Castanea 66:3–23.
Frost CC (1993) Four centuries of changing landscape patterns in the longleaf pine ecosystem. In: Hermann SM (ed) Proceedings of the Tall Timbers Fire Ecology Conference. Tall Timbers Research Station, Tallahassee, FL, pp 17–37
Frost CC, Walker J, Peet RK, et al. (1986) Fire‐dependent savannas and prairies of the Southeast: Original extent, preservation and status. Wilderness and Natural Aeas in the eastern United States: A Management Challenge. Center for Applied Studies, School of Foresty, Stephen F. Austin St. University, Nacogdoches, Texas, pp 348–357
Glitzenstein JS, Streng DR, Wade DD (2003) Fire frequency effects on longleaf pine (Pinus palustris P. Miller) vegetation in South Carolina and northeast Florida, USA. Natural Areas Journal 23:
Goman M, Leigh DS (2004) Wet early to middle Holocene conditions on the upper Coastal Plain of North Carolina, USA. Quaternary Research 61:256–264.
Gough L, Shaver GR, Carroll J, et al. (2000) Vascular plant species richness in Alaskan arctic tundra: the importance of soil pH. Journal of Ecology 88:54–66.
Grime JP (1973) Competitive exclusion in herbaceous vegetation. Nature 242:344–347.
6 Hubbell SP (2001) The unified neutral theory of biodiversity and biogeography. Princeton University Press, Princeton N.J.
Huffman JM (2006) Historical fire regimes in the southeastern pine savannas. Loiusisana State University
Jackson ST, Williams JW (2004) Modern analogs in Quaternary paleoecology: Here today, gone yesterday, gone tomorrow? Annual Review of Earth and Planetary Sciences 32:495–537.
Leigh DS (2006) Terminal Pleistocene braided to meandering transition in rivers of the Southeastern USA. Catena 66:155–160.
Leopold LB, Wolman MG, Miller JP (1964) Fluvial processes in geomorphology. WH Freeman, San Francisco
Myers RL, White DL (1987) Landscape history and changes in sandhill vegetation in north‐central and south‐central Florida. Bulletin of the Torrey Botanical Club 114:21–32.
Overpeck JT, Webb RS, Webb III T (1992) Mapping eastern North American vegetation change of the past 18 ka: No‐analogs and the future. Geology 20:1071.
Pimm SL, Russell GJ, Gittleman JL, Brooks TM (1995) The future of biodiversity. Science 269:347– 350.
Prentice IC, Bartlein PJ, Webb III T (1991) Vegetation and climate change in eastern North America since the last glacial maximum. Ecology 2038–2056.
Silvertown J (2004) Plant coexistence and the niche. Trends in Ecology & Evolution 19:605–611.
Sohl N, Owens J (1991) Cretaceous stratigraphy of the Carolina coastal plain. In: Horton JW, Zullo VA (eds) Geology of the Carolinas: Carolina Geological Society fiftieth anniversary volume. University of Tennessee Press, Knoxville, pp 191–220
Soltis DE, Morris AB, McLachlan JS, et al. (2006) Comparative phylogeography of unglaciated eastern North America. Molecular Ecology 15:4261–4293.
Sorrie BA, Weakley AS (2001) Coastal plain vascular plant endemics: Phytogeographic patterns. Castanea 66:50–82.
Takhtajan A (1986) Floristic Regions of the World. University of California Press, Berkeley, CA
Tal M, Gran K, Murray AB, et al. (2004) Riparian vegetation as a primary control on channel characteristics in multi‐thread rivers. In: Bennett SJ, Simon A (eds) Riparian vegetation and fluvial geomorphology: Hydraulic, hydrologic, and geotechnical interactions. pp 43–58
7 Tilman D (1982) Resource competition and community structure. Princeton University Press, Princeton, N.J.
Walker J, Peet RK (1983) Composition and species diversity in pine‐wiregrass savannas of the Green Swamp, North Carolina. Vegetatio 55:166–179.
Watts WA (1979) Late Quaternary vegetation of central Appalachia and the New Jersey coastal plain. Ecological Monographs 427–469.
Watts WA (1980) The late Quaternary vegetation history of the southeastern United States. Annual Review of Ecology and Systematics 11:387–409.
Williams RJ, Duff GA, Bowman D, Cook GD (1996) Variation in the composition and structure of tropical savannas as a function of rainfall and soil texture along a large‐scale climatic gradient in the Northern Territory, Australia. Journal of Biogeography 23:747–756.
8 EFFECTS OF LIGHT AND TEMPERATURE ON GERMINATION OF PYXIDANTHERA BREVIFOLIA WELLS
(DIAPENSIACEAE)
*Previously published in the Journal of the Torrey Botanical Society, 137(4): 348‐354. 2010. Used by permission
2.1 Abstract
Pyxidanthera brevifolia is an evergreen semi‐woody cushion plant endemic to the Sandhills of North and South Carolina, with the majority of populations occurring on Fort Bragg Military Reservation in
North Carolina. Currently the species is listed as Endangered in North Carolina and is designated as a
Species at Risk (SAR) by the US Department of Defense. Previous studies have suggested that seeds may not be viable because they failed to germinate under controlled conditions. Our objectives in this study were to attempt germination of Pyxidanthera brevifolia seeds, determine the best temperature conditions for germination, and understand more about germination requirements to aid in future restoration efforts. Using seeds that had been stored at room temperature for six months, we performed a germination experiment at the NCSU Phytotron with six treatments, all combinations of three temperature regimes (low (18oC day/14°C night), medium (22/18°C), and high
(26/22°C)) and two light conditions (light and dark). We monitored the experiment for 13 weeks, recording the number of seeds germinating per dish and the number of days to germination for seeds in each treatment. We found that Pxyidanthera brevifolia produces germinable seeds and that there are significant effects of light and temperature on germination. Highest germination occurred under low temperature and high light conditions (78%); the combination of high temperature and
9 no light produced the lowest germination (6%). Seeds exposed to light germinated significantly earlier at the coolest temperature, compared to medium and high temperatures. These results indicate that it is possible to germinate seeds of this rare plant and suggest that germination of
Pyxidanthera brevifolia likely occurs in late fall and is dependent on adequate light availability.
2.2 Introduction
Developing effective conservation strategies for rare plant species demands a comprehensive understanding of the numerous biotic and abiotic factors that can limit population growth. Efforts to conserve rare plants can benefit from ex situ, or off‐site, propagation (Schemske et al. 1994; Guerrant et al. 2004), which can be used to provide a source of plants for population augmentation (Brumback et al. 2004; Mooney and McGraw 2007), reintroduction (Bowles and
McBride 1996), or for studying questions relevant to species conservation. Ex situ propagation of rare plants has aided studies of physiological tolerance (Wang et al. 2006; Kimball and Campbell
2008; Marchin et al. 2009), phenotypic plasticity (Picotte et al. 2007), pollination (Hackney and
McGraw 2001), mating system evolution (Moeller and Geber 2005), and conservation genetics
(Levin et al. 1979). Population augmentation and reintroduction can ameliorate negative consequences of commercial over‐harvesting of wild populations (Kharkwal et al. 2008) and introduction of exotic pathogens and pests (Lee et al. 1995; Hebard 2001). Information gleaned from studies of ex situ propagation can also provide insights about the response of a species to management (Schwartz and Hermann 1999; Rhoades et al. 2009) and global climate change
(Maschinski et al. 2006). Often little is known about the propagation of rare plants because of a lack of study or underreporting of failed attempts. In the case of Pyxidanthera brevifolia (Sandhills
10 pixiemoss) there is a long history of failed propagation attempts and observations of the species’ apparent lack of recruitment from sexual reproduction (Wells 1929; Reynolds 1966; Primack and
Wyatt 1975).
Pyxidanthera brevifolia is a winter‐flowering evergreen cushion plant (Fig. 1) that occurs in a narrow geographic range within the Sandhills region of North and South Carolina (Wells 1929;
Weakley 2011). Currently, P. brevifolia is found in six North Carolina counties and two South
Carolina counties, with approximately 90% of the known populations occurring on Fort Bragg
Military Reservation, NC (Buchanan and Finnegan 2008). The species is listed by the state of North
Carolina as Endangered, designated by the US Department of Defense as a Species At Risk, and ranked by NatureServe as S2 (imperiled) and S3 (vulnerable) in South Carolina and North Carolina, respectively. Pyxidanthera brevifolia occupies Xeric Sandhill Scrub (Schafale and Weakley 1990), a habitat characterized by an exposed topographic position and excessively well‐drained soils that create harsh growing conditions, especially during the summer season (Wells and Shunk 1931). The more wide‐ranging congeneric P. barbulata is found in the outer Atlantic Coastal Plain from northeastern South Carolina to southeastern Virginia, with disjunct populations in the Pine Barrens of New Jersey and older dunes on Long Island, New York. The two species co‐occur on Fort Bragg, but are separated ecologically, with P. brevifolia occurring in the xeric uplands and P. barbulata occurring in ecotones between streamhead pocosins and adjacent pine‐dominated woodlands
(Schafale and Weakley 1990; Sorrie et al. 2006).
Successful seed germination in the field has not been observed for P. brevifolia, and there has been some question about seed production and viability. Wells and Shunk (1931) observed low seed‐set, and the seeds produced appeared to be inviable. Moreover, they observed no seedlings in
11 the field and speculated that P. brevifolia populations were clonal relicts maintained by vegetative propagation, though there is scant evidence for this. A subsequent study also observed low seed‐set for the species and production of seeds of questionable viability (Reynolds 1966). More recently,
Primack and Wyatt (1975) did not find low seed‐set; percentages of seed‐producing capsules collected from six populations ranged from 22‐83%. Nevertheless, all studies conducted thus far have reported a failure to germinate P. brevifolia seeds in situ (Wells and Shunk 1931; Reynolds
1966; Primack and Wyatt 1975). Attempts at transplanting P. brevifolia have also largely failed
(Primack and Wyatt 1975) or have not yet documented survival beyond four years (Hohmann et al. unpublished data).
Successful propagation of P. brevifolia is important because of its limited range, relatively low population numbers, and the pressures that it faces both currently and in the near future.
Because of base realignments, Fort Bragg Military Reservation has recently experienced unprecedented expansion, with more than 32,000 new troops being assigned to United States Army
Forces Command (FORSCOM). In addition, the nearby city of Fayetteville has also been expanding, and several road projects have led to the destruction of P. brevifolia populations. Our objectives in this study were to germinate P. brevifolia seeds to aid in possible future population augmentation, restoration, or establishment efforts, to identify the optimal germination conditions for P. brevifolia seeds, and to identify possible environmental controls on seed germination for this endemic
Sandhills species.
12 2.3 Methods
Seeds were collected during the last two weeks of April 2008 from at least 5 individuals in each of 24 randomly selected P. brevifolia subpopulations representing 19 populations (as delineated by the North Carolina Natural Heritage Program) on Fort Bragg, NC. Seeds were stored in paper envelopes under ambient laboratory conditions at North Carolina State University until
November 2008. Seeds from the 24 subpopulations were thoroughly mixed and placed on moistened filter paper in 8.5 cm diameter Petri dishes. Each dish was randomly assigned to one of six treatments of a factorial experiment consisting of three temperature settings (low, 18°C day
/14°C night; medium, 22/18°C; and high, 26/22°C) and two light settings (complete darkness and 12 hr daily exposure to fluorescent light). The temperature settings selected represented available growth chambers. We included five replicate Petri dishes per treatment with 10 seeds per replicate.
To simulate complete darkness, Petri dishes in the dark treatment were placed in breathable fabric bags custom designed for the North Carolina State University Phytotron to be both breathable and eliminate all light; seeds in the dark treatment were only light‐exposed when monitoring for germination and watering. We monitored the experiment for 13 weeks, recording germination success and time to germination for each seed in the experiment on average every 4.3 days. The experiment was conducted at the North Carolina State University Phytotron.
We tested for main effects of light and temperature on seed germination percentages using a generalized linear model with a binomial error structure and a logit link function as implemented in the statistical program R (R Development Core Team 2012). We tested for significant effects of light and temperature on the seed germination percentage with a Bonferroni correction using the
13 glht() function from the R package multcomp (Hothorn et al. 2008). We tested for differences in average time to germination using a linear mixed effects model in the R package NLME.
2.4 Results
Percentage of germinating seeds across all treatments was 47%, but it varied considerably among treatments. Overall, germination percentage in the light was twice that in darkness (60% vs