Asymmetrical gene flow, which has frequently been documented in naturally occurring hybrid zones, can result from various genetic and demographic factors. Understanding these factors is important for determining the ecological conditions that permitted hybridization and the evolutionary potential inherent in hybrids. Here, we characterized morphological, nuclear, and chloroplast variation in a putative hybrid zone between Schiedea menziesii and S. salicaria, endemic Hawaiian species with contrasting breeding systems. Schiedea menziesii is hermaphroditic with moderate selfing; S. salicaria is gynodioecious and wind-pollinated, with partially selfing hermaphrodites and largely outcrossed females. We tested three hypotheses: 1) putative hybrids were derived from natural crosses between S. menziesii and S. salicaria, 2) gene flow via pollen is unidirectional from S. salicaria to S. menziesii and 3) in the hybrid zone, traits associated with wind pollination would be favored as a result of pollen-swamping by S. salicaria. Schiedea menziesii and S. salicaria have distinct morphologies and chloroplast genomes but are less differentiated at the nuclear loci. Hybrids are most similar to S. menziesii at chloroplast loci, exhibit nuclear allele frequencies in common with both parental species, and resemble S. salicaria in pollen production and pollen size, traits important to wind pollination. Additionally, unlike S. menziesii, the hybrid zone contains many females, suggesting that the nuclear gene responsible for male sterility in S. salicaria has been transferred to hybrid plants. Continued selection of nuclear genes in the hybrid zone may result in a population that resembles S. salicaria, but retains chloroplast lineage(s) of S. menziesii.
Citation: Wallace LE, Culley TM, Weller SG, Sakai AK, Kuenzi A, Roy T, et al. (2011) Asymmetrical Gene Flow in a Hybrid Zone of Hawaiian Schiedea (Caryophyllaceae) Species with Contrasting Mating Systems. PLoS ONE 6(9): e24845. https://doi.org/10.1371/journal.pone.0024845
Editor: Jeff Demuth, University of Texas Arlington, United States of America
Received: June 13, 2011; Accepted: August 22, 2011; Published: September 19, 2011
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Funding: The authors acknowledge support from NSF grants DEB-0445410 to MN and LW and DEB-9207724 to AKS and SGW as well as the McBryde endowment of National Tropical Botanical Garden to MN and WLW. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Analyses of hybrid zones have provided a wealth of information on factors that favor interspecific hybridization, the nature of pre- and post-zygotic barriers to hybridization, and the stability of hybrid zones –. These studies provide insights into the role of hybridization in contemporary natural populations as well as past hybridization events. The structure and stability of hybrid zones depend on the extent to which parental species are genetically and ecologically distinct, their dispersal abilities, and the fitness of hybrid offspring . When hybrids have lower fitness compared to their parents, a tension zone is formed and repeated hybridization is needed to maintain the hybrid zone , . In contrast, when hybrids have higher fitness than their parents, they are likely to increase in frequency, and may displace adjacent parental populations to expand their range , .
Contrasts between maternally-inherited chloroplast DNA and nuclear DNA inherited through both parents can provide confirmation of hybridization and evidence for the direction of gene flow –. Asymmetric introgession is much stronger in organelle genes than nuclear genes because the former experiences lower rates of gene flow due to uniparental inheritance. Selection is often invoked to explain asymmetric patterns of introgression , . For example, selection for greater pollen production or a more effective means of dispersing pollen could result in preferential gene flow to a species with lower pollen production or a less effective means of pollen dispersal. Demographic factors may also play an important role in determining the direction of introgression as Currat et al.  found strong evidence of asymmetrical introgression in a wide variety of taxa where gene flow occurred from a local to an invading species.
When hybridization occurs between selfing and outcrossing species, a rapidly advancing hybrid zone favoring the outcrossing species is expected to develop. For example, when dioecious, wind-pollinated populations of Mercurialis annua (Euphorbiaceae) came into contact with monoecious, more highly selfing populations of M. annua, the monoecious populations were swamped by pollen from the dioecious wind-pollinated populations . Rapid movement of the hybrid zone favoring the dioecious form of M. annua was thought to result from differences in pollen flow related to the differing breeding systems within this species. Variation in pollinator preference could have similar effects on the outcome of hybridization. Campbell et al.  described an ‘advancing wave’ dynamic for Ipomopsis aggregata and I. tenuituba (Polemoniaceae). In the hybrid zone between these two species, traits of I. aggregata were favored because hummingbirds preferentially visit I. aggregata and are much more common pollinators than the hawkmoths that favor I. tenuituba. In experimental arrays, I. aggregata received much more pollen from conspecifics and hybrids than I. tenuituba . Similar array studies demonstrated that hybridization between species of Castilleja (Orobanchaceae) was also influenced by pollinator constancy, which was in turn determined by the frequency of hybrids, leading to the possibility of a self-reinforcing mechanism creating areas where hybrids might be very common . Finally, in a hybrid zone between Penstemon davidsonii and P. newberryi (Plantaginaceae), gene flow from the high-elevation P. davidsonii to the lower elevation P. newberryi was mediated primarily through frequent visits by hummingbirds and other pollinators to both hybrids and P. newberryi . Consequently, knowledge of pollinator behavior and the reproductive system are important for predicting the outcome of hybridization and the movement of current hybrid zones across the landscape.
We investigated the potential for hybridization between two endemic Hawaiian species of Schiedea (Caryophyllaceae), considering how variation in reproductive systems and pollination biology could influence genetic structure of the putative hybrid zone. Schiedea, the fifth largest plant radiation in the Hawaiian Islands, is notable for its high diversity in breeding systems and associated variation in outcrossing rates , . Dioecious species are obligate outcrossers, while hermaphroditic species range from highly outcrossing  to autogamous species . Species with intermediate outcrossing rates (mixed mating systems) may be hermaphroditic (e.g., S. menziesii) , gynodioecious (e.g., S. salicaria) , or subdioecious (e.g., S. globosa) , . All sexually dimorphic (including gynodioecious, subdioecious and dioecious) species of Schiedea are wind-pollinated and occur in dry habitats . In contrast, hermaphroditic species of Schiedea, even those occurring in dry habitats, are not adapted for wind-pollination.
Morphological and molecular traits were analyzed in a putative hybrid zone between hermaphroditic, moderately selfing Schiedea menziesii and gynodioecious, wind-pollinated Schiedea salicaria. Schiedea menziesii is protandrous and believed to be self-pollinated through geitonogamous pollinations . No potential biotic pollinators have been observed in natural populations of this species ; thus, limited pollen movement leading to selfed progeny is expected to occur through wind pollination. Schiedea salicaria shows adaptations to wind pollination, including an abundance of pollen and smaller pollen grains than those of biotically pollinated species . This species produces hermaphroditic individuals that are partially selfing and female plants that are outcrossing (Table 1). The putative hybrid nature of plants that exhibit a wide range of morphological variation resembling S. menziesii and S. salicaria was suggested by the earliest plant collectors on West Maui . Thus, we hypothesized that these individuals with intermediate morphologies represent hybrids from the two hypothesized parental species rather than variants of one of the parental species. Furthermore, we hypothesized that S. salicaria served as the pollen parent of the putative hybrids because it is outcrossing and wind-pollinated, characteristics that should allow for more efficient pollen transfer compared to S. menziesii, the more selfing species. As a result of this difference in pollination, the putative hybrids are hypothesized to resemble S. salicaria in morphological traits due to the greater influx of genes from S. salicaria and selection for traits of this species. Given these hypotheses, we expected the following: (1) Plants will contain a mixture of molecular markers and morphological traits from both parents; (2) Plants will retain the maternal genome of S. menziesii, which will be reflected by similarity of chloroplast markers between the hybrids and S. menziesii; (3) Neutral nuclear molecular markers in the putative hybrids, in contrast, will be represented by a combination of those found in both parental species; (4) Morphological traits associated with wind pollination will resemble S. salicaria because wind has selected for characteristics that promote higher rates of pollen dispersal exhibited by S. salicaria. This pattern is similar to the ‘advancing wave’ dynamic described by Campbell et al.  for Ipomopsis, where the agent of selection is biotic, rather than abiotic.
Hybrid individuals were variable, with some individuals possessing elongate leaves of intermediate width with a single vein, and others producing broader, weakly three-veined leaves. The first and second principal components of the multivariate analysis of morphological traits accounted for 34% and 20.5% of the observed variability (Fig. 1). The first principal component loaded most heavily on leaf and pubescence traits, as well as the relative shape of the inflorescence (narrow vs. broad due to the length of lateral branches). The second principal component loaded most heavily on internode length. Canonical discriminant analysis confirmed the major conclusions based on the principal components analysis. Traits having the highest values for the first canonical variable were leaf length and width, the number of veins in the leaves, inflorescence pubescence, pedicel length, and lateral branch length in the inflorescence. Comparison of values for the first canonical variable of the parental species and plants from the hybrid zone (Fig. 2) indicated that morphological traits in most of the hybrids were more similar in morphology to S. salicaria than S. menziesii, although four of the 14 hybrids more closely resembled S. menziesii.
Although differences were not significant, hybrids were intermediate between the parental species in pollen production and pollen size, with values for both traits closer to S. salicaria, the wind-pollinated parent. Based on seven hybrid individuals, pollen production per flower for hybrids averaged 14,645 pollen grains per flower (SE = 1,113), compared to values of 14,756 (SE = 1,689) for S. salicaria, and 13,935 (SE = 1,645) for S. menziesii . Pollen size of the hybrids averaged 34.0 µm (SE = 0.47), compared to 33.4 µm (SE = 1.76) for S. salicaria and 34.8 µm (SE = 0.71) for S. menziesii. Average pollen fertility of the hybrids was 95.9% (SE = 1.29).
The hybrid zone included plants that were female. Of the 25 flowering plants surveyed for sex expression in the field, 7 (28%) were female, not significantly different from the 12–13% frequency of females observed in populations of S. salicaria . Schiedea menziesii is primarily hermaphroditic in the field. Only one out of several hundred plants examined in three field populations was female.
Despite the rarity of these taxa, a substantial amount of genetic variation was detected within taxa and populations (Table 2). Significant deviations from Hardy-Weinberg equilibrium were detected for locus SA30 in S. salicaria population 842, S. menziesii population 950, and the hybrids, and for locus SA31 in S. menziesii population 950 and the hybrids. Possible null alleles were detected in several populations for locus SA30 (S. menziesii 849, 949, 950; S. salicaria 842; hybrids), but this locus did not differ significantly in allele frequencies among the parents or hybrids (Table 3). Additional null alleles were detected at locus SA04 in population S. menziesii 950 and at locus SA06 in population S. salicaria 842. Over 85% of the loci were polymorphic in both species and their hybrids, and expected heterozygosity was high in all taxa (S. salicaria: Ho = 0.45, He = 0.53; S. menziesii: Ho = 0.35, He = 0.45; hybrids: Ho = 0.39, He = 0.49; Table 2). The hybrid population contained more alleles on average (A = 3.67; Ap = 4.00) than S. salicaria (A = 3.56; Ap = 3.88) or S. menziesii (A = 3.26; Ap = 3.64). Schiedea menziesii and S. salicaria shared 28 alleles, and the hybrids harbored 22 of these. The hybrids shared four alleles with S. menziesii that were not present in S. salicaria, and two alleles shared between the hybrids and S. salicaria were not found in S. menziesii. Five rare alleles were unique to the hybrids (mean frequency = 0.064). Despite the lack of high frequency species-specific alleles, S. menziesii and S. salicaria exhibited significantly different allele frequencies at six of the nine loci (FST = 0.117; P<0.0001; Table 3).
No evidence of linkage disequilibrium was detected in the hybrid population, but we did find significant FST values (P<0.05) for four of the six loci that also exhibit significantly different allele frequencies in the parental species (Table 3). The Bayesian assignment of population ancestry based on only two groupings followed species designations for many S. menziesii and S. salicaria individuals (Fig. 3). Sixteen of the 36 samples of S. menziesii were assigned to a single cluster and all had probabilities of assignment greater than 91% (see black bars in Fig. 3). None of the eight samples from population 949 could be assigned to a single cluster, but this may reflect intraspecific variation that is also apparent in the cpDNA data set (see results below). Sixteen of the 25 samples of S. salicaria were also assigned to a single cluster with greater than 91% probability, and this cluster was distinct from S. menziesii (see white bars in Fig. 3). One S. salicaria plant out of 20 samples in population 842 was assigned to the S. menziesii cluster with a probability of 92%. Six hybrid samples were assigned to a single cluster (three to each parental cluster) with a probability of at least 91%, whereas the remaining eight hybrid plants exhibited shared ancestry between the two clusters (Fig. 3).
For each individual the probability of assignment to the two clusters is indicated. A cut-off of 90% probability of assignment to a single cluster was used to assign individuals; this cut-off is indicated by the horizontal lines at the top (for black bars) and bottom (for white bars) of the plot. Black bars correspond to the primarily–S. menziesii cluster, and white bars to the primarily-S. salicaria cluster. Individuals with mixed ancestry exhibit a combination of black and white bars with neither bar reaching 90% probability.
The five loci exhibited different levels of variation within each taxon (Table S1). In the concatenated data set, 15 unique multilocus chloroplast haplotypes were identified, including eight in S. menziesii, five in S. salicaria, and two in the hybrids (Table 4). No hybrid individual contained a haplotype that was identical to either parental species, but the hybrid haplotypes were more similar to S. menziesii than to S. salicaria, including having single locus haplotypes identical to those of S. menziesii at four of the loci (rpS16, petN-psbM, psbE-petL, trnL microsatellite) and differing by a minimum of four steps from S. menziesii haplotype 2 (Fig. 4). This topology is robust to different methods of gap coding and analysis as similar results were produced when we employed multistate gap coding  and maximum parsimony analyses  (results not shown).
Gaps were included in the analysis. Observed haplotypes are numbered and indicated as diamonds (S. menziesii), squares (S. salicaria), or triangles (hybrids). Small circles represent unobserved inferred haplotypes. The line connecting each haplotype indicates one mutational change, except where otherwise indicated. Closed connections among haplotypes indicate an alternative relationship between haplotypes. The 24 steps between haplotypes 7 and 4 include a 23 base indel in ndhJ-trnF, the 10 steps between an inferred haplotype and haploytpe 7 include a 10 base indel in psbM-trnD, and a 5 base indel in rpL16 distinguishes S. menziesii and hybrid haplotypes from those of S. salicaria.
Gene diversity was highest in S. menziesii (H = 0.640), with S. salicaria and the hybrids exhibiting no nucleotide diversity (Table 4). However, variation was observed in S. salicaria and the hybrids in insertions/deletions. The two S. salicaria populations did not share haplotypes, although the haplotypes in these two populations were distinct from hybrid and S. menziesii haplotypes. Within S. menziesii, populations 849 and 949 also did not share haplotypes, but population 950 shared haplotypes with the other two S. menziesii populations (Table 4). AMOVA indicated that S. menziesii and S. salicaria were genetically distinct (FST = 0.68; P<0.0001) and that the hybrids were also distinct from both parents (S. menziesii: FST = 0.29, P = 0.002; S. salicaria: FST = 0.97, P<0.0001).
The dynamic landscape of oceanic islands, where active volcanoes continue to generate new land masses, suggests a wealth of opportunity for interspecific hybridization through the erosion of ecological barriers or the persistence of hybrid individuals in habitats separated from parental species. Indeed, instances of historical and on-going hybridization are increasingly reported , . On the Hawaiian Islands, 210 instances of interspecific hybridization spanning 59 genera have been documented . Among these are the Hawaiian silversword alliance –, Cyrtandra , Metrosideros , Portulaca , Rubus , and Scaevola . In addition to the S. menziesii- S. salicaria hybrid zone discussed here, Willyard et al.  uncovered several instances of cryptic chloroplast introgression throughout Schiedea. As expected of recently evolved systems, many of these hybridization events occur between species that are not widely divergent genetically , and ecological barriers are typically stronger than genetic barriers to gene flow , . Given the increased detection of interspecific hybridization in continental and island plant systems, a more detailed understanding of the adaptive nature of hybridization  is warranted. Studies that utilize multiple sources of data, as we have here, serve as the best tests of the dynamics of hybridization and the adaptive potential inherent in hybrids.
Status of the Putative Hybrids
Plants exhibiting a wide range of morphological variation resembling S. menziesii and S. salicaria were found on West Maui by the earliest plant collectors and have been suggested to be putative hybrids between these species  or variants of S. salicaria . The combination of morphological traits and nuclear and chloroplast genetic markers described in the present study suggests that all plants from Halepohaku, the putative hybrid zone, are hybrids of S. menziesii and S. salicaria, rather than variants of S. salicaria sensu stricto  or a variety of S. salicaria . Whereas the two parental species are morphologically and genetically distinct, the hybrid population exhibits evidence of ancestry from both parents. The hybrids most closely resembled S. salicaria in inflorescence and vegetative traits (Figs. 2, 3), but they have chloroplast genomes that are more similar to those in S. menziesii (Fig. 4). Although few species-specific alleles were found at the nuclear microsatellite loci, there is also evidence in these data to suggest a hybrid origin for these plants. The parental species are significantly differentiated at six of the nine loci examined (FST = 0.117; P<0.05), and the hybrids exhibit significant differences from just one parent at four of these loci (Table 3), suggesting that the other parent was the predominant donor of alleles at those loci. If the purported hybrids were variants of a single species, then they would be expected to show the same patterns of allelic differentiation when compared to conspecific populations. The hybrids also have a greater number of alleles per locus than either parental species, consistent with the expectation that both species have contributed nuclear alleles to the hybrid zone. Bayesian assignment of hybrid plants to the two clusters, which roughly correspond to the two parental species, also points to a hybrid origin for many of the plants in this population. Welch and Rieseberg  found a similar lack of species-specific alleles, but divergent allele frequencies, in the diploid parents of the hybrid species, Helianthus paradoxus. According to Chapman and Burke , low levels of genetic divergence between parental species are more likely to result in homoploid hybrids than allopolyploids after interspecific hybridization. Although chromosome counts have not been done for these species of Schiedea, the high level of fertility between hybrid zone plants and S. salicaria and between S. menziesii and S. salicaria, suggests that there are not ploidal differences between the parents and plants from the hybrid zone (S. Weller and A. Sakai, unpubl. data). No more than two alleles were ever found at the microsatellite loci, which also suggests the hybrids are diploid. Thus, the lack of species-specific alleles at nuclear loci in S. menziesii and S. salicaria is consistent with findings in other hybrid systems involving closely related parental taxa.
The hybrid population appears to be a naturally occurring contact zone between S. menziesii and S. salicaria, given that human encroachment in the area is limited. In light of Hillebrand's early collections, the S. menziesii - S. salicaria hybrid zone is likely a stable population that has existed for more than a century. Although the nuclear loci were not sufficiently variable to differentiate among early and late generation hybrids, the wide range of morphological and genotypic variability among the hybrids suggests these individuals are not first generation hybrids. Viable hybrids of S. menziesii and S. salicaria form readily in the greenhouse and are morphologically similar to naturally occurring hybrids (S. Weller & A. Sakai, unpublished data). The high fertility of the hybrids under these conditions indicates that there are unlikely to be intrinsic barriers to gene exchange among the hybrids in the field, consistent with the hypothesis that plants in the hybrid zone are F2 or later generation hybrids. Additionally, we did not find significant linkage disequilibrium among any of the loci in the hybrid population, which further suggests that this is an older stable hybrid population in which certain allelic combinations have been broken up and a range of genotypes are now present in the population. Thus, this hybrid zone is most similar to the stable zones predicted when hybrids have equal or greater fitness than their parents , .
Asymmetric Gene Flow Occurs from S. salicaria to S. menziesii
Hybrid zones can be generated from unidirectional gene flow from one species to another, or through bidirectional gene flow with both species serving as maternal and paternal parents. The direction of gene movement in hybrid zones has important implications for understanding the evolutionary potential and sustainability of hybrid progeny. For example, both parents have contributed to the cytoplasmic diversity of the hybrid sunflower Helianthus deserticola, leading to high levels of differentiation among populations . In the polyploid Tragopogon complex, T. miscellus exhibits different floral morphology depending on whether T. dubius or T. pratensis is the maternal parent . In contrast, other studies of natural hybrid zones have demonstrated unidirectional gene flow , , . Actually, instances of asymmetric gene flow in hybrid zones are commonly documented in other natural systems , including Hawaiian taxa. For example, Friar et al.  identified ongoing hybridization from Hawaiian Dubautia ciliolata to D. arborea leading to the formation of an active hybrid zone. However, there was little evidence of introgression into parental populations . Randell et al.  also found asymmetrical hybridization between Rubus hawaiiensis and R. rosifolius on Maui based on the presence of shared cpDNA haplotypes between the hybrids and R. rosifolius but not R. hawaiiensis. Howarth and Baum's  study of Scaevola demonstrates that homoploid hybrid speciation may be particularly common in island systems where congeneric species are often not strongly diverged. We have also found evidence that hybridization occurs readily between S. menziesii and S. salicaria and that gene flow in this natural hybrid zone has been asymmetrical. Assuming that the cpDNA genome is maternally inherited in Schiedea, an assumption that is supported across many flowering plant families ,  as well by our sequence data on maternal families (results not shown), the cpDNA data are most consistent with S. menziesii being the seed parent of the extant hybrids. Although haplotypes were not shared between species or with the hybrids, the hybrid haplotypes were more genetically similar to those of S. menziesii from which they differed a minimum of four mutational steps from S. menziesii haplotypes on the central ridge of Lihau (849; Fig. 4 haplotypes 1–3) and those from the gulch below this ridge (950; Fig. 4 haplotypes1–3 and 8), but they differed at least 24 steps from S. salicaria haplotypes (Fig. 4 haplotypes 9–13). This implies that the pollen parent has been S. salicaria, and this is consistent with the occurrence of wind pollination in S. salicaria and its absence in S. menziesii .
The conditions that facilitated the initial formation of hybrids certainly required that S. menziesii and S. salicaria be in closer physical contact than they are today, and this likely occurred in conjunction with past environmental and/or climatic changes on the Hawaiian Islands , . For example, the island of Maui was part of Maui Nui, an island that reached a maximum size of 14,000 km2  during the Pleistocene. It is expected that contact between the parental species occurred during this time as they expanded their ranges to new areas of Maui Nui. Presently, the parental species are not in close physical contact. Schiedea menziesii occurs in shrublands situated on ledges and cliffs from sea level up to 670 m, while S. salicaria is found in shrublands along steep slopes from 180 to 670 m. The extant hybrid zone is in an area closer to the range of S. menziesii, and other hybrid zones have also been noted in this area on West Maui . Trade winds blow across the Hawaiian Islands from a northeasterly direction, but wind direction on the scale necessary for pollen movement is variable and depends upon local topography (S. Weller & A. Sakai, pers. observ.). While S. salicaria populations are currently situated to the east of S. menziesii, they are too distant to serve as the source of pollen. Thus, we suggest that populations of S. salicaria likely occurred closer to the hybrid zone in the past, and this permitted interspecific hybridization to occur. Contact between the parental species may have been limited geographically, given that that S. menziesii from the Hahakea population 949, located 10.3 km north of the hybrid zone (Fig. 5) is genetically distinct with cpDNA haplotypes that are nearly as distinct from haplotypes in conspecific populations 849 and 950 as they are from S. salicaria and the hybrid haplotypes. Population 949 also exhibited genetic differences in the nuclear data and could not readily be assigned to the S. menziesii cluster in the Bayesian analysis.
Schiedea menziesii, diamonds; S. salicaria, squares; hybrids, triangles. The inset map shows the location of Maui relative to the other Hawaiian Islands.
We found no evidence of introgression of S. menziesii plastid haplotypes into extant S. salicaria populations. However, there may be some evidence of bidirectional gene flow within the nuclear data set. Based on the Bayesian assignment analysis, one of the S. salicaria samples has a higher probability of ancestry from S. menziesii, and 29 other parental samples (20 of S. menziesii and 9 of S. salicaria; Fig. 3) exhibit shared ancestry. While this could be due to low polymorphism and a lack of species-specific alleles at these loci, many individuals were confidently assigned to a single cluster, and the species have significantly different allele frequencies (Table 3). Additional nuclear data from more variable loci are needed to validate this inference, but these data suggest cryptic remnants of gene flow from S. menziesii into S. salicaria. Similar patterns of heterogeneous gene flow across the genome have been reported in Dubautia . The widespread lack of chloroplast introgression from S. menziesii into S. salicaria could be due to selection against individuals with this cytonuclear combination or pollen-swamping by S. salicaria due to its greater pollen production and more effective means of pollen dispersal compared to S. menziesii.
Currat et al.  predict that unidirectional cytoplasmic introgression should occur from the local species to the invading species. Early on, bidirectional gene flow could have occurred during the initial contact between parental species even in the face of greater pollen production by S. salicaria if S. menziesii outnumbered invading S. salicaria plants. If F1 hybrids containing cpDNA genomes of S. salicaria subsequently served as pollen parents in backcrosses to S. menziesii and continued to increase in frequency, then all traces of S. salicaria serving as the seed parent would be lost. It is likely that the extant hybrids do not resemble S. salicaria in their chloroplast genome because they are late-generation hybrids and there has been widespread pollen-swamping by S. salicaria or hybrids that resemble this species. By having a distinct advantage in greater pollen production than S. menziesii, hybrids that were more like S. salicaria would have been more likely to swamp S. menziesii, thereby yielding hybrids with cpDNA haplotypes that are more similar to S. menziesii.
The absence of identical haplotypes in the hybrids and S. menziesii can be explained by 1) unsampled S. menziesii populations, including extinct populations, containing those haplotypes, 2) origin from S. menziesii haplotypes followed by mutational changes in the chloroplast loci since the formation of the hybrid zone, or 3) the presence of heteroplasmy and recombination within the chloroplast genome, which have been demonstrated for mitochondrial genomes in other members of the Caryophyllaceae . We do not believe that the hybrid haplotypes are recombinants because there is little homoplasy in the haplotype network and none associated with the hybrid haplotypes. Additionally, if they were recombinants, we would expect the hybrids to exhibit variation in how similar they are to each of the parents at different loci. Instead, we see a greater similarity to S. menziesii at six of the seven loci, suggesting a direct origin from this species rather than recombination. Evaluation of the other two hypotheses requires comparison of additional genetic markers to quantify genome-wide variation between the hybrids and the parental species.
Wind Pollination Favors Traits of S. salicaria in the Hybrid Zone
When hybridization occurs between selfing and outcrossing species, traits of the outcrossing species are expected to be favored among hybrids. Moreover, when there are differences in pollination mechanism, it is expected that selection will favor the species with more effective pollen dispersal. Because S. salicaria disperses pollen via wind while S. menziesii shows no indication of wind pollination , we hypothesized that selection should favor morphological traits most similar to S. salicaria in the hybrids as a result of pollen-swamping by this species. If wind is the selective agent in the hybrid zone, traits that promote more effective pollen transfer will be favored. Thus, in each generation, S. salicaria traits promoting wind pollination (i.e., small pollen size, greater pollen production, inflorescence traits) should increase in frequency in the hybrid zone because individuals with these traits are most likely to serve as pollen donors. Other S. salicaria traits controlled by nuclear genes should also increase as wind pollination favors the nuclear genome of S. salicaria. We found a high degree of similarity between the hybrids and S. salicaria in the traits associated with wind pollination, including greater pollen production, smaller pollen size, more flowers per inflorescence, and the details of inflorescence shape (Figs. 1, 2). Thus, these data are consistent with evolution of morphological traits associated with wind-pollination among the hybrid plants.
Male-sterility is controlled by nuclear genes in Schiedea , and as a result, breeding system traits were also expected to be influenced by hybridization between S. menziesii and S. salicaria. The presence of unisexual flowers is often seen in conjunction with wind pollination –. In Schiedea, once male-sterility alleles have been introduced into a population, wind pollination may be necessary for females to become established in populations because of pollinator limitation , . We predicted that some hybrid individuals would exhibit male sterility as a result of gene flow from gynodioecious S. salicaria to hermaphroditic S. menziesii. The occurrence of male sterility in the hybrid zone (28% females) suggests that gene flow from S. salicaria has prominently shaped the reproductive system of the hybrid zone. In conjunction with the morphological adaptations to wind pollination, females may have a selective advantage in the ecological setting of the hybrid zone. In contrast, females are very rare in the field for S. menziesii, presumably because of the absence of both pollinators and traits favoring wind pollination , . Hybrids between S. menziesii and S. salicaria may be evolving towards more effective mechanisms for wind pollination and outcrossing, including the occurrence of females, in light of the high levels of inbreeding depression (Table 1) found in populations of both parental species , , and the occurrence of high levels of heritable variation for traits associated with wind pollination , .
We have documented the presence of an interspecific hybrid zone between two species of Schiedea on West Maui and found evidence of unidirectional cytoplasmic gene flow contributing to the diversity of the current hybrid population. Although bidirectional gene flow may have occurred in the past, as evidenced by mixed ancestry of some individuals from the parental species, we suggest that the selective advantage of wind pollination in the hybrid zone has driven phenotypes towards S. salicaria, the wind pollinated species. Because S. menziesii is not adapted to wind pollination and seed dispersal occurs over very short distances in all Schiedea species, introgression from S. menziesii is unlikely to alter the breeding system of S. salicaria . However, the hybrid zone could influence the evolution of dimorphism in S. menziesii by introducing genes from S. salicaria associated with male sterility and wind pollination if these plants come in contact with hybrids of S. salicaria in the future. Hybridization has facilitated the evolution of dimorphic breeding systems from hermaphroditism in other species . Recent phylogenetic analyses of Schiedea have revealed widespread hybridization events leading to chloroplast capture . Given the dynamics of the hybrid zone between S. menziesii and S. salicaria uncovered here, interspecific hybridization may have the potential to promote breeding system evolution in other species of Schiedea, just as it has promoted radiations of other notable Hawaiian endemic groups , .
Materials and Methods
Schiedea menziesii and S. salicaria are rare shrubs endemic to the Hawaiian Islands and inhabiting remnant dry shrubland of the West Maui Mountains , . Schiedea salicaria is currently known only on the eastern side of West Maui. One historic collection of S. salicaria near Lahaina on the west side of West Maui is likely to represent a hybrid with S. menziesii . Both S. menziesii and S. salicaria occur on dry, north-facing steep slopes at 30–670 m and 180–670 m elevation, respectively . Both species were more widespread historically and one population of S. menziesii on Lana'i has been extirpated. Leaves of S. menziesii are typically falcate and linear to linear-lanceolate while those of S. salicaria are shorter, symmetrical, and elliptic . S. salicaria leaves have a single vein, while leaves of S. menziesii almost always have three veins running the length of the leaf. Inflorescences of both species are similar, although those of S. menziesii are slightly more laterally compressed and have fewer, larger flowers. Flowers are apetalous and appear during the wetter winter months when most plant growth occurs. Based on wind tunnel studies, S. salicaria is wind-pollinated, while S. menziesii is not adapted to wind pollination and has larger pollen grains and reduced pollen production compared to S. salicaria . Hermaphrodites of both species have substantial levels of self pollination (s = 0.398–0.607, Table 1) , . Schiedea salicaria is gynodioecious with 12–13% females ,  while S. menziesii has previously been considered hermaphroditic . Females of S. menziesii have been produced from field-collected seed, however, and a single female individual has been observed in a natural population (Table 1; T. Culley, pers. obs.). Sex in Schiedea is under nuclear control with male sterility determined at a single locus . Females are homozygous recessive (hh) and hermaphrodites are either heterozygous (Hh) or homozygous dominant (HH). Vigorous hybrids have been produced between the two species by hand-pollinations under greenhouse conditions (S. Weller & A. Sakai, unpublished data). Putative natural hybrid populations have also been documented on West Maui . Plants within these putative hybrid populations exhibit a wide range of morphological variation in vegetative traits.
The hybrid zone sampled for this study occurs on Halepohaku, on the southwest side of West Maui between the ranges of the two parental species (Fig. 5). Sex ratios were recorded in February, 1995. All plants located in this area appeared to be of hybrid origin based on their morphological traits. Schiedea salicaria was sampled from two populations on the east side of West Maui (Fig. 5; Table 5), the only area where S. salicaria is known to occur. Three populations of S. menziesii were sampled, all located north of Halepohaku (Fig. 5; Table 5).
Individuals of both species and hybrids sampled for molecular analysis originated as field-collected seeds or cuttings from several different populations (Fig. 5; Table 5; vouchers at US) and were grown in the greenhouse at the University of California, Irvine. All necessary permits were obtained for the described field studies. These permits covered the collection of S. menziesii plants in the Lihau Natural Area Reserve on W. Maui (granted by B. Gagne), putative hybrid plants from the Pu'u 'Ula'ula area on W. Maui collected by S. Perlman and K. Wood (granted by the State of Hawaii), and S. salicaria plants (granted by Wailuku Agribusiness) prior to its listing as endangered. Permission to study a population of Schiedea menziesii was granted by J. Falconer at the Pioneer Mill Company on Maui, on land owned by Pioneer Mill.
Traits measured in the greenhouse were those distinguishing the two species , including inflorescence length, number of inflorescence nodes, length of the lateral inflorescence branches, total number of flowers in the inflorescence, internode lengths of vegetative portions of flowering branches, leaf length, leaf width, and leaf curvature, number of leaf veins and degree of leaf pubescence, sepal length and extent of sepal pubescence, and pedicel length and pedicel pubescence. Principal components analysis (PCA)  was used to detect patterns of similarity between hybrid individuals and the presumed parental species. Additionally, canonical discriminant analysis (Proc Candisc in SAS) was used to calculate raw canonical coefficients of the first canonical variable for the parental species. Hybrids were not included in this latter calculation because their inclusion could obscure differences between the parental species. Subsequently, these coefficients were used to calculate the first canonical variable for the parental species and the hybrids.
Pollen production of each putative hybrid individual was measured by placing undehisced anthers of 2–6 flowers per genotype in a lactophenol-aniline blue solution . Five to seven genotypes per species were used to estimate pollen production and pollen size. Ten hemacytometer counts per genotype were used to estimate pollen production per flower, and genotype averages were then averaged to obtain values for each of the species and plants from the hybrid zone. Pollen size was estimated by measuring the diameter of 50 pollen grains per sample. Individual means of each plant were used as replicates in the hybrid population, and compared to values for the parental species reported earlier . Fully formed pollen grains that stained dark blue in aniline blue were assumed to be viable.
Leaf tissue was collected from plants maintained in the greenhouse. DNA was extracted from frozen or silica-dried tissue using either a modified CTAB-based technique of Doyle and Doyle  at the University of Cincinnati or with the DNeasy Plant Mini Extraction Kit (Qiagen, Valencia, CA, USA) at the University of South Dakota. DNA was stored at −20°C until further analysis. We genotyped individuals at nine nuclear microsatellite loci and two chloroplast microsatellite loci, and sequenced individuals for five chloroplast regions.
Nine of the 12 microsatellite primers originally developed for S. adamantis  were used in the current study of S. salicaria and S. menziesii. These primers amplified di- (SA30), tri- (SA04, SA05, SA15, SA32, SA36, SA38), and pentanucleotide (SA06, SA31) repeats consistently across all populations. DNA samples were analyzed in multiplexed reactions which also included fluorescent labeling of the 5′ end of each fragment during the PCR process . Each sample was evaluated using two different primer groups: Group 1 (SA04, 05, 15, 30, 36, 38) and Group 2 (SA06, 31, 32). PCR was performed in 10 µL reaction volumes with the Qiagen Multiplex Kit (Qiagen Inc., Valencia, CA) as follows: 5 µL Multiplex PCR Master Mix, 1 µL 10× Primer Mix (each primer at 2 µM), 3.8 µL dH20, and 0.2 µL DNA. The thermal cycler program consisted of 95°C for 15 min, followed by 30 cycles of 94°C for 30 s, 57°C for 45 s, and 72°C for 45 s, and then with 8 cycles each of 94°C for 30 s, 53°C for 45 s, and 72°C for 45 s. A final extension consisted of 72°C for 10 min. PCR was conducted at the University of Cincinnati and PCR products were run on a 3730 XL Sequencer at the Cornell University BioResource Center using the LIZ 500 (Applied Biosystems, Foster City, CA, USA) internal size standard with resulting data analyzed in Genemarker (SoftGenetics LLC, State College, PA, USA).
We tested for the presence of null alleles using Micro-Checker . Genetic variability at the microsatellite loci was characterized within the two parental species and putative hybrids by the mean number of alleles per loci (A), percentage of polymorphic loci (P; 95% level), observed heterozygosity (Ho), and the expected proportion of heterozygous loci per individual (He) using the GDA software package . No alleles were fixed in either of the parental species, thus we looked for evidence of admixture in allele frequencies of the putative hybrid population. We tested for significant differences in allele frequencies at each locus between the hybrids and each of the parental species using AMOVA  to assign parental origin of the predominant alleles at each microsatellite locus. These analyses were conducted on allele frequencies and assuming an infinite alleles model (i.e., FST) in Arlequin v. 3.11 . Significance of FST was tested through 10,000 permutations of the data. A hybrid population may be expected to exhibit allele frequencies that are more similar to one parent at loci that are distinct between the parental species. Furthermore, when looking across all loci, a mosaic of differentiation is expected in the hybrid zone. By contrast, if the putative hybrids are variants of one parental species, then we expect allele frequencies of the hybrids to only be distinct from the non-parental species. We also tested for linkage disequilibrium (LD) within the putative hybrid population using the Burrows composite estimate based on genotypic frequencies calculated in Popgene v. 1.31 . Recently hybridized populations should exhibit strong LD if they are derived from two gene pools with differing allele frequencies. Linkage disequilibrium should decay with subsequent generations of hybrids as associations between loosely linked loci are reduced . If the hybrids are primarily composed of later generation hybrids, then we would not expect to find high estimates of LD. Lastly, we used a Bayesian analysis in Structure v. 2.1  to assign individuals to one of two populations based on recent ancestry. We chose K = 2 and a model of correlated allele frequencies. This analysis was conducted using 100,000 burn-in steps followed by 500,000 steps in the Markov chain from which the posterior probabilities were derived. Individuals were assigned to a single cluster if their probability of membership was greater than 90% and the confidence intervals for inclusion in each cluster were non-overlapping. Otherwise, an individual was considered to be of mixed ancestry.
Sequence data were collected at five non-coding chloroplast regions, including four intergenic spacers (petL-psbE, ndhJ-trnF, petN-psbM, psbM-trnD), and the intron of rpS16. The regions were amplified using universal primers for petL-psbE, ndhJ-trnF , petN-psbM and psbM-trnD  and rpS16 . Reaction volumes of 25 or 50 µl included 1 µl of template DNA; 250 µM of dNTP (Promega, Madison, WI, USA); 1× Go-Taq Flexi Buffer (Promega, Madison, WI, USA); 2.5 mM of MgCl2; 0.2 µM of each primer (Integrated DNA Technologies Inc., IA, USA); 1.0 µL Bovine Serum Albumin (Promega, Madison, WI, USA) and 2.5 units of Go-Taq Flexi DNA Polymerase (Promega, Madison, WI, USA). The thermal cycler program consisted of 1 cycle of 95°C for 2 min, followed by 35 cycles of 95°C for 1 min, 50°C for 1 min, 65°C for 4 min with a 1°C/8 s ramp from the annealing temperature to the extension temperature . PCR products were verified on 1% agarose gels before cleaning with the Qiaquick PCR cleaning kit (Qiagen, Valencia, CA, USA.).
DNA sequencing reactions were performed with the ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction kit, v3.1 and subjected to capillary electrophoresis on an Avant-3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) at the University of South Dakota. Cycle sequencing reactions were carried out in 15 µL volumes with Big Dye Terminator v. 3.1 (Applied Biosystems, Foster City, CA, USA) and the PCR primers. Cycle sequencing reactions were purified using precipitation in 3 M NaOH and 95% EtOH. Both forward and reverse directions were sequenced. Sequence chromatograms were edited, and forward and reverse sequences were assembled to produce a single sequence for each sample using Sequencher 4.1 (GeneCodes, Ann Arbor, MI, USA). The data were aligned using Se-al v2.09a  with the inclusion of gaps to accommodate insertions/deletions across taxa. Sequences are deposited in GenBank (accessions HM596477-HM596500).
The samples were also genotyped at two single-nucleotide microsatellite regions of the chloroplast genome (i.e., trnL intron and trnT-trnD intergenic spacer region) using primers developed by A. Kuenzi and T. Culley (unpublished data) based on a trnL chloroplast sequence for Schiedea in the NCBI database (GenBank accession DQ907871) and unpublished sequences for the trnT-trnD intergenic spacer region (GenBank accession GQ224409; L. Wallace & M. Nepokroeff, unpublished data). The primers were fluorescently tagged and used in multiplexed PCR reactions of all available S. salicaria, S. menziesii, and hybrid samples using the same conditions as described above for the nuclear microsatellites. The results were then analyzed in Genemarker (SoftGenetics LLC, State College, PA, USA) to identify a single allele for each locus in each sample.
The five sequenced gene regions were concatenated for each individual, and the microsatellite genotypes were also translated into sequence data and included with the concatenated sequences for each individual. A haplotype network was constructed from the fully concatenated data set using statistical parsimony in TCS . Gaps were included as fifth character states. Genetic diversity was determined for each population as well as for S. menziesii, S. salicaria, and the hybrids using DnaSP v. 5 . We calculated the number of unique haplotypes with and without indels, haplotype diversity, and nucleotide diversity. An AMOVA , performed in Arlequin vers. 3.11 , was used to test for significant differences in haplotype frequency among S. menziesii, S. salicaria, and hybrids using pairwise difference as an estimate of genetic distance. Significance of F-statistics was determined using 10,000 permutations of the data.
Number of unique haplotypes and nucleotide diversity observed at each chloroplast locus in each population. For number of haplotypes, the value in parentheses is the number of unique haplotypes when indels are considered in addition to nucleotide substitutions. For nucleotide diversity, standard deviation is reported in parentheses.
We thank Bob Hobdy for showing us the hybrid zone on Halepohaku, Steve Perlman, Ken Wood, and Hank Oppenheimer for their collections and help in the field, the curators of PTBG and US for allowing us to use their collections, Betsy Gagné for access to the Lihau Natural Area Reserve, Kimo Falconer for advice and access to Hahakea, Wailuku Agribusiness for access to populations of S. salicaria, and Yvonne Alliman, Weigang Yang, and Richard Basile for cultivation of plants in the greenhouse. We thank Jennifer Hawks for assistance with DNA extractions and data analyses at the University of Cincinnati, and Jeff Engemann, Tina Clausen, Alyssa Hotz, Bridget Jacobs and Ann Willyard for assistance in the laboratory at the University of South Dakota.
Conceived and designed the experiments: TMC SGW MN. Performed the experiments: LEW TMC SGW AK TR. Analyzed the data: LEW TMC SGW AK. Contributed reagents/materials/analysis tools: LEW TMC SGW AKS WLW MN. Wrote the paper: LEW TMC SGW AKS WLW MN.
- 1. Rieseberg LH, Carney SC (1998) Tansley Review – Plant hybridization. New Phytol 140: 598–624.LH RiesebergSC Carney1998Tansley Review – Plant hybridization.New Phytol140598624
- 2. Buggs RJA (2007) Empirical study of hybrid zone movement. Heredity 99: 301–312.RJA Buggs2007Empirical study of hybrid zone movement.Heredity99301312
- 3. Currat M, Ruedi M, Petit RJ, Excoffier L (2008) The hidden side of invasions: massive introgression by local genes. Evolution 62: 1908–1920.M. CurratM. RuediRJ PetitL. Excoffier2008The hidden side of invasions: massive introgression by local genes.Evolution6219081920
- 4. Harrison RG (1993) Hybrid Zones and the Evolutionary Process. Oxford: Oxford University Press. 376 p.RG Harrison1993Hybrid Zones and the Evolutionary ProcessOxfordOxford University Press376
- 5. Barton N (1979) The dynamics of hybrid zones. Heredity 43: 341–359.N. Barton1979The dynamics of hybrid zones.Heredity43341359
- 6. Barton N, Hewitt GM (1985) Analysis of hybrid zones. Ann Rev Ecol Syst 16: 113–148.N. BartonGM Hewitt1985Analysis of hybrid zones.Ann Rev Ecol Syst16113148
- 7. Rieseberg LH, Beckstrom-Sternberg S, Doan K (1990) Helianthus annuus ssp. texanus has chloroplast DNA and nuclear ribosomal genes of Helianthus debilis ssp. cucumerifolius. P Natl Acad Sci USA 87: 593–597.LH RiesebergS. Beckstrom-SternbergK. Doan1990Helianthus annuus ssp. texanus has chloroplast DNA and nuclear ribosomal genes of Helianthus debilis ssp. cucumerifolius.P Natl Acad Sci USA87593597
- 8. Rieseberg LH, Ellstrand NC (1993) What can molecular and morphological markers tell us about plant hybridization? CRC Cr Rev Plant Sci 12: 213–241.LH RiesebergNC Ellstrand1993What can molecular and morphological markers tell us about plant hybridization?CRC Cr Rev Plant Sci12213241
- 9. Martinsen GD, Whitham TG, Turek RJ, Keim P (2001) Hybrid populations selectively filter gene introgression between species. Evolution 55: 1325–1335.GD MartinsenTG WhithamRJ TurekP. Keim2001Hybrid populations selectively filter gene introgression between species.Evolution5513251335
- 10. Wu CA, Campbell DR (2005) Cytoplasmic and nuclear markers reveal contrasting patterns of spatial genetic structure in a natural Ipomopsis hybrid zone. Mol Ecol 14: 781–792.CA WuDR Campbell2005Cytoplasmic and nuclear markers reveal contrasting patterns of spatial genetic structure in a natural Ipomopsis hybrid zone.Mol Ecol14781792
- 11. Wallace LE (2006) Spatial genetic structure and frequency of interspecific hybridization in Platanthera aquilonis and P. dilatata (Orchidaceae) occurring in sympatry. Am J Bot 93: 1001–1009.LE Wallace2006Spatial genetic structure and frequency of interspecific hybridization in Platanthera aquilonis and P. dilatata (Orchidaceae) occurring in sympatry.Am J Bot9310011009
- 12. Lehman N, Eisenhawer A, Hansen K, Mech LD, Peterson RO, et al. (1991) Introgression of coyote mitochondrial DNA into sympatric North American gray wolf populations. Evolution 45: 104–119.N. LehmanA. EisenhawerK. HansenLD MechRO Peterson1991Introgression of coyote mitochondrial DNA into sympatric North American gray wolf populations.Evolution45104119
- 13. Borge T, Lindroos K, Nadvornik P, Syvanen AC, Saetre GP (2005) Amount of introgression in flycatcher hybrid zones relfects regional differences in pre and post-zygotic barriers to gene exchange. J Evolution Biol 18: 1416–1424.T. BorgeK. LindroosP. NadvornikAC SyvanenGP Saetre2005Amount of introgression in flycatcher hybrid zones relfects regional differences in pre and post-zygotic barriers to gene exchange.J Evolution Biol1814161424
- 14. Buggs RJA, Pannell JR (2006) Rapid displacement of a monoecious plant lineage is due to pollen swamping by a dioecious relative. Curr Biol 16: 996–1000.RJA BuggsJR Pannell2006Rapid displacement of a monoecious plant lineage is due to pollen swamping by a dioecious relative.Curr Biol169961000
- 15. Campbell DR, Waser NM, Meléndez-Ackerman EJ (1997) Analyzing pollinator-mediated selection in a plant hybrid zone: hummingbird visitation patterns on three spatial scales. Am Nat 149: 295–315.DR CampbellNM WaserEJ Meléndez-Ackerman1997Analyzing pollinator-mediated selection in a plant hybrid zone: hummingbird visitation patterns on three spatial scales.Am Nat149295315
- 16. Campbell DR, Waser NM, Pederson GT (2002) Predicting patterns of mating and potential hybridization from pollinator behavior. Am Nat 159: 438–450.DR CampbellNM WaserGT Pederson2002Predicting patterns of mating and potential hybridization from pollinator behavior.Am Nat159438450
- 17. Hersch EI, Roy BA (2007) Context-dependent pollinator behavior: an explanation for patterns of hybridization among three species of Indian paintbrush. Evolution 61: 111–124.EI HerschBA Roy2007Context-dependent pollinator behavior: an explanation for patterns of hybridization among three species of Indian paintbrush.Evolution61111124
- 18. Kimball S (2008) Links between floral morphology and floral visitors along an elevational gradient in a Penstemon hybrid zone. Oikos 117: 1064–1074.S. Kimball2008Links between floral morphology and floral visitors along an elevational gradient in a Penstemon hybrid zone.Oikos11710641074
- 19. Wagner WL, Weller SG, Sakai A (2005) Monograph of Schiedea (Caryophyllaceae-Alsinoideae). Systematic Botany Monographs 72: 1–169.WL WagnerSG WellerA. Sakai2005Monograph of Schiedea (Caryophyllaceae-Alsinoideae).Systematic Botany Monographs721169
- 20. Sakai AK, Weller SG, Wagner WL, Nepokroeff M, Culley TM (2006) Adaptive radiation and evolution of breeding systems in Schiedea (Caryophyllaceae), an endemic Hawaiian genus. Ann Mo Bot Gard 93: 49–63.AK SakaiSG WellerWL WagnerM. NepokroeffTM Culley2006Adaptive radiation and evolution of breeding systems in Schiedea (Caryophyllaceae), an endemic Hawaiian genus.Ann Mo Bot Gard934963
- 21. Culley TM, Weller SG, Sakai AK, Rankin AE (1999) Inbreeding depression and selfing rates in a self-compatible, hermaphroditic species, Schiedea membranacea (Caryophyllaceae). Am J Bot 86: 980–987.TM CulleySG WellerAK SakaiAE Rankin1999Inbreeding depression and selfing rates in a self-compatible, hermaphroditic species, Schiedea membranacea (Caryophyllaceae).Am J Bot86980987
- 22. Weller SG, Sakai AK, Thai DA, Tom J, Rankin AE (2005) Inbreeding depression and heterosis in populations of Schiedea viscosa, a highly selfing species. J Evolution Biol 18: 1434–1444.SG WellerAK SakaiDA ThaiJ. TomAE Rankin2005Inbreeding depression and heterosis in populations of Schiedea viscosa, a highly selfing species.J Evolution Biol1814341444
- 23. Rankin AE, Weller SG, Sakai AK (2002) Mating system instability in Schiedea menziesii (Caryophyllaceae). Evolution 56: 1574–1585.AE RankinSG WellerAK Sakai2002Mating system instability in Schiedea menziesii (Caryophyllaceae).Evolution5615741585
- 24. Weller SG, Sakai AK (2005) Selfing and resource allocation in Schiedea salicaria (Caryophyllaceae), a gynodioecious species. J Evolution Biol 18: 301–308.SG WellerAK Sakai2005Selfing and resource allocation in Schiedea salicaria (Caryophyllaceae), a gynodioecious species.J Evolution Biol18301308
- 25. Sakai AK, Karoly K, Weller SG (1989) Inbreeding depression in Schiedea globosa and S. salicaria (Caryophyllaceae), subdioecious and gynodioecious Hawaiian species. Am J Bot 76: 437–444.AK SakaiK. KarolySG Weller1989Inbreeding depression in Schiedea globosa and S. salicaria (Caryophyllaceae), subdioecious and gynodioecious Hawaiian species.Am J Bot76437444
- 26. Sakai AK, Weller SG (1991) Ecological aspects of sex expression in subdioecious Schiedea globosa (Caryophyllaceae). Am J Bot 78: 1280–1288.AK SakaiSG Weller1991Ecological aspects of sex expression in subdioecious Schiedea globosa (Caryophyllaceae).Am J Bot7812801288
- 27. Weller SG, Sakai AK, Rankin AE, Golonka A, Kutcher B, et al. (1998) Dioecy and the evolution of wind pollination in Schiedea and Alsinidendron (Caryophyllaceae: Alsinoideae) in the Hawaiian Islands. Am J Bot 85: 1377–1388.SG WellerAK SakaiAE RankinA. GolonkaB. Kutcher1998Dioecy and the evolution of wind pollination in Schiedea and Alsinidendron (Caryophyllaceae: Alsinoideae) in the Hawaiian Islands.Am J Bot8513771388
- 28. Weller SG, Sakai AK, Wagner WL, Herbst DR (1990) Evolution of dioecy in Schiedea (Caryophyllaceae: Alsinoideae) in the Hawaiian Islands: biogeographical and ecological factors. Syst Bot 15: 266–276.SG WellerAK SakaiWL WagnerDR Herbst1990Evolution of dioecy in Schiedea (Caryophyllaceae: Alsinoideae) in the Hawaiian Islands: biogeographical and ecological factors.Syst Bot15266276
- 29. Cassens I, Mardulyn P, Milinkovitch MC (2005) Evaluating intraspecific “network” construction methods using simulated sequence data: do existing algorithms outperform the global maximum parsimony approach? Syst Biol 54: 363–372.I. CassensP. MardulynMC Milinkovitch2005Evaluating intraspecific “network” construction methods using simulated sequence data: do existing algorithms outperform the global maximum parsimony approach?Syst Biol54363372
- 30. Müller K (2006) Incorporating information from length-mutational events into phylogenetic analysis. Mol Phylogenet Evol 38: 667–676.K. Müller2006Incorporating information from length-mutational events into phylogenetic analysis.Mol Phylogenet Evol38667676
- 31. Ellstrand NC, Whitkus R, Rieseberg LH (1996) Distribution of spontaneous plant hybrids. P Natl Acad Sci USA 93: 5090–5093.NC EllstrandR. WhitkusLH Rieseberg1996Distribution of spontaneous plant hybrids.P Natl Acad Sci USA9350905093
- 32. Whitney KD, Ahern JR, Campbell LG, Albert LP, King MS (2010) Patterns of hybridization in plants. Perspect Plant Ecol 12: 175–182.KD WhitneyJR AhernLG CampbellLP AlbertMS King2010Patterns of hybridization in plants.Perspect Plant Ecol12175182
- 33. Baldwin BG, Robichaux RH (1995) Historical biogeography and ecology of the Hawaiian silversword alliance (Asteraceae): new molecular phylogenetic perspectives. In: Wagner WL, Funk VA, editors. Hawaiian biogeography: evolution on a hot spot archipelago. Washington, D.C.: Smithsonian Institution Press. pp. 257–287.BG BaldwinRH Robichaux1995Historical biogeography and ecology of the Hawaiian silversword alliance (Asteraceae): new molecular phylogenetic perspectives.WL WagnerVA FunkHawaiian biogeography: evolution on a hot spot archipelagoWashington, D.C.Smithsonian Institution Press257287
- 34. Barrier M, Baldwin BG, Robichaux RH, Purugganan MD (1999) Interspecific hybrid ancestry of a plant adaptive radiation: allopoplyploidy of the Hawaiian silversword alliance (Asteraceae) inferred from floral homeotic gene duplications. Mol Biol Evol 16: 1105–1113.M. BarrierBG BaldwinRH RobichauxMD Purugganan1999Interspecific hybrid ancestry of a plant adaptive radiation: allopoplyploidy of the Hawaiian silversword alliance (Asteraceae) inferred from floral homeotic gene duplications.Mol Biol Evol1611051113
- 35. Friar EA, Cruse-Sanders JM, McGlaughlin ME (2007) Gene flow in Dubautia arborea and D. ciliolata: the roles of ecology and isolation by distance in maintaining species boundaries despite ongoing hybridization. Mol Ecol 16: 4028–4038.EA FriarJM Cruse-SandersME McGlaughlin2007Gene flow in Dubautia arborea and D. ciliolata: the roles of ecology and isolation by distance in maintaining species boundaries despite ongoing hybridization.Mol Ecol1640284038
- 36. Friar EA, Prince LM, Cruse-Sanders JM, McGlaughlin ME, Butterworth CA, et al. (2008) Hybrid origin and genomic mosaicism of Dubautia scabra (Hawaiian Silversword Alliance; Asteraceae, Madiinae). Syst Bot 33: 589–597.EA FriarLM PrinceJM Cruse-SandersME McGlaughlinCA Butterworth2008Hybrid origin and genomic mosaicism of Dubautia scabra (Hawaiian Silversword Alliance; Asteraceae, Madiinae).Syst Bot33589597
- 37. Remington DL, Robichaux RH (2007) Influences of gene flow on adaptive speciation in the Dubautia arborea-D. ciliolata complex. Mol Ecol 16: 4014–4027.DL RemingtonRH Robichaux2007Influences of gene flow on adaptive speciation in the Dubautia arborea-D. ciliolata complex.Mol Ecol1640144027
- 38. Smith JF, Burke CC, Wagner WL (1996) Interspecific hybridization in natural populations of Cyrtandra (Gesneriaceae) on the Hawaiian Islands: evidence from RAPD markers. Plant Syst Evol 200: 61–77.JF SmithCC BurkeWL Wagner1996Interspecific hybridization in natural populations of Cyrtandra (Gesneriaceae) on the Hawaiian Islands: evidence from RAPD markers.Plant Syst Evol2006177
- 39. Harbaugh DT, Wagner WL, Percy DM, James HF, Fleischer RC (2009) Genetic structure of the polymorphic Metrosideros (Myrtaceae) complex in the Hawaiian Islands using nuclear microsatellite data. Plos One 4: e4698.DT HarbaughWL WagnerDM PercyHF JamesRC Fleischer2009Genetic structure of the polymorphic Metrosideros (Myrtaceae) complex in the Hawaiian Islands using nuclear microsatellite data.Plos One4e4698
- 40. Kim I, Carr GD (1990) Cytogenetics and hybridization of Portulaca in Hawaii. Syst Bot 15: 370–377.I. KimGD Carr1990Cytogenetics and hybridization of Portulaca in Hawaii.Syst Bot15370377
- 41. Randell RA, Howarth DG, Morden CW (2004) Genetic analysis of natural hybrids between endemic and alien Rubus (Rosaceae) species in Hawai'i. Conserv Genet 5: 217–230.RA RandellDG HowarthCW Morden2004Genetic analysis of natural hybrids between endemic and alien Rubus (Rosaceae) species in Hawai'i.Conserv Genet5217230
- 42. Howarth DG, Baum DA (2005) Genealogical evidence of homoploid hybrid speciation in an adaptive radiation of Scaevola (Goodeniaceae) in the Hawaiian Islands. Evolution 59: 948–961.DG HowarthDA Baum2005Genealogical evidence of homoploid hybrid speciation in an adaptive radiation of Scaevola (Goodeniaceae) in the Hawaiian Islands.Evolution59948961
- 43. Willyard A, Wallace LE, Wagner WL, Weller SG, Sakai AK, et al. (2011) Estimating the species tree for Hawaiian Schiedea (Caryophyllaceae) from multiple loci in the presence of reticulate evolution. Mol Phylogenet Evol 60: 29–48.A. WillyardLE WallaceWL WagnerSG WellerAK Sakai2011Estimating the species tree for Hawaiian Schiedea (Caryophyllaceae) from multiple loci in the presence of reticulate evolution.Mol Phylogenet Evol602948
- 44. Chapman MA, Burke JM (2007) Genetic divergence and hybrid speciation. Evolution 61: 1773–1780.MA ChapmanJM Burke2007Genetic divergence and hybrid speciation.Evolution6117731780
- 45. McGlaughlin ME, Friar EA (2011) Evolutionary diversification and geographical isolation in Dubautia laxa (Asteraceae), a widespread member of the Hawaiian silversword alliance. Ann Bot- London 107: 357–370.ME McGlaughlinEA Friar2011Evolutionary diversification and geographical isolation in Dubautia laxa (Asteraceae), a widespread member of the Hawaiian silversword alliance.Ann Bot- London107357370
- 46. Seehausen O (2004) Hybridization and adaptive radiation. Trends Ecol Evol 19: 198–207.O. Seehausen2004Hybridization and adaptive radiation.Trends Ecol Evol19198207
- 47. Hillebrand W (1888) Flora of the Hawaiian Islands. Heideberg: Carl Winter; London: Williams and Norgate, New York: B. Westermann and Co. 673 p.W. Hillebrand1888Flora of the Hawaiian IslandsHeidebergCarl Winter; London: Williams and Norgate, New York: B. Westermann and Co673
- 48. Sherff EE (1945) Revision of the genus Schiedea Cham. & Schlecht. Brittonia 5: 308–335.EE Sherff1945Revision of the genus Schiedea Cham. & Schlecht.Brittonia5308335
- 49. Welch ME, Rieseberg LH (2002) Patterns of genetic variation suggest a single, ancient origin for the diploid hybrid species Helianthus paradoxus. Evolution 56: 2126–2137.ME WelchLH Rieseberg2002Patterns of genetic variation suggest a single, ancient origin for the diploid hybrid species Helianthus paradoxus.Evolution5621262137
- 50. Gross BL, Schwarzbach AE, Rieseberg LH (2003) Origin(s) of the diploid hybrid species Helianthus deserticola (Asteraceae). Am J Bot 90: 1708–1719.BL GrossAE SchwarzbachLH Rieseberg2003Origin(s) of the diploid hybrid species Helianthus deserticola (Asteraceae).Am J Bot9017081719
- 51. Tate JA, Symonds VV, Doust AN, Buggs RJA, Mavrodiev E, et al. (2009) Synthetic polyploids of Tragopogon miscellus and T. mirus (Asteraceae): 60 Years after Ownbey's discovery. Am J Bot 96: 979–988.JA TateVV SymondsAN DoustRJA BuggsE. Mavrodiev2009Synthetic polyploids of Tragopogon miscellus and T. mirus (Asteraceae): 60 Years after Ownbey's discovery.Am J Bot96979988
- 52. Wolfe AD, Elisens WJ (1995) Evidence of chloroplast capture and pollen-mediated gene flow in Penstemon sect. Peltanthera (Scrophulariaceae). Syst Bot 20: 395–412.AD WolfeWJ Elisens1995Evidence of chloroplast capture and pollen-mediated gene flow in Penstemon sect. Peltanthera (Scrophulariaceae).Syst Bot20395412
- 53. Hamzeh M, Sawchyn C, Perinet P, Dayanandan S (2007) Asymmetrical natural hybridization between Populus deltoides and P. balsamifera (Salicaceae). Can J Bot 85: 1227–1232.M. HamzehC. SawchynP. PerinetS. Dayanandan2007Asymmetrical natural hybridization between Populus deltoides and P. balsamifera (Salicaceae).Can J Bot8512271232
- 54. Corriveau JL, Coleman AW (1988) Rapid screening method to detect potential biparental inheritance of plastid DNA and results for over 200 angiosperm species. Am J Bot 75: 1443–1458.JL CorriveauAW Coleman1988Rapid screening method to detect potential biparental inheritance of plastid DNA and results for over 200 angiosperm species.Am J Bot7514431458
- 55. Zhang Q, Liu Y, Sodmergen (2003) Examination of the cytoplasmic DNA in male reproductive cells to determine the potential for cytoplasmic inheritance in 295 angiosperm species. Plant Cell Physiol 44: 941–951.Q. ZhangY. LiuSodmergen2003Examination of the cytoplasmic DNA in male reproductive cells to determine the potential for cytoplasmic inheritance in 295 angiosperm species.Plant Cell Physiol44941951
- 56. Hewitt G (2000) The genetic legacy of the Quaternary Ice Ages. Nature 405: 907–913.39.G. Hewitt2000The genetic legacy of the Quaternary Ice Ages.Nature405907913.39
- 57. Price J, Elliott-Fisk D (2004) Topographic history of the Maui Nui complex, Hawaii, and its implications for biogeography. Pac Sci 58: 27–45.J. PriceD. Elliott-Fisk2004Topographic history of the Maui Nui complex, Hawaii, and its implications for biogeography.Pac Sci582745
- 58. Pearl SA, Welch ME, McCauley DE (2009) Mitochondrial heteroplasmy and paternal leakage in natural populations of Silene vulgaris, a gynodioecious plant. Mol Biol Evol 26: 537–545.SA PearlME WelchDE McCauley2009Mitochondrial heteroplasmy and paternal leakage in natural populations of Silene vulgaris, a gynodioecious plant.Mol Biol Evol26537545
- 59. Weller SG, Sakai AK (1991) The genetic basis of male sterility in Schiedea (Caryophyllaceae), an endemic Hawaiian genus. Heredity 87: 265–273.SG WellerAK Sakai1991The genetic basis of male sterility in Schiedea (Caryophyllaceae), an endemic Hawaiian genus.Heredity87265273
- 60. Renner SS, Ricklefs RE (1995) Dioecy and its correlates in the flowering plants. Am J Bot 82: 596–606.SS RennerRE Ricklefs1995Dioecy and its correlates in the flowering plants.Am J Bot82596606
- 61. Culley TM, Weller SG, Sakai AK (2002) The evolution of wind pollination in angiosperms. Trends Ecol Evol 17: 361–369.TM CulleySG WellerAK Sakai2002The evolution of wind pollination in angiosperms.Trends Ecol Evol17361369
- 62. Vamosi JC, Otto SP, Barrett SCH (2003) Phylogenetic analysis of the ecological correlates of dioecy in angiosperms. J Evolution Biol 16: 1006–1018.JC VamosiSP OttoSCH Barrett2003Phylogenetic analysis of the ecological correlates of dioecy in angiosperms.J Evolution Biol1610061018
- 63. Friedman J, Barrett SCH (2008) A phylogenetic analysis of the evolution of wind pollination in the angiosperms. Int J Plant Sci 169: 49–58.J. FriedmanSCH Barrett2008A phylogenetic analysis of the evolution of wind pollination in the angiosperms.Int J Plant Sci1694958
- 64. Weller SG, Sakai AK, Culley TM, Campbell DR, Dunbar-Willis AK (2006) Predicting the pathway to wind pollination: heritabilities and genetic correlations of inflorescence traits associated with wind pollination in Schiedea salicaria (Caryophyllaceae). J Evolution Biol 19: 331–342.SG WellerAK SakaiTM CulleyDR CampbellAK Dunbar-Willis2006Predicting the pathway to wind pollination: heritabilities and genetic correlations of inflorescence traits associated with wind pollination in Schiedea salicaria (Caryophyllaceae).J Evolution Biol19331342
- 65. Sakai AK, Weller SG, Culley TM, Campbell DR, Dunbar-Wallis AK, et al. (2008) Sexual dimorphism and the genetic potential for evolution of sex allocation in the gynodioecious plant, Schiedea salicaria. J Evolution Biol 21: 18–29.AK SakaiSG WellerTM CulleyDR CampbellAK Dunbar-Wallis2008Sexual dimorphism and the genetic potential for evolution of sex allocation in the gynodioecious plant, Schiedea salicaria.J Evolution Biol211829
- 66. Obbard DJ, Harris SA, Buggs RJA, Pannell JR (2006) Hybridization, polyploidy, and the evolution of sexual systems in Mercurialis (Euphorbiaceae). Evolution 60: 1801–1815.DJ ObbardSA HarrisRJA BuggsJR Pannell2006Hybridization, polyploidy, and the evolution of sexual systems in Mercurialis (Euphorbiaceae).Evolution6018011815
- 67. Wagner WL, Herbst DR, Sohmer SH (1990) Manual of the Flowering Plants of Hawai'i. Honolulu: University of Hawaii Press & Bishop Museum. WL WagnerDR HerbstSH Sohmer1990Manual of the Flowering Plants of Hawai'iHonoluluUniversity of Hawaii Press & Bishop Museum1,853
- 68. SAS Institute Inc. (2002–2003) SAS. Version 9.1. Cary: SAS Institute, Inc. SAS Institute Inc.2002–2003SAS. Version 9.1CarySAS Institute, Inc
- 69. Doyle JJ, Doyle JL (1987) A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochemical Bulletin 19: 11–15.JJ DoyleJL Doyle1987A rapid DNA isolation procedure for small amounts of fresh leaf tissue.Phytochemical Bulletin191115
- 70. Culley TM, Weller SG, Sakai AK, Putnam KA (2008) Characterization of microsatellite loci in the Hawaiian endemic shrub Schiedea adamantis (Caryophyllaceae) and amplification in related species and genera. Mol Ecol Resour 8: 1081–1084.TM CulleySG WellerAK SakaiKA Putnam2008Characterization of microsatellite loci in the Hawaiian endemic shrub Schiedea adamantis (Caryophyllaceae) and amplification in related species and genera.Mol Ecol Resour810811084
- 71. Van Oosterhout C, Hutchinson WF, Derek P, Wills M, Shipley P (2004) Micro-Checker: software for identifying and correcting genotyping errors in microsatellite data. Mol Ecol Notes 4: 535–538.C. Van OosterhoutWF HutchinsonP. DerekM. WillsP. Shipley2004Micro-Checker: software for identifying and correcting genotyping errors in microsatellite data.Mol Ecol Notes4535538
- 72. Lewis PO, Zaykin D (2001) Genetic Data Analysis: Computer program for the analysis of allelic data, Version 1.0 (d16c). PO LewisD. Zaykin2001Genetic Data Analysis: Computer program for the analysis of allelic data, Version 1.0 (d16c).Available via http://hydrodictyon.eeb.uconn.edu/people/plewis/software.php. Accessed 24 August 2011. Available via http://hydrodictyon.eeb.uconn.edu/people/plewis/software.php. Accessed 24 August 2011.
- 73. Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular variance inferred from metric distances among DNA haplotypes: application of human mitochondrial DNA restriction data. Genetics 131: 479–491.L. ExcoffierPE SmouseJM Quattro1992Analysis of molecular variance inferred from metric distances among DNA haplotypes: application of human mitochondrial DNA restriction data.Genetics131479491
- 74. Excoffier L, Laval G, Schneider S (2005) Arlequin ver. 3.0: An integrated software package for population genetics data analysis. Evol Bioinform 1: 47–50.L. ExcoffierG. LavalS. Schneider2005Arlequin ver. 3.0: An integrated software package for population genetics data analysis.Evol Bioinform14750
- 75. Yeh FC, Boyle TJB (1997) Population genetic analysis of co-dominant and dominant markers and quantitative traits. Belg J Bot 129: 157.FC YehTJB Boyle1997Population genetic analysis of co-dominant and dominant markers and quantitative traits.Belg J Bot129157
- 76. Forbes SH, Allendorf FW (1991) Associations between mitochondrial and nuclear genotypes in cutthroat tout hybrid swarms. Evolution 45: 1332–1349.SH ForbesFW Allendorf1991Associations between mitochondrial and nuclear genotypes in cutthroat tout hybrid swarms.Evolution4513321349
- 77. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945–959.JK PritchardM. StephensP. Donnelly2000Inference of population structure using multilocus genotype data.Genetics155945959
- 78. Shaw J, Lickey EB, Schilling EE, Small RL (2007) Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: the tortoise and the hare III. Am J Bot 94: 275–288.J. ShawEB LickeyEE SchillingRL Small2007Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: the tortoise and the hare III.Am J Bot94275288
- 79. Lee C, Wen J (2004) Phylogeny of Panax using chloroplast trnC-trnD intergenic region and the utility of trnC-trnD in interspecific studies in plants. Mol Phylogenet Evol 31: 894–903.C. LeeJ. Wen2004Phylogeny of Panax using chloroplast trnC-trnD intergenic region and the utility of trnC-trnD in interspecific studies in plants.Mol Phylogenet Evol31894903
- 80. Popp M, Oxelman B (2001) Inferring the history of the polyploid Silene aegaea (Caryophyllaceae) using plastid and homoeologous nuclear DNA sequences. Mol Phylogenet Evol 20: 474–481.M. PoppB. Oxelman2001Inferring the history of the polyploid Silene aegaea (Caryophyllaceae) using plastid and homoeologous nuclear DNA sequences.Mol Phylogenet Evol20474481
- 81. Small RL, Ryburn J, Cronn RC, Seelanan T, Wendel JF (1998) The tortoise and the hare: Choosing between noncoding plastome and nuclear Adh sequences for phylogeny reconstruction in a recently derived group. Am J Bot 85: 1301–1315.RL SmallJ. RyburnRC CronnT. SeelananJF Wendel1998The tortoise and the hare: Choosing between noncoding plastome and nuclear Adh sequences for phylogeny reconstruction in a recently derived group.Am J Bot8513011315
- 82. Rambaut A (1996) Se-Al: Sequence alignment editor software. A. Rambaut1996Se-Al: Sequence alignment editor software.Available via http://tree.bio.ed.ac.uk/software/seal. Accessed 24 August 2011. Available via http://tree.bio.ed.ac.uk/software/seal. Accessed 24 August 2011.
- 83. Clement M, Posada D, Crandall KA (2000) TCS: a computer program to estimate gene genealogies. Mol Ecol 9: 1657–1659.M. ClementD. PosadaKA Crandall2000TCS: a computer program to estimate gene genealogies.Mol Ecol916571659
- 84. Librado P, Rozas J (2009) DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25: 1451–1452.P. LibradoJ. Rozas2009DnaSP v5: A software for comprehensive analysis of DNA polymorphism data.Bioinformatics2514511452