| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2 Institut für Umweltwissenschaften der Universität Zürich, Winterthurerstr. 190, CH-8057 Zürich, Switzerland; 3 Botanisches Institut der Universität Basel, Schönbeinstrasse 6, CH-4056 Basel, Switzerland; and 4 Laboratoire d'Evolution et Systématique, CNRS URA 2154, Université de Paris-Sud, Bât. 362, F-91405 Orsay Cedex, France
Received for publication June 3, 1999. Accepted for publication October 5, 1999.
ABSTRACT
Hybrid zones may be structured by environmentally independent selection against intrinsically unfit hybrids (tension zone models), or by environmentally dependent fitness differences among parental species and hybrids (ecological selection-gradient models). A 30-m slope in a mountain grassland harbors a hybrid zone of the clonal perennials, Prunella grandiflora and P. vulgaris (Lamiaceae), with P. grandiflora in the upper, P. vulgaris in the lower, and both parental species and P. grandiflora x P. vulgaris Hybrids in a narrow middle part. We found gradients for soil depth and water content, and vegetation height and biomass along the slope. A reciprocal transplant experiment yielded crossing reaction norms for vegetative reproduction. Parental species were locally adapted to their home sites, while the three taxa did not differ in vegetative reproduction in the Hybrid transplant site. Local adaptation for vegetative reproduction of P. grandiflora was mediated through higher survival and that of P. vulgaris through higher ramet number, indicating adaptation of their clonal growth strategies (phalanx vs. guerrilla) to the different habitats. Hybrid performance was intermediate between that of the parental species in all three sites, although Hybrids flowered more often than the parental species in the Hybrid site. Our results support ecological selection-gradient rather than tension zone models.
Key Words: clonal growth • hybrid superiority • Lamiaceae • local adaptation • phalanx vs. guerrilla growth • Prunella • species barriers
The evolutionary significance of hybridization between different species and the nature of hybrid zones has been debated for several decades (Stebbins, 1959
; Barton and Hewitt, 1985
; Abbott, 1992
; Harrison, 1993
; Arnold, 1997
). Species can be defined as groups of interbreeding individuals that are reproductively isolated from other such groups (Mayr, 1942
). However, for many closely related species, particularly plants, reproductive isolation is not complete and hybridization occurs where their geographical ranges overlap. If individuals from different species mate and reproduce freely and hybridization has little or no fitness disadvantage, then the species are expected to fuse into a single interbreeding species (Karlin and McGregor, 1972
; Slatkin, 1985
; Gavrilets, 1997
). Conversely, sufficiently strong selection against hybrids may facilitate the evolution of assortative mating and thus reinforce reproductive isolation between the hybridizing species (Fisher, 1930
; Dobzhansky, 1940
). Consequently, hybrid zones should ultimately either broaden or disappear over time.
Premating isolation is relatively unlikely to be completed by reinforcement in hybrid zones (Barton and Hewitt, 1985
; Spencer, McArdle, and Lambert, 1986
; Butlin, 1987
; Hostert, 1997
). Modifiers of the mating system building up premating barriers may be costly and therefore not easily spread to fixation (Spencer, McArdle, and Lambert, 1986
), or establishment of tight linkage between these modifier genes and genes linked with fitness will be counteracted by recombination (Felsenstein, 1981
). Nevertheless, there are many examples of hybrid zones stably existing even over evolutionarily relevant periods of time, and this without fusing the parental species (Hewitt, 1988
).
Currently, there are two groups of hypotheses for how barriers between species can be maintained without complete reproductive isolation. Tension zone or dynamic equilibrium (Moore, 1977
) models (Key, 1968
; Bazykin, 1969
; Barton and Hewitt, 1985
) assume that morphological or genetic clines along geographical transects in a hybrid zone represent a balance between dispersal and endogenous selection against hybrids. In this case, low migration rates and/or low intrinsic hybrid fitness reduces diffusion of genes between different species. Intrinsic hybrid disadvantage arises if parental genomes are incompatible, or if recombination disrupts coadapted gene complexes, resulting in hybrid inviability and sterility (Barton and Hewitt, 1985
; Hewitt, 1988
; Arnold and Hodges, 1995
), or increased susceptibility to parasites (Strauss, 1994
). Tension zone models are ecologically neutral models. Because hybrids are intrinsically unfit, clines develop wherever parental species come into contact, i.e., independent of environmental conditions (Barton and Hewitt, 1989
).
Alternatively, hybrid zones may be structured by environmentally dependent selection, e.g., along environmental gradients or ecotones (Endler, 1977
; Moore, 1977
; Moore and Price, 1993
). Ecological selection-gradient models assume that clines in morphology or allele frequencies represent equilibria between gene flow and differential selection along the environmental gradient (Haldane, 1948
; Endler, 1973
; Slatkin, 1973
; May, Endler, and McMurtrie, 1975
). That is, diffusion of alleles between taxa is prevented because the alleles are in the wrong environment rather than together with the wrong genes (Barton and Hewitt, 1989
). Consequently, both position and width of the hybrid zone are determined by exogenous selection along this gradient. A special case of this type of models is the bounded hybrid superiority model (Moore, 1977
; Moore and Buchanan, 1985
; Moore and Price, 1993
). It assumes that hybrids are more fit than the parental species in the middle part of the environmental gradient, but less fit than the parental species towards the edges of or outside the hybrid zone.
At present, the relative importance of intrinsic vs. environmental factors contributing to hybrid disadvantage is under debate (Barton and Hewitt, 1989
; Arnold and Hodges, 1995
; Arnold, 1997
). Endogenous selection against unfit hybrids is well documented (Barton and Hewitt, 1985
), but probably less common than generally thought (Arnold and Hodges, 1995
; Rieseberg, 1995
; Burke, Carney, and Arnold, 1998
). The position of many hybrid zones correlates with potential environmental selective factors (e.g., Moore, 1977
; Rand and Harrison, 1989
; Cruzan and Arnold, 1993
; Moore and Price, 1993
; Graham, Freeman, and McArthur, 1995
; Shoemaker, Ross, and Arnold, 1996
; MacCallum et al., 1997
; Freeman et al., 1999
), but reciprocal transplant experiments testing the relative fitness of both parental species and hybrids are still relatively rare. For example, hybrids did either not have substantial fitness disadvantages (Levin and Schmidt, 1985
), at least none in the hybrid sites (Janson, 1983
; Rolán-Alvarez, Johannesson, and Erlandsson, 1997
), or were even fitter in the hybrid sites (Emms and Arnold, 1997
; Wang et al., 1997
).
The perennial clonal grassland plants, Prunella vulgaris L. and Prunella grandiflora L. (Lamiaceae) can easily hybridize (Böcher, 1949
; Hess, Landolt, and Hirzel, 1972
; Hegi, 1979
), but in sympatry they are often clearly spatially separated, with hybrid plants only occurring where P. vulgaris and P. grandiflora patches adjoin (F. Fritsche, personal observation). Clonal growth strategies of the two species differ fundamentally. Prunella vulgaris plants produce loosely aggregated branches with long internodes (Schmid and Harper, 1985
). This guerrilla strategy (Lovett Doust, 1981
; Schmid, 1985
) allows foraging for resources and rapid biomass accumulation under favorable conditions (Grime, 1979
; Schmid, 1990
; De Kroon and Huchings, 1995
). In contrast, P. grandiflora has compact stands of physiologically tightly integrated branches (phalanx type), advantageous in less productive environments characterized by abiotic stress (Grime, 1979
; Jónsdóttir and Watson, 1997
). Thus, we hypothesized that Prunella hybrid zones have underlying environmental gradients favoring the growth strategy of P. grandiflora at one end, that of P. vulgaris at the other, and probably a combination of the two (i.e., a hybrid strategy) in the middle part.
Here, we studied a 30-m hybrid zone of P. grandiflora, putative hybrids, and P. vulgaris along a slope in a Swiss Jura Mountain grassland. First, we used morphological and floral traits to characterize this cline and test whether plants, visually characterized as Hybrids, could be distinguished from the parental species. Second, we measured soil and vegetation parameters to describe the environmental variation along the slope. Finally, we reciprocally transplanted P. grandiflora, Hybrids, and P. vulgaris genotypes and measured survival, clonal reproduction, and flowering probability for the first year after transplanting. According to the tension zone model, parental species should be equally fit at all sites and Hybrids always less fit than both parental species. Ecological selection-gradient models would be supported if relative performance of the plant groups varied with transplant site, such that parental species performed best at their native sites and Hybrids were as fit as at least one of the parental species. The bounded hybrid superiority model predicts that all three groups perform best in their own site.
MATERIALS AND METHODS
Study organisms
Prunella vulgaris is found in a variety of habitats (Hegi, 1979
). In lawns and pastures, such as our study population, plants are perennial and exhibit a guerrilla strategy (Lovett Doust, 1981
; Schmid, 1985
). Creeping vegetative branches of P. vulgaris root in autumn and initiate new daughter shoots. These daughter ramets overwinter as small rosettes and elongate in the following spring. Physiological connections between daughter ramets often decay within one growing season (Schmid, 1985
). About 20% of the ramets produce erect stems with one to several inflorescences (spikes), flowering from June to September (F. Fritsche, unpublished data). Flowers are self-compatible with abundant seed set, even in the absence of pollinators (Winn and Werner, 1987
).
Prunella grandiflora L. is restricted to the dry, nutrient-poor soil of grasslands on the European mainland (Hegi, 1979
). Branches are produced at the base of the primary shoot that dies back in autumn. The remaining daughter ramets overwinter as small rosettes and, in turn, produce none to several daughter ramets before they die after the following season (F. Fritsche, unpublished data). Ramets spread only little and develop a rhizomatous main axis. Larger rhizomes may eventually disconnect into clonal fragments. From July to September, nearly half of the shoots of a plant flower. Prunella grandiflora generally produces larger flowers than does P. vulgaris (Binz, 1986
).
Prunella grandiflora x P. vulgaris hybrids have been reported as intermediate in morphological and floral traits characterizing the parental species (Hegi, 1979
). A narrow area between two pure stands of P. grandiflora and P. vulgaris at our study site contains plants that appeared intermediate for several characters (see below) and clear-cut classification according to a standard key of the Swiss Flora (Binz, 1986
) was impossible. Therefore, we considered these plants as hybrids.
Study site
This study was performed in a calcareous grassland in the Jura Mountains, 30 km south of Basel, Switzerland. This south-southwest-exposed pasture has been grazed by horses for at least 15 yr, but was already managed as a pasture in 1953 (Zoller, 1954
). Observations of Hybrids in this grassland date back until the 1980s (S. Fiechter, personal communication, University of Basel). The upper part of the pasture has a slope of over 20°, while the lower part is nearly flat and partly shaded by a row of trees. Soil depth, water capacity, and nutrients diminish from the flat area towards the steeper upper part of the slope (Birrer, 1993
). Species composition changes from more mesophilic species in the lower part of the slope to species representative of unfertilized grasslands in the upper part (Fiechter, 1989
). Prunella vulgaris is restricted to the lower, flat area of the pasture where it dominates vegetation, whereas P. grandiflora is found almost over the entire area but at much lower densities. At the upper border of the partly shaded area (~10 x 30 m) Hybrid plants (proportion: 0.48) coexist together with both P. vulgaris (0.16) and P. grandiflora (0.36).
Classification of parental species and hybrid
We used nine morphological and floral traits to test whether the plants collected from the three areas in our study site and propagated for the transplant experiment indeed differentiated into parental species and hybrids. Five of these traits were determined on the plants removed from the field: (1) growth habit (i.e., whether the plants produced prostrate, ascending, or erect stems); (2) shoot compactness (proportion of internodia <7 mm, measured on the longest (with up to eight internodes) shoot of the plant); (3) leaf structure ("smooth" or "crumpled"); (4) leaf form (leaf length divided by width, averaged over the first pair of fully expanded leaves on the longest shoot); (5) stem color (all four sides of the stem red, all four green, or two red/two green). These plants had not yet initiated flowering. Therefore floral traits were determined for plants derived as cuttings from these plants and transplanted back into the field. We measured another four traits on flowering experimental plants: (6) distance between inflorescence and uppermost leaf; (7) inflorescence length; (8) flower angle (i.e., bending of the corolla tube); and (9) corolla tube length (measured on one flower randomly chosen from each plant). Because inflorescences elongate during the season, inflorescence length was measured at each monitoring date and the maximum value used for analysis. We further averaged measurements of all four floral traits across experimental plants (up to six plants) derived from the same original plant. We compared the three plant groups separately for each of the nine traits, using one-way ANOVAs for continuous and G tests for categorical variables. Student-Newman-Keuls tests were used for a posteriori multiple comparisons (Zar, 1984
). Comparisons between pairs of plant groups for categorical traits were performed by separate G tests for pairwise combinations of groups (Sokal and Rohlf, 1981
). Analyses were carried out with the JMP 3.0 statistical package (SAS, 1994
).
Environmental variation along the slope
We used two soil and two vegetative parameters to characterize environmental conditions in the habitats of P. grandiflora in the upper, Hybrids in the middle, and P. vulgaris in the lower part of the slope. In each of the three plant sampling areas (see below) we randomly chose five patches containing Prunella plants. We determined soil depth and height of vegetation in these patches, and harvested all aboveground biomass from a 25 x 25 cm area, which was dried for 2 wk at 80°C and then weighed. In two of the five patches, we also took soil samples to measure soil humidity. Samples were dried for 3 d at 105°C and percent water content was determined (Scheffer and Schachtschabel, 1989
). As above, to test for differences among the three areas, we carried out one-way ANOVAs and a posteriori multiple comparisons for each parameter separately.
Transplant experiment
In April 1993, we collected P. vulgaris plants from the lower, Hybrids from the middle, and P. grandiflora plants from the upper part of the slope. We marked all available plants in a 10 x 30 m area in each of the three parts and dug out up to 35 (randomly chosen) of these marked plants. To increase the chance of sampling different genotypes, plants had to be separated from each other by at least 2 m. The largest connected part of a plant was considered a "genet" (Harper and White, 1974
) and used for propagation of plant material in the greenhouse. From each genet we generated up to eight cuttings, i.e., shoot tips of ~3 cm length with four leaves and no roots. All cuttings were numbered and randomly planted in plastic containers filled with a mixture of 90% sand and 10% soil from the original site to ensure mycorrhizal infection, typically found in both species (Streitwolf-Engel et al, 1997
). Until new roots had formed the plants were raised for 5 wk under uniform greenhouse conditions to minimize potential carryover effects from the field. The six plants most similar in size from each of 20 clones from each of the three groups were chosen for transplanting.
Plants were transplanted as follows. Along a vertical transect through the slope we determined a patch of 3 x 3 m (referred to as "site" hereafter) in each of the three sampling areas. These transplanting sites visually represented the typical habitat; in the Hybrid site both parental species and Hybrids were present. Within each site, we determined two blocks each with 60 planting positions, arranged in five horizontal rows with 12 planting positions per row. Rows were separated by 10 cm and planting positions within rows by 20 cm. We removed aboveground biomass in a 3-cm radius around each planting position. At the end of May 1993, 360 plants were transplanted into the prepared positions. Within each site and block we planted one plant from each of 20 genets from each of the two parental species and the Hybrids (total N in each site = 3 x 20 x 2 = 120), randomized across planting positions. The sites were fenced against horses. To facilitate initial establishment, we regularly watered the plants during the first 2 wk after transplantation until the soil around the planting positions was recolonized by vegetation. To make conditions more similar to those outside the fences, we cut back the vegetation within the fenced sites if its height exceeded that of the vegetation outside the fences. Plants were monitored in 3-wk intervals from 16 June to 17 November 1993 and from 7 March to 7 May 1994.
We analyzed the following traits: (1) survival—whether or not the plant survived until May 7 1994; (2) ramet number—the number of ramets of those plants that survived until 7 May 1994; this is an often used fitness measure for clonal plants (Harper, 1977
; De Kroon and Huchings, 1995
) and correlates with plant biomass or leaf area in P. vulgaris (Schmid and Harper, 1985
) (we did not harvest biomass because the study was initially designed as a long-term experiment); (3) vegetative reproduction—the number of ramets of all plants, both dead (hence with a value of 0) and alive at the end of May 1994. Combining traits (1) and (2), this parameter is the product of survival until a given age and (vegetative) fecundity at that age—a classical definition of fitness (Stearns, 1992
); (4) flowering status—whether or not plants produced flowers with successful fruit set during summer and autumn 1993 (plants had not yet flowered again by the end of the monitoring period in spring 1994).
Data were analyzed using maximum likelihood approaches. To allow for binomial error structure of binary survival or flowering data ("dead"/"alive"; "flowering"/"not flowering") we carried out logistic regressions using the logit link function of the GLIM statistical package (Baker, 1987
). To analyze variation in ramet number and vegetative reproduction, we employed the GLIM log link function, appropriate for poisson error structure (because many plants had few ramets). Logistic regression estimates the change in deviance (= twice the log-likelihood ratio) caused by a given model term; this deviance approximates and is tested against the 
2 distribution. However, for terms to be tested against other terms in the model we calculated mean deviance changes (= deviance divided by degrees of freedom) to produce pseudo-F statistics that approximate and are therefore tested against the F distribution (McCullagh and Nelder, 1989
). This type of analysis is referred to as analysis of deviance (Schmid and Dolt, 1994
). For each of the four variables we tested fully factorial models with the effects of "group" (= the three plant types: P. vulgaris, P. grandiflora, and the P. vulgaris x P. grandiflora hybrid), "genet" nested within group (= the family of six cuttings derived from the same plant), transplanting "site" and "block" nested within site. Model fitting was analogous to calculation of SAS Type II sums of squares (SAS, 1988
). To calculate the deviance change caused by a particular factor in the model, we first fitted all other terms not containing this factor, then the factor in question. Genet(group) and block(site) were considered as random factors and all other effects as fixed. To investigate in more detail the variation among the three plant groups within sites we also performed analyses for each site separately. In these analyses, we first fitted the effect of block, then group and genet(group), with the latter serving as error term for the group effect. Breaking down the group effect into the deviance explained by the three possible pairwise comparisons between groups allowed us to test for differences in their ranking within sites.
RESULTS
Morphological and floral traits of parental species and hybrids
We obtained significant differences among P. grandiflora, P. vulgaris, and the putative P. grandiflora x P. vulgaris hybrids (= Hybrids) for the nine morphological and floral traits (Table 1). Multiple comparisons showed that P. grandiflora could be clearly distinguished from P. vulgaris (Table 1). For all but one trait (corolla tube length), Hybrids were intermediate between the parental species, although in some tests they were not significantly different from one of the two parental species. Floral traits of Hybrids were consistently more similar to those of P. grandiflora than P. vulgaris.
|
|
|
|
Vegetative reproduction
Combining survival and ramet number into a composite parameter revealed that the Grandiflora site was less favorable for vegetative reproduction (mean ± 1 SE = 0.68 ± 0.11) than were the other two sites (Hybrid site: 1.37 ± 0.11; Vulgaris site: 1.29 ± 0.18; significant effect of site, Table 3). We found rank reversals in vegetative reproduction for the two parental species across sites (significant site x group interaction; Table 3). Prunella vulgaris and P. grandiflora each ranked first in their own site. This home site advantage was more pronounced for P. grandiflora over P. vulgaris (Grandiflora site: G vs. V: F1,57 = 9.59, P = 0.0030) than vice versa (Vulgaris site: V vs. G: F1,57 = 4.49, P = 0.0385). Unlike for the Grandiflora (F2,57 = 7.25, P = 0.0016) and Vulgaris (F2,57 = 6.25, P = 0.0033) sites, there was no significant variation in vegetative reproduction among the three plant groups in the Hybrid site (F2,57 = 2.10, P = 0.1318). Hybrid vegetative reproduction was intermediate between that of the parental species in all three sites, and only in the Grandiflora site was there a (marginally) significant difference between Hybrids and one of the two parental species (H vs. V: F1,57 = 4.15, P = 0.0463, all other comparisons between Hybrids and parental species: F1,57 < 1.60, P >0.2; Fig. 1C).
Flowering status
Only 18.7% of the plants flowered and produced fruits during 1993. Therefore, to balance the data set, we combined the number of flowering plants across genets for each of the three plant groups within each block and site; we did not consider plants that died without flowering before the last monitoring date in 1993. Overall, P. grandiflora flowered more often (29%) than did the Hybrids (20%) and P. vulgaris (7%; significant effect of group, Table 3). Less than 10% of the plants flowered in the Grandiflora site, whereas ~25% of the plants flowered in the each of the other two sites (significant effect of site, Table 3). There also was a significant site x group interaction (Table 3). Flowering propensity of P. vulgaris was constantly low across all sites. In contrast, Prunella grandiflora increased in flowering propensity down the slope, with almost half of its plants flowering in the Vulgaris site, while the Hybrids did not flower at all in the Grandiflora site and flowered most often in their own site (Fig. 1D). Additional analysis revealed significant variation in flowering propensity among groups in all three sites (all tests: 2 > 6.15, df = 2, P < 0.05). Both P. grandiflora and Hybrids tended to have a flowering home-site advantage over non-native plants, although they did not flower statistically more often than both non-native groups (Grandiflora site: G vs. V: 2 = 1.24, P = 0.2655, G vs. H: 2 = 8.26, P = 0.0041; Hybrid site: H vs. G: 2 = 1.15, P = 0.2835, H vs. V: 2 = 6.27, P = 0.0123; all df = 1), probably because of the low statistical power due to the low number of flowering plants in these sites. In its own site, P. vulgaris flowered consistently less often than did both non-native groups (V vs. H: 2 = 4.88, P = 0.0272; V vs. G: 2 = 12.26, P = 0.0005, df = 1; Fig. 1D).
DISCUSSION
Plants considered as Prunella grandiflora x P. vulgaris hybrids (= Hybrids) were, on average, intermediate between the parental species in all but one of the nine characters examined (Table 1). This is consistent with previous descriptions of Prunella hybrids (Hess, Landolt, and Hirzel, 1972
; Hegi, 1979
). Morphology is widely used to identify hybrids, and clines of morphology often correlate with those based on chemical or genetic characters (Moore and Buchanan, 1985
; Freeman et al., 1991
; Nürnberger et al., 1995
; Shoemaker, Ross, and Arnold, 1996
; Howard et al., 1997
). Of course, without genetic analysis, it is still possible that our Hybrids were in fact phenotypically plastic variants of the parental species. However, half of the plants in the Hybrid area were clearly either P. grandiflora or P. vulgaris, and ambiguous phenotypes grow only in this narrow area. Hence, we think that we have identified true hybrids.
The two main types of hybrid zone models differ in their assumptions about the nature of selection within the zone, which can be tested by reciprocal transfer experiments (Moore and Price, 1993
). The tension zone model assumes intrinsic, universal hybrid unfitness and no variation in fitness between parental species across the hybrid zone (Barton and Hewitt, 1985, 1989
), whereas ecological selection-gradient models assume genotype-by-environment interactions, i.e., that the relative fitness of parental species and hybrids varies across the hybrid zone (Endler, 1977
; Moore, 1977
; Moore and Price, 1993
).
Our results do not support the tension zone model. First of all, Prunella hybrids were not uniformly less fit than the parental species, and their performance in the different transplanting sites was almost always intermediate between that of the parental species (Fig. 1). To be able to transplant genetically identical replicates of naturally occurring genotypes, we had chosen to use cuttings from field plants rather than (seed) progeny from artificial crosses. Hence, we may have missed hybrid unfitness at seed or seedling stages, where developmental stability is important. However, even if intrinsic hybrid disadvantage exists, it may be overcome relatively easily in clonal plants (Emms and Arnold, 1997
), because a few successful initial hybrid individuals may be able to spread clonally, thereby establishing a hybrid population. Offspring from intra-hybrid mating or backcrossing may then suffer less from hybrid unfitness (Rieseberg, 1995
).
A further inconsistency with the tension zone model is the finding of plant group x transplanting site interactions. In fact, our results support the presence of an environmentally mediated selection gradient structuring this hybrid zone. In particular, we obtained rank reversals of the two parental species, such that each species had a home-site advantage over nonnative plants, at least for vegetative reproduction measured in this one-year period (Fig. 1C). The home-site advantage of P. grandiflora was mediated by differential mortality, while that of P. vulgaris resulted from differential shoot production (Fig. 1A, B). This pattern appears to reflect adaptation of their clonal growth strategies to the environmental gradient along this slope.
In the upper, steep Grandiflora area, soil is less deep, contains fewer nutrients (Birrer, 1993
), less water, and has sparser vegetative coverage than in the lower areas of the slope (Table 2), suggesting that growth and survival in this area are limited by abiotic conditions. This explains not only the high overall mortality in this transplant site, but also the strong survival advantage of the conservative phalanx growth strategy of P. grandiflora—little clonal reproduction and investment into storage organs instead of ramet spread—over P. vulgaris, and to a lesser degree over the Hybrids (Fig. 1A). In contrast, the less steep Vulgaris area, with its denser vegetation, partial shading, and better nutrient and water availability, may provide an environment where fitness is determined by competition rather than abiotic stress (Law, Bradshaw, and Putwain, 1977
; Caswell, 1985
; Herben and Hara, 1997
). Such conditions can be expected to favor the opportunistic guerilla growth strategy of P. vulgaris, capable of rapid production and spread of new ramets without the need for expensive storage organs (Grime, 1979
; Schmid and Harper, 1985
; De Kroon and Huchings, 1995
). Indeed, in this site, the three plant groups survived equally well, but P. vulgaris produced considerably more shoots than non-native plants (Fig. 1A, B).
If parental species are adapted towards either end of the selection gradient, hybrids are likely to be stably maintained in that part of the gradient where parental reaction norms cross. Depending on the shape of the reaction norms of all three groups in this part, hybrids may then be less fit than the parental species, equally fit, or even fitter (Moore, 1977
; Moore and Buchanan, 1985
). Reciprocal seed and seedling transfer in a big sagebrush (Artemisia tridentata ssp.) hybrid zone revealed home site fitness advantages for both parental species and hybrids (Wang et al., 1997
). Similarly, reciprocal rhizome transfer in an iris (Iris fulva x I. hexagona) hybrid zone, found hybrid superiority in the hybrid sites, although hybrids also performed very well in parental sites (Emms and Arnold, 1997
).
In our experiment, the parental species' reaction norms for vegetative reproduction indeed crossed in the Hybrid transplant site, but there was no evidence for bounded hybrid superiority. In fact, all three reaction norms crossed in the Hybrid site (or near to it, Fig. 1C), such that there was no significant variation among parental species and hybrids. These results are similar to those for a marine snail (Littorina saxatilis) hybrid zone (Rolán-Alvarez, Johannesson, and Erlandsson, 1997
), where parental ecotypes were favored in the upper and lower shore zone, respectively, whereas both parental forms and hybrids survived equally well in the midshore zone.
Bounded hybrid superiority may arise in environments that have distinct, novel features, to which hybrid genotypes become adapted (Emms and Arnold, 1997
; Wang et al., 1998
; Freeman et al., 1999
; Wang, McArthur, and Freeman, 1999
). This hybrid zone marks a transition from a typical nutrient-poor mountain grassland community to a more mesophyllic community, obviously with an underlying environmental gradient. We have no indication that the transition area has unique features, but clearly this would require more detailed analysis, e.g., of the plant or mycorrhizal community composition (see, Streitwolf-Engel et al., 1997
). From plant performance in the Hybrid site, at least, it appears that this area is intermediate and represents aspects of both parental environments rather than a novel habitat. For parental species, the P. grandiflora's survival and P. vulgaris's ramet number advantage cancelled one another. Hybrids, although flowering more often (see below), did not combine parental growth strategies to benefit from higher survival and more growth; instead, their performance was quite similar to P. grandiflora. Consequently, no single growth strategy may be able to outcompete the others in this area, resulting in coexistence of parental species and Hybrids.
Our study has focused on aspects of vegetative reproduction in these clonal plants. However, Hybrids were more prone to flower and set seed in their own site than were nonnative plants. Similar to their natural field counterparts, P. vulgaris plants flowered far less often than did P. grandiflora (or Hybrids) (Fig. 1D). Clearly, differences in sexual reproduction among Hybrids and parental species can affect our conclusions about their relative fitness. Recruitment of new plants from seeds may further enhance the P. grandiflora home-site advantage, counteract the P. vulgaris home-site advantage, but also generate hybrid superiority in the Hybrid site. (Bounded) hybrid superiority for reproductive traits was also observed in the big sagebrush and iris systems (Graham, Freeman, and McArthur, 1995
; Wang et al., 1997
; Burke, Carney, and Arnold, 1998
).
Furthermore, variation in flowering behavior is likely to influence amount and direction of pollen transfer within this hybrid zone. For example, given their similarity in flowering frequencies and floral traits (Table 1), one may expect more backcrossing events between P. grandiflora and Hybrids, particularly if pollinators discriminate between flower types (see, e.g., Young, 1996
; Campbell, Waser, and Meléndez-Ackerman, 1997
; Campbell, Waser, and Wolf, 1998
). Asymmetric backcrossing may further help in preventing gene flow between P. grandiflora and P. vulgaris.
Sexual reproduction is only one of two options for clonal plants, and it is often argued that vegetative reproduction can be more important for population dynamics and growth, especially in grassland perennials (Lovett Doust, 1981
; Schmid and Harper, 1985
; Snaydon, 1985
; Schmid, 1990
). Recruitment from seed varies considerably among P. vulgaris populations (Schmid, 1985
; Winn and Werner, 1987
). In this grassland, P. vulgaris reproduced mainly by vegetative growth, although some seedlings established over a 2-yr period (F. Fritsche, unpublished data). Clearly, to assess the relative contribution of sexual vs. asexual reproduction to plant fitness in this hybrid zone, we need more detailed information about clonal spread and seed production, seedling establishment, or pollinator behavior, ideally over longer periods than in the present study.
In conclusion, our results support the assumption of an environmental selection-gradient model. The position and width of this hybrid zone appear to be maintained by diverging selection pressures on the gradient, with local adaptation of the parental species at either end. Although we cannot fully exclude the possibility of bounded hybrid superiority, parental species and hybrids seem to be equally fit and therefore coexist in the center of the gradient. Clearly, generalizations from one particular hybrid zone must be made with caution. Nonetheless, it is conceivable that, at larger geographical scales, niche differentiation of the two Prunella species leads to a discontinuous, mosaic distribution of sympatric populations and hybrid zones, as suggested for other systems (Harrison and Rand, 1989
; Shoemaker, Ross, and Arnold, 1996
; Howard et al., 1997
). In particular, specialization of P. grandiflora to a rare habitat a priori restricts opportunities for hybridization and thus facilitates the maintenance of separated gene pools without the need for complete reproductive isolation.
FOOTNOTES
1 The authors thank Bernhard Schmid, Dani Prati, Markus Fischer, Jacqui Shykoff, Jacob Koella, and two anonymous reviewers for discussion and comments on earlier versions of the manuscript; Dres Birrer for introducing us to Prunella life history and input in designing the experiment; and Michael Fry for invaluable help during field work. This work was funded by the Swiss National Science Foundation (Swiss Priority Program on Environment, grant 5001–035229 to B. Schmid). 
5 Address for correspondence: McGill University, Biology Department, 1205, Avenue Docteur Penfield, Montreal, Quebec, Canada H3A 1B1.
LITERATURE CITED
Abbott, R. J. 1992 Plant invasions, interspecific hybridization and the evolution of new plant taxa. Trends in Ecology and Evolution 7: 401–404.[CrossRef]
Arnold, M. L. 1997 Natural hybridization and evolution, Oxford University Press, Oxford, UK.
———, and S. A. Hodges. 1995 Are natural hybrids fir or unfit relative to their parents? Trends in Ecology and Evolution 10: 67–71.
Baker, R. J. 1987 GLIM 3.77 reference manual, 2nd ed. Numerical Algorithms Group, Oxford, UK.
Barton, N. H., and G. M. Hewitt. 1985 Analysis of hybrid zones. Annual Review of Ecology and Systematics 16: 113–148.[CrossRef][ISI]
———, and ———. 1989 Adaptation, speciation and hybrid zones. Nature 341: 497–503.[CrossRef][Medline]
Bazykin, A. D. 1969 Hypothetical mechanism of speciation. Evolution 23: 685–687.[CrossRef][ISI]
Binz, A. 1986 Schul- und Exkursionsflora für die Schweiz, 18th ed., Schwabe, Basel.
Birrer, D. 1993 Genetisch-ökologischer Vergleich zwischen einem Spezialisten (Prunella grandiflora, L.) und einem Generalisten (Prunella vulgaris, L.). Diploma Thesis, Universität Basel, Bazel, Switzerland.
Böcher, T. W. 1949 Racial divergences in Prunella vulgaris in relation to habitat and climate. New Phytologist 48: 285–314.[ISI]
Burke, J. M., S. E. Carney, and M. L. Arnold. 1998 Hybrid fitness in the Louisiana irises: analysis of parental and F1 performance. Evolution 52: 37–43.
Butlin, R. K. 1987 Speciation by reinforcement. Trends in Ecology and Evolution 2: 8–13.
Campbell, D. R., N. M. Waser, and E. J. Meléndez-Ackerman. 1997 Analyzing pollinator-mediated selection in a plant hybrid zone: hummingbird visitation patterns on three spatial scales. American Naturalist 149: 295–315.[CrossRef][ISI]
———, ———, and P. G. Wolf. 1998 Pollen transfer by natural hybrids and parental species in an Ipomopsis hybrid zone. Evolution 52: 1602–1611.[CrossRef][ISI]
Caswell, H. 1985 The evolutionary demography of clonal reproduction. In J. B. C. Jackson, L. W. Buss, and R. E. Cook [eds.], Population biology and evolution of clonal organisms, 187–224. Yale University Press, New Haven, Connecticut, USA.
Cruzan, M. B., and M. L. Arnold. 1993 Ecological and genetic associations in an Iris hybrid zone. Evolution 47: 1432–1445.[CrossRef][ISI]
De Kroon, H., and M. J. Huchings. 1995 Morphological plasticity in clonal plants: the foraging concept reconsidered. Journal of Ecology 83: 143–152.[CrossRef]
Dobzhansky, T. 1940 Speciation as a stage in evolutionary divergence. American Naturalist 74: 312–321.[CrossRef][ISI]
Emms, S. K., and M. L. Arnold. 1997 The effect of habitat on parental and hybrid fitness: transplant experiments with Louisiana Irises. Evolution 51: 1112–1119.[CrossRef][ISI]
Endler, J. A. 1973 Gene flow and population differentiation. Science 179: 243–250.
———. 1977 Geographic variation, speciation and clines. Princeton University Press, Princeton, New Jersey, USA.
Felsenstein, J. 1981 Skepticism towards Santa Rosalia, or why are there so few kinds of animals? Evolution 35: 124–138.[CrossRef][ISI]
Fiechter, S. 1989 Veränderungen von Trespen- und Halbtrockenrasen (Mesobromion) zwischen Delémont und Lützeltal. Diploma Thesis, University of Basel, Basel, Switzerland.
Fisher, R. A. 1930 The genetical theory of natural selection. Clarendon Press, Oxford, UK.
Freeman, D. C., W. A. Turner, E. D. McArthur, and J. H. Graham. 1991 Characterization of a narrow hybrid zone between two subspecies of big sagebrush (Artemisia tridentata: Asteraceae). American Journal of Botany 78: 805–815.[CrossRef][ISI]
———, H. Wang, S. Sanderson, and E. D. McArthur. 1999 Characterization of a narrow hybrid zone between two subspecies of big sagebrush (Artemisia tridentata: Asteraceae): VII. Community and demographic analyses. Evolutionary Ecology Research 1: 487–502.
Gavrilets, S. 1997 Hybrid zones with Dobzhansky-type epistatic selection. Evolution 51: 1027–1035.[CrossRef][ISI]
Graham, J. H., D. C. Freeman, and E. D. McArthur. 1995 Narrow hybrid zone between two subspecies of Big Sagebrush (Artemisia tridentata: Asteraceae). II. Selection gradients and hybrid fitness. American Journal of Botany 82: 709–716.[CrossRef][ISI]
Grime, J. P. 1979 Plant strategies and vegetation processes, John Wiley & Sons, New York, New York, USA.
Haldane, J. B. S. 1948 The theory of a cline. Journal of Genetics 48: 277–284.[ISI]
Harper, J. L. 1977 Population biology of plants, Academic Press, London, UK.
———, and J. White. 1974 The demography of plants. Annual Review of Ecology and Systematics 5: 419–463.
Harrison, R. G. [ed.]. 1993 Hybrid zones and the evolutionary process. Oxford University Press, New York, New York, USA.
———, and D. M. Rand. 1989 Mosaic hybrid zones and the nature of species boundaries. In D. Otte, and J. Endler [eds.], Speciation and its consequences, 111–133. Sinauer, Sunderland, Massachusetts, USA.
Hegi, G. 1979 Illustrierte Flora von Mittel-Europa, Band 4. Hanser Verlag, München, Germany.
Herben, T., and T. Hara. 1997 Competition and spatial dynamics of clonal plants. In H. De Kroon and J. van Groenendael [eds.], The ecology and evolution of clonal plants, 331–357. Backhuys Publishers, Leiden, The Netherlands.
Hess, H. E., E. Landolt, and R. Hirzel. 1972 Flora der Schweiz. Birkhäuser Verlag, Basel, Switzerland.
Hewitt, G. M. 1988 Hybrid zones—natural laboratories for evolutionary studies. Trends in Ecology and Evolution 3: 158–167.[CrossRef]
Hostert, E. E. 1997 Reinforcement: a new perspective on an old controversy. Evolution 51: 697–702.[CrossRef][ISI]
Howard, D. J., R. W. Preszler, J. Williams, S. Fenchel, and W. J. Boecklen. 1997 How discrete are oak species? Insights from a hybrid zone between Quercus grisea and Quercus gambelii. Evolution 51: 747–755.
Janson, K. 1983 Selection and migration in two distinct phenotypes of Littorina saxatilis in Sweden. Oecologia 59: 58–61.
Jónsdóttir, I. S., and M. A. Watson. 1997 Extensive physiological integration: an adaptive trait in resource-poor environments. In H. De Kroon, and J. van Groenendael [eds.], The ecology and evolution of clonal plants, 109–136. Backhuys Publishers, Leiden, The Netherlands.
Karlin, S., and J. McGregor. 1972 Application of method of small parameters in multi-niche population genetics models. Theoretical Population Biology 3: 180–209.
Key, K. H. L. 1968 The concept of stasipatric speciation. Systematic Zoology 17: 14–22.[CrossRef]
Law, R., A. D. Bradshaw, and P. D. Putwain. 1977 Life-history variation in Poa annua. Evolution 31: 233–246.
Levin, D. A., and K. P. Schmidt. 1985 Dynamics of a hybrid zone in Phlox: an experimental demographic investigation. American Journal of Botany 72: 1404–1409.[CrossRef][ISI]
Lovett Doust, L. 1981 Population dynamics and local specialization in a clonal perennial (Ranunculus repens). 1. The dynamic of ramets in contrasting habitats. Journal of Ecology 69: 743–755.[CrossRef]
MacCallum, C., B. Nürnberger, N. H. Barton, and J. M. Szymura. 1997 Habitat preference in the Bombina hybrid zone in Croatia. Evolution 52: 227–239.[CrossRef][ISI]
May, R. M., J. A. Endler, and R. E. McMurtrie. 1975 Gene frequency clines in the presence of selection opposed by gene flow. American Naturalist 109: 659–676.[CrossRef][ISI]
Mayr, E. 1942 Systematics and the origin of species. Columbia University Press, New York, New York, USA.
McCullagh, P., and J. A. Nelder. 1989 Generalised linear models, 2nd ed. Chapman and Hall, London, UK.
Moore, W. S. 1977 An evaluation of narrow hybrid zones in vertebrates. Quarterly Review of Biology 52: 263–277.[CrossRef][ISI]
———, and D. B. Buchanan. 1985 Stability of the northern flicker hybrid zone in historical times: implications for adaptive speciation theory. Evolution 39: 135–151.[CrossRef][ISI]
———, and J. T. Price. 1993 Nature of selection in the northern flicker hybrid zone and its implications for speciation theory. In R. G. Harrison [ed.], Hybrid zones and the evolutionary process, 196–225. Oxford University Press, New York, New York, USA.
Nürnberger, B., N. Barton, C. MacCallum, J. Gilchrist, and M. Appleby. 1995 Natural selection on quantitative traits in the Bombina hybrid zone. Evolution 49: 1224–1238.[CrossRef][ISI]
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 43: 432–439.[CrossRef][ISI]
Rieseberg, L. H. 1995 The role of hybridization in evolution: old wine in new skins. American Journal of Botany 82: 944–953.[CrossRef][ISI]
Rolán-Alvarez, E., K. Johannesson, and J. Erlandsson. 1997 The maintenance of a cline in the marine snail Littorina saxatilis: the role of home site advantage and hybrid fitness. Evolution 51: 1838–1847.[CrossRef][ISI]
SAS. 1988 SAS/STAT User's guide, Release 6.03. SAS Institute. Cary, North Carolina, USA.
———. 1994 JMP statistics and graphics guide. SAS Institute, Cary, North Carolina, USA.
Scheffer, F., and P. Schachtschabel. 1989 Lehrbuch der Bodenkunde, Ferdinand Enke Verlag, Stuttgart, Germany.
Schmid, B. 1985 Clonal growth in grassland perennials. II. Growth form and fine-scale colonizing ability. Journal of Ecology 73: 809–818.[CrossRef]
———. 1990 Some ecological and evolutionary consequences of modular organization and clonal growth in plants. Evolutionary Trends in Plants 4: 25–34.
———, and C. Dolt. 1994 Effects of maternal and paternal environment and genotype on offspring phenotype in Solidago altissima L. Evolution 48: 1525–1549.[CrossRef][ISI]
———, and J. L. Harper. 1985 Clonal growth in grassland perennials. I. Density and pattern dependent competition between plants with different growth forms. Journal of Ecology 73: 793–808.[CrossRef]
Shoemaker, D. D., K. G. Ross, and M. L. Arnold. 1996 Genetic structure and evolution of a fire ant hybrid zone. Evolution 50: 1958–1976.[CrossRef][ISI]
Slatkin, M. 1973 Gene flow and selection in a cline. Genetics 75: 733–756.
———. 1985 Gene flow in natural populations. Annual Review of Ecology and Systematics 16: 393–430.[ISI]
Snaydon, R. W. 1985 Aspects of the ecological genetics of pasture species. In J. Haeck and J. W. Woldendorp [eds.], Structure and functioning of plant populations, 127–152. North-Holland, Amsterdam, The Netherlands.
Sokal, R. R., and F. J. Rohlf. 1981 Biometry, W. H. Freeman, New York, New York, USA.
Spencer, H. G., B. H. McArdle, and D. M. Lambert. 1986 A theoretical investigation of speciation by reinforcement. American Naturalist 128: 241–262.[CrossRef][ISI]
Stearns, S. C. 1992 The evolution of life histories. Oxford University Press, Oxford, UK.
Stebbins, G. L. 1959 The role of hybridization in evolution. Proceedings of the American Philosophical Society 103: 231–251.
Strauss, S. Y. 1994 Levels of herbivory and parasitism in host hybrid zones. Trends in Ecology and Evolution 9: 209–214.[CrossRef]
Streitwolf-Engel, R., T. Boller, A. Wiemken, and I. R. Sanders. 1997 Clonal growth traits of two Prunella species are determined by co-occurring arbuscular mycorrhizal fungi from a calcareous grassland. Journal of Ecology 85: 181–191.[CrossRef][ISI]
Wang, H., D. W. Byrd, J. L. Howard, E. D. McArthur, J. H. Graham, and D. C. Freeman. 1998 Narrow hybrid zone between two subspecies of big sagebrush (Artemisia tridentata: Asteraceae). V. Soil properties. International Journal of Plant Sciences 159: 139–147.[CrossRef]
———, E. D. McArthur, and D. C. Freeman. 1999 Narrow hybrid zone between two subspecies of Big Sagebrush (Artemisia tridentata: Asteraceae). IX. Elemental uptake and niche separation. American Journal of Botany 86: 1099–1107.
———, ———, S. C. Sanderson, J. H. Graham, and D. C. Freeman. 1997 Narrow hybrid zones between two subspecies of big sagebrush (Artemisia tridentata: Asteraceae). IV. Reciprocal transplant experiments. Evolution 51: 95–102.[CrossRef][ISI]
Winn, A. A., and P. A. Werner. 1987 Regulation of seed yield within and among populations of Prunella vulgaris. Ecology 68: 1224–1233.
Young, N. D. 1996 An analysis of the causes of genetic isolation in two Pacific Coast iris hybrid zones. Canadian Journal of Botany 74: 2006–2013.[CrossRef]
Zar, J. H. 1984 Biostatistical analysis, 2nd ed. Prentice-Hall, Englewood Cliffs, New Jersey, USA.
Zoller, H. 1954 Die Typen der Bromus erectus-Wiesen des Schweizer Juras. Beiträge zur Geobotanischen Landesaufnahme der Schweiz Bern, Heft 33.
This article has been cited by other articles:
|
|
 |
|
  A. P. LORENZ-LEMKE, V. C. MUSCHNER, S. L. BONATTO, A. C. CERVI, F. M. SALZANO, and L. B. FREITAS Phylogeographic Inferences Concerning Evolution of Brazilian Passiflora actinia and P. elegans (Passifloraceae) Based on ITS (nrDNA) Variation Ann. Bot., April 1, 2005; 95(5): 799 - 806. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. P. Hauser, C. Damgaard, and R. B. Jorgensen Frequency-dependent fitness of hybrids between oilseed rape (Brassica napus) and weedy B. rapa (Brassicaceae) Am. J. Botany, April 1, 2003; 90(4): 571 - 578. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Schweitzer, G. D. Martinsen, and T. G. Whitham Cottonwood hybrids gain fitness traits of both parents: a mechanism for theirlong-term persistence? Am. J. Botany, June 1, 2002; 89(6): 981 - 990. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||
