| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
(American Journal of Botany. 2008;95:1072-1078.) doi: 10.3732/ajb.0800087 © 2008 Botanical Society of America, Inc. |
What's this? |
Ecology |
Plant Ecology and Systematics, Department of Ecology, Sölvegatan 37, S-223 62 Lund, Sweden
Received for publication 6 March 2008. Accepted for publication 15 July 2008.
ABSTRACT
Despite evidence that both pollinators and nonpollinator agents of selection can shape the evolution of floral characters, there have been few attempts to compare the strengths and directions of selection from pollinators and other agents in the same study system. In this investigation of Leucanthemum vulgare, a self-incompatible composite known for its conspicuous white rays, I obtained data from a ray removal experiment in the field and from a segregating F2 population in an experimental garden to assess the role of pollinator and nonpollinator selection as stabilizing factors on floral evolution in this species. Removal of all rays reduced the pollination success of heads by 31–35%, but did not significantly affect the level of infestation by larvae of the fly Tephritis neesii. Data from F2 plants indicated a potential for indirect selection on ray morphology, mediated through links between ray morphology and measures of vegetative size and plant vigor. The results of this study show that individuals of the normal, rayed phenotype have a clear selective advantage, in terms of both pollinator attraction and general plant vigor. Thus, there were no conflicting selection pressures between the pollinators and the other selective agents considered in this study.
Key Words: Asteraceae • floral evolution • Leucanthemum vulgare • pollination • selection
Understanding the selective mechanisms underlying the great diversity of floral sizes, shapes and rewards in the angiosperms, remains a primary challenge for plant evolutionary biologists. There is ample evidence that both pollinators and nonpollinator agents of selection can shape the evolution and diversification of floral characters (Grant, 1949
; Stebbins, 1974; Faegri and van der Pijl, 1979; Galen, 1999b; Fenster et al., 2004; Strauss and Whittall, 2006). In the case of floral size variables, the patterns of selection are determined by (1) the extent to which large floral organs enhance the plant’s visual display to floral mutualists and antagonists (Bell, 1985; Brody, 1992; Galen, 1999a), (2) the importance of perianths as protective organs during the bud stage (Delph et al., 1996), (3) the existence of tradeoffs between the allocation of resources to floral advertising vs. fruit or seed production (Charlesworth and Charlesworth, 1987; Andersson, 2005), and (4) the pattern of genetic association between floral and nonfloral size characters (Primack, 1987; Andersson, 1993, 1997). Thus, it is necessary to consider a broad variety of selective pressures and fitness consequences to understand how pollinators and other ecological forces interact to determine the optimum floral phenotype of a species or population. Despite growing evidence for the importance of both pollinator and nonpollinator selection in floral evolution (Galen, 1999a; Strauss and Whittall, 2006), there have been relatively few attempts to compare the strengths and directions of selection from pollinators and other agents in the same study system (Strauss and Whittall, 2006). In this regard, it seems particularly relevant to study plants with generalist pollinators, given the inferred role of conflicting selection pressures in preventing the evolution of specialization in generalized pollination systems (Fenster et al., 2004).
Characters reflecting flower morphology often show high phenotypic stability within and among plants of the same species or population (Stebbins, 1974), as evidenced by the frequent use of floral features as "key characters" of individual species, especially in taxonomic groups where pollination by animals predominates (Grant, 1949). While the relative constancy of flowers may be a consequence of unifying selection imposed by local pollinators (Berg, 1960; Stebbins, 1974; Wolfe and Krstolic, 1999), there are other possible causes, for example, intrinsic developmental factors that limit the range of phenotypes that can be expressed in a floral character and negative (purifying) selection arising from the physiologically adverse effects that may be associated with major changes in flower development (Stebbins, 1974; Fenster and Galloway, 1997; Cresswell, 1998).
Identifying agents of unifying selection is difficult when all or almost all individuals express the same phenotype. One way to circumvent this problem is to experimentally enhance the variation, either by artificial manipulation or the use of segregating progenies derived from crosses between phenotypically distinct genotypes. Data from manipulative experiments afford great confidence when inferring the functional (causal) relationship between phenotype and fitness, provided that the experimental units can be randomized across treatment categories (Mitchell-Olds and Shaw, 1987). Segregating populations allow detailed analyses of selection when the expanded variation is more continuous, i.e., when hybrid intermediates bridge the gap between the extreme phenotypes. Although measures of pollination success generally decrease for plants with artificially manipulated corollas (e.g., Nilsson, 1988; Sandvik and Totland, 2003; for exceptions, see Wilson, 1995; Herrera, 2001), there is still too little experimental data to draw general conclusions about the relative importance of pollinator and nonpollinator selection in preventing the invasion of aberrant floral morphologies.
Specialized ray florets (hereafter referred to as rays), a striking feature of many species in the Asteraceae family, play a major role as pollinator attractants (Leppik, 1977; Bertin and Kerwin, 1998), as evidenced by the positive association between pollination success and the possession of rays found in ray removal experiments (e.g., Stuessy et al., 1986) and in populations polymorphic for rayed and discoid (rayless) individuals (Marshall and Abbott, 1982; Abbott and Irwin, 1988). Nevertheless, there have been several parallel reversals to discoid heads in this family (Bremer and Humphries, 1993), and genetic data indicate a considerable potential for selective forces other than pollinators to influence the evolution of ray morphology, especially in self-fertilizing species (Abbott, 1986; Comes, 1998; Oxford et al., 1996; Abbott et al., 1998).
In oxeye daisy (Leucanthemum vulgare Lam.), a normally rayed species with a generalized pollination system, the discoid type is rare throughout the geographic range of the species (Tutin et al., 1976; Bogle, 1983), suggesting a history of unifying selection in favor of the ancestral, rayed phenotype. In this study, I obtained data from a ray removal experiment and a segregating F2 population, established in a near-natural garden environment, to evaluate the role of pollinators, floral antagonists and intrinsic, developmental links with overall size and performance variables in preventing the spread of the derived, discoid phenotype. Specifically, I asked: Are fully rayed plants of oxeye daisy more successful in terms of pollinator attraction or more likely to escape floral enemies than those with no or short rays? And, are measures of ray morphology correlated with stem height or measures of plant vigor?
MATERIALS AND METHODS
The plant
Oxeye daisy is an insect-pollinated, perennial plant found in open, grassy areas, both in its native geographic range (Eurasia) and in regions where the species occurs as a naturalized weed, e.g., North America. It flowers from June to August when the branched rhizome develops a variable number of stalks, each having one or a few heads on long terminal peduncles. Heads of the normal, rayed type are up to 7 cm in diameter and consist of hundreds of 4 mm long hermaphroditic disc florets surrounded by a variable number of pistillate ray florets, each having a 10–35 mm long, white ligule surmounting a short corolla tube. Fertilized florets develop into one-seeded, indehiscent fruits (achenes) with no special mechanism for dispersal (Howarth and Williams, 1968).
Plants of oxeye daisy are more or less self-incompatible and require insect visitation to set fruit (see Results, The effects of ray removal). Flower visitors include a variety of generalist pollinators, mainly members of Coleoptera, Diptera, Hymenoptera and Lepidoptera. Several herbivores and florivores are associated with oxeye daisy (Howarth and Williams, 1968), including larvae of the fly Tephritis neesii Meigen (Diptera: Tephritidae). Larvae of Tephritis make large mines in the receptacles causing the early arrest of flowers in parts of the inflorescences.
Ray removal experiment
The ray removal experiment was done in the summer of 2006 at two sites in the northern part of the city of Lund (southern Sweden). The first site (the "meadow site") is an artificially established hay meadow outside a university building (Department of Ecology, Lund University) about 1.6 km NE of the central railway station. The second site (the "roadside site") is a grassy roadside about 1 km NE of the meadow site. Oxeye daisy is common at both sites, forming patches of varying size within communities dominated by grasses (e.g., Festuca rubra, Poa pratense) and other herbaceous perennials (e.g., Hypericum performatum).
At the start of the experiment (early June), I marked 60 pairs of adjacent heads at each site, removed all the rays from one of the heads in each pair (before any of the florets had exposed their stigmas) and assigned the other head as a control. The two heads in each pair were separated by 10–30 cm depending on plant density, and different pairs of heads were separated by a minimum of 1 m. The patchy distribution of stems, together with the close similarity in head size and phenology within patches (S. Andersson, personal observation), indicated that heads in the same pair could represent the same clone. About 4 wk later, I collected all heads and placed each head in a sealed paper bag.
Many heads were infested by Tephritis larvae, as evidenced by the presence of large mines and a number of florets that failed to develop beyond the early bud stage (easily distinguished by their small size and dark color), and the emergence of adult flies from the heads stored in the sealed paper bags. To assess the level of infestation, I recorded the number of flies in each paper bag.
A sample of 30 "nonarrested" florets from each head was used to determine pollination efficiency, quantified as the proportion of florets that developed into achenes (hereafter called fruit set). Achenes from ray florets were excluded from the calculations of fruit set to allow comparison with heads on which all rays had been removed.
A crossing experiment was carried out to determine whether fruit set is dependent on outcrossing and can thus serve as a measure of the extent of cross-pollination achieved. In the autumn of 2006, I sowed a bulked sample of achenes from heads used in the ray removal experiment, planted 30 of the resulting seedlings in separate pots, and placed the pots in an unheated, pollinator-proof greenhouse. In 2007, when the majority of the plants reached anthesis, I marked two terminal heads on each individual and assigned the two heads to different pollination treatments: (1) hand-outcrossing with pollen from one or two other plants, and (2) self-pollination with pollen from the same head or another head on the same plant. Pollen was transferred with separate cotton swabs. Slow floral development made it necessary to repeat the pollination procedures every second or third day for 2 weeks. Fruit set was determined as in the field study.
Segregating population
The segregating F2 population was established to investigate, first, whether ray morphology covaries with vegetative size and performance variables, and second, how pollinators respond to differences in ray morphology when the floral variation is continuous rather than discrete, as in the ray removal experiment. This plant material originated from crosses between 10 greenhouse-grown plants, derived from a bulked sample of achenes from a few, scattered individuals of the discoid phenotype in a natural population of the rayed variant (about 75 km ESE of Lund). The progeny from the field-collected achenes had rayed flower heads, but with varying degrees of reduction in ray length and number (S. Andersson, personal observation), indicating that their discoid parents in the field had been outcrossed with pollen from plants of the normal, rayed variant. I assumed that the rayed, greenhouse-grown plants were heterozygous at loci with large effects on ray morphology and considered the outcrossed progeny of these (F1) parents as F2 segregants.
In 2004, I planted about 730 F2 seedlings from a bulked sample of achenes, obtained by mixing achenes from all the successful F1 crosses, in separate pots and placed the plants in a sunny part of an experimental garden (University of Lund), where they were exposed to natural levels of pollination and flower herbivory. The potted plants were arranged in five adjacent rows along a fence facing south on a bed of fine sand (
15 cm between rows, 10 cm between pots within rows). Water was supplied as needed, but no fertilizer was applied. The garden is situated 200 m S of the meadow site in a region where only the rayed morph has been recorded (S. Andersson, personal observation).
In 2005, each F2 plant was classified into one of four performance categories: (1) dead before flowering, (2) alive but remaining vegetative, (3) bolted but all heads wilted immediately before or during anthesis, and (4) bolted and at least one head developed to the fruit stage. Plants in category 4 were classified as rayed or discoid, and rayed plants were scored for ray length, based on the longest ray in the head terminating the tallest stem. This head was almost always the first to flower. Early initiation of rays during the bud stage enabled the identification of the floral morph (rayed vs. discoid) for plants in category 3.
In 2006, I scored each individual for performance (using the categories used in 2005) and floral morph and recorded the total number of heads initiated and the length and number of rays for the head terminating the tallest stem. If more than one flowering stem was available, I also obtained data on ray length, ray number, and stem height for the shortest flowering stem on each individual. The paired data for these variables made it possible to relate ray morphology and stem height both between and within individuals.
To investigate the relationship between ray morphology and pollination success, I scored one head on each of 80 F2 plants for fruit set, as in the ray removal experiment. For 20 of these plants, I also assessed the fruit set of a head that had been outcrossed by hand to determine whether female fertility was limited by insufficient pollination in the garden environment. Infestation by Tephritis was too infrequent to allow assessment of floral herbivory in the F2 population.
Statistical analysis
Following arcsine square-root transformation of fruit set, I performed a two-way ANOVA without replication for each field site using treatment (rayed vs. rayless) as a fixed factor and the identity of the head pair as a random block factor. A similar approach was used to assess the effects of open vs. hand pollination (fixed factor) and plant identity (random block factor) in the F2 population. The F2 data were also analyzed with one-way ANOVA to compare the mean fruit set of rayed and discoid individuals and with multiple regression to determine whether ray morphology and other characters were under direct selection in the F2 population. A standardized regression coefficient (b) quantified the direct relationship between each phenotypic character and fruit set, holding all the other characters constant. The role of pollinators as selective agents in the F2 population was also evaluated by regressing the degree of pollen limitation (quantified as the difference in fruit set between hand- and open-pollinated heads) on ray length and ray number.
The fruit set of self- and cross-pollinated heads in the greenhouse was compared using Wilcoxon’s matched-pairs signed-ranks test because transformation failed to improve normality. The data on floral herbivory (the level of infestation by Tephritis) from the ray removal experiment were too skewed to be used in ANOVA. For these data, rayed and rayless heads were compared using Mann–Whitney’s U test (data pooled over head pairs).
Associations between floral and nonfloral variables in the F2 population were studied in three ways. First, ANOVAs or goodness-of-fit tests were performed to compare rayed and discoid plants with respect to head number, stem height, survival rate and flowering percentage. Second, the data from the rayed individuals were subjected to product-moment correlation analysis to search for subtle relationships between the quantitative variables. This analysis also provided tests for correlated reductions in stem height and ray morphology, quantified as the difference between the tallest and shortest stem within those plants that initiated more than one stem in the 2006 season. Third, I performed paired t tests (using plants as blocks) to compare floral measurements across years (ray length) and across tall and short stems on the same individual (ray length, ray number). Relevant analyses included row number (1–5) as a covariate to provide statistical control for spatial variation within the garden plot. The assumption of normality in ANOVAs and correlation analyses was fulfilled by log-transformation of head number.
Descriptive statistics were based on data measured in their original scale. All analyses were carried out with SuperANOVA (1989) on a Macintosh computer.
RESULTS
The effects of ray removal
Greenhouse data show that oxeye daisy is self-incompatible: cross-pollinated heads had a much higher fruit set (mean 57%, range 0–93%, N = 20) than heads subjected to self-pollination (mean 2.5%, range 0–20%, N = 20; Z = –3.83, P < 0.001, Wilcoxon’s matched-pairs signed-ranks test). Two-way ANOVAs on data from the ray removal experiment revealed significant differences in fruit set between head pairs (meadow: F49,49 = 4.57; roadside: F49,49 = 4.31, P < 0.001 in both cases) and a highly significant decrease in fruit set following ray removal, both at the meadow site (mean 42% vs. 65% for control heads; F1,49 = 69.18, P < 0.001) and the roadside site (mean 41% vs. 59% for control heads; F1,49 = 37.06, P < 0.001).
The number of Tephritis larvae per head was higher and more variable at the meadow site (mean 1.95, range 0–8) and the roadside site (mean 0.46, range 0–5) than in the F2 population (mean 0.19, range 0–2), but was not significantly affected by ray removal in the field experiment (meadow: U = 1174.50, P = 0.593; roadside: U = 1197.50, P = 0.628).
Patterns of variation in the F2 population
As can be seen in Table 1, relatively few F2 individuals died before flowering (0–1%), remained vegetative (12–13%) or produced heads that wilted before the fruit stage (6–12%). Most of the plants (86–88%) produced heads that could be scored for the presence or absence of morphologically distinct ray florets (rayed vs. discoid). The percentage of rayed plants increased from 83% in 2005 (529 of 640 plants) to 91% in 2006 (570 of 628 plants). This increase was almost entirely due to individuals that shifted from being discoid in the first year to rayed in the second year.
|
|
21 = 296.28, P < 0.001) and 2006 (100% vs. 32% for discoid plants, 21 = 398.42, P < 0.001). In 2006, the rayed plants had the tallest stems (mean 48.4 cm vs. 18.9 cm for discoid plants; F1,626 = 226.20, P < 0.001, ANOVA) and the largest number of heads (mean 8.77 vs. 3.22 for discoid plants; F1,626 = 27.92, P < 0.001). The differences in stem height and head number remained significant after the inclusion of row number as a covariate (P < 0.001 in both cases).
The possession of rays in 2005 was associated with a higher reproductive performance in 2006, especially when quantified as total head number (9.35 vs. 4.78 for discoid plants; F1,579 = 12.44, P < 0.001), but to some extent also when measured as flowering percentage (95% vs. 81% for discoid plants, 21 = 29.10, P < 0.001) and the percentage of plants that developed their heads to the fruit stage (100% vs. 68% for discoid plants; 21 = 159.55, P < 0.001). The fraction of plants that remained alive in 2006 was not significantly affected by the presence or absence of rays in the 2005 season (>99% in both groups).
Correlation analysis of quantitative data from rayed F2 individuals revealed positive associations between variables (Table 3). Ray length and ray number per head were positively correlated with stem height and, to a lesser extent, total head number. All associations became weaker (especially those involving head number) but generally remained significant after accounting for the effect of row number. Ray length and ray number also remained significantly positively correlated with stem height after adjusting for variation in head number (r > 0.20, P < 0.001–0.01).
|
Judging from the results of two-way ANOVA, controlling for a significant effect of plant identity (F24,24 = 7.01, P < 0.001), the F2 plants had a significant increase in fruit set after hand pollination (mean 46%, range 0–87%, N = 25) relative to open pollination (mean 15%, range 0–57%, N = 79; F1,24 = 61.90, P < 0.001). Data from open-pollinated heads revealed a significantly higher fruit set for rayed F2 plants (mean 17%, range 0–57%, N = 65) than for discoid F2 plants (mean 5%, range 0–30%, N = 14; F1,77 = 8.24, P < 0.01). Multiple regression analyses revealed a positive relationship between ray length and fruit set for open-pollinated heads (b = 0.67, P < 0.01), whereas the relationship for hand-pollinated heads failed to reach significance (b = 0.70, P = 0.10). Ray number, stem height and total head number did not significantly affect fruit set, regardless of pollination treatment (|b| < 0.20, P > 0.34). Neither ray length nor ray number were significantly correlated with the degree of pollen limitation (|b| < 0.15, P > 0.49).
DISCUSSION
Despite growing evidence that both pollinator and nonpollinator selection can drive the evolution of floral characters (Grant, 1949; Faegri and van der Pijl, 1979; Galen, 1999b; Fenster et al., 2004; Strauss and Whittall, 2006), only a few investigations have explored how pollinators and other ecological forces operate simultaneously in exerting selection on floral size characters (Strauss and Whittall, 2006). In the study presented here, I have obtained data from a ray removal experiment in the field and a segregating F2 population in a near-natural garden environment, to assess the roles of pollinators and nonpollinator agents of selection as stabilizing factors on floral evolution in oxeye daisy, a self-sterile composite known for its conspicuous white rays. My study not only provided insights into the direct effects of rays on pollination success and floral herbivory, but also enabled me to assess the potential for indirect selection through intrinsic, developmental links with vegetative size and performance variables.
Plants of oxeye daisy almost always possess rayed flower heads, both in the native Eurasian range and in North America (Tutin et al., 1976; Bogle, 1983). The sparse and scattered occurrence of the discoid oxeye daisy morph suggests that the normal, rayed variant is optimal and that selection operates in a way that prevents the establishment and spread of aberrant floral phenotypes. The results of the present investigation demonstrate that the rayed phenotype has a clear selective advantage, in terms of both pollinator attraction and plant vigor and that it would be difficult for discoid mutants to establish and spread in the area studied. Interactions with floral herbivores (Tephritis) seem to play a minor role in exerting selection on ray morphology; thus, there were no conflicting selection pressures between the generalist pollinators and the nonpollinator agents of selection considered in this study.
Ray morphology vs. pollinator attraction and flower herbivory
The results of this study agree with the inferred advantage of floral organs that enhance the plant’s visual display to pollinators (Leppik, 1977; Bell, 1985; Bertin and Kerwin, 1998). The percentage fruit set of open-pollinated heads decreased by 31–35% after ray removal and was 71% lower for discoid plants in the segregating F2 population, which showed continuous rather than discontinuous variation between the discoid and the fully rayed phenotype. Thus, the reproductive advantage of possessing rayed flower heads was consistent not only across locations, but also between populations with widely different arrays of floral phenotypes. These consistencies point toward a predominant role for pollinator-mediated selection in maintaining the ancestral, rayed phenotype of oxeye daisy.
The decrease in percentage fruit set after ray removal falls within the range observed for composites with few or solitary heads (23% in Scalesia affinis, Rostgaard Nielsen et al., 2002; 64% in Helianthus grosseserratus, Stuessy et al., 1986), but exceeds values reported in species with a large number of heads massed together in dense, corymbiform inflorescences (0% in Senecio jacobaea, Andersson, 1996; 12% in Achillea ptarmica, Andersson, 1991; 15% in S. integrifolius, Andersson and Widén, 1993). The current study therefore provides additional support for the idea that the possession of rays is most favorable when the basic attraction units are individual heads rather than clusters of small heads in large compound inflorescences (Andersson, 1996).
Consistent differences were found in percentage fruit set between pairs of adjacent stems in the field (ray removal experiment) and between different plants in the F2 population, which agrees with results regarding other composites (Andersson 1991, 1996; Andersson and Widén, 1993). Thus, the use of head pairs or plants as blocks enhanced the statistical power to detect a functional relationship between ray morphology and pollination success by adjusting for variation in unmeasured parameters such as plant density, resource status, and degree of cross-fertility. In this context, it is necessary to stress that hand pollination significantly increased fruit set in the garden plot; thus, female fertility was at least partly determined by the amount of cross-pollination achieved, a necessary assumption when using fruit set as a proxy for pollinator visitation. Moreover, the positive relationship between ray production and fruit set among F2 plants remained significant when measures of plant vigor (e.g., total head number) were held constant in a multiple regression analysis. Because the amount of pollen donation (male success) is expected to be an increasing function of the visitation rate, it is also reasonable to assume that male and female fertility respond similarly to differences in ray morphology and that the results concerning pollination efficiency therefore apply to both sexual functions (Andersson, 1991).
Plants in the F2 population had no significant relationship between ray length and the degree of pollen limitation, as would be expected if long-rayed heads had the highest visitation rates. Although this result argues against direct, pollinator-mediated selection on ray morphology, the overall response to ray removal or hand-pollination demonstrates clearly that the possession of rays is associated with increased pollinator visitation. The individual-level analyses probably lacked the power to detect a relationship between ray morphology and pollen limitation, given the small number of F2 plants subjected to both open- and hand-pollination.
Floral antagonists, for example, flower herbivores, predispersal seed predators, and nectar thieves, sometimes act as strong selective agents on floral features that make plants more attractive to pollinators (Brody, 1992; Galen, 1999a; Strauss and Whittall, 2006). For instance, ovipositing females of the sunflower moth Homoeosoma electellum have been found to discriminate against plants with small flower heads (Pilson, 2000). In this study of oxeye daisy, no detectable effect of ray removal was found on the level of infestation by Tephritis, despite the drastic reduction in floral display arising from the manipulation. As found in a previous study of this and other composites, the incidence of larval infestation seems to be strongly determined by the diameter of the disc (Fenner et al., 2002), raising the possibility that the floral antagonists use attributes of the disc, for example, the number, color, or scent of disc florets, rather than the possession of rays, when searching for their host plants.
The potential for indirect selection
The data from the F2 population confirm previous observations from other plants that there is a potential for selective pressures other than pollinators and floral antagonists to impose selection on floral morphology through developmental associations between floral and nonfloral characters (e.g., Primack, 1987; Andersson, 1997; Strauss and Whittall, 2006). First, plants with discoid or short-rayed heads initiated fewer heads, their heads were less likely to develop to the fruit stage, and they had a significantly lower flowering percentage and inflorescence production in the subsequent year than those with well-developed rays. Second, I detected positive associations between stem height, ray length, and ray number per head among F2 plants in the rayed category, regardless of whether position (row number) or head number was held constant. The latter associations were also manifested as a correlated decrease in stem height and ray morphology within those plants that initiated more than one stem in the 2006 season. These findings indicate the existence of two partly independent axes of covariation, one involving ray morphology and general plant vigor and another involving ray morphology and plant stature. Given that general performance variables such as head number are under positive selection and that the relationship between low plant vigor and "raylessness" can be generalized to other populations, there is considerable potential for indirect selection pressures to prevent the spread of the derived, discoid phenotype.
The high interyear consistencies observed for ray morphology suggest that the floral variation had a strong genetic basis, confirming results from a previous study of oxeye daisy (Bogle, 1983). On the basis of the observed frequency of the discoid morph (about 10%), it is also reasonable to conclude that rayed heads are dominant over discoid heads, and that the suppression of rays is controlled by a relatively small number of genes, as observed in other species (Ford and Gottlieb, 1990; Andersson, 2001b; Gillies et al., 2002). However, it is not known whether the absence of rays represents a pleiotropic side effect of genes with deleterious effects on plant vigor, whether the reduction in vigor represents a side effect of genes suppressing ray development, or whether the observed associations stem from linkage equilibrium between loci controlling flower development and loci affecting vegetative growth parameters. In view of the strong linear relationship between stem height and quantitative measures of ray morphology, which remained significant after adjusting for variation in plant vigor, it seems reasonable to attribute some of the covariation to loci expressed during stem and flower development, not solely to loci with general effects on plant vigor. Loci affecting levels of endogenous gibberellin activity are one possible source of stem height–flower size associations, given the strong influence of gibberellin on the growth of both floral and vegetative structures (Jones, 1973; Koning, 1984).
A small fraction of F2 plants shifted from being discoid in the first season to being rayed in the second season, and there was a slight (though significant) increase in ray length between the two flowering seasons. These observations indicate that the overall plant vigor—and the ability to produce rays—increased with increasing age in the F2 population. This age effect was too weak to obscure the high interyear consistencies documented for ray morphology, but raises some concern about the "discoid" plants used as parents for this F2 population. However, the detection of discoid F2 segregants shows that the original parents must have been homozygous or heterozygous for allele(s) with negative effects on ray development.
In this regard, it must be emphasized that the F2 plants represent a small number of parents and that the observed character associations may be specific to the population from which the parents originated. Bogle (1983), whose intermorph crosses involved oxeye daisy plants from a population in North America, presented no quantitative data on nonfloral variables, but noted that the discoid progeny plants generally flowered earlier than those with rayed flower heads, a pattern not observed in the present investigation (S. Andersson, personal observation). It will therefore be necessary to perform additional crosses (involving several unrelated genotypes of the discoid variant) before any broad generalizations can be made regarding patterns of covariance between ray morphology and other characters in oxeye daisy.
Realized patterns of selection on ray morphology
This study of oxeye daisy revealed no conflicting selection pressures between the pollinators, the florivores, and the indirect, pleiotropic associations inferred from the correlation analyses; consequently, one would expect a strong persistent relationship between fitness and the extent to which naturally occurring plants approach the fully rayed phenotype in this study system. These patterns contrast with the more variable or complex relationships documented in other species of the Asteraceae family. Larger, more conspicuous heads of sunflower (Helianthus) species have been found to enhance the attraction of both pollinators (Stuessy et al., 1986) and floral antagonists (Pilson, 2000), suggesting that the optimum ray size is smaller than that favored by pollinators alone. A similar interpretation may be invoked in cases where the possession of rays entails a cost in terms of flower or fruit production, as observed in Achillea ptarmica (Andersson, 1999), Crepis tectorum (Andersson, 2006), and Madia sativa (Celedón-Neghme et al., 2007). Data for C. tectorum also indicate the existence of positive genetic correlations between plant height, achene size, and flower size, a pattern that should reduce the selective optimum at sites selecting for short stature and (or) small propagule size (Andersson, 1993). In Senecio jacobaea, in which the frequency of rayed and discoid individuals varies ecogeographically (van der Meijden, 1976), the optimum floral phenotype may be a compromise between several selection pressures, involving not only pollinators but also physiologically adverse effects of raylessness and spatially varying selection on germination characteristics (Andersson, 2001a), mediated through the close developmental link between ray and fruit dimorphism in this species (McEvoy, 1984). One would also expect morph-specific differences in germination behavior, as well as other characters, to affect the frequency of rayed and discoid individuals of S. vulgaris, given the strong, persistent linkage relationships between floral and nonfloral characters in certain populations of this largely selfing species (Abbott, 1986; Oxford et al., 1996; Abbott et al., 1998; Comes, 1998). The existence of such associations, coupled with the reduced selection pressures on floral size characters in selfing taxa, suggests that nonpollinator selection has played a significant role in the establishment and spread of discoid genotypes in S. vulgaris. With the results from other wild plant species (Galen, 1999b; Strauss and Whittall, 2006), the available data indicate a considerable potential for selective forces other than pollinators to shape the evolution and diversification of floral characters.
FOOTNOTES
1 The author thanks M. Wirén for collection of plant material, H. Persson and B. Jacobsson for technical assistance in the experimental garden, R. Bygebjerg for identification of the floral herbivore, H. Sheppard for linguistic advice, and C. Fenster and two anonymous reviewers for valuable comments on the manuscript. Financial support was provided by the Swedish Research Council. 
2 e-mail: stefan.andersson{at}ekol.lu.se
LITERATURE CITED
Abbott, R. J. 1986. Life-history variation associated with the polymorphism for capitulum type and outcrossing rate in Senecio vulgaris. Heredity 56: 381–391.[CrossRef][Web of Science]
Abbott, R. J., F. C. Bretagnolle, AND C. Thébaud. 1998. Evolution of a polymorphism for outcrossing rate in Senecio vulgaris: Influence of germination behavior. Evolution 52: 1593–1601.[CrossRef][Web of Science]
Abbott, R. J., AND J. A. Irwin. 1988. Pollinator movements and the polymorphism for outcrossing rate at the ray floret locus in groundsel, Senecio vulgaris. Heredity 60: 295–298.[CrossRef][Web of Science]
Andersson, S. 1991. Floral display and pollination success in Achillea ptarmica (Asteraceae). Holarctic Ecology 14: 186–191.
Andersson, S. 1993. Population differentiation in Crepis tectorum (Asteraceae): Patterns of correlation among characters. Biological Journal of the Linnean Society 49: 185–194.[CrossRef][Web of Science]
Andersson, S. 1996. Floral display and pollination success in Senecio jacobaea (Asteraceae): Interactive effects of head and corymb size. American Journal of Botany 83: 71–75.[CrossRef][Web of Science]
Andersson, S. 1997. Genetic constraints on floral evolution in Nigella. Biological Journal of the Linnean Society 62: 519–532.[Web of Science]
Andersson, S. 1999. The cost of floral attractants in Achillea ptarmica (Asteraceae): Evidence from a ray removal experiment. Plant Biology 1: 569–572.[CrossRef]
Andersson, S. 2001a. Fitness consequences of floral variation in Senecio jacobaea (Asteraceae): Evidence from a segregating hybrid population and a resource manipulation experiment. Biological Journal of the Linnean Society 74: 17–24.[CrossRef][Web of Science]
Andersson, S. 2001b. The genetic basis of floral variation in Senecio jacobaea (Asteraceae). Journal of Heredity 92: 409–414.
Andersson, S. 2005. Floral costs in Nigella sativa (Ranunculaceae): Compensatory responses to perianth removal. American Journal of Botany 92: 279–283.
Andersson, S. 2006. Experimental demonstration of floral allocation costs in Crepis tectorum. Canadian Journal of Botany 84: 904–909.[CrossRef]
Andersson, S., AND B. Widén. 1993. Pollinator-mediated selection on floral traits in a synthetic population of Senecio integrifolius (Asteraceae). Oikos 66: 72–79.[CrossRef][Web of Science]
Bell, G. 1985. On the function of flowers. Proceedings of the Royal Society of London, B, Biological Sciences 224: 223–265.
Berg, R. L. 1960. The ecological significance of correlation pleiades. Evolution 14: 171–180.[CrossRef][Web of Science]
Bertin, R. I., AND M. A. Kerwin. 1998. Floral sex ratios and gynomonoecy in Aster (Asteraceae). American Journal of Botany 85: 235–244.[Abstract]
Bogle, A. L. 1983. Discoid heads: An inheritable trait in Chrysanthemum leucanthemum L. Rhodora 85: 461–464.[Web of Science]
Bremer, K., AND C. J. Humphries. 1993. Generic monograph of the Asteraceae-Anthemideae. Bulletin of the Natural History Museum (London) 23: 73–179.
Brody, A. K. 1992. Oviposition choices by a predispersal seed predator (Hylemya sp.). I. Correspondence with hummingbird pollinators and the role of plant size, density and floral morphology. Oecologia 91: 56–62.[Web of Science]
Celedón-Neghme, C., W. L. Gonzáles, AND E. Gianoli. 2007. Costs and benefits of attractive floral traits in the annual species Madia sativa (Asteraceae). Evolutionary Ecology 21: 247–257.[CrossRef][Web of Science]
Charlesworth, D., AND B. Charlesworth. 1987. The effect of investment in attractive structures on allocation to male and female functions in plants. Evolution 41: 948–968.[CrossRef][Web of Science]
Comes, H. P. 1998. Major gene effects during weed evolution: Phenotypic characters cosegregate with alleles at the ray floret locus in Senecio vulgaris L. (Asteraceae). Journal of Heredity 89: 54–61.
Cresswell, J. E. 1998. Stabilizing selection and the structural variability of flowers within species. Annals of Botany 81: 463–473.
Delph, L. F., L. F. Galloway, AND M. L. Stanton. 1996. Sexual dimorphism in flower size. American Naturalist 148: 299–320.[CrossRef][Web of Science]
Faegri, K., AND L. van der Pijl. 1979. The principles of pollination biology. Pergamon Press, Oxford, UK.
Fenner, M., J. E. Cresswell, R. A. Hurley, AND T. Baldwin. 2002. Relationship between capitulum size and pre-dispersal seed predation by insect larvae in common Asteraceae. Oecologia 130: 72–77.[Web of Science]
Fenster, C. B., W. S. Armbruster, P. Wilson, M. R. Dudash, AND J. D. Thomson. 2004. Pollination syndromes and floral specialization. Annual Review of Ecology and Evolution 35: 375–403.[CrossRef]
Fenster, C. B., AND L. F. Galloway. 1997. Developmental homeostasis and floral form: Evolutionary consequences and genetic basis. International Journal of Plant Sciences 158: S121–S130.[CrossRef][Web of Science]
Ford, V. S., AND L. D. Gottlieb. 1990. Genetic studies of floral evolution in Layia. Heredity 64: 29–44.[CrossRef][Web of Science]
Galen, C. 1999a. Flowers and enemies: Predation by nectar-thieving ants in relation to variation in floral form of an alpine wildflower, Polemonium viscosum. Oikos 85: 426–434.[CrossRef][Web of Science]
Galen, C. 1999b. Why do flowers vary? The functional ecology of variation in flower size and form within natural plant populations. Bioscience 49: 631–640.[CrossRef][Web of Science]
Gillies, A. C. M., P. Cubas, E. S. Coen, AND R. J. Abbott. 2002. Making rays in the Asteraceae: Genetics and evolution of radiate versus discoid flower heads. In Q. C. B. Cronk, R. M. Bateman, and J. A. Hawkins [eds.], Developmental genetics and plant evolution, 233–246. Taylor & Francis, London, UK.
Grant, V. 1949. Pollination system as isolating mechanisms in angiosperms. Evolution 3: 82–97.[CrossRef][Web of Science][Medline]
Herrera, C. M. 2001. Deconstructing a floral phenotype: Do pollinators select for corolla integration in Lavandula latifolia? Journal of Evolutionary Biology 14: 574–584.[CrossRef][Web of Science]
Howarth, S. E., AND J. T. Williams. 1968. Biological flora of the British Isles. Chrysanthemum leucanthemum L. Journal of Ecology 56: 585–595.[CrossRef][Web of Science]
Jones, R. L. 1973. Gibberellins: Their physiological role. Annual Review of Plant Physiology 24: 571–598.[Web of Science]
Koning, R. E. 1984. The roles of plant hormones in the growth of the corolla of Gaillardia grandiflora (Asteraceae) ray flowers. American Journal of Botany 71: 1–8.[CrossRef][Web of Science]
Leppik, E. E. 1977. The evolution of capitulum types of the Compositae in the light of insect–flower interaction. In V. H. Heywood, and J. B. Harborne [eds.], The biology and chemistry of the Compositae, I, 61–89. Academic Press, London, UK.
Marshall, D. F., AND R. J. Abbott. 1982. Polymorphism for outcrossing frequency at the ray floret locus in Senecio vulgaris 1. Evidence. Heredity 48: 227–235.[CrossRef][Web of Science]
McEvoy, P. B. 1984. Dormancy and dispersal in dimorphic achenes of tansy ragwort, Senecio jacobaea. Oecologia 61: 160–168.[CrossRef][Web of Science]
Mitchell-Olds, T., AND R. G. Shaw. 1987. Regression analysis of natural selection: Statistical inference and biological interpretation. Evolution 41: 1149–1161.[CrossRef][Web of Science]
Nilsson, L. A. 1988. The evolution of flowers with deep corolla tubes. Nature 334: 147–149.[CrossRef]
Oxford, G. S., T. J. Crawford, AND K. Pernyes. 1996. Why are capitulum morphs associated with other characters in natural populations of Senecio vulgaris (groundsel)? Heredity 76: 192–197.[CrossRef][Web of Science]
Pilson, D. 2000. Herbivory and natural selection on flowering phenology in wild sunflower, Helianthus annuus. Oecologia 122: 72–82.[CrossRef][Web of Science]
Primack, R. B. 1987. Relationships among flowers, fruits and seeds. Annual Review of Ecology and Systematics 18: 409–430.[CrossRef][Web of Science]
Rostgaard Nielsen, L., M. Philipp, AND H. R. Siegismund. 2002. Selective advantage of ray florets in Scalesia affinis and S. pedunculata (Asteraceae), two endemic species from the Galápagos. Evolutionary Ecology 16: 139–153.[CrossRef][Web of Science]
Sandvik, S. M., AND Ø. Totland. 2003. Quantitative importance of staminodes for female reproductive success in Parnassia palustris under contrasting environmental conditions. Canadian Journal of Botany 81: 49–56.[CrossRef]
Stebbins, G. L. 1974. Flowering plants: Evolution above the species level. Edward Arnold, London, UK.
Strauss, S. Y., AND J. B. Whittall. 2006. Non-pollinator agents of selection on floral traits. In L. D. Harder, and S. C. H. Barrett [eds], Ecology and evolution of flowers, 120–138. Oxford University Press, Oxford, UK.
Stuessy, T. F., D. M. Spooner, AND K. A. Evans. 1986. Adaptive significance of ray corollas in Helianthus grosseserratus (Compositae). American Midland Naturalist 115: 191–197.[CrossRef][Web of Science]
Tutin, T. G., V. H. Heywood, N. A. Burges, D. M. Moore, D. H. Valentine, S. M. Walters, AND D. A. Webb. 1976. Flora europaea, vol. 4. Cambridge University Press, Cambridge, UK.
van der Meijden, E. 1976. Het verspreidingsgebied van Senecio jacobaea L. var. nudus Weston. Gorteria 8: 57–61.
Wilson, P. 1995. Selection for pollination success and the mechanical fit of Impatiens flowers around bumblebee flowers. Biological Journal of the Linnean Society 55: 355–383.[Web of Science]
Wolfe, L. M., AND J. L. Krstolic. 1999. Floral symmetry and its influence on variance in flower size. American Naturalist 154: 484–488.[CrossRef][Medline]
CiteULike Complore Connotea Del.icio.us Digg Facebook Reddit Technorati Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  Compiled by, F. Tooke, T. Chiurugwi, and N. Battey Flowering Newsletter bibliography for 2008 J. Exp. Bot., June 23, 2009; (2009) erp154v1. [Full Text] [PDF]
|
|
|
|
 |
|
 R. Torices and A. A. Anderberg Phylogenetic analysis of sexual systems in Inuleae (Asteraceae) Am. J. Botany, May 1, 2009; 96(5): 1011 - 1019. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
