ABSTRACT MORPHOLOGY and INTROGRESSIVE HYBRIDIZATION in NORTH AMERICAN DIPHASIASTRUM by Laura L. Klein Although Interspecific

Home , Diphasiastrum, Diphasiastrum digitatum, Diphasium, Lycopodiaceae, Lycopodium

ABSTRACT

MORPHOLOGY AND INTROGRESSIVE HYBRIDIZATION IN NORTH AMERICAN DIPHASIASTRUM

by Laura L. Klein

Although interspecific hybridization in plants often results in sterility, occasionally hybrids show evidence of fertility. Because of their reproductive competence, such F1’s have the ability to interbreed with each other and with their parents, forming introgressive offspring that are morphologically intermediate between the original parental forms. North American Diphasiastrum complanatum, D. digitatum, and D. tristachyum exhibit evidence of such introgressive hybridization. To test for introgression, morphological measurements for parental and putative hybrid taxa were collected and analyzed using principle components analyses and hybrid indices. Spore fertility was also analyzed within the study set. Results support the presence of introgression among all three pair combinations. Although not infrequent, the numbers of hybrid forms are far exceeded by parental forms and so there are no compelling reasons not to recognize these three taxa at the species level.

MORPHOLOGY AND INTROGRESSIVE HYBRIDIZATION IN NORTH AMERICAN DIPHASIASTRUM

A Thesis

Submitted to the

Faculty of Miami University

in partial fulfillment of

the requirements for the degree of

Master of Science

Department of Botany

Laura L. Klein

Miami University, Oxford, Ohio

2012

Advisor ______R. James Hickey

Reader ______Richard C. Moore

Reader ______Michael A. Vincent

CONTENTS

Introduction ...... 1

Materials and Methods ...... 8

Results ...... 12

Discussion ...... 28

Appendix A ...... 39

Appendix B ...... 43

Appendix C ...... 45

Appendix D ...... 46

Appendix E ...... 47

Appendix F ...... 48

ii LIST OF TABLES

Table 1: List of characters used in analysis of morphological variation...... 11

Table 2: Summary character statistics for parental taxa...... 14

Table 3: Summary character statistics for hybrid taxa...... 17

Table 4: Summary of spore fertility data...... 26

iii LIST OF FIGURES

Figure 1: Lycopodiaceae phylogeny...... 4

Figure 2: North American distribution ranges for Diphasiastrum complanatum, D. digtatum, and D. tristachyum...... 5

Figure 3: Bayesian majority rule concensus phylogram of Diphasiastrum using chloroplast data...... 7

Figure 4: Illustration of characters used in present study...... 10

Figure 5: Principle components analysis (PCA) of morphological characters measured for Diphasiastrum complanatum, D. digitatum, and D. tristachyum...... 16

Figure 6: PCA of morphological characters measured for Diphasiastrum complanatum, D. digitatum, D. tristachyum, D. xhabereri, D. xverecundum, and D. xzeilleri...... 20

Figure 7: PCA of morphological characters measured for Diphasiastrum complanatum, D. digitatum, D. tristachyum, and D. xhabereri...... 22

Figure 8: PCA of morphological characters measured for Diphasiastrum complanatum, D. digitatum, D. tristachyum, and D. xverecundum...... 23

Figure 9: PCA of morphological characters measured for Diphasiastrum complanatum, D. digitatum, D. tristachyum, and D. xzeilleri...... 24

Figure 10: Hybrid indices of parental and putative hybrid taxa for all species pairs...... 25

Figure 11: Percent fertility for individuals scored using hybrid indices for each of the three species pairs as seen in Fig. 10...... 27

Figure 12: Taxonomic distribution of the 1423 herbarium specimens borrowed for this study...... 28

iv ACKNOWLEDGEMENTS

My time at Miami University has truly been one of the most educational and rewarding experiences of my life. I attribute this to all the wonderful people I’ve met along the way. I am indebted to my undergraduate mentors, Tom Lammers and Bob Wise, who introduced me to the world of botany, shared tips for being a successful grad student, and taught me an important lesson: “be curious.” Such wisdom has been necessary to my success in academia.

Thank you to Barb Wilson and Vickie Sandlin for making my transition into and residence in the botany department easy and enjoyable. My fellow botanists and I were lucky to have such dedicated, helpful women on our side!

Mike Vincent and Rich Moore have been insightful committee members, as well as invaluable resources. Mike, a master taxonomist and excellent educator, has shared his perspicuity as a steadfast lunch companion. Rich has helped cultivate my knowledge as an evolutionary biologist and shared his perspective as a newly tenured professor with good humor.

I am fortunate to have worked with my magnificent lab group. Special thanks goes to Mirabai McCarthy, who helped me transition into graduate life at Miami, but also became my surrogate family. Sushma Shrestha, Li Zheng, and Tia Ahlquist have also been valued friends and colleagues, always available to discuss plants, commiserate over research difficulties, or offer encouragement. Thanks especially to Stephen Barr, who helped me collect spore fertility data. My thesis would not be as impactful without his hard work.

An occasional social life helped keep me focused and sane during my Masters. A big thanks goes out to my biking buddies and my cohort broomball team – go Rad Scientists! I’m especially grateful for my frequent dinner companions: Michael Oxendine and Samantha Rumschlag. Whether it was dancing, swimming, or sharing the struggles all young scientists go through, we are successful with help from each other.

I attribute my strength and dedication to the support of my loving family. Care packages assembled by Grandma Lois, my sister, Jessica, and my aunts provided remembrances of home. Grandma Rita’s faithful letters were a weekly reminder that I am missed. I have boundless gratitude for my parents’ unwavering love, comfort, and generosity, which eased the burdens that seemed to build during graduate school. I love you all very much!

Lastly, thank you, Jim Hickey, for luring me into the world of Lycopodium. More than just imparting on me an appreciation for ancestral plant lineages, you’ve taught me how to think like a scientist and educator. You have helped to develop my confidence as a researcher and imparted on me the importance of frequent reflection (sans cigarettes). Most importantly, I will take your advice as I face future challenges, and JUST DO IT.

INTRODUCTION To provide context for speciation, evolutionary studies, or biological discussions in general, biologists use species concepts. The biological species concept (Mayr 1969 and DeQueiroz 2007) is one of the most commonly used species concepts, as it clearly defines species as populations that are reproductively isolated, an intuitively comprehensible construct. However, the boundaries established by the biological species concept are critically questioned when two species are able to interbreed. Many botanists have shown that hybridization is widespread among plant species (Anderson 1936, Conant and Cooperdriver 1980, Rieseberg 1997, Walker 1958, Wagner 1951). Hybrids can often be identified by intermediate morphology, abortive pollen grains or spores, failure to produce fruit or reach sporangial maturation, or other indications of disrupted meiosis (Anderson 1936, 1949, 1953, Barrington et al. 1989, Wagner 1962, 1968, 1987, Wagner et al. 1986). Additive biochemical signatures in isozymes and flavanoids (Barrington et al. 1989) or intermediate DNA content (Aagaard 2009 and Bennert et al. 2011) have also been used as evidence of hybridization. Although most hybrids are sterile because of reproductive abnormalities, many fern species are theorized to have originated via hybridization (Barrington et al. 1989, Caluff 2002, Knobloch 1976, Manton 1950, Soltis and Soltis 2009). Fertility is often achieved through polyploidy. By doubling their chromosome set, allopolyploid species reestablish complimentary homologues (Barrington et al. 1989). In some cases, interspecific hybrids may retain full or partial fertility (Conant and Cooperdriver 1980,

Rieseberg et al. 1996, Walker 1958, Wilce 1965). Geographic isolation of the fertile F1 may form new allohomoploid species (Barrington et al. 1989, Caluff 2002, Conant and Cooper-Driver 1980). Anderson (1936, 1949, 1953) noted that many fertile hybrids not only reproduce but may also backcross with parent species. Most of these hybrids with intermediate characteristics were in populations near parent species in newly created niches, such as man-made or natural secondary successional zones. Anderson and Hubricht (1938) described this phenomenon as introgressive hybridization: “…the gradual infiltration of the germplasm of one species into another as the result of hybridization and repeated backcrossing” (Anderson 1936). Anderson (1936, 1949) further concluded that the result

of this crossing was increased variation in the parental genome. Introgressive hybrids typically have homoploid parental species and repeated backcrossing often results in populations that intergrade indistinguishably into the parental taxa (Anderson 1949, Heiser 1949). To aid in the study and identification of introgressive hybrids, Anderson developed a hybrid index, which describes and quantifies morphological intermediacy (Anderson 1936). The first step in using this index is to identify a set (N) of clearly differentiating characters, whether quantitative or qualitative, for the parental species. All character states for one parent are arbitrarily designated as 0, while the character states for the other parent are scored as 1. In this way, an individual specimen of the first species would have a summative character score of 0, whereas a specimen of the other parent would have a summative score of N. Hybrid individuals would be predicted to have a summative score of N/2 and backcrossing specimens should have summary total character scores that are closer to the parent species with which it backcrossed. By quantifying characters in this way, the frequency of parental and hybrid forms can be estimated for a population (Anderson 1936). Hybrid indices have been used with a variety of organisms to document hybridity (Anderson 1936, 1949, 1953, Buerkle 2005, Hardig et al. 2000, Nolte et al. 2009, Watano et al. 2004). In the pteridophytes, Barrington (1986) compared Polystichum xpotteri to its progenitors P. acrostichoides and P. braunii. This use of the hybrid index allowed Barrington to highlight intermediate and unique character states that aid in the identification of both parents and hybrids. Wilce (1965) also utilized a hybrid index to distinguish Lycopodium sabinifolium from its parent species, L. sitchense and L. tristachyum. Wilce’s comparison showed that L. sabinifolium has many intermediate character states as well as some morphological similiarities with the progenitor species. In these studies, hybrids represent a morphological continuum, often displaying intermediate character states as well as ancestral morphology. Lycopodium, existing since the Lower Jurassic, is a cosmopolitan genus within the monophyletic Lycopodiaceae (Wilce 1965). The genus is comprised of at least forty species distributed among nine sections that have been recognized based on distinctive character discontinuities among the species (∅llgaard 1987). One of these groups,

Lycopodium section Complanata Victorin, is characterized by carrot-shaped gametophytes, typically anisophyllous, decussate leaves (a dorsal, a ventral, and two lateral ranks of leaves), and n=23 (∅llgaard 1987, Wagner and Beitel 1992, 1993, and Wilce 1965). Comprised of approximately twenty species, present distributions suggest that living members of this group have radiated from an evolutionary origin in southeastern Asia (Wilce 1965). Wagner and Beitel (1992) followed earlier generic revisions of Lycopodium (Holub 1975, ∅llgaard 1987) and accepted nine genera, including Diphasiastrum Holub, to better reflect morphology and shared traits of these presumed phylogenetic lineages (Fig. 1). In eastern North America, three frequently encountered Diphasiastrum species are D. complanatum (L.) Holub, D. digitatum (Dill.) Holub, and D. tristachyum (Pursh) Holub. These taxa have partially overlapping distributions (Fig. 2) and have been known to produce at least partially fertile hybrids (Wagner and Beitel 1992, 1993, Wilce 1965). Morphological discontinuities among the three are subtle, but each possesses character states that are generally recognized as species specific. These three Diphasiastrum species are typically found in habitats of dry to mesic hardwood or coniferous forests, generally on acidic soils. All three are also found in open, successional fields where they are slightly shaded by tall weeds and other pteridophytes (Wilce 1965). Diphasiastrum complanatum has a nearly circumboreal distribution, with populations in Europe, northern Asia, and stretches along the North American continent from Alaska to Newfoundland, no further south than the Great Lakes and upper New England regions. Diphasiastrum digitatum is endemic to eastern North America, extending from Newfoundland southward to northern Alabama and Georgia and spreading as far west as eastern Missouri. Unlike the other two species, D. digitatum can also be found on more basic substrates. The third species, D. tristachyum, has a North American distribution similar to that of D. digitatum; however, populations extend north into Quebec and Ontario and are typically not found south of Tennessee. Diphasiastrum tristachyum tolerates more xeric habitats than the other two species and can often be found in dry upland forests, at least is eastern North America. Diphasiastrum tristachyum is also found throughout Europe and Asia, perhaps suggesting

Figure 1. Lycopodiaceae phyologeny taken from Wikström and Kenrick 2001. A box is placed around the genus Diphasiastrum (represented as Lycopodium section Complanata in the phylogeny).

a. b.

Figure 2. North American distribution ranges for a) Diphasiastrum complanatum, b) Diphasiastrum digitatum, and c) Diphasiastrum tristachyum. Taken from Wagner and Beitel 1993.

a larger, circumboreal distribution prior to glaciation events (Wagner and Beitel 1993, Wilce 1965). In the Great Lakes and New England regions, intermediate forms have been noted where these three species’ distributions overlap (Gilman 2003, Haines 2003, Wagner and Beitel 1993, Wilce 1965). These intermediates are reported as incompletely fertile, and some of the studied specimens have been identified as F2 recombinants or as backcrosses with parental species (Wilce 1965). If those identifications are correct, then introgressive hybridization is occurring between some or all of these species. Evidence for the recognition of these three taxa at the species level comes from the relatively few intermediate forms in regions of allopatry and their higher rates of spore abortion (Aagaard 2009, Bennert et al. 2011, Haines 2003, ∅llgaard 1987, Wagner and Beitel 1992, and Wilce 1965). However, some authors (Holub 1975, Kukkonen 1967) recognize Diphasiastrum digitatum and/or D. tristachyum at the subspecific level. Representative collections of Diphasiastrum have been analyzed with flow cytometry to determine DNA content (Bennert et al. 2011). In that work, D. complanatum, D. digitatum, and D. tristachyum were shown to be almost universally diploid with statistically different total nuclear DNA contents. In addition, specimens of morphological intermediacy, putative hybrids, showed intermediate DNA contents. Intermediate morphology, high spore abortion rates, and cytological evidence are all accepted means of identifying fern hybrids. Finally, a recent gene tree of Diphasiastrum (Aagaard 2009, Fig. 3) reveals that D. complanatum, D. digitatum, and D. tristachyum fall out in distinct lineages. Holub (1975) recognized the intermediate forms of D. complanatum x D. tristachyum as the hybrid Diphasiastrum xzeilleri (Rouy) Holub and of D. digitatum x D. tristachyum as Diphasiastrum xhabereri (House) Holub. Gilman (2003) recently named the previously recognized (Wagner and Beitel 1992 and Wilce 1965) but unnamed intermediate between D. complanatum and D. digitatum as Diphasiastrum xverecundum Gilman. The goal of this study was to determine if there are clear boundaries around D. complanatum, D. digitatum, and D. tristachyum and if there is evidence of introgression or alternatively to see if we are looking at one variable species with three extreme

Figure 3. Bayesian majority rule consensus phylogram of Diphasiastrum using non- recombinant, maternally inherited chloroplast loci. Taken from Aagaard 2009.

morphologies. To answer this question, the null hypothesis tested was: Diphasiastrum complanatum, D. digitatum, and D. tristachyum represent variant forms of a single, morphologically continuous species. If these three taxa represent one species, continuous and frequent morphological variation would be evidenced using multivariate analyses and hybrid indices. Individual OTU’s, including morphological intermediates, would be expected to be fully fertile, in accordance with the biological species concept, and intermediate morphotypes should be as abundant or more so than the extreme forms. If Diphasiastrum complanatum, D. digitatum, and D. tristachyum are three distinct species undergoing introgressive hybridization, their intermediate forms should be less abundant than the parental taxa, and exhibit morphologies that lie between those of putative parental species. High spore abortion rates are also to be expected in the intermediate forms. Finally, putative hybrids or backcrosses should be confined to regions of parental sympatry (Anderson 1948).

MATERIALS AND METHODS 1423 herbarium specimens of North American parental taxa and presumptive hybrids were obtained from Harvard University (GH), University of Michigan (MICH), Missouri Botanical Garden (MO), Miami University (MU), and University of Wisconsin (WIS). In addition, field collections were undertaken in July 2011. Populations of Diphasiastrum digitatum, D. tristachyum, and an intermediate Diphasiastrum form were found in Northern Wisconsin. Voucher and duplicate specimens were deposited at the Miami University Herbarium. Sets of 25 specimens per species were chosen as ‘species standards’ based on initial assessment of character completeness, identity, and geographic origin. Two Diphasiastrum digitatum specimens were considered to be intermediate taxa after data analysis. The selected D. complanatum specimens were collected a minimum of 400 miles outside the other two species’ ranges. Diphasiastrum digitatum as well as D. tristachyum specimens from the most southern portions of their respective ranges were chosen to represent populations farthest away from areas of overlap with D. complanatum, the intention being to examine the “purest” forms of each species. Because D. digitatum and D. tristachyum have very similar distributions, criteria for selection of these were made on the basis of unique character states, i.e., D. digitatum

species have opposite branching arrangements without annual constrictions, whereas D. tristachyum species have alternate branching arrangements and have annual constrictions. 60 putative hybrids were selected based on completeness and initial identity (annotation labels or personal identification; see Appendix A for list of OTUs). A list of measured characters (Fig. 4) is presented in Table 1. Branchlet widths, secondary and tertiary peduncle lengths, strobili lengths and widths, and leaf length and widths were measured using the ocular micrometer of a Wild M3C dissecting scope. Branching pattern was established by measuring the distances between first lateral branches. If the mean distance between branches was less than four millimeters, the specimen was considered to have an opposite branching habit, and was given a score of 0. Alternate branching, scored as 1, was defined as the condition wherein the second branch was positioned more than halfway between the first and third branches. Second branches positioned less than halfway between the first and third branches but more than four millimeters above the first were considered to be sub-opposite with a score of 0.5 (Fig. 4c). Primary peduncle lengths were measured using a Ward’s Natural Science metric ruler; branchlet angles were measured using a protractor. Several characters were excluded after preliminary analyses because they lacked statistical significance. Unused characters include distance between upright shoots, number of strobili per upright shoot, number of peduncles per upright shoot, number of dividing strobilus tips, number strobili with sterile tips, quaternary peduncle lengths, and length of lateral leaf free tips. Rhizome diameter and number of dichotomies per first season branch were also measured; however, they were excluded due to lack of consistent material. Despite their exclusion some of these characters may be significant across species. For example, rhizome diameter is much larger in D. tristachyum, and D. digitatum produces several dichotomies during first season branch, whereas D. complanatum and D. tristachyum produce few or no dichotomies during the first season. Characters were analyzed among Diphasiastrum complanatum, D. digitatum, and D. tristachyum to identify the most diagnostic characters. Mean, standard deviation, range, and 95% confidence intervals were established for each character per species set. ANOVA and Tukey tests were also performed to establish significant mean differences (P < 0.05) between species using R. Principle Components Analysis, created in R, was

Figure 4. Illustration of characters used in the present study (see Table 1 for list of characters). a) Composite diagrammatic representation of a “Diphasiastrum.” b) Dorsal (left) and ventral (right) views of lateral branchlets. c) Diagrammatic interpretation of (from top to bottom) opposite, alternate, and sub-opposite branching along upright aerial stem.

Table 1. List of characters used in analysis of morphological variation. The n values represent sample sizes per OTU used in statistical analyses.

Measurement n Unit of Measurement

1. Annual Branchlet Constrictions all present = 1, absent = 0 2. Branching Pattern all alternate (≥3mm) = 1, opposite (<3mm) = 0 3. Branchlet Width 20 mm 4. Angle of Branch Dichotomy 10 degrees 5. Number of Strobili per Peduncle all count 6. Primary Peduncle Length all mm 7. Tertiary/Secondary Peduncle Length all mm Ratio 8. Strobilus Length/Width Ratio all mm 9. Dorsal Leaf Length 10 mm 10. Dorsal Leaf Width 10 mm 11. Ventral Leaf Tip Length 10 mm 12. Ventral Leaf Width at Midpoint 10 mm 13. Ventral Leaf Width at Base 10 mm 14. Lateral Leaf Length 10 mm 15. Lateral Leaf Width 10 mm

employed to summarize overall morphological variation among and between species (see Appendix B for R code). To assess hybridization, three separate pair-wise Anderson hybrid indices were established. Quantitative characters were only employed when the character states’ means were significantly different and the 95% confidence intervals for those means showed no overlap. Any OTU character condition between those confidence intervals was scored as 0.5. Spore abortion rates were also estimated during this study. Although staining techniques are rarely successful with Lycopodium, the difference between viable and nonviable spores is discernable by analysis of shape and opacity (Wagner and Chen 1965, Wagner et al. 1986, Wilce 1965). “Normal” spores were determined by size, shape, and presence of oil bodies; “bad” or abortive spores were identified by their irregular shape and/or by their lack of visible contents. Spore samples were extracted from the lower to mid region of mature strobili of study specimens by gently probing sporangia with a dissecting needle and collecting spores on the tip. The harvested spores were then mounted in Hoyer’s Mounting Medium and dispersed as evenly as possible on a slide before adding a cover slip. Because study specimens with immature or senescent sporangia could not be sampled, additional fertile OTUs were sampled to increase sample size. Percentage of viable spores was determined from a count of 50 spores per specimen.

RESULTS Character state analyses identified a number of characters (Table 2) with significant differences among the three Diphasiastrum species. Angle of branch dichotomy, tertiary/secondary peduncle length ratio, strobilus length/width ratio, ventral leaf tip length, and ventral leaf width at base are significant and distinctive for each of the three taxa. The remaining characters are sufficient to distinguish individual species from the other two. Principle Components Analysis (PCA) of the parental species using these characters reinforces their utility in discriminating taxa. This is clearly evidenced by the level of variation explained by the first and second axes (Appendix C) and the distribution of the three species in morphometric space (Fig. 5). The biplot shows three distinct clusters within the ordination space, each representing a Diphasiastrum species.

12 Table 2. Summary character statistics for parental taxa. Values presented are mean ± standard deviation, range, and 95% confidence interval, respectively. Character states whose means are significantly different (P < 0.05) from those of other taxa and have distinct 95% CIs are in bold font and identified by differently lettered superscript.

Character D. complanatum D. digitatum D. tristachyum n=25 n=23 n=25 1. Annual Constriction 1.00 ± 0.00a 0.00 ± 0.00b 1.00 ± 0.00a (presence=1, absence=0) 1.00 – 1.00 0.00 – 0.00 1.00 – 1.00 (1.00, 1.00) (0.00, 0.00) (1.00, 1.00) 2. Branching Pattern 0.88 ± 0.22a 0.04 ± 0.14b 0.96 ± 0.14a (alternate=1, sub- 0.50 – 1.00 0.00 – 0.50 0.50 – 1.00 opposite=0.5, opposite=0) (0.79, 0.97) (-0.02, 0.11) (0.90, 1.02) 3. Lateral Branchlet Width 1.62 ± 0.19a 1.57 ± 0.19a 1.13 ± 0.06b (mm) 1.23 – 1.91 1.16 – 1.92 1.05 – 1.27 (1.54, 1.70) (1.49, 1.66) (1.10, 1.16) 4. Angle of Branch 37.20 ± 6.21a 32.72 ± 3.37b 29.82 ± 3.27b Dichotomy (°) 26.00 – 53.00 28.00 – 40.50 24.00 – 36.50 (34.64, 39.76) (31.26, 34.17) (28.47, 31.17) 5. Number of Strobili per 1.03 ± 0.08a 3.67 ± 0.58b 3.80 ± 0.57b Peduncle 1.00 – 1.35 2.50 – 4.50 2.00 – 4.70 (1.00, 1.06) (3.42, 3.92) (3.56, 4.04) 6. Primary Peduncle Length 21.17 ± 6.71a 63.73 ± 16.15b 64.57 ± 13.41b (mm) 4.29 – 46.80 35.25 – 94.67 30.53 – 86.00 (16.59, 25.76) (56.75, 70.71) (59.03, 70.11) 7. Tertiary/Secondary 0.00 ± 0.00a 3.24 ± 1.65b 1.55 ± 0.49c Peduncle Length Ratio 0.00 – 0.00 1.06 – 8.64 0.79 – 2.44 (mm) (0.00, 0.00) (2.52, 3.95) (1.35, 1.75) 8. Strobilus Length/ Width 6.74 ± 1.59a 11.31 ± 2.95b 8.53 ± 1.34c Ratio (mm) 3.75 – 10.50 7.07 – 18.62 6.38 – 10.92 (6.08, 7.39) (10.03, 12.59) (7.97, 9.08) 9. Dorsal Leaf Length (mm) 3.36 ± 0.61a 2.76 ± 0.30b 2.86 ± 0.27b 2.16 – 4.72 2.13 – 3.38 2.39 – 3.32 (3.11, 3.61) (2.64, 2.89) (2.75, 2.97) 10. Dorsal Leaf Width (mm) 0.70 ± 0.08a 0.56 ± 0.06b 0.57 ± 0.05b 0.57 – 0.86 0.45 – 0.66 0.51 – 0.70 (0.66, 0.73) (0.53, 0.58) (0.55, 0.59) 11. Ventral Leaf Tip Length 1.10 ± 0.07a 1.00 ± 0.11b 1.29 ± 0.09c (mm) 0.95 – 1.28 0.79 – 1.19 1.05 – 1.43 (1.07, 1.13) (0.95, 1.04) (1.25, 1.33)

13 Table 2., continued 12. Ventral Leaf Width at 0.24 ± 0.03a 0.14 ± 0.02b 0.25 ± 0.03a Midpoint (mm) 0.20 – 0.30 0.12 – 0.18 0.21 – 0.32 (0.23, 0.26) (0.14, 0.15) (0.24, 0.26) 13. Ventral Leaf Width at 0.54 ± 0.09a 0.36 ± 0.05b 0.41 ± 0.03c Base (mm) 0.23 – 0.71 0.26 – 0.44 0.35 – 0.46 (0.50, 0.57) (0.34, 0.38) (0.39, 0.42) 14. Lateral Leaf Length 3.83 ± 0.65a 3.17 ± 0.40b 3.51 ± 0.32b (mm) 2.70 – 5.38 2.47 – 3.83 2.98 – 4.18 (3.56, 4.10) (3.00, 3.34) (3.38, 3.64) 15. Lateral Leaf Width 0.66 ± 0.09a 0.63 ± 0.08a 0.50 ± 0.04b (mm) 0.46 – 0.80 0.46 – 0.77 0.43 – 0.60 (0.63, 0.70) (0.60, 0.67) (0.49, 0.52)

Figure 5. Principle components analysis (PCA) of morphological characters measured for Diphasiastrum complanatum (points “C”), D. digitatum (points “D”), and D. tristachyum (points “T”). Arrows represent the direction of separation and relative strengths (character loadings) of each of the numbered characters (Table 1).

15 Table 3. Summary character statistics for hybrid taxa. Values presented are mean ± standard deviation, range, and 95% confidence interval, respectively. Character states whose means are significantly different (P < 0.05) from those of other taxa and have distinct 95% CIs are in bold font and identified by differently lettered superscript. Double letters represent a taxon whose 95% CI overlaps with the other taxa.

Character D. xhabereri D. xverecundum D. xzeilleri n=20 n=15 n=23 1. Annual Constriction 0.70 ± 0.34a 0.94 ± 0.17b 1.00 ± 0.00b (presence=1, absence=0) 0.00 – 1.00 0.50 – 1.00 1.00 – 1.00 (0.54, 0.86) (0.85, 1.03) (1.00, 1.00) 2. Branching Pattern 0.28 ± 0.34a 0.75 ± 0.37b 0.88 ± 0.27b (alternate=1, sub- 0.00 – 1.00 0.00 – 1.00 0.00 – 1.00 opposite=0.5, opposite=0) (0.11, 0.44) (0.56, 0.94) (0.76, 0.99) 3. Lateral Branchlet 1.38 ± 0.13a 1.51 ± 0.19b 1.36 ± 0.11a Width (mm) 1.14 – 1.62 1.24 – 1.98 1.19 – 1.59 (1.32, 1.44) (1.41, 1.61) (1.32, 1.41) 4. Angle of Branch 35.23 ± 5.43a 36.91 ± 5.83a 38.10 ± 4.06a Dichotomy (°) 27.00 – 50.00 28.50 – 46.00 31.50 – 49.00 (32.69, 37.76) (33.80, 40.01) (36.39, 39.82) 5. Number of Strobili per 3.64 ± 0.71a 3.31 ± 0.84ab 2.93 ± 0.69b Peduncle 1.83 – 4.70 1.33 – 4.60 1.63 – 4.33 (3.31, 3.97) (2.86, 3.75) (2.64, 3.23) 6. Primary Peduncle 69.37 ± 17.38a 53.97 ± 14.54b 55.67 ± 18.40b Length (mm) 36.83 – 97.25 31.06 – 80.50 27.07 – 97.33 (61.23, 77.50) (46.22, 61.72) (47.91, 63.44) 7. Tertiary/Secondary 2.10 ± 1.25a 1.54 ± 1.17ab 1.11 ± 0.67b Peduncle Length Ratio 0.62 – 6.12 0.00 – 4.68 0.00 – 2.56 (mm) (1.52, 2.69) (0.92, 2.16) (0.82, 1.39) 8. Strobilus Length/ Width 9.83 ± 1.92a 8.28 ± 1.16b 7.61 ± 1.27b Ratio (mm) 7.09 – 14.79 6.47 – 10.13 4.87 – 10.20 (8.93, 10.73) (7.66, 8.90) (7.08, 8.15) 9. Dorsal Leaf Length 3.03 ± 0.34a 3.27 ± 0.54a 3.23 ± 0.37a (mm) 2.41 – 3.80 2.45 – 4.57 2.49 – 3.90 (2.87, 3.19) (2.98, 3.56) (3.07, 3.38) 10. Dorsal Leaf Width 0.57 ± 0.05a 0.62 ± 0.07b 0.60 ± 0.04ab (mm) 0.47 – 0.66 0.51 – 0.74 0.53 – 0.66 (0.55, 0.59) (0.58, 0.65) (0.58, 0.61) 11. Ventral Leaf Tip 1.11 ± 0.15a 0.90 ± 0.14b 0.98 ± 0.13b Length (mm) 0.74 – 1.38 0.50 – 1.18 0.77 – 1.30 (1.04, 1.18) (0.82, 0.98) (0.92, 1.03)

16 Table 3., continued 12. Ventral Leaf Width at 0.19 ± 0.04a 0.21 ± 0.05a 0.23 ± 0.03a Midpoint (mm) 0.13 – 0.27 0.16 – 0.34 0.19 – 0.34 (0.18, 0.21) (0.19, 0.24) (0.21, 0.24) 13. Ventral Leaf Width at 0.40 ± 0.04a 0.46 ± 0.08b 0.46 ± 0.04b Base (mm) 0.31 – 0.46 0.37 – 0.68 0.39 – 0.54 (0.38, 0.42) (0.42, 0.50) (0.45, 0.48) 14. Lateral Leaf Length 3.54 ± 0.48a 3.89 ± 0.64a 3.89 ± 0.44a (mm) 2.75 – 4.42 3.10 – 5.57 3.13 – 4.91 (3.32, 3.77) (3.54, 4.23) (3.70, 4.07) 15. Lateral Leaf Width 0.59 ± 0.06a 0.67 ± 0.11b 0.62 ± 0.05a (mm) 0.48 – 0.75 0.54 – 1.00 0.52 – 0.74 (0.56, 0.61) (0.62, 0.73) (0.59, 0.64)

The first two PCA axes account for 43.0 and 26.0% of the morphological variation in parental taxa, respectively. Character vectors, represented by arrows (Fig. 5), depict the directionality and strength of each of the fifteen characters. OTUs (represented as points) with large values for a given character are distributed in the direction of the corresponding arrows on the graph. Branch width, angle of branch dichotomy, and lateral leaf width are important in separating Diphasiastrum tristachyum from D. complanatum and D. digitatum. Annual branchlet constriction and opposite branching (both measured as presence or absence) separate D. digitatum from the other parental taxa, as does ventral leaf base at midpoint and ventral leaf tip length. Diphasiastrum complanatum is demarcated by strobilus length/width ratio, tertiary/secondary peduncle ratio, first peduncle length, and number of strobili per peduncle. Three individuals of D. complanatum cluster near D. tristachyum; these individuals have unusually small values for branch width, lateral leaf width, and angle of branch dichotomy. A D. digitatum individual is oriented lower, more closely towards the D. tristachyum cluster, than other specimens because of slightly narrower branch widths and smaller leaf measurements. A summary of character state distribution for the putative Diphasiastrum hybrids is presented in Table 3. Presence or absence of annual constriction, branching pattern, branchlet width, number of strobili per peduncle, primary peduncle length, tertiary/secondary peduncle ratio, strobilus length/width ratio, dorsal leaf width, ventral leaf tip length, ventral leaf width at base, and lateral leaf width differ significantly between hybrid taxa in various pair-wise comparisons. However, no characters are statistically different among all three Diphasiastrum hybrids. The remaining characters are statistically insignificant and insufficient to distinguish hybrid taxa from one another. Certain characters appear to be diagnostic for each of the hybrids. D. xhabereri is unique from the other hybrid taxa as annual constrictions are more commonly absent or infrequent, branching pattern is general opposite or sub-opposite, primary peduncle length, strobilus length/width ratios, and ventral leaf tip lengths are large, and ventral leaf widths at base are small. Diphasiastrum xverecundum generally have annual constrictions and larger values for branchlet width and lateral leaf width measurements than the other hybrids. Finally, D. xzeilleri specimens have annual constrictions,

Figure 6. Principle components analysis (PCA) of morphological characters measured for Diphasiastrum complanatum (area “C”), D. digitatum (area “D”), D. tristachyum (area “T”) D. xhabereri (points “H”), D. xverecundum (points “V”), and D. xzeilleri (points “Z”). Arrows represent the direction of separation and relative strengths (character loadings) of each of the numbered characters (Table 1).

generally alternate branching, small tertiary/secondary peduncle length ratios, low strobilus length/width ratios, and large values for ventral leaf width at midpoint measurements. When putative hybrid taxa were included with parental taxa, the PCA was inverted relative to the ordination of individual species (Fig. 6); however, the clustering of parental species remains essentially identical. For the most part, hybrid OTUs occupy intermediate morphospace in this analysis, with little incursion into parental taxa space. In this analysis, axes 1 and 2 account for 37.4 and 20.2% of morphological variation, respectively; character loadings are provided in Appendix D. Although D. complanatum, D. digitatum, and D. tristachyum occupy spatially similar clusters in this biplot, there is some compaction. For the most part, Diphasiastrum xhabereri occupies space between D. digitatum and D. tristachyum; its position is most influenced by variation in branch width, angle of branch dichotomy,dorsal leaf length, dorsal leaf width, ventral leaf width at base, lateral leaf length, and lateral leaf width (Fig. 7). Two D. xhabereri individuals are positioned within the D. digitatum cluster due to opposite branching patterns, and values for strobilus length/width ratio and ventral leaf width at midpoint and base similar to that of D. digitatum. Another xhabereri individual borders the D. tristachyum cluster; its position is due to small branchlet width and angle of branch dichotomy measurements. Diphasiastrum xverecundum individuals occupy the space between D. complanatum and D. digitatum (Fig. 8). Annual branchlet constrictions, branching patterns, and small values for ventral leaf tip length and width at midpoint and base affect the orientation of these taxa. Two D. xverecundum OTUs are very close to the D. complanatum cluster due to large measurements for branch width, angle of branch dichotomy, dorsal leaf length and width, ventral leaf width at base, and lateral leaf length and width. Branch width, lateral leaf width, and angle of branch dichotomy place D. xzeilleri between D. complanatum and D. tristachyum, but number of strobili per peduncle, primary peduncle length, tertiary/secondary peduncle length ratio, and strobilus length/width ratio shift these OTUs closer to D. complanatum (Fig. 9). Hybrid indices of paired parental taxa and their corresponding hybrids are shown in Figure 10 (Appendix E). Each index shows that parental taxa are positioned towards the extreme scores, towards zero or the maximum score, with hybrids in a medial

Figure 7. Principle components analysis (PCA) of morphological characters measured for Diphasiastrum complanatum (area “C”), D. digitatum (area “D”), D. tristachyum (area “T”), and D. xhabereri (points “H”). Arrows represent the direction of separation and relative strengths (character loadings) of each of the numbered characters (Table 1).

Figure 8. Principle components analysis (PCA) of morphological characters measured for Diphasiastrum complanatum (area “C”), D. digitatum (area “D”), D. tristachyum (area “T”), and D. xverecundum (points “V”). Arrows represent the direction of separation and relative strengths (character loadings) of each of the numbered characters (Table 1).

Figure 9. Principle components analysis (PCA) of morphological characters measured for Diphasiastrum complanatum (area “C”), D. digitatum (area “D”), D. tristachyum (area “T”), and D. xzeilleri (points “Z”). Arrows represent the direction of separation and relative strengths (character loadings) of each of the numbered characters (Table 1).

Figure 10. Hybrid indices of parental and putative hybrid taxa for all species pairs; value ranges for each species pair vary due to different numbers of characters used in the separate analyses. Hybrid specimens (gray) received scores according to how closely characters resembled a parental state. a) Characters typical of Diphasiastrum digitatum (white) were scored as zero and D. tristachyum (black) were scored as 9.0; D. xhabereri individuals were scored appropriately by character. b) D. complanatum (white) characters were scored as zero and D. digitatum (black) were scored as 13.0; D. xverecundum individuals were scored appropriately by character. c) D. complanatum (white) characters were scored as zero and D. tristachyum (black) characters were scored as 11.0; D. xzeilleri individuals were scored appropriately by character.

Table 4. Numbers of individuals falling within estimated fertility ranges for all Diphasiastrum taxa.

D. D. D. D. D. D. complanatum digitatum tristachyum xhabereri xverecundum xzeilleri Percent n=20 n=27 n=24 n=19 n=12 n=12 Fertility 90 – 100% 10 25 16 9 4 6 80 – 89% 4 2 7 9 2 2 60 – 79% 4 0 0 1 5 0 <60% 2 0 1 0 1 4

Figure 11. Percent fertility for individuals scored using hybrid indices for each of the three species pairs (Figure 10). a) Diphasiastrum digitatum, D. tristachyum and D. xhabererei. b) D. complanatum, D. digitatum, and D. xverecundum. c) D. complanatum, D. tristachyum, and D. xzeilleri.

D. complanatum n = 129

c > d c > t n = 2 n = 2

D. D. xzeilleri xverecundum n = 103 n = 74

d > c t > c n = 24 n = 7

D. digitatum d > t t > d D. tristachyum n = 588 n = 68 D. n = 5 n = 264 xhabereri n = 157

Figure 12. Taxonomic distribution of the 1423 herbarium specimens borrowed for this study.

position. Spore fertility (Table 4) is added to each hybrid index in Figure 11, provided data was available for the OTU. The three Diphasiastrum species demonstrate high spore viability; the majority of specimens had a fertility percentage above 90, with D. digitatum having the highest viability, followed respectively by D. tristachyum and D. complanatum. Hybrid taxa show a greater range of fertility, with generally higher rates of abortive spores. Diphasiastrum xverecundum spores had the lowest spore viability, whereas D. xhabereri showed relatively high levels of fertility. Figure 12 (Appendix F) provides an overall assessment of our identifications of the 1423 herbarium collections borrowed for this study. The most common taxa among these collections are D. digitatum and D. tristachyum. Diphasiastrum xhabereri, D. complanatum, and D. xzeilleri specimens were represented slightly less often, whereas D. xverecundum and individuals with morphologies favoring one parent more strongly than one would expect for an F1 hybrid were the least frequent.

DISCUSSION Historically, Diphasiastrum taxa have been difficult to separate. Many botanists agree that the variations seen among and within the species are often subtle and must be carefully assessed and quantified (Bennert et al. 2011, Kukkonen 1967, Wagner and Beitel 1992, Wilce 1965); therefore, character selection is extremely important for determining species. The list of characters used in this study clearly defines the three parental taxa (Table 2), as evidenced by the distinct clustering in PCA (Fig. 5). My emphasis and analyses of these selected characters have led me, in some cases, to stricter species circumscriptions than previous authors (e.g., Wilce 1965). Diphasiastrum digitatum is strictly defined as always lacking annual constrictions, even on lowermost branches, displanying an opposite branching pattern, and as having four strobili umbellate on the peduncle. These distinctions exclude a number of individuals that would be included under D. digitatum by Wagner and Beitel (1992) and Wilce (1965). These authors consider D. complanatum to have one to four strobili per peduncle (usually one to two) and only rarely possessing a subterranean rhizome. In this study, though, pure D. complanatum is interpreted as having one, rarely two, strobili per peduncle and terrestrial rhizomes. D. tristachyum is easily identified by its narrow branch widths, divergent secondary and tertiary peduncle dichotomies, and subterranean rhizomes.

Distinct species clusters in a PCA confirm the utility of this suite of identifying characters and the associated stricter species definitions. The broad space occupied by Diphasiastrum complanatum in the PCA biplot (Fig. 5) suggests the characters used to define the species are phenotypically labile. Many researchers have had difficulty defining D. complanatum, as it varies greatly in morphology; individuals must be assessed carefully, using multiple characters for correct identification (Bennert et al. 2011, Kukkonen 1967, Wagner and Beitel 1992, Wilce 1965). Important characters for D. complanatum identification agreed upon by most taxonomists include the angle of branching dichotomy, branch width, primary peduncle length, number of strobili per peduncle, and dorsal leaf length (Kukkonen 1967, Wagner and Beitel 1992, Wilce 1965). Wagner and Beitel (1992) and Wilce (1965) followed a rather broad definition of the taxon, whereas others have divided the species into varieties and subspecies. Victorin (1925) named several varieties of Lycopodium complanatum to account for phenotypic and geographic distinctions. Kukkonen (1967) classified Diphasiastrum complanatum as D. complanatum subsp. complanatum. The current study uses a more narrow interpretation of D. complanatum, recognizing the role of extreme habitats in modifying morphology; much of the morphological variation is attributed to hybridization and introgression. Kukkonen (1967) identified individuals with or without subterranean rhizomes with multiple peduncles per upright shoot as D. complanatum var. polystachyum; we considered such individuals to be of hybrid origin. The dwarfed, alpine forms Kukkonen (1967) classified as D. complanatum var. montellii were interpreted as pure D. complanatum in this study. Kukkonen (1967) was unable to unambiguously distinguish var. montellii from typical D. complanatum in his analysis. Dwarf phenotypes have often been attributed to extreme climates (Holub 1975, Kukkonen 1967, and Wagner et al. 1986, Wilce 1965), and so should not be considered unusual. Three individuals with this phenotype can be seen in the parent PCA (Fig. 5). These OTUs are shifted towards the D. tristachyum cluster, in fact one individual falls within the D. tristachyum group. These specimens were examined repeatedly during the study and were unable to be interpreted as anything other than extreme dwarf forms of D. complanatum. This phenotypic plasticity effected the quantitative measurements, but otherwise seem to be representative of the D. complanatum.

Although parental taxa can be readily indentified, intermediate taxa are problematic and proper identification requires considerable experience with the genus (Bennert et al. 2011, Kukkonen 1967, Wilce 1965). Based on PCA results of parental and putative hybrid taxa (Fig. 6), we suggest hybridization as well as introgression is occurring. The assemblage of OTU’s occupying the space between the parental taxa is consistent with a history of hybridization. The intermediate positioning of putative hybrids in PCA space (biplots) is clear and consistent in each of the three parents and hybrid analyses (Figs. 7, 8 and 9). The broadly distributed morphology of the OTUs examined in this study suggest not only the presence of putative F1 hybrids but of introgressive individuals as well. Introgression is evidenced by the presence of some hybrid taxa, which are more closely associated, phenotypically, to one particular parent or the other (Aagaard 2009 and this study). Aagaard (2009) also detected patterns of first generation backcrosses in European D. xzeilleri. Although Bennert et al. (2011) were unable to confirm Aagaard’s findings in their work on European Diphasiastrum, they demonstrated, using nuclear DNA content that triploid hybrids had backcrossed with parental taxa. Bennert et al. (2011) also noted that all hybrids were initially identified by morphology, thus confirming that morphological variation is adequate to detect hybridization in Diphasiastrum. It is apparent that problems with Diphasiastrum identification are complicated by hybridization and introgression. Morphological evidence of introgressive hybridization is also provided by the hybrid indices (Fig. 10). As expected, parental taxa generally acquire extreme scores, whereas individuals identified as hybrids largely receive intermediate scores. Some putative hybrids received scores decidedly close to one of the postulated parents (Figs. 10a, 10b); these individuals are interpreted to be introgressants. Each of the hybrid indices also show parental taxa encroaching into the middle of the hybrid index. This is particularly true for D. complanatum, a species that shows considerable morphological variation. Diphasiastrum digitatum, much like D. complanatum responds to extreme environments in ways that affect morphology; for example, high sun forms develop narrower branch widths and branching angles. All individuals with scores approaching median values were re-evaluated to determine that they did indeed represent their respective taxon.

The addition of spore fertility data to the hybrid indices (Fig. 11) further confirms the presence of introgression and hybridization. Figures 11b and 11c show that those specimens towards the low and high scores, i.e., parent species, generally exhibited higher rates of spore viability compared to the morphologically intermediate individuals. Some D. complanatum individuals have reduced spore viability; however, most of these specimens come from very cold regions and might be susceptible to environmental interference during meiosis. Some D. tristachyum individuals also have low spore viability, which may be the result of drought or other, extreme environmental conditions. Diphasiastrum digitatum specimens show the greatest robustness in spore viability, with only one OTU having less than 90% fertile spores. It also appears (Fig. 11a) that hybrids of D. digitatum, the parental taxon with the highest spore viability, and D. tristachyum, with the second highest, produce hybrid offspring with relatively high spore viability. When D. digitatum hybridizes with D. complanatum, the parental taxon with the lowest rates of spore fertility, to produce D. xverecundum, there is a reduction in spore viability relative to that seen in D. digitatum (Fig. 11b). Diphasiastrum xzeilleri hybrids have the lowest spore fertility of all the hybrid taxa, perhaps an expectation give the slightly lower fertility rates of both D. complanatum and D. tristachyum. Why hybrid taxa would have spore viability values intermediate between their respective parent species is unclear, but this pattern of spore fertility is evident. The taxonomic distribution of the herbarium specimens borrowed during this study (Fig. 12) is a reflection of the commonality of each Diphasiastrum species and their intermediate forms, tempered by the frequency of collecting within their geographic ranges. Diphasiastrum digitatum far exceeds all other members of the genus in the current study; it is also a parent species of the most frequently collected hybrids and introgressants. This is likely due to the fact that D. digitatum’s distribution lies within dense human populations and is easily accessed. Conversely, the northern distribution of D. complanatum is less commonly frequented by people, which results in fewer collections. D. xzeilleri is likely well represented in that D. complanatum and D. tristachyum are sympatric in commonly visited locations (New England and Great Lakes regions). D. xverecundum specimens are less commonly collected; perhaps due to D.

digitatum’s more southern distribution as opposed to D. complanatum, which resides in northern habitats. The majority of putative hybrids obtained for use in this study were collected in regions of sympatry of the parental species, but some geographic outliers are present within the study set. Anderson (1936) suggests that introgressants will most commonly be found in areas of sympatry, a statement largely supported by the current study. Aagaard (2009), Bennert et al. (2011), Kukkonen (1967), and Wilce (1965) also noted that hybrid populations are in areas of sympatry, often in once heavily glaciated landscapes. Bennert et al. (2011) also suggested that multiple hybridization events have occurred. However, extant sympatry is not always in evidence; within the current sample set an individual identified as D. xzeilleri was found in western Canada, well outside the nearest reported location of D. tristachyum. Diphasiastrum spores are capable of long distance wind dispersal and secondary transport via rainwater (Bennert et al. 2011). Thus it is plausible that spores from a distant hybrid clone were dispersed to this locality and became established (Bennert et al. 2011). Hybrid viability increases the possibility that hybrid populations can be established outside the normal geographic range of their parental taxa (Bennert et al. 2011). Aagaard (2009), Kukkonen (1967), and Bennert et al. (2011) have noted the appearance of hybrids outside regions of sympatry. Similarly, Rumsey (2012) has identified D. tristachyum as well as intermediate forms outside of what has been considered to be their known distribution in Great Britain. It is possible such individuals were founded from populations on the European continent. Beyond Diphasiastrum, other accounts of hybrid taxa growing beyond their parental species’ distributions have been documented. Conant and Cooperdriver (1980) demonstrate the hybrid species Nephelea polystichoides (Christ) Tryon can be found outside of the region of sympatry of its’ parental species. They suggest that gene recombination in the hybridization process may allow for ecological adaptation different from the parental taxa. Such evidence validates the presence of hybrid spores outside areas of parental sympatry. Evolutionary relationships among the parental Diphasiastrum taxa infer that putative hybrid taxa are in fact hybrids, and not the result of incomplete speciation. Kukkonen (1967) suggested that D. complanatum and D. tristachyum are of hybrid

origin; he treats both to be subspecies within D. complanatum, and interprets morphological intermediacy and higher rates of spore abortion as evidence. Kukkonen (1967) also suggests that the varieties polystachyum and montellii are the result of hybridization between ssp. complanatum and ssp. tristachyum. While this study considers var. polystachyum to be synonymous with D. xzeilleri, the case for var. montellii, being outside the range of sympatry and in areas of extreme climates, is less compelling. Species relationships depicted in the gene tree provided by Aagaard (2009, Fig. 3) suggests that intermediates within Diphasiastrum are not a function of incomplete divergence, but instead represent post-speciation introgression. In that phylogeny, the parental taxa are not sister to one another; D. complanatum and D. tristachyum represent distinct lineages, and D. digitatum appears to be most closely related to South American D. thyoides. Wilce (1965) hypothesized that D. digitatum shares a common ancestor with Central and South American Diphasiastrum taxa (D. thyoides s.l.) based on morphology; Aagaard’s (2009) phylogeny appears to confirm this relationship. Aagaard (2009, Fig. 3) depicts an individual of D. xzeilleri in the clade with D. digitatum. It is unlikely that this specimen is correctly identified because parental relationships of D. xzeilleri do not support such a position in the phylogeny. Also of note in her work is the position of D. complanatum ssp. montellii as separate from other D. complanatum individuals. This taxon, considered valid in Europe, is generally interpreted as a synonym of D. complanatum in the US. Either this is also the result of misidentification, or future investigations into the relationship between typical and dwarfed, alpine individuals of D. complanatum are needed. It seems clear from Aagaard’s (2009) phylogeny that the parental taxa are not sister taxa to one another, and incomplete speciation cannot be used to explain the presence of the putative hybrids analyzed here. Introgression within Diphasiastrum complanatum, D. digitatum, and D. tristachyum is apparent and well supported on morphological grounds. At the current time, the frequency of these intermediates compared with the numbers of parental forms is insufficient to cloud the recognition of the three parental taxa under review. If future climate change continues to warm northern climates, these species’ ranges and habitats

might be expected to change in ways that will bring populations closer together, inducing more introgression. Hybridization can accelerate adaptation (Aagaard 2009), which will undoubtedly be dramatically affected by climate change. Although some populations in Europe and North America are threatened by habitat loss, Diphasiastrum species are adept at colonizing early successional zones (Bennert et al. 2011). Once established, colonies can attain great sizes and ages (Bennert et al. 2011). Combining various traits inherited from different parental species may also make hybrids more successful at establishing populations due to hybrid vigor. For example, D. complanatum’s cold- hardiness could facilitate D. xverecundum migrating into colder climes, or alternatively the D. digitatum genome may allow for persistence during global warming. The subterranean rhizome of D. tristachyum confers fire resistance and drought tolerance (Kukkonen 1967, Rumsey 2012), which if inherited by hybrid progeny, will undoubtedly aid in fitness and success. Wilce (1965) suggests that these species’ ranges have been slowly returning to pre-Pleistocene distributions, which, in and of itself, may increase parental contact. Currently, D. complanatum, D. digitatum, and D. tristachyum are considered good species due to their distinct morphologies, unique distributions, reproductive competence, and unique genome size (Bennert et al. 2011). However, increased introgression may homogenize character states, making the taxa increasingly harder to recognize.

LITERATURE CITED

Aagaard, SMD. 2009. Reticulate Evolution in Diphasiastrum (Lycopodiaceae). Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 623. Web: http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva- 99584 Anderson, E. 1936. Hybridization in American Tradescantias. Annuals of the Missouri Botanical Garden. 23(3): 511-525. ______. 1948. Hybridization of the Habitat. Evolution 2(1): 1-9. ______. 1949. Introgressive Hybridization. Wiley and Sons, New York. ______. 1953. Introgressive Hybridization. Biological reviews of the Cambridge Philosophical Society 28(3): 280-307. Anderson, E and L Hubricht. 1938. Hybridization in Tradescantia. III. The evidence for introgressive hybridization. American Journal of Botany 25(6): 396-402. Barrington, DS. 1986. The morphology and cytology of Polystichum x potteri hybr. nov. (=P. acrostichoides x P. braunii). Rhodora 88(855): 297-313. Barrington, DS, CH Haufer, CR Werth. 1989. Hybridization, reticulation, and species concepts in the ferns. American Fern Journal 79(2): 55-64. Bennert, W, K Horn, M Kauth, J Fuchs, ISB Jakobsen, B ∅llgaard, M Schnittler, M Steinberg, and R Viane. 2011. Flow cytometry confirms reticulate evolution and reveals triploidy in central European Diphasiastrum taxa (Lycopodiaceae, Lycophyta). Annuals of Botany 108(5): 867-876. Buerkle, CA. 2005. Maximum-likelihood estimation of a hybrid index based on molecular markers. Molecular Ecology Notes 5(3): 684-687. Caluff, MG. 2002. xCyathidaria, a new notogenus in the Cyantheaceae (Pteridophyta). Willdenowia 32(2): 281-283. Conant, DS and G Cooper-Driver. 1980. Autogamous allohomoploidy in Alsophila and Nephelea (Cyatheaceae): A new hypothesis for speciation in homoploid and homosporous ferns. American Journal of Botany 67(9): 1269-1288. DeQueiroz, K. 2007. Species concepts and species delimitation. Systematic Biology 56(6): 879-886.

Gilman, AV. 2003. Diphasiastrum xverecundum (Lycopodiaceae), nothosp. nov. Rhodora 105: 136-142. Haines, A. 2003. The Families Huperziaceae and Lycopodiaceae of New England: A taxonomic and ecological reference. V.F. Thomas Co., Southwest Harbor, ME. Hardig, TM, SJ Brunsfeld, RS Fritz, M Morgan, CM Orianss. 2000. Morphological and molecular evidence for hybridization and introgression in a will (Salix) hybrid zone. Molecular Ecology 9(1): 9-24. Heiser, CB. 1949. Natural hybridization with particular reference to introgression. The Botanical Review 15(10): 645-687. Holub, J. 1975. Diphasiastrum, a new genus in Lycopodiaceae. Preslia, Praha, 47(2): 97-110. Knobloch, IW. 1976. Pteridophyte hybrids. Publ. Mus. Michigan State Univ., Biol. Ser. 5(4): 273-352. Kukkonen, I. 1967. Studies on the variability of Diphasium (Lycopodium) complanatum. Annales Botanici Fennici 4(4): 441-470. Lexer, C, Z Lai, and LH Rieseberg. 2003. Candidate gene polymorphisms associated with salt tolerance in wild sunflower hybrids: implications for the origin of Helianthus paradoxus, a diploid hybrid species. New Phytologist 161(1): 225- 233. Manton, I. 1950. Problems of cytology and evolution in the pteridophyta. Cambridge University Press, London. Mayr, E. 1969. The biological meaning of species. Biological Journal of the Linnean Society, 1: 311-320. Nolte, AW, Z Gompert, CA Buerkle. 2009. Variable patterns of introgression in two sculpin hybrid zones suggest that genomic isolation differs among populations. Molecular Ecology 18(12): 2615-2627. ∅llgaard, B. 1987. A revised classification of the Lycopodiaceae s. lat. Opera Botanica 92:153-178. Rieseberg, LH. 1997. Hybrid origins of plant species. Annual Review of Ecology and Systematics 28: 359-389.

Rieseberg, LH, J Whitton, and R. Linder. 1996. Molecular marker discordance in plant hybrid zones and phylogenetic trees. Acta Botanica Neerlandica 45(3): 243-262. Rumsey, FJ. 2012. Diphasiastrum tristachyum (Pursh) Holub (Lycopodiaceae: Lycopodiophyta) – an overlooked extinct British native. The Fern Gazatte 19(2): 55-62. Soltis, PS and DE Soltis. 2009. The role of hybridization in plant speciation. Annual Review of Plant Biology 60: 561-588. Victorin, FM. 1925. Les lycopodinées du Quebec et leurs formes mineures. Contrib. Lab. Bot. Univ. Montréal 3. Montreal. Wagner, WH. 1951. Cytotaxonomic analysis of evolution in pteridophyta. Evolution 5(2): 177-181. ______. 1962. Irregular morphological development in hybrid ferns. Phytomorphology 12(1): 87-100. ______. 1968. Hybridization, taxonomy, and evolution. Modern methods in plant taxonomy. Academic Press, London and New York: 113-138. ______. 1987. Some questions about natural hybrids in ferns. Botanica helvetica 97(2): 195-205. Wagner, WH and JM Beitel. 1992. Generic classification of modern north American Lycopodiaceae. Annals of the Missouri Botanical Garden 79(3): 676-686. Wagner, WH and JM Beitel. 1993. Lycopodiaceace. In FNA Editorial Committee. Flora of North America 2, Pteridophytes and Gymnosperms. Oxford University Press, New York: 29-32. Wagner, WH and KL Chen. 1965. Abortion of spores and sporangia as a tool in the detection of Dryopteris hybrids. American Fern Journal 55(1): 9-29. Wagner, WH, FS Wagner, and WC Taylor. 1986. Detecting abortive spores in herbarium specimens of sterile hybrids. American Fern Journal 76(3): 129-140. Walker, TG. 1958. Hybridization in some species of Pteris L. Evolution 12(1): 82-92. Watano, Y, A Kanai, N Tani. 2004. Genetic structure of hybrid zones between Pinus pumila and P. parviflora var. pentaphylla (Pinaceae) revealed by molecular hybrid index analysis. American Journal of Botany 91(1): 65-72.

Wikström, N and P Kenrick. 2001. Evolution of Lycopodiaceae (Lycopsida): estimating divergence times from rbcL gene sequences by use of nonparametric rate smoothing. Molecular Phylogenetics and Evolution 19(1): 65-72. Wilce, JH. 1965. Section Complanata of the genus Lycopodium. Beihefte Nova Hedwigia 19. 1-233. Wilce, JH. 1972. Lycopod spores, I. General spore patterns and the generic segregates of Lycopodium. American Fern Journal 62(3): 65-79

Appendix A. List of Diphasiastrum OTUs used in study.

Species Herbarium Accession No. Collector/No. Locality D. complanatum MICH NA M.G. Duman/ 70-533 YT D. complanatum MICH NA J.W. Thompson/ 8922 WA D. complanatum MICH NA H.T. Shacklette / 2968 NT D. complanatum MICH NA S.G. Shetler, K.J. Stone/ YT 3095 D. complanatum MICH NA J. Wilce/ 187-5 NL D. complanatum MO 1772822 E. Scamman/ 1783 AK D. complanatum MO 1772698 H. Petersen/ SN BC D. complanatum WIS 0261216 A.L. Lewicki/ 1 BC D. complanatum MICH NA M.G. Duman/ 71-219 SK D. complanatum MICH NA E.G. Voss/ 8995 MB D. complanatum MICH NA J. Wilce/ 187-3 NL D. complanatum MICH NA R.W. Woodward, C.H. ME Bissell/ SN D. complanatum MICH NA J. Wilce/ 187-4 NL D. complanatum MICH 875857 D.A. Watt/ SN ON D. complanatum MICH NA J.F. Collins, E.B. ME Chamberlain/ 222 D. complanatum MO 3665885 W.K.W. Baldwin, J. SK MacPherson/ 10509 D. complanatum MICH NA E.J. Winslow/ SN VT D. complanatum MO 923217 Marie-Victorin, QC Rolland-Germain/ 18020 D. complanatum MO 970241 Y. Mexia/ 2253 AK D. complanatum WIS 0261215 B. Awerson/ 861 MB D. complanatum MICH NA W.K.W. Baldwin/ 2246 MB D. complanatum WIS 0261230 G.F. Ledingham/ SN SK D. complanatum MICH SN L.H. Jordal/ 1902 AK D. complanatum MICH NA H.M. Raup, D.S. YT Correll/ 11477 D. complanatum MICH NA H.T. Shacklette/ 2769 NT D. digitatum MO 5409129 T.N. McCoy/ SN KY D. digitatum MO 5794035 C.F. Nixon, R.D. AL Whetstone, C.L. Lawler/ 3463 D. digitatum MO 5397323 P.G. & L.L. Reed/ 1982- WV 65 D. digitatum MO 1142644 E.M. Fling Roush/ SN WV D. digitatum WIS 0261207 S.B. Jones/ 23101 GA D. digitatum MICH NA S.B. Jones/ 23101 GA D. digitatum MO 5791613 J.R. MacDonald, R. MS Warren/ 10913 D. digitatum MO 5794049 C.F. & M.J. Nixon/ AL 2080 D. digitatum MO 5690751 S. Femmer/ SN MO D. digitatum MO 5402912 C.F. Reed/ 53447 VA D. digitatum MO 5402949 C.F. Reed/ 53621-A VA D. digitatum MU 171057 Small, Randall/ 64 TN D. digitatum MICH NA R.W. Gettman/ 710 SC D. digitatum WIS 0261191 G.W. Ramsey, B. VA Jackson, R.S. Freer, H.

Schenkle, Jr./ 24549 D. digitatum MO 5409127 C.F. Reed/ 71386 KY D. digitatum MICH 1198685 R.F.C. Naczi, L.A. KY Heeg/ 6853 D. digitatum MU 178087 D.L. Banks/ 7713 KY D. digitatum MU 258023 MA & MW Vincent/ KY 13646 D. digitatum WIS 0261192 S. & G. Jones/ 4990 NC D. digitatum WIS 0261189 T. Daggy/ 4031 VA D. digitatum MO 3844299 S.R. Hill/ 16667 SC D. digitatum MU 162414 E. Sundell, S. Barnes, S. NC McDougald, J. Pagan D. digitatum MO 5794062 J.S. Miller, J.K. Myers/ NC 5344 D. tristachyum MU NA E.S. Steele/ SN VA D. tristachyum MICH NA A.W. Cusick/ 23936 WV D. tristachyum MU 111649 H.K. Svenson/ SN MA D. tristachyum MO 2609929 D.H. Lorence, D. KY Boufford, K. West/ 1793 D. tristachyum MO 1056246 J.R. Churchill/ SN MA D. tristachyum MICH NA F.W. Hunnewell/ 4635 MA D. tristachyum WIS 0251845 R. LeClair/ SN WI D. tristachyum MO 5422073 C.F. Reed/ 31324 MD D. tristachyum MO 5422068 C.F. Reed/ SN MD D. tristachyum MO 2491540 A.B. Seymour/ SN MD D. tristachyum MO 46878 NA/ 2663a NC D. tristachyum MO 828587 E.B. Harger/ 6017 CT D. tristachyum MO 5422083 C.F. Reed/ 70075 WV D. tristachyum WIS 0225981 N.C. Fasset/ 7557 WI D. tristachyum WIS 0261247 L.J. Musselman/ 2928 VA D. tristachyum WIS 0261242 G.W. & K. Ramsey, VA W.A. Sherwood/ 22585 D. tristachyum MO 5422072 C.F. Reed/ 31324 MD D. tristachyum MICH NA E.L. Morris/ 1169 WV D. tristachyum MU 160183 R.J. Hickey/ 279 MA D. tristachyum WIS 0261250 M. Frost/ SN MA D. tristachyum MO 1848488 NA/ 621 MA D. tristachyum MO 5422062 C.F. Reed/ 17409-B KY D. tristachyum WIS 0261241 R. Parker/ 68271 CT D. tristachyum WIS 0261263 S.R. Hill/ 18590 RI D. tristachyum MICH NA A.N. Leeds/ 3518 PA D. xhabereri WIS 0261186 F.W. Gray/ 854 WV D. xhabereri MU 111260 T. Antonio/ 115 NC D. xhabereri GH NA J.V. Haberer/ 3022 NY D. xhabereri GH 214580 W.H. Blanchard/ SN VT D. xhabereri GH 214573 E.F. Williams/ SN ME D. xhabereri GH 214575 J.F. Collins, E.B. ME Chamberlain/ SN D. xhabereri GH 214581 E.M. Kittredge/ SN VT D. xhabereri MO 3820241 R.D. Whetstone/ 13017 AL D. xhabereri WIS 0261251 C.H. Knowlton/ 23626 MA, ME D. xhabereri WIS 0261233 B.N. Goli/ 25447 MA D. xhabereri MU 195094 A.M. Navaro/ 339 ON D. xhabereri WIS 0226004 R.C. Koeppen. 367 WI D. xhabereri MICH 1103364 C.O. Grassl/ 3123 MI

D. xhabereri MICH 11103373 T.J. Mazzci/ SN MI D. xhabereri MICH 1103365 J.R. Bruce/ 73108 MI D. xhabereri GH NA W.H. Wagner, Hagenah, MI Evans, Lellinger/ 9370 D. xhabereri MU 102362 W.H. Wagner Jr., F.S. VA Wagner/ 68083 D. xhabereri MO 3579908 R.C. Moran/ 2387 VA D. xhabereri MU 4396 A.W. Cusick, R. OH Gardner/ 36031 D. xhabereri MICH NA W.H. Wagner Jr., F.S. VA Wagner/ 68041 D. xverecundum GH 216780 A.S.P./ SN NH D. xverecundum GH 216781 A.S.P./ SN NH D. xverecundum GH NA M.L. Grant/ 3118 MN D. xverecundum GH 216785 E.M. Kittredge/ SN VT D. xverecundum GH 216783 A.V. Gilman VT D. xverecundum GH 216775 A.S. Pease, R.C. Bean/ ME 29026 D. xverecundum GH 267859 A.V. Gilman/ 97395 ME D. xverecundum GH 216777 E.F. Williams/ SN ME D. xverecundum GH NA S.F. Blake/ 5420 NB D. xverecundum GH 216776 E.F. Williams/ SN NH D. xverecundum GH NA F. Rolland-Germain/ QC 19183 D. xverecundum MO 1772704 M.E. Jones/ SN MT D. xverecundum MO 1772703 M.E. Jones/ SN MT D. xverecundum MICH NA J.F. Collins/ 226 ME D. xverecundum MICH 1311615 Gilman/ 97395 ME D. xverecundum WIS 0225989 L.S. Cheney/ 7672 WI D. xzeilleri GH NA M.L. Fernald, K.M. NL Wiegand/ 4394 D. xzeilleri GH NA H.M. Raup/ 6797 SK D. xzeilleri GH NA J.F. Collins, M.L. QC Fernald/ SN D. xzeilleri GH NA A.S. Pease, R.C. Bean/ ON 20484 D. xzeilleri GH NA E.C. & T.G. Yuncker/ ON 5461 D. xzeilleri GH NA N. Hotchkiss, P. Jones/ MN 440 D. xzeilleri GH 216769 A.R. Hodgden, F. ME Steele, A.S. Pease, A. Lincoln Jr./ 10273 D. xzeilleri GH NA N. Hotchkiss, P. Jones/ MN 440 D. xzeilleri GH NA J.C. Parlin/ 1833 ME D. xzeilleri GH 216774 E.M. Kittredge/ SN VT D. xzeilleri GH 216771 H.A. Allard/ 6960 ME D. xzeilleri GH 216768 A.S. Pease/ 10457 NH D. xzeilleri GH 216766 A.S. Pease/ 29379 VT D. xzeilleri GH 216767 C.H. Knowlton/ SN NH D. xzeilleri GH 214579 A.V. Gilman/ 93321 VT D. xzeilleri MO 1720031 C.C. Loan/ 78 AB D. xzeilleri MICH NA W.J. Cody, J.M. Matte/ NT 9171 D. xzeilleri MO 5422087 C.F. Reed/ 33224 WV

D. xzeilleri MO 2374979 D. Pittillo/ 3164a NC D. xzeilleri MU 156353 S. Navaro/ 195 ON D. xzeilleri MICH 1331602 E.G. Voss/ 14514 MI D. xzeilleri MO 3579914 R.C. Moran/ 1320 WI D. xzeilleri MO 857281 R. Hoffmann/ SN WI

Appendix B. R code used to calculate mean, range, standard deviation, and 95% confidence intervals, and perform ANOVAs, Tukey Tests, and Principle Components Analysis (PCA). Analyses were done for all 15 characters, therefore “#character#” designates a line of code that would change for an individual character.

Mean, Range, Standard allhh.df<- Deviation, and 95% read.table("/Users/lauraklein/Documents/Thesis/Dat Confidence Interval a/MeansNAcharsp_hybrids.csv",header=TRUE,sep=",") all.df[c(1)] summary(all.df) str(all.df)

# (species) (character) #character#.df <- all.df[,2] str(#character#.df) summary(#character#.df) length(#character#.df) mean(#character#.df) sd(#character#.df) error <- qt(0.975,df=length(#character#.df)- 1)*sd(#character#.df)/sqrt(length(#character#.df)) error left <- mean(#character#.df)-error left right <- mean(#character#.df)+error right

ANOVA and Tukey all.df<- Tests read.table("/Users/lauraklein/Documents/Thesis/Dat a/parentanovas.csv",header=TRUE,sep=",") str(all.df) library(ggplot2) library(multcomp)

#character#.df <- all.df[,c(1,2)] str(#character#.df) qplot(specimen, #character#, data=#character#.df, stat="summary", fun.y = "mean") + geom_point(shape=1, colour="red") #character#.fit <- lm(#character# ~ specimen, data=#character#.df) anova(#character#.fit) summary(#character#.fit) par(mfrow=c(1,2)) plot(predict(#character#.fit),resid(#character#.fi t)) abline(h=0) qqnorm(resid(#character#.fit))

#character#.Tukey <- glht(#character#.fit, linfct=mcp(specimen="Tukey")) summary(#character#.Tukey) confint(#character#.Tukey) par(mar=c(6,11.5,4,2)) plot(#character#.Tukey)

PCA all.df<- read.table("/Users/lauraklein/Documents/Thesis/Dat a/MeansNAcharsp.csv",header=TRUE,sep=",") all.df[c(1)] #exclude variable 1 (specimen) allr.df <- all.df[c(-1)] str(allr.df)

arc.pca.all <- princomp(allr.df, scores=TRUE, cor=TRUE) summary(arc.pca.all) plot(arc.pca.all) plot(arc.pca.all, type="lines") biplot(arc.pca.all) arc.pca.all$loadings arc.pca.all$scores

Appendix C. Character loadings for PCAs 1 and 2 of parental taxa. Loading scores that are very small values not provided in the analysis output are represented as >0.

Character PCA 1 PCA 2 1. Annual Branchlet Constriction 0.303 -0.296 2. Branching Pattern 0.284 -0.272 3. Branch Width >0 0.461 4. Angle of Branching Dichotomy 0.169 0.280 5. Number of Strobili per Peduncle -0.311 -0.190 6. Primary Peduncle Length -0.291 -0.118 7. Tertiary/Secondary Peduncle Length Ratio -0.313 0.111 8. Strobilus Length/Width Ratio -0.255 0.131 9. Dorsal Leaf length 0.262 0.165 10. Dorsal Leaf Width 0.316 0.151 11. Ventral Leaf Tip Length 0.102 -0.390 12. Ventral Leaf Width at Midpoint 0.288 -0.293 13. Ventral Leaf Width at Base 0.324 >0 14. Lateral Leaf Length 0.251 >0 15. Lateral Leaf Width 0.122 0.417

Appendix D. Character loadings for PCAs 1 and 2 of parental and hybrid taxa. Loading scores that are very small values not provided in the analysis output are represented as >0.

Character PCA 1 PCA 2 16. Annual Branchlet Constriction -0.282 0.316 17. Branching Pattern -0.269 0.242 18. Branch Width -0.151 -0.489 19. Angle of Branching Dichotomy -0.178 -0.243 20. Number of Strobili per Peduncle 0.314 >0 21. Primary Peduncle Length 0.270 >0 22. Tertiary/Secondary Peduncle Length Ratio 0.301 -0.174 23. Strobilus Length/Width Ratio 0.255 -0.201 24. Dorsal Leaf length -0.260 -0.218 25. Dorsal Leaf Width -0.320 >0 26. Ventral Leaf Tip Length >0 0.304 27. Ventral Leaf Width at Midpoint -0.274 0.326 28. Ventral Leaf Width at Base -0.351 >0 29. Lateral Leaf Length -0.252 -0.137 30. Lateral Leaf Width -0.189 -0.438

Appendix E. List of characters with significantly different 95% CIs used to create hybrid indices. Diphasisatrum species are listed in the second and third columns with their total score (N).

Character D. digitatum (0) D. tristachyum (9) Annual Branchlet Constriction (-0.02, 0.10) (0.84, 1.00) Branching Pattern (-0.02, 0.10) (0.90, 1.02) Branchlet Width (1.50, 1.65) (1.10, 1.16) Tertiary/Secondary Peduncle Length Ratio (2.61, 4.01) (1.35, 1.75) Strobilus Length/Width Ratio (10.07, 12.41) (7.97, 9.08) Ventral Leaf Tip Length (0.96, 1.04) (1.25, 1.33) Ventral Leaf Width at Midpoint (0.14, 0.15) (0.24, 0.26) Ventral Leaf Width at Base (0.34, 0.38) (0.39, 0.42) Lateral Leaf Width (0.60, 0.67) (0.49, 0.52)

Character D. complanatum (0) D. digitatum (13) Annual Branchlet Constriction (1.00, 1.00) (-0.02, 0.10) Branching Pattern (0.79, 0.97) (-0.02, 0.10) Angle of Branching Dichotomy (34.64, 39.76) (31.54, 34.34) Number of Strobili per Peduncle (1.00, 1.06) (3.49, 3.99) Primary Peduncle Length (16.59, 25.76) (57.90, 70.81) Tertiary/Secondary Peduncle Length Ratio (0.00, 0.00) (2.61, 4.01) Strobilus Length/Width Ratio (6.08, 7.39) (10.07, 12.41) Dorsal Leaf Length (3.11, 3.61) (2.68, 2.92) Dorsal Leaf Width (0.66, 0.73) (0.53, 0.58) Ventral Leaf Tip Length (1.07, 1.13) (0.96, 1.04) Ventral Leaf Width at Midpoint (0.23, 0.26) (0.14, 0.15) Ventral Leaf Width at Base (0.50, 0.57) (0.34, 0.38) Lateral Leaf Length (3.56, 4.10) (3.03, 3.36)

Character D. complanatum (0) D. tristachyum (11) Branchlet Width (1.54, 1.70) (1.10, 1.16) Angle of Branching Dichotomy (34.64, 39.76) (28.47, 31.17) Number of Strobili per Peduncle (1.00, 1.06) (3.56, 4.04) Primary Peduncle Length (16.59, 25.76) (59.03, 70.11) Tertiary/Secondary Peduncle Ratio (0.00, 0.00) (1.35, 1.75) Strobilus Length/Width Ratio (6.08, 7.39) (7.97, 9.08) Dorsal Leaf Length (3.11, 3.61) (2.75, 2.97) Dorsal Leaf Width (0.66, 0.73) (0.55, 0.59) Ventral Leaf Tip Length (1.07, 1.13) (1.25, 1.33) Ventral Leaf Width at Base (0.50, 0.57) (0.39, 0.42) Lateral Leaf Width (0.63, 0.70) (0.49, 0.52)

Appendix F. Total species counts out of a total of 1423 specimens barrowed for this study from each herbarium. Hybrid individuals more strongly introgressed with one parent species is represented by the first letter of the dominant parent’s specific epithet as greater than (>) the first letter of the other parent species’ specific epithet.

Taxon GH MICH MO MU WIS Total D. complanatum 1 35 57 14 22 129 D. digitatum 1 19 394 145 29 588 D. tristachyum 0 29 138 39 58 264 D. xhabereri 20 37 58 20 22 157 d > t 2 4 25 25 12 68 t > d 1 2 2 0 0 5 D. xverecundum 14 19 19 10 12 74 c > d 0 1 1 0 0 2 d > c 0 8 9 3 4 24 D. xzeilleri 40 32 23 12 19 103 c > t 0 0 1 1 0 2 t > c 0 5 2 0 0 7