| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2 Biology Department, Humboldt State University, Arcata, California 95521 USA
Received for publication June 22, 1999. Accepted for publication October 22, 1999.
ABSTRACT
Morphological characters, chloroplast DNA, and allozymes were used to analyze the distribution of individuals within a hybrid population of the ferns Polystichum munitum and P. imbricans in northwestern California. Microsites within the population were characterized according to soil moisture and light levels reaching the plants. In sites with low soil moisture and high light levels, all of the ferns were genetically and morphologically like P. imbricans. In contrast, ferns with the genetic and morphological identity of P. munitum predominated in moist shady sites. Intermediate sites supported very few P.munitum, a wide variety of hybrid recombinants, and a majority of ferns with P. imbricans characteristics. The pattern of variation within the population is noteworthy because of the close proximity of the habitat extremes and the long-range dispersal of fern spores. We conclude that natural selection along environmental gradients must be a major factor in determining the ecological and genetic associations within the hybrid zone. The results of this study are evaluated in the context of the fern life cycle and compared to the assumptions of models explaining the establishment and maintenance of hybrid zones, which vary in the role attributed to environmentally mediated natural selection.
Key Words: cpDNA • Dryopteridaceae • ferns • hybridization • Polystichum • pteridophytes
Natural hybridization can accelerate the presentation of a wide variety of genetic combinations to selection (Anderson, 1949
; Anderson and Stebbins, 1954
; Stebbins, 1959
). This observation has spurred some researchers to investigate hybrid zones as sources of adaptive genotypes that could become favored by selection in certain environments (e.g., Anderson and Hubricht, 1938
; Cruzan and Arnold, 1993
). In contrast, some biologists working to understand clines (long and usually narrow regions of overlap between hybridizing populations) have viewed hybrid zones as opportunities to learn how reproductive barriers arise between incompletely isolated taxa. The persistence of clinal hybrid zones is explained by the Tension Zone model as a balance between dispersal of the parental genotypes into the region of overlap and selection against the hybrids (Barton and Hewitt, 1985
). In this model clines are not maintained by differential adaptation to environmental conditions (Barton and Hewitt, 1985
). Instead, there is endogenous selection against the hybrids in the form of decreased viability and/or fertility resulting from the disruption of the parental genomes (Arnold, 1997
). The opposing view is held by the Mosaic (Rand and Harrison, 1989
), the Bounded Hybrid Superiority (Moore, 1977
), and the Evolutionary Novelty models (Arnold, 1997
), which predict that environment-dependent (exogenous) selection determines the distribution of individuals within hybrid zones. Here, we present the results of correlative population genetic and environmental analyses within a population of hybrid ferns, which suggest that natural selection along environmental gradients is responsible for the observed distribution of individuals.
Although natural hybridization is common in ferns, the hybrids are usually sterile unless fertility is restored through changes in ploidy. In comparison to numerous examples from angiosperms (Rieseberg and Wendel, 1993
), very few cases of fertile diploid hybrids have been reported in homosporous ferns (Mayer and Mesler, 1993
). Both morphological and molecular analyses have shown that the two closely related ferns Polystichum munitum (Kaulf.) C. Presl. and Polystichum imbricans (D. Eaton) D.H. Wagner ssp. imbricans hybridize beyond the F1 generation in the Klamath Mountains of northwestern California (Mayer, 1989
; Mayer and Mesler, 1993
; Mullenniex, Hardig, and Mesler, 1999
). The two species are long-lived perennials with different habitat preferences. Polystichum munitum is most common in mesic forests and along stream banks, while P. imbricans is typically found in more exposed sites such as rocky cliffs and outcrops. In environmentally intermediate areas, field identification of the two species is often difficult due to hybridization, yet nearby in more extreme habitats (i.e., either moist and shady or dry and exposed sites), individuals fitting the descriptions of the species predominate. This observation has led to speculation that the species remain distinct because of their opposing habitat requirements (Mayer, 1989
; Mullenniex, 1996
; Soltis, Soltis, and Wolf, 1990
).
The purpose of this study was to examine the extent to which habitat selection for various parental and hybrid genotypes occurs within the Polystichum hybrid zone by looking at the pattern of ecological and genetic associations at a single site. Using morphological characters in conjunction with allozyme and chloroplast DNA markers, we show that the ferns within our study site are not randomly distributed relative to soil moisture and light-level gradients and argue that the most likely cause for the patterns we observed is natural selection. The dynamics of the hybrid zone are evaluated in the context of the fern life cycle and compared to the assumptions of models explaining the establishment and maintenance of hybrid zones.
MATERIALS AND METHODS
We selected four Polystichum sites for our analysis. Three of these were single-species sites sampled to polarize the morphological and genetic characters that distinguish the species. These included a Polystichum munitum site located in the Arcata Community Forest in Humboldt County, California, USA, and two previously sampled Polystichum imbricans sites in Trinity County, California, USA, Burnt Ranch and Mile Marker 13 (Mayer, 1989
; Mullenniex, 1996
). The single hybrid site located along Swede Creek in Trinity County, California, USA, consisted of ferns morphologically like both species in close proximity to individuals with hybrid morphology. At each single-species site, we tagged 25 ferns typical of either P. munitum or P. imbricans. At Swede Creek, we tagged every reproductively mature Polystichum within the boundaries of the study area, which we subjectively delineated to include the range of microhabitats there. We included a total of 136 plants from Swede Creek in our analysis.
In early June 1996, we collected two young frond-tips from each tagged individual at the four sites for genetic analysis. We kept the samples on ice during transportation, then refrigerated a set at 4°C for allozyme analysis, and froze another at -80°C for nucleic acid extraction. In July, we collected and pressed a single frond from each tagged individual at the Swede Creek, Burnt Ranch, and Mile Marker 13 sites for vouchers and for scoring morphometric characters. We collected vouchers of the P. munitum plants in the Arcata Community Forest in early February 1997. The vouchers are in the Humboldt State University Herbarium (HSC).
Morphological data
We collected data on three morphological features of the ferns: the width of the rachis scales, the width/length ratio of the pinnae, and the indusium margin shape. Previous studies have shown these characters to be particularly useful in distinguishing the species (Wagner, 1979
; Mayer, 1989
; Mayer and Mesler, 1993
; Mullenniex, Hardig, and Mesler, 1999
). We measured the longest pinna of each pressed frond from its apex to its point of attachment to the rachis. The width of the same pinna was measured from the apex of the acroscopic serrature immediately distal from the auricle to the apex of the serrature directly across from it (Wagner, 1979
). To score the rachis scale width, we made wet-mounts of several scales from the base of the blade of each pressed frond, then measured the width of the widest scale at 50x with an ocular micrometer. We scored the shape of the indusium margin subjectively on a scale from 1 to 4. Indusia receiving a score of 1 were entire or variously toothed, but not ciliate. Indusia with many long, multicellular projections, or with shorter gland-tipped projections received a score of 4. We scored indusia with a single multicellular projection with a 2 and indusia with many short multicellular projections as 3's. We rescored indusia at random to check for consistency in scoring. Only individuals with long-ciliate or with glandular-tipped trichomes were observed in the P. munitum population, but we found several ferns having indusium margins with combinations of glands and cilia in the Swede Creek site. These also received a score of 4.
Molecular markers
The genotypes of the ferns were determined for a single allozyme locus and for a chloroplast DNA restriction fragment length polymorphism (RFLP). We scored each individual from the four sites for the fluorescent esterase locus Fe-3, for which P. imbricans and P. munitum do not share alleles (Mayer and Mesler, 1993
). The analysis was carried out generally following the procedures of Soltis et al. (1983)
, using starch gels and a 1.0 mol/L sodium acetate buffer at pH 5.0, followed by staining with an agar overlay.
For the DNA analysis, we used a standard CTAB (hexadecyltrimethylammonium bromide) extraction procedure (Doyle and Doyle, 1987
, as modified by Soltis et al., 1991
) to extract the nucleic acids from the frozen leaf tissue of each fern. Approximately 0.1 g of leaf tissue was homogenized with sterile glass rods and liquid nitrogen directly in 1.5-mL microcentrifuge tubes. The nucleic acids were resuspended in distilled water and treated with ribonuclease "A" before PCR (polymerase chain reaction) amplification of a segment of the chloroplast encoded rbcL gene. PCR reactions were carried out in 25-µL reaction volumes containing 3 µL of unstandardized template DNA, 20 mmol/L Tris-HCL pH 8.3, 50 mmol/L KCl, 1.5 mmol/L MgCl, 5% DMSO (dimethyl sulfoxide), 0.001% Tween-20, 0.2 mmol/L each dNTP, 0.73 units of Taq polymerase (in storage buffer B; Promega, Madison, Wisconsin, USA), and 6.25 pmol each of the forward and reverse primers. We used a Perkin-Elmer PE 2400 thermocycler programmed for 2 min at 94°C, 30 cycles of 1.5 min at 94°C, 2 min at 48°C, and 3 min at 72°C, with a 16-min hold cycle at 72°C followed by storage at 4°C. The primers (Mullenniex, Hardig, and Mesler, 1999
) were developed by Paul Wolf at Utah State University. Nla III restriction digests of the amplified products were separated on 1.4% agarose gels containing ethidium bromide and scored for the RFLP developed by Mullenniex (1996)
.
Environmental variables
Our field observations within the hybrid zone suggest that the species are separated by strongly differing tolerances to light and moisture levels. It is likely that the greatest selective pressures due to water stress occur in the late summer when the weather is typically dry, and daytime temperatures often exceed 100° F (~38° C). To estimate the percentage of shade directly above the fronds of each fern, we used diazo paper booklet photometers (Friend, 1961
; see also, Lubbers, 1982
; Lubbers and Christenson, 1986
). Metal rods were used to support the photometers in a horizontal position slightly above the tallest fronds and directly above each rosette. We set out the booklets after dark on 19 August 1996, and collected them after dark the following night. The day during which the photometers were exposed was completely cloudless. We kept the photometers cool and in light-proof bags during transportation, then stored them at 4°C before developing them with ammonia vapor. We counted the number of exposed pages in each booklet, then scored the "between paper" readings (Friend, 1961
) subjectively by sorting the photometers into three piles based on the intensity of the bright spot on the last exposed page. We checked for consistency by comparing the booklets within and between the piles during the sorting procedure. To each photometer in the pile with the least exposed last pages, we added a value of 0.25 pages to the total number of exposed pages in the booklet. Values of 0.50 and 0.75 pages were added to the total number of exposed pages of each photometer in the piles with intermediate and maximally exposed last pages, respectively. To generate the variable "shade index," we scored all of the photometers relative to the photometers that received the most light at the site, and scaled the scores between zero and one. Although the actual relationship between the number of exposed pages and the amount of light received by a booklet is exponential (Friend, 1961
), the shade index values can be roughly interpreted as the fraction of shade at the site of each plant, with a value of zero corresponding to full sunlight, and a value of one corresponding to total shade. While the diazo paper photometers provide a relative index of light availability (Kearns and Inouye, 1993
), they respond to ultraviolet light and cannot be used to estimate photosynthetically active radiation.
To estimate available soil moisture within the range of habitats at Swede Creek, we measured the xylem water potential (Scholander et al., 1965
; Ball et al., 1983
) of individual ferns with a PMS pressure bomb (PMS Instrument Company, Corvallis, Oregon, USA). We delineated eight areas (i.e., "strata") within the population that seemed relatively homogeneous in soil moisture conditions based on our subjective judgments of where the natural changes in exposure, slope, plant species composition, and other factors would necessitate drawing the boundaries. Within each of the strata we measured at least 25% of the mature Polystichum, drawing tag numbers at random. The work was carried out on two consecutive nights in August (on the same nights as the photometers were placed or collected), beginning after about midnight, and continuing until dawn. We measured a total of 45 ferns from the eight strata (Fig. 1).
|
We used the results of the multiple comparison procedures to pool strata that were not significantly different in shade index or xylem water potential. This initial pooling resulted in three groups of strata based on light intensity, and in three groups based on soil moisture levels. Combining the original strata into groups with different combinations of soil moisture and light levels resulted in the following five groups (of the nine possible): (1) high moisture–low light (HM/LL); (2) high moisture–intermediate light (HM/IL); (3) intermediate moisture–intermediate light (IM/IL); (4) intermediate moisture–high light (IM/HL); and (5) low moisture–high light (LM/HL).
We used the five soil moisture/light groups of pooled strata ("H2O/light groups") to investigate the correlation between genotype and environment at Swede Creek. We used a 
2 test of goodness-of-fit to investigate any deviation of the Fe-3 alleles from Hardy-Weinberg equilibrium within the entire Swede Creek sample, and 2 tests of independence to look at the distribution of the cpDNA types and indusium margin scores across the five groups. Cytonuclear disequilibrium values (Table 1) between the Fe-3 alleles and the cpDNA haplotypes were computed as in Asmussen, Arnold, and Avise (1987)
using a computer program supplied by C. Basten at North Carolina State University. These values test for nonrandom associations between nuclear and cytoplasmic (i.e., cpDNA) genomes and are useful in assessing the directionality of crosses between hybridizing populations, levels of gene flow and assortative mating, and the action of selection within hybrid zones (Arnold, 1993
).
|
Single-species sites
Our morphological and genetic analyses of the ferns from the single-species sites confirmed the utility of each of the characters that we used in distinguishing the species. The allele Fe-3-c is essentially fixed at the pure Polystichum munitum site (the rare allele Fe-3-a was present in one individual), and Fe-3-d occurred at high frequency at the two P. imbricans sites. The rare allele Fe-3-b, which we have found only at P. imbricans and hybrid sites, occurred at low frequency at Burnt Ranch and Mile Marker 13 (Table 2). For the disequilibrium calculations and the figures, we pooled the four individuals at Swede Creek with Fe-3-ac genotypes into the Fe-3-cc genotypic class, and the two individuals with Fe-3-bd and Fe-3-de genotypes into the Fe-3-dd class. All of the ferns we sampled from the Community Forest had the cpDNA haplotype diagnostic for P. munitum, and all ferns from Burnt Ranch and Mile Marker 13 scored for the P. imbricans cpDNA haplotype (Table 2).
|
). These scales are typically narrower than the ovate-lanceolate scales of P. munitum. The ranges of the two continuous morphological variables, rachis scale width and pinna width/length ratio, did not overlap in the P. imbricans and P. munitum plants we sampled with the exception of four individuals from Mile Marker 13, which had P. munitum-like pinna width/length ratios. We followed Mayer (1989)
and Mullenniex (1996)
in using these characters for the axes of pictorialized scatter diagrams. These diagrams represent the species identities of hybrid plants more accurately than hybrid index histograms, in which various combinations of characters can result in the same hybrid index score (Grant, 1981
). We used the diagrams instead of attempting to classify individuals into parental and various hybrid types, because the classification can depend on which markers are used, and information can be lost in the categorization (Barton and Gale, 1993
). The diagrams in Fig. 2 show the character combinations of the ferns from the single-species sites and of the ferns from the five H2O/light groups present at Swede Creek.
|
|
2 = 82.44, df = 12, P < 0.0001). Because of the large number of cells in the contingency table with expected values <5 (50%), we used a randomization procedure to verify the significance of our 2 statistic (Manly, 1997
). We computed the overall 2 for the table after each of 10 000 rounds of distributing the observed indusium margin scores randomly among the H2O/light groups and obtained no values equaling or exceeding our observed value of 82.44 (i.e., P < 0.0001). Ferns with P. imbricans indusium morphology were completely absent from the high moisture/low light strata, and ferns with P. munitum type indusia did not occur in the low moisture/high light strata (Figs. 1, 2; Table 2).
The pattern of molecular marker variation within the Swede Creek site was also correlated with the environmental strata, and for most of the individuals, concordant with the morphological data (Figs. 1, 2; Table 2). Only individuals homozygous for the allozyme allele typical of P. imbricans were present in the LM/HL group, and ~94% of the ferns in the wet and shady habitat contained only alleles typical of P. munitum (Table 2). Intermediate conditions (HM/IL, IM/IL, IM/HL) supported a high proportion of individuals homozygous for the P. imbricans allozyme marker (~60%), fewer heterozygotes (~30%), and only ~10% P. munitum homozygotes. Chloroplast haplotype frequencies were also correlated with the H2O/light groups (2 = 60.81, df = 4, P < 0.0001). About 91% of the ferns in the high moisture/low light habitat had the P. munitum cpDNA marker, and all of the ferns in low-moisture, high-light habitat had P. imbricans chloroplasts (Figs. 1, 2; Table 2).
We found significant genetic disequilibria within the Swede Creek site. The 2 test for goodness-of-fit (pooling individuals with the rare Fe-3 alleles "a," "b," and "e") showed significant deviation from Hardy-Weinberg equilibrium proportions (2 = 49.3, df = 1, P < 0.0001), with the greatest contribution to the overall 2 due to the less-than-expected number of heterozygotes. The population also showed a significant overall level of cytonuclear disequilibrium (Table 1, "D"), and significant disequilibria between the cpDNA markers and the homozygote classes (Table 1, "D1" and "D3"). A random association existed between the cpDNA makers and the Fe-3-cd genotype (Table 1, "D2"), indicating a lack of directionality in interspecific matings.
DISCUSSION
Our morphological, molecular, and environmental analyses of the ferns at Swede Creek revealed a strong correlation between genotype and environment. With only three exceptions, individuals scoring as Polystichum munitum were restricted to moist shady sites, whereas P. imbricans occupied both the intermediate and xeric sites. While we detected a few later generation hybrids in the HM/LL stratum, most of the ferns that we were able to identify as hybrids occurred in the environmentally intermediate strata (HM/IL, IM/IL, and IM/HL) and were completely absent from the xeric sites (Table 2). By our criteria, more of the individuals in the intermediate habitats scored as "pure" P. imbricans than as hybrids (50 vs. 41, respectively), but for reasons described below, our estimate of the number of ferns of hybrid ancestry in these areas is likely to be biased downward.
The spatial distribution of individuals within the hybrid zone at Swede Creek is best explained as the result of environmentally mediated natural selection. Because of the small size of the study area, it is likely that all of the strata within the study boundaries at Swede Creek have received spore rain from both species and the hybrids, and that only individuals with genotypes suited for the various microclimates within the site have survived to reproductive maturity. Long-range dispersal of Polystichum spores has been suggested in an allozyme survey of two P. munitum populations by Soltis and Soltis (1987)
in which they found an "extremely high degree of gene flow among populations" (Nm = 24.0) and high levels of intrapopulational gene flow among subpopulations in both populations surveyed (Nm = 4.17 and 7.96). Spores dispersed to the study site are therefore likely to originate both from ferns within the site and from the surrounding area, where both species of Polystichum and hybrids are very common. Neither species spreads or reproduces clonally, and both species are likely to be outcrossing and reproduce mostly by intergametophytic mating (Soltis and Soltis, 1987
; Haufler, 1987
). The significant environment-by-genotype associations that we measured are indirect tests of hybrid fitness. However, the results of the present analysis, the mosaic nature of the Polystichum hybrid zone, the dispersal biology of the ferns, and our observations (E. Kentner and M. Mesler, personal observations) of similar environment-by-genotype associations in other hybrid zones between these two species suggest the influence of exogenous selection.
We chose to measure light levels and water stress in the late summer because the differences between the strata are likely to be greatest during this period and because we felt that the seasonally extremely hot and dry conditions on the slopes were likely to be selecting against ferns without suitable adaptations. Because fern gametophytes are free living and must survive in the environment long enough to produce new sporophytes, selection may act at this stage. The gametophytes of two apogamous species of the fern genus Pellaea have been shown to be tolerant to desiccation for periods of time up to several months (Pickett and Manuel, 1926
). While recognizing that differential desiccation tolerance of gametophytes may be important in determining the distribution of ferns at Swede Creek, we feel that it is unlikely that the ephemeral and morphologically simple Polystichum gametophytes are differentially adapted to xeric conditions because water is required for successful fertilization. Lacking detailed knowledge of this aspect of the biology of the gametophytes, we assume that spores germinate and develop into gametophytes during periods of ample moisture leaving the sporophytes generated by successful fertilizations to withstand the hot and dry conditions of the summer and early fall.
Because sexual reproduction occurs between gametophytes, this portion of the fern life cycle must play an important role in the dynamics of hybridization between the two species, although very little is known regarding the biology and ecology of this stage of the life cycle in nature. Despite claims to the contrary (Yatskievych, 1993
), Polystichum gametophytes employ antheridiogen as an intergametophytic mating mechanism (M. Mesler, personal observation), and intragametophytic selfing is rare or nonexistent (Soltis and Soltis, 1987
; see also, Haufler, 1987
). Experiments with gametophytes in culture are needed to determine the desiccation tolerances of the two species, to investigate potential barriers to hybrid formation, and to verify that chloroplasts are maternally inherited in Polystichum, as has been reported for the fern genus Asplenium (Vogel et al., 1998
). Although no data on prezygotic barriers to hybridization have yet been collected, the cytonuclear disequilibrium statistic (D2 = 0.002 ± 0.017, P = 0.9893) indicating a random association of the heterozygous Fe-3 genotypes with the cpDNA haplotypes suggests that either species is equally likely to be the maternal parent in first-generation interspecific crosses. Putative F1s show reduced pairing during meiosis (Wagner, 1987
), and low spore germination rates, which seem to increase with backcrossing (Mayer and Mesler, 1993
).
Which hybrid zone model best explains our observations within the Polystichum hybrid zone? It is clear that the environmental dependence and dispersal biology of the ferns precludes the Tension Zone (Barton and Hewitt, 1985
) as a valid explanation for the hybrid zone at Swede Creek. The distribution of ferns clearly fits the pattern described by the Mosaic model (Harrison, 1986
; Howard, 1986
), and although clinal patterns of variation have been described in mosaic hybrid zones (Arnold, 1997
), clinal variation, if present at Swede Creek, would have to occur at a scale that was much too small relative to the dispersal of Polystichum spores for the Tension Zone model to hold. We lack at least two pieces of information that are crucial in evaluating the other models. First, we require additional species-specific nuclear markers. With only a single-species specific nuclear marker and no information on the genetic basis of the morphological traits, it is impossible to determine how many ferns scoring as "pure" parentals by our criteria are actually backcross or later generation hybrids. Our results therefore underestimate the frequency of hybrids within the H2O/light groups and make it impossible to recalculate the cytonuclear disequilibrium statistics with the parentals excluded from the analysis. This recalculation would help determine whether dispersal of parental types into the hybrid zone is responsible for the significant disequilibria or whether other factors such as selection or assortative mating are likely to be involved. Second, our study lacks estimates of relative fitness for the various genotypes, a major component of most of the hybrid zone models. The Bounded Hybrid Superiority model, for example, holds that the hybrids are more fit than either of the parental species in certain habitats, yet because our only estimate of fitness is the frequency of individuals, which is itself an underestimate for the hybrids, we cannot fairly evaluate the model. Although the original formulation of the Mosaic model suggested that hybrids would have lower fitness than the parentals throughout the hybrid zone (Harrison, 1986
; Howard, 1986
), it is also possible for mosaics to form in situations in which hybrids are more fit in intermediate environments, or in which they have intermediate fitness in all patches (D. Howard, New Mexico State University, personal communication). We argue that the differential fitness of the parental taxa in the habitat extremes at Swede Creek must be the major factor contributing to the mosaic pattern of the hybrid zone, regardless of the fitness of the hybrids. The Evolutionary Novelty model may be the most realistic in describing the processes that maintain hybrid zones because it allows for a variety of different outcomes to interspecific hybridization and includes the possibility that hybrids may be fit and evolutionarily stable (Arnold, 1997
). The reduced fertility of putative F1s (Mayer and Mesler, 1993
) and the environment dependence indicated by our results are consistent with the components of endogenous and exogenous selection predicted to occur in hybrid zones by this model (Arnold, 1997
). The great variety of hybrid Polystichum that exist in nature may be providing the "hybrid bridge" necessary for the introgession of genetic material between the taxa with the potential for adaptive evolution. Additional research should shed light on the evolutionary potential of the novel genotypes that are being generated by this unique example of homoploid hybridization in Polystichum, a genus well known for its numerous examples of allopolypoid speciation.
FOOTNOTES
1 The authors thank Jenny Moore for her help in the field and lab and for preparing the map of Swede Creek, Anne Mullenniex for help with the cpDNA markers, Robert Douglas for help in the field and lab, John McRory for help in the field, Renate Wesselingh for writing the randomization program, and Mike Arnold, Amy Bouck, Mark Bulger, John Burke, Jill Johnston, Tim Lawlor, Beth McCoy, John Sawyer, Chris Spencer, Jacob Varkey, and Joe Williams for valuable discussions and comments on earlier versions of this manuscript. 
3 Author for correspondence, current address: Department of Genetics, University of Georgia, Athens, Georgia 30602 USA (e-mail: ekentner{at}arches.uga.edu
).
LITERATURE CITED
Anderson, E. 1949 Introgressive hybridization. John Wiley & Sons, New York, New York, USA.
———, and L. Hubricht. 1938 Hybridization in Tradescantia III. The evidence for introgressive hybridization. American Journal of Botany 25: 396–402.[CrossRef][Web of Science]
———, and G. L. Stebbins, Jr. 1954 Hybridization as an evolutionary stimulus. Evolution 8: 378–388.
Arnold, J. 1993 Cytonuclear disequilibria in hybrid zones. Annual Review of Ecology and Systematics 24: 521–554.
Arnold, M. L. 1997 Natural hybridization and evolution. Oxford University Press, New York, New York, USA.
———, and B. D. Bennett. 1993 Natural hybridization in Louisiana irises: genetic variation and ecological determinants. In R. G. Harrison [ed.], Hybrid zones and the evolutionary process, 115–139. Oxford University Press, Oxford, UK.
Asmussen, M. A., J. Arnold, and J. C. Avise. 1987 Definition and properties of disequilibrium statistics for associations between nuclear and cytoplasmic genotypes. Genetics 115: 755–768.
Ball, C. T., J. Keeley, H. Mooney, J. Seemann, and W. Winner. 1983 Relationship between form, function, and distribution of two Arctostaphylos species (Ericaceae) and their putative hybrids. Acta Oecologia/Oecologia Plantarum 4: 153–164.
Barton, N. H., and K. S. Gale. 1993 Genetic analysis of hybrid zones. In R. G. Harrison [ed.], Hybrid zones and the evolutionary process, 13–45. Oxford University Press, Oxford, UK.
———, and G. M. Hewitt. 1985 Analysis of hybrid zones. Annual Review of Ecology and Systematics 16: 113–148.
Cruzan, M. B., and M. L. Arnold. 1993 Ecological and genetic associations in an Iris hybrid zone. Evolution 47: 1432–1445.[CrossRef][Web of Science]
Doyle, J. J., and J. L. Doyle. 1987 A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15.
Friend, D. T. C. 1961 A simple method for measuring integrated light values in the field. Ecology 42: 577–580.[CrossRef][Web of Science]
Grant, V. 1981 Plant speciation. Columbia University Press, New York, New York, USA.
Hardig, T. M. 1995 A morphometric and molecular investigation of hybridization between Polystichum munitum, and P. imbricans in Northern California. Master's thesis, Humboldt State University, Arcata, California, USA.
Harrison, R. G. 1986 Pattern and process in a narrow hybrid zone. Heredity 56: 337–349.[Web of Science]
Haufler, C. H. 1987 Electrophoresis is modifying our concepts of evolution in homosporous pteridophytes. American Journal of Botany 74: 953–966.[CrossRef][Web of Science]
Howard, D. J. 1986 A zone of overlap and hybridization between two ground cricket species. Evolution 40: 34–43.[CrossRef][Web of Science]
Kearns, C. A., and D. W. Inouye. 1993 Techniques for pollination biologists. University Press of Colorado, Niwot, Colorado, USA.
Lubbers, A. E. 1982 Spatial and temporal variation in the reproductive characteristics of Thalictrum thalictroides (L.) Eames and Boivin, a forest herbaceous perennial. Ph.D. dissertation, Duke University, Durham, North Carolina,USA.
———, and N. L. Christensen. 1986 Intraseasonal variation in seed production among flowers and plants of Thalictrum thalictroides (Ranunculaceae). American Journal of Botany 73: 190–203.[CrossRef][Web of Science]
Manly, B. F. J. 1997 Randomization, bootstrap and monte carlo methods in biology. Chapman and Hall, London, UK.
Mayer, M. S. 1989 A morphometric investigation of hybridization between between Polystichum munitum, and P. imbricans in Northern California. Master's thesis, Humboldt State University, Arcata, California, USA.
———, and M. R. Mesler. 1993 Morphometric evidence of hybrid swarms in mixed populations of Polystichum munitum and P. imbricans (Dryopteridaceae). Systematic Botany 18: 248–260.[CrossRef][Web of Science]
Moore, W. S. 1977 An evaluation of narrow hybrid zones in vertebrates. Quarterly Reviews in Biology 52: 263–277.[CrossRef][Web of Science]
Mullenniex, A. F. 1996 A molecular and morphometric investigation of hybridization between the two fern species Polystichum munitum and P. imbricans (Dryopteridaceae). Master's thesis, Humboldt State University, Arcata, California, USA.
———, T. M. Hardig, and M. R. Mesler. 1999 Molecular confirmation of hybrid swarms in the fern genus Polystichum (Dryopteridaceae). Systematic Botany 23: 421–426.[CrossRef][Web of Science]
Pickett, F. L, and M. E. Manuel. 1926 An ecological study of certain ferns: Pellaea atropurpurea (L.) Link and Pellaea glabella Mettenius. Bulletin of the Torrey Botanical Club 53: 1–5.[CrossRef]
Rand, D. M., and R. G. Harrison. 1989 Ecological genetics of a mosaic hybrid zone: mitochondrial, nuclear, and reproductive differentiation of crickets by soil type. Evolution 42: 432–449.
Rieseberg, L. H., and J. F. Wendel. 1993 Introgression and its consequences in plants. In R. G. Harrison [ed.], Hybrid zones and the evolutionary process,70–109. Oxford University Press, Oxford, UK.
Scholander, P. F., H. T. Hammel, E. D. Bradstreet, and E. A. Hemminsen. 1965 Sap pressure in vascular plants. Science 148: 339–346.
Soltis, D. E., C. H. Haufler, D. C. Darrow, and G. J. Gastony. 1983 Starch gel electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers, and staining schedules. American Fern Journal 73: 9–27.[CrossRef][Web of Science]
———, P. S. Soltis, T. G. Collier, and M. L. Edgerton. 1991 Chloroplast DNA variation within and among genera of the Heuchera group (Saxifragaceae): evidence for chloroplast transfer and paraphyly. American Journal of Botany 78: 1091–1112.[CrossRef][Web of Science]
Soltis, P. S., and D. E. Soltis. 1987 Population structure and estimates of gene flow in the homosporous fern Polystichum munitum. Evolution 41: 620–629.
———, ———, and P. G. Wolf. 1990 Allozymic divergence in North American Polystichum (Dryopteridaceae). Systematic Botany 15: 205–215.[CrossRef][Web of Science]
Stebbins, G. L., Jr. 1959 The role of hybridization in evolution. Proceedings of the American Philosophical Society 103: 231–251.
Vogel, J. C., S. J. Russell, F. J. Rumsey, J. A. Barrett, and M. Gibby. 1998 Evidence for maternal transmission of chloroplast DNA in the genus Asplenium (Aspleniaceae, Pteridophyta). Botanica Acta 111: 247–249.[Web of Science]
Wagner, D. H. 1979 Systematics of Polystichum in Western North America north of Mexico. Pteridologia 1.
———. 1987 Meiotic pairing success affected by environment in a Polystichum hybrid. American Journal of Botany 74: 762–763.
Yatskievych, G. 1993 Antheridiogen response in Phanerophlebia and related fern genera. American Fern Journal 83: 30–36[CrossRef]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
