| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
What's this? |
Ecology |
Department of Biology, Indiana University, 1001 E. 3rd Street, Bloomington, Indiana 47405 USA
Received for publication November 2, 2005. Accepted for publication March 20, 2006.
ABSTRACT
Natural selection should favor the integration of floral traits that enhance pollen export and import in plant populations that rely upon pollinators. If this is true, then phenotypic correlations between floral traits should weaken in self-fertilizing groups that do not require pollinator visitation to produce seed. We tested this hypothesis in Leavenworthia, a plant genus in which there have been multiple independent losses of the sporophytic self-incompatibility system found throughout the Brassicaceae. In particular, we conducted phylogenetically independent contrasts of floral trait correlations between two pairs of self-incompatible (SI) and self-compatible (SC) sister taxa. In support of the hypothesis that pollinator-mediated selection integrates floral traits, we found that both SC Leavenworthia taxa have weaker overall floral correlations in comparison to sister taxa that rely upon pollinators. The two independently derived SC Leavenworthia flowers have significantly weaker stamen–petal or pistil–petal correlations, respectively, whereas the stamen–pistil correlation remains constant. These patterns suggest that relaxation of pollinator-mediated selection weakens the integration of traits associated with pollen export and import. The retention of high stamen–pistil correlations in the SC taxa of Leavenworthia further implies that the integration of these traits is either constrained or maintained by selection favoring the successful transfer of pollen within flowers to secure self-pollination.
Key Words: Brassicaceae • developmental constraints • flowers • Leavenworthia • mating system • phenotypic correlations • pollination
Organisms are not random collections of traits, and natural selection should cause certain traits to covary quantitatively or to exhibit "phenotypic integration" (Pigliucci and Preston, 2004
). At the population and species level, the evolution of phenotypic correlation patterns is facilitated when natural selection integrates traits that interact to perform a given function and is constrained when traits share a common genetic or developmental basis (Olson and Miller, 1958
; Cheverud, 1984
; Falconer and MacKay, 1996
). Although natural selection is likely to play a large role in the evolution of morphological adaptations, it is important to consider genetic and developmental constraints. The role of these constraints in the process of adaptation is perhaps one of the most complex and intriguing questions in animal and plant quantitative genetics because it limits the number and diversity of phenotypic outcomes of evolution by natural selection (Cheverud, 1984
; Arnold, 1992
).
In flowering plants, patterns of variation and covariation for many floral traits involved with pollinator attraction, pollen donation, and pollen receipt are thought to be the consequences of pollinator-mediated selection (for review, see Cresswell, 1998
; Medel et al., 2003
; and references therein). While pollinators exert directional selection that can change the mean trait value and the amount of variation maintained in a population, they can also affect the extent to which certain combinations of traits influence fitness (Berg, 1960
; Gomez, 2000
; Maad, 2000
). In theory, natural selection should produce the floral morphology that maximizes male and female fitness through the export and import of pollen, respectively (Morgan, 1992
). With certain floral morphologies, the more precise the fit must be between a pollinator and the floral morphology to ensure pollen donation and receipt, the higher the degree of phenotypic integration within flowers is expected (Conner and Via, 1993
; Conner and Sterling, 1995
; Armbruster et al., 2004
; but see Ushimaru and Nakata, 2001
). Because pollinator-mediated selection favors particular combinations of floral traits that increase fitness, such as larger petals for pollinator attraction in combination with a specific stamen length for efficient pollen delivery, correlations between traits may evolve as a consequence of this correlational selection among individuals within natural populations. The consequences of pollinator-mediated selection on the degree of floral integration may therefore be revealed by comparing populations, species, or higher taxonomic groups with divergent pollination systems (Armbruster, 1991
; Armbruster and Schwaegerle, 1996
; Armbruster et al., 1999
).
In addition to the patterns produced by selection, floral traits that share developmental precursors and components of their developmental genetic system are also expected to be highly integrated (Diggle, 1992
; Conner and Sterling, 1995
). Related flowering taxa that share developmental patterns but undergo different selection regimes can be compared to understand the basis of phenotypic integration and to differentiate developmental and selective causes of integration (Kudoh et al., 2001
). Developmental integration is expected to be constant across closely related populations and species, while selective integration is expected to be species- and possibly population-specific, as indicated by environmental factors influencing male and female fitness (Conner and Sterling, 1995
; Kudoh et al., 2001
; Herrera et al., 2002
). Regardless of the specific selective regime operating in nature, the degree of integration may be revealed by quantifying the degree to which traits are genetically correlated. Empirical studies further suggest that genetic correlations can be approximated by measuring phenotypic correlations in morphological plant traits, an approach that has been used to address a variety of evolutionary hypotheses (Roff, 1996
; Waitt and Levin, 1998
; see discussion in Herrera et al., 2002
). This has been shown to be the case specifically for morphological traits (Roff, 1996
), plant traits at the species level (Waitt and Levin, 1998
), floral traits at the intraspecific level (Herrera et al., 2002
), and for floral traits in Raphanus raphanistrum in the Brassicaceae (Conner, 2002
). Therefore, in this study, phenotypic integration is measured to approximate genetic correlations and constraints.
In our approach the null hypothesis is that developmental constraints should manifest themselves as conserved phenotypic correlation patterns in floral traits among related taxa. Alternatively, a change in the selection regime should override these conserved patterns. Recently, this approach has been used to determine the consequences of shifts of correlational pollinator-mediated selection on phenotypic correlation patterns in related outcrossing taxa (e.g., the effects of different pollinators on the floral correlation patterns of outcrossing taxa at the species level, Conner and Sterling, 1995
; Armbruster et al., 2004
; and at the population level, Herrera et al., 2002
). Similarly, the consequences of the relaxation of correlational pollinator-mediated selection are also predicted to manifest themselves in patterns of floral-trait covariation in related self-incompatible (SI) and self-compatible (SC) taxa (Thompson et al., 1998
). If developmental constraints are strong, self-fertilizing flowers are expected to retain all of the floral correlations historically generated by pollinator-mediated selection before the SI system was lost in environments with pollen limitation.
Comparisons of floral-trait correlations in related outcrossing and selfing taxa allow for a test of several general predictions against the null hypothesis (Ushimaru and Nakata, 2002
). First, selfing taxa are predicted to show weaker overall floral correlations than related outcrossing taxa with specialized pollinators (a component of Berg's [1959
, 1960
] hypothesis). This is predicted because all floral traits directly or indirectly involved with efficient pollen transfer in selfing taxa (e.g., pistil length, stamen length, sepal length, and petal length), as well as traits contributing to overall flower size (all floral trait lengths), are no longer under the correlational selection experienced by their outcrossing ancestors (Berg, 1960
; Armbruster et al., 1999
). If pollen and ovule number are associated with stamen and pistil size in outcrossing taxa (e.g., Yang and Guo, 2004
), then these traits are also expected to have weaker correlations in small-flowered selfing taxa (but see Mazer and Delesalle, 1998
). Second, the specific trait correlations that are expected to weaken to the greatest degree are those that are functionally involved in pollen pickup and deposition by pollinators. In order to identify functionally relevant traits, it is important to make predictions based on species-specific pollination biology and floral structure (Armbruster et al., 1999
, 2004
).
Specific predictions for pollinator-mediated selection on floral-trait correlations have previously been identified in the Brassicaceae for the outcrossing wild radish, Raphanus raphanistrum (Conner and Via, 1993
; Conner and Sterling, 1995
). The basic floral structure found in this species, and also in Leavenworthia, consists of four petals (with a "limb" and "claw" component), a central pistil with the stigma projecting slightly above the tubular portion of the petals surrounded by four paired stamens of a similar height to the pistil and flanked by two single shorter stamens above the nectaries (Lloyd, 1965
). It is hypothesized that the primary pollinators of wild radish select for placement of the anthers at the top of the stamen to be at the edge of the corolla tube as well as the stigma in order to facilitate pollen pickup and deposition. This correlational selection may lead to a strong correlation between the tubular portion of the petal and the stamen height and the tubular portion of the petal and pistil height (Conner and Via, 1993
; Conner and Sterling, 1995
).
We investigated the consequences of relaxed pollinator-mediated selection in Leavenworthia, a plant genus in which there have been multiple independent losses of self-incompatibility. Because of the pollinator activity observed in outcrossing Leavenworthia taxa (see Materials and Methods for a detailed description), we predict that the phenotypic correlations between petal size and both sex organs (stamen and pistil) will be reduced in SC Leavenworthia taxa because these correlations are not being actively favored by pollinators (Lloyd, 1965
; Conner and Sterling, 1995
). A related prediction is that correlational selection will increase covariation between traits involved with self-fertilization in SC species, in contrast to the expected decline in degree of covariation in traits involved with pollen pickup and deposition and pollinator attraction (Ushimaru and Nakata, 2002
). There should be continuous natural selection favoring the functional integration of stamen–pistil traits in SC Leavenworthia taxa because the correct placement of the stamen relative to the pistil is necessary for successful self-pollination (Rollins, 1963
; Ushimaru and Nakata, 2002
). Thus, we do not expect the correlation between the pistil and the stamen to be reduced in SC Leavenworthia taxa. In this report, we compare floral-trait correlations between two pairs of SI and SC sister taxa within the genus Leavenworthia with a highly conserved process of floral development (Rollins, 1963
). By utilizing phylogenetically independent contrasts of plants that differ in their reliance on pollinators to produce seed, we are therefore able to directly determine whether relaxed pollinator-mediated selection in derived self-fertilizing taxa decreases the phenotypic integration of flowers.
MATERIALS AND METHODS
Leavenworthia
Leavenworthia (Brassicaceae) is a small genus consisting of eight species that are restricted to the cedar glades of the southeastern United States (Rollins, 1963
). These habitats are ecologically distinctive because they are characterized by exposed fields of dolomitic limestone covered by an extremely thin and moist layer of soil (Baskin et al., 1995
). Leavenworthia species are winter annuals that are similar morphologically but possess considerable variation in their capacity to self-fertilize. In the genus Leavenworthia alone, there have been at least four independent losses of the sporophytic self-incompatibility (SI) system commonly found in the Brassicaceae (Bateman, 1955
; Rollins, 1963
; Lloyd, 1967
; Beck et al., 2006
). Species of Leavenworthia share the same basic floral structure of R. raphanistrum and have similar pollinators (Lloyd, 1965
). They are pollinated primarily by native solitary bees and by honey bees (Lloyd, 1965
). Pollinator observations of L. stylosa and L. crassa suggest that native bees "wallow" around the long stamen and stigma of the pistil, while introduced honey bees place their heads in direct contact with the long stamen and the stigma. Honey bees also interact with the short stamen, as the force of their landing on the petal pulls the petal and short stamen down and away from the nectaries, where they then forage (Lloyd, 1965
).
We studied two pairs of sister Leavenworthia taxa that varied in their possession of the self-incompatibility system (Rollins, 1963
). The two pairs of sister taxa used in this study belong to subgeneric clades differentiated by chromosome number (n = 11 and 15; Baldwin, 1945
; Beck et al., 2006
). Leavenworthia stylosa and L. torulosa (n = 15) are restricted to limestone cedar glades in and around the Central Basin of Tennessee, USA (Rollins, 1963
). Leavenworthia stylosa is found entirely within this geographic region, whereas L. torulosa may be found in southern portions of Kentucky and northern portions of Alabama. Although L. stylosa and L. torulosa are sympatric throughout their range, there is complete pre- and postzygotic reproductive isolation between these species (Rollins, 1963
). Leavenworthia stylosa is a self-incompatible species with large flowers, whereas L. torulosa primarily self-fertilizes and has smaller flowers (Rollins, 1963
). Apart from the loss of self-incompatibility and reduced flower size in L. torulosa, fruit and seed morphology are the only reliable characters that differentiate the species. In particular, L. torulosa has wingless seeds and torulose siliques.
Leavenworthia alabamica (n = 11) is endemic to the Moulton Valley of northern Alabama, where it occurs in only four counties (Rollins, 1963
). This species consists of geographically isolated self-incompatible and self-compatible populations (Lloyd, 1965
). In particular, the taxon found at the center of the species range is self-incompatible, whereas peripherally located populations of the sister taxon are smaller and primarily self-fertilizing (Lloyd, 1965
; Busch, 2005a
). Excluding the loss of self-incompatibility and associated reductions in floral size in this taxon, the only trait that reliably differentiates the taxa is silique length (Rollins, 1963
; Lloyd, 1965
). Given the limited morphological divergence between these taxa, small-flowered L. alabamica likely represent a recent loss of self-incompatibility coupled with the evolution of floral traits facilitating self-pollination and subsequent self-fertilization.
Floral morphology
The flowers of all Leavenworthia species are very similar morphologically. Flowers have four sepals and petals that surround the male and female sex organs (Fig. 1). Enclosed within the petals are two short stamens on both sides of the pistil and two pairs of long stamens. The anthers of the long stamens often reach or exceed the height of the stigmatic surface. Although these aspects of floral morphology are conserved throughout the genus, there are multiple parallel adaptations within the derived self-fertilizing taxa of Leavenworthia. In particular, petal size and the length of floral organs have been reduced and the degree of scent produced by flowers has declined (Rollins, 1963
; Lloyd, 1965
). In concert with these changes in floral size and scent, the anther sacs are rotated toward (rather than away from) the pistil in self-fertilizing Leavenworthia. These alterations in floral morphology are thought to be responses to both relaxed pollinator-mediated selection and natural selection to ensure successful autonomous self-pollination in pollen-limited environments (Lloyd, 1965
; Busch 2005a
).
|
|
. Seeds were collected from five large-flowered, self-incompatible populations and four geographically isolated, self-compatible populations with small flowers (for locality data, see Busch, 2005b
; Table 1). Approximately 5–20 seeds were collected from 50 maternal families in all nine populations, though only 25 maternal plants were sampled in the smallest population. Seeds were germinated and grown according to the protocols outlined for L. stylosa and L. torulosa. Plants of L. alabamica were reared in Indiana University greenhouses during the late fall and winter of 2003. Although the plants used in this experiment were grown in different years, they were reared at the same time of the year in the same greenhouse room to minimize environmental effects on the traits of interest.
Measurement of floral traits
Six floral traits were measured on one of the first three flowers of 180 L. stylosa plants total (N = 45, 52, 47, and 38 individual plants from four populations) and 83 L. torulosa plants (N = 87 and 17 plants from two populations). Sepal length, petal length, short-stamen length, long-stamen length, and pistil length were measured to the nearest 0.01 mm using digital calipers. Ovule number was also counted in three populations of L. stylosa and all populations of L. torulosa. Five floral traits were measured on one of the first three flowers of 86 self-incompatible L. alabamica plants (N = 20, 20, 17, 16, and 13 from five SI populations, as reported in Busch, 2005a
) and 67 self-compatible L. alabamica plants (N = 21, 17, 16, and 13 from four SC populations). In addition to petal length, long-stamen length, pistil length, and ovule number, the number of pollen grains per flower was estimated. Pollen grains from flowers were submerged in a 3:1 solution of lactic acid and glycerol. The number of pollen grains in a known volume of sample was counted and then multiplied to estimate the total number of pollen grains per flower. Autonomous selfing rate, or fruit set under pollinator exclusion, was estimated in two L. stylosa populations, one L. torulosa population, and all nine L. alabamica populations by dividing the number of flowers that set fruit in a pollinator-free greenhouse by the total number of flowers produced by a plant (Lloyd, 1992
). Rates of autonomous self-fertilization were compared between SI and SC sister taxa using a Wilcoxon two-sample test (Sokal and Rohlf, 1995
). These nonparametric tests were employed because there was a preponderance of zero autonomy rates in both samples of self-incompatible Leavenworthia taxa.
It is possible that any observed differences in floral correlations between each pair of sister taxa may be influenced by systematic changes in the variance associated with the degree of inbreeding (Charlesworth and Charlesworth, 1995
). It was therefore necessary to test this hypothesis in order to rule out the possibility that changes in floral integration may be a byproduct of changes in the variability of the traits under study. We calculated the means, standard deviations, phenotypic variances, and coefficients of variations for each floral trait using SPSS version 11.0 for Windows (SPSS, Chicago, Illinois, USA). The coefficient of variation (CV) standardizes the variance independent of scale, and is equal to the standard deviation of a trait, multiplied by 100, divided by the trait mean (Sokal and Rohlf, 1995
). This standardized estimate of the variance allowed traits with divergent means to be compared in order to determine whether they are relatively more or less variable. Differences in floral trait coefficients of variation between each pair of sister taxa (L. stylosa–L. torulosa and SI L. alabamica–SC L. alabamica) were evaluated with paired-samples t tests (Sokal and Braumann, 1980
; Ushimaru and Nakata, 2001
).
Floral correlations
Phenotypic correlations were computed as Pearson product-moment correlations (r). All correlations between floral traits were z-transformed to ensure that the sample correlations were normally distributed (Sokal and Rohlf, 1995
, pp. 575–583). To test the idea that there is a greater degree of overall integration between floral traits in SI vs. SC sister taxa, we conducted a paired t test comparing all of the z-transformed phenotypic correlations. We also assessed which correlations were not significantly different from zero with a sequential Bonferroni correction for multiple comparisons (Rice, 1989
; Moran, 2003
). In addition to the general degree of floral integration, we were primarily interested in testing the hypothesis that specific floral-trait correlations (petal–stamen, petal–pistil, and stamen–pistil) are reduced in self-fertilizing Leavenworthia. To test the hypothesis that the petal–stamen and petal–pistil correlations would be weaker in the SC taxa, while the pistil–stamen correlation would not, we used parametric t tests to compare these specific correlations observed in SI and SC sister taxa (Sokal and Rohlf, 1995
, p. 582). We evaluated the null hypothesis of no difference between taxa with one-sided significance tests because of our a priori hypothesis that self-fertilizing Leavenworthia will have smaller floral correlations. Because the number of observations exceeded 50 in each taxon, we were able to compute 95% confidence intervals of the z-transformed correlations (Sokal and Rohlf, 1995
). The lower and upper limits were then back-transformed to the r scale.
RESULTS
Leavenworthia stylosa and L. torulosa
Leavenworthia torulosa has higher autonomous selfing rates than L. stylosa (Z = 9.844; P < 0.001; Table 1). Floral trait coefficients of variation did not differ between the selfing and outcrossing species (t = –1.437, df = 5, P = 0.210; Table 2). Floral correlations in L. stylosa ranged from 0.237 (ovule–sepal) to 0.796 (long stamen–petal), with a mean correlation of 0.526 (0.1741). In L. torulosa, floral trait correlations ranged from –0.077 (ovule–sepal) to 0.736 (long stamen–petal) with a mean correlation of 0.335 (0.2160). The overall magnitude of floral trait correlations was smaller in L. torulosa than in L. stylosa (t = 6.051; df = 14; P < 0.001) showing a decreased integration of flowers from the selfing species (Table 3). Despite decreased overall integration in L. torulosa and a trend for all functionally related trait correlations to be smaller, only the petal–pistil correlation was significantly smaller in the selfing L. torulosa (Fig. 2A).
|
|
|
|
|
Overall integration patterns among floral traits
We report that both SC Leavenworthia taxa have overall floral-trait correlations of a significantly smaller magnitude in comparison with sister taxa that rely upon pollinators. These patterns are not influenced by changes in the variance of floral traits associated with the degree of inbreeding and are consistent with the hypothesis that flowering taxa with unspecialized or selfing pollination systems will have weaker floral correlations than taxa with specialized pollination systems (a corollary of Berg's [1959
, 1960
] hypothesis; Armbruster et al., 1999
). Further, these results are similar to the results of Berg's influential empirical study in which she compared species based on mating system as well as pollination system (Berg, 1960
). As a general trend, patterns of developmental and morphological floral-trait covariation have been reported to differ between related outcrossing and selfing taxa in Arenaria, Clarkia, Mimulus, and Collinsia (Hill et al., 1992
; Fenster et al., 1995
; Runions and Geber, 2000
; Armbruster et al., 2002
; Mazer et al., 2004
) and between pairs of floral traits in sister outcrossing and selfing Hosta and Mazus species (Ushimaru and Nakata, 2002
).
The degree to which the relaxation of pollinator-mediated selection leads to decreased correlations in selfing species among the lengths of all floral traits may depend upon the extent to which pleiotropic relationships are complete among floral organs (Conner, 2002
). Further, floral-trait correlation patterns in SC taxa may be generated by the relaxation of pollinator-mediated selection in some cases and by new or indirect selection pressures in others (e.g., correlational selection on pistil and stamen length for efficiency of autonomous self-fertilization, Ushimaru and Nakata 2002
; strong stabilizing selection on the pollen to ovule ratio effecting increased correlations between pollen and ovule number, Mazer, 1996
; Mazer and Delesalle, 1998;
indirect selection, Andersson, 2005
), as well as a mix of relaxed selection and new selection forces (Lloyd, 1965
). For example, further tests of Berg's (1959
, 1960
) hypothesis for decreased overall correlations in floral traits reported that floral traits did not have a consistent correlational pattern in neotropical plants with generalized or specialized pollinators (Armbruster et al., 1999
). In order to support the specific hypothesis that relaxed selection may be responsible for the observed correlation patterns, it is important to test predictions based on how functionally important combinations of floral traits influence male and female fitness in nature.
Evidence for relaxed selection on specific floral trait correlations
Interestingly, the two independently derived SC Leavenworthia taxa exhibit significantly weaker stamen–petal or pistil–petal correlations than the SI taxa. In contrast, correlations among the lengths of the other specific pairs of functionally related floral organs were conserved between taxa. This result supports the prediction that traits involved with plant-to-plant movement of pollen by pollinators will show declines in integration in SC taxa. In the SI/SC L. alabamica comparison, the petal–stamen correlation was significantly weaker in the self-fertilizing taxon. An analogous result was reported in related Hosta species that differed in mating system (Ushimaru and Nakata, 2002
). In Hosta, the tepal–filament correlation (equivalent to the petal–stamen correlation in Leavenworthia) was found to be (marginally) significantly weaker in the selfing species, while all other floral-trait correlations measured remained constant. Similarly, floral-trait correlations reported for other outcrossing species in the Brassicaceae, such as wild radish, have strong correlations between the petal and stamen (Conner and Sterling, 1995
). For example, the petal–stamen correlation was reported to be stronger than background floral correlations in two species of the Brassicaceae and a related outcrossing species (Conner and Sterling, 1995
). Evidence for correlational selection on male function in wild radish from selection studies lends further support to the hypothesis that in SI Brassicaceae taxa floral-correlation patterns are generated by pollinator-mediated selection (Morgan and Conner, 2001
). Relaxed pollinator-mediated selection should therefore account for the weaker stamen–petal correlation in SC L. alabamica relative to SI L. alabamica.
In the SI L. stylosa/SC L. torulosa comparison, we report that the pistil–petal correlation is significantly weaker in the selfing species. This decline in the observed correlation between these floral traits across species suggests that, in the outcrossing species, there may be selection via female function to increase the correlation between the pistil and petal. Although manipulative ecological studies suggest that stabilizing selection via pollen deposition is critical to patterns of variation in female floral traits in Brassica napus (Cresswell, 2000
), the hypothesis of pollinator-mediated selection on female function has not found as much empirical support in the Brassicaceae as selection on male function (e.g., Conner and Via, 1993
; Conner and Sterling, 1995
). It may be that there are differences in the ecological context of the SI–SC transition in the two Leavenworthia taxa pairs and that the strength of correlational or stabilizing selection though male function or through female function depends on this difference in ecological and pollinator context (Ashman and Morgan, 2004
; Maad and Alexandersson, 2004
).
Potential selection for efficient self-pollination
In both of the SC Leavenworthia taxa, we found that the stamen–pistil correlation remains constant and does not weaken in comparison with related SI taxa; thus, this conserved correlation is either maintained by selection or developmentally constrained. Similarly, a stronger stamen–pistil correlation was reported for the selfing species in comparison with a related outcrossing species of Mazus (Ushimaru and Nakata, 2002
). In Leavenworthia, self-pollination occurs when the height of the anthers and stigmatic surface are equivalent and when the anther sacs are rotated to facilitate pollen deposition (Lloyd, 1965
). Thus, the important effect of anther rotation on self-pollination may be one reason we observed a conserved stamen–pistil correlation rather than a larger correlation in SC Leavenworthia taxa, as reported for Mazus (Ushimaru and Nakata, 2002
). It is also possible that in Leavenworthia, as reported for wild radish, the pistil may continue to elongate as a component of ovary and fruit development and increase measurement error or decrease the magnitude of this correlation coefficient (e.g., see methods in Conner and Sterling, 1995
). Alternatively, the stamen–pistil correlation in selfing taxa of Leavenworthia may be of a sufficient degree for efficient self-pollination.
The role of developmental constraint
If the null hypothesis of developmental constraint had held, it could have manifested itself in two ways. First, the correlation patterns would have been conserved between related sister taxa, which we did not observe. Second, developmentally related trait pairs would have had higher correlations than trait pairs that did not share developmental genetic components. If this were true, trait pairs from adjacent floral whorls should be more developmentally constrained than trait pairs from nonadjacent floral whorls. Neither component of this second prediction was observed consistently, and future studies undertaking a more thorough examination of floral correlation patterns among all floral whorls would be useful in evaluating this prediction (I. A. Anderson, unpublished manuscript). Although it has been revealed empirically that pleiotropy is the underlying genetic cause of floral-trait correlations in wild radish (Conner, 2002
) and in some studies the developmental null hypothesis has explained the data most convincingly (C. Herrera, 2001
; Herrera et al., 2002
), manipulative selection studies suggest that the evolution of floral-trait correlations is not entirely constrained (Stanton and Young, 1994
; Delph et al., 2004
).
Summary
The relaxation of correlational selection mediated by pollinators will have consequences for patterns of trait covariation in selfing taxa unless there are developmental or genetic constraints (Berg, 1960
), just as the relaxation of directional pollinator-mediated selection leads to a reduction in petal and flower size, an often-observed phenomenon in selfing taxa relative to sister outcrossing taxa (Rollins, 1963
; Lloyd, 1965
; Ornduff, 1969
; Jain, 1976
; Wyatt, 1988
; Barrett, 2002
). Relaxed pollinator-mediated stabilizing selection has been demonstrated to affect correlation patterns in several contexts (island plants, Thompson et al., 1998
; traits involved with sex in clonal plants, Dorken et al., 2004
; gender traits in dioecious or gender-specialized systems, Ushimaru et al., 2003
). For example, in Menyanthes trifoliata, a distylous species, variable herkogamy and loss of the reciprocal arrangement of stigma and anther heights between morphs is attributed to the reduced pollinator fauna of islands and therefore to the relaxation of stabilizing selection on floral morphs (Thompson et al., 1998
). The present study suggests that relaxed correlational selection by pollinators weakens the integration of traits associated with pollen export and import. In contrast, the placement of the stamen with respect to the pistil is either constrained or maintained by selection favoring successful self-pollination in SC Leavenworthia taxa. Our results are some of the first to suggest that changes in pollinator-mediated correlational selection may cause evolutionary change in floral form in Leavenworthia.
FOOTNOTES
1 The authors thank L. Delph, C. Herlihy, E. Osnas, D. Schoen, and J. Steven for comments on analyses and the manuscript. W. S. Armbruster and an anonymous reviewer provided comments that much improved an earlier version of this paper. I.A.A. thanks J. Beck, D. Estes, and A. Shea for help with location of populations in Tennessee. This work was supported by a U.S. National Science Foundation IGERT Fellowship to I.A.A.; the Indiana Academy of Sciences, Indiana University, and NSF to J.W.B., and NSF grant DEB-0075318 to L. Delph. 
2 Author for correspondence (inanders{at}indiana.edu
; fax: 812-855-6705)
3 Present address: Department of Biology, McGill University, Stewart Building, 1205 Docteur Penfield, Montréal, Quebec H3A 1B1 Canada
LITERATURE CITED
Andersson S.. 2005. Floral costs in Nigella sativa (Ranunculaceae): compensatory responses to perianth removal. American Journal of Botany 92: 279-283.
Armbruster W. S.. 1991. Multilevel analysis of morphometric data from natural plant populations: insights into ontogenetic, genetic, and selective correlations in Dalechampia scandens. Evolution 45: 1229-1244.[CrossRef][Web of Science]
Armbruster W. S. Di Stilio V. S. Tuxill J. D. Flores T. C. Runk J. L. V.. 1999. Covariance and decoupling of floral and vegetative traits in nine neotropical plants: a re-evaluation of Berg's correlation-pleiades concept. American Journal of Botany 86: 39-55.
Armbruster W. S. Mulder C. P. H. Baldwin B. G. Kalisz S. Wessa B. Nute H.. 2002. Comparative analysis of late floral development and mating-system evolution in tribe Collinsieae (Scrophulariaceae s.l). American Journal of Botany 89: 37-49.
Armbruster W. S. Pelabon C. Hansen T. F. Mulder C. P. H.. 2004. Floral integration, modularity, and accuracy: distinguishing complex adaptations from genetic constraints. In M. Pigliucci and K. Preston [eds.] Phenotypic integration: studying the ecology and evolution of complex phenotypes 23-49 Oxford University Press, New York, New York, USA.
Armbruster W. S. Schwaegerle K. E.. 1996. Causes of covariation of phenotypic traits among populations. Journal of Evolutionary Biology 9: 261-276.[CrossRef][Web of Science]
Arnold S. J.. 1992. Constraints on phenotypic evolution. American Journal of Botany 140: S85-S107.
Ashman T. L. Morgan M. T.. 2004. Explaining phenotypic selection on plant attractive characters: male function, gender balance or ecological context?. Proceedings of the Royal Society of London, B, Biological Sciences 271: 553-559.
Baldwin J. T. Jr.. 1945. Chromosomes of Cruciferae II. Cytogeography of Leavenworthia. Bulletin of the Torrey Botanical Club 72: 367-378.[CrossRef]
Barrett S. C. H.. 2002. The evolution of plant sexual diversity. Nature Reviews Genetics 3: 274-284.[CrossRef][Web of Science][Medline]
Baskin J. M. Webb D. H. Baskin C. C.. 1995. A floristic plant ecology study of the limestone glades of northern Alabama. Bulletin of the Torrey Botanical Club 122: 226-242.[CrossRef][Web of Science]
Bateman A. J.. 1955. Self-incompatibility systems in the angiosperms. III. Cruciferae. Heredity 9: 53-68.
Beck J. B. Al-Shehbaz I. A. Schaal B. A.. 2006. Leavenworthia (Brassicaceae) revisited: testing classic systematic and mating system hypotheses. Systematic Botany 31: 151-159.[CrossRef][Web of Science]
Berg R. L.. 1959. A general evolutionary principle underlying the origin of developmental homeostasis. American Naturalist 93: 103-105.[CrossRef][Web of Science]
Berg R. L.. 1960. The ecological significance of correlation pleiades. Evolution 14: 171-180.[CrossRef][Web of Science]
Busch J. W.. 2005a. The evolution of self-compatibility in geographically peripheral populations of Leavenworthia alabamica (Brassicaceae). American Journal of Botany 92: 1503-1512.
Busch J. W.. 2005b. Inbreeding depression in self-incompatible and self-compatible populations of Leavenworthia alabamica. Heredity 94: 159-165.[CrossRef][Web of Science][Medline]
Charlesworth D. Charlesworth B.. 1995. Quantitative genetics in plants—the effect of the breeding system on genetic-variability. Evolution 49: 911-920.[CrossRef][Web of Science]
Cheverud J. M.. 1984. Quantitative genetics and developmental constraints on evolution by selection. Journal of Theoretical Biology 110: 155-171.[Web of Science][Medline]
Conner J. K.. 2002. Genetic mechanisms of floral trait correlations in a natural population. Nature 420: 407-410.[CrossRef][Medline]
Conner J. K. Sterling A.. 1995. Testing hypotheses of functional-relationships—a comparative survey of correlation patterns among floral traits in five insect-pollinated plants. American Journal of Botany 82: 1399-1406.[CrossRef][Web of Science]
Conner J. K. Via S.. 1993. Patterns of phenotypic and genetic correlations among morphological and life-history traits in wild radish, Raphanus raphanistrum. Evolution 47: 704-711.[CrossRef][Web of Science]
Cresswell J. E.. 1998. Stabilizing selection and the structural variability of flowers within species. Annals of Botany 81: 463-473.
Cresswell J. E.. 2000. Manipulation of female architecture in flowers reveals a narrow optimum for pollen deposition. Ecology 81: 3244-3249.[CrossRef][Web of Science]
Delph L. F. Frey F. M. Steven J. C. Gehring J. L.. 2004. Investigating the independent evolution of the size of floral organs via G-matrix estimation and artificial selection. Evolution & Development 6: 438-448.
Diggle P. K.. 1992. Development and the evolution of plant reproductive characters. In R. Wyatt [ed.] Ecology and evolution of plant reproductive characters 326-355 Chapman & Hall, New York, New York, USA.
Dorken M. E. Neville K. J. Eckert C. G.. 2004. Evolutionary vestigialization of sex in a clonal plant: selection versus neutral mutation in geographically peripheral populations. Proceedings of the Royal Society of London, B, Biological Sciences 271: 2375-2380.
Falconer D. S. MacKay T. F. C.. 1996. Introduction to quantitative genetics Prentice Hall, New York, New York, USA.
Fenster C. B. Diggle P. K. Barrett S. C. H. Ritland K.. 1995. The genetics of floral development differentiating 2 species of Mimulus (Scrophulariaceae). Heredity 74: 258-266.[Web of Science]
Gomez J. M.. 2000. Phenotypic selection and response to selection in Lobularia maritima: importance of direct and correlational components of natural selection. Journal of Evolutionary Biology 13: 689-699.[CrossRef][Web of Science]
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]
Herrera C. M. Cerda X. Garcia M. B. Guitian J. Medrano M. Rey P. J. Sanchez-Lafuente A. M.. 2002. Floral integration, phenotypic covariance structure and pollinator variation in bumblebee-pollinated Helleborus foetidus. Journal of Evolutionary Biology 15: 108-121.[CrossRef][Web of Science]
Herrera J.. 2001. The variability of organs differentially involved in pollination, and correlations of traits in Genisteae (Leguminosae: Papilionoideae). Annals of Botany 88: 1027-1037.
Hill J. P. Lord E. M. Shaw R. G.. 1992. Morphological and growth rate differences among outcrossing and self-pollinating races of Arenaria uniflora (Caryophyllaceae). Journal of Evolutionary Biology 5: 559-573.[CrossRef][Web of Science]
Jain S. K.. 1976. The evolution of inbreeding in plants. Annual Review of Ecology and Systematics 7: 469-495.
Kudoh H. Sugawara T. Wu S. G. Murata J.. 2001. Morph-specific correlations between floral traits in a distylous Ophiorrhiza napoensis (Rubiaceae) population in southern China. Journal of Tropical Ecology 17: 719-728.[Web of Science]
Lloyd D. G.. 1965. Evolution of self-compatibility and racial differentiation in Leavenworthia (Cruciferae). Contributions from the Gray Herbarium of Harvard University 195: 3-134.
Lloyd D. G.. 1967. The genetics of self-incompatibility in Leavenworthia crassa Rollins (Cruciferae). Genetica 38: 227-242.[CrossRef][Web of Science]
Lloyd D. G.. 1992. Self- and cross-fertilization in plants. II. The selection of self-fertilization. International Journal of Plant Sciences 153: 370-380.[CrossRef]
Maad J.. 2000. Phenotypic selection in hawkmoth-pollinated Platanthera bifolia: targets and fitness surfaces. Evolution 54: 112-123.[Web of Science][Medline]
Maad J. Alexandersson R.. 2004. Variable selection in Platanthera bifolia (Orchidaceae): phenotypic selection differed between sex functions in a drought year. Journal of Evolutionary Biology 17: 642-650.[CrossRef][Web of Science][Medline]
Mazer S. J.. 1996. Covariation among floral traits in Spergularia marina (Caryophyllaceae): geographic and temporal variation in phenotypic and among-family correlations. Journal of Evolutionary Biology 9: 993-1015.[CrossRef][Web of Science]
Mazer S. J. DeLesalle V. A.. 1998. Contrasting variation within and covariation between gender-related traits in autogamous versus outcrossing species: alternative evolutionary predictions. Evolutionary Ecology 12: 403-425.[CrossRef][Web of Science]
Mazer S. J. Paz H. Bell M. D.. 2004. Life history, floral development, and mating system in Clarkia xantiana (Onagraceae): do floral and whole-plant rates of development evolve independently?. American Journal of Botany 91: 2041-2050.
Medel R. Botto-Mahan C. Kalin-Arroyo M.. 2003. Pollinator-mediated selection on the nectar guide phenotype in the Andean monkey flower, Mimulus luteus. Ecology 84: 1721-1732.[CrossRef][Web of Science]
Moran M. D.. 2003. Arguments for rejecting the sequential Bonferroni in ecological studies. Oikos 100: 403-405.[CrossRef][Web of Science]
Morgan M. T.. 1992. The evolution of traits influencing male and female fertility in outcrossing plants. American Naturalist 139: 1022-1051.[CrossRef][Web of Science]
Morgan M. T. Conner J. K.. 2001. Using genetic markers to directly estimate male selection gradients. Evolution 55: 272-281.[CrossRef][Web of Science][Medline]
Olson E. C. Miller R. L.. 1958. Morphological integration University of Chicago Press, Chicago, Illinois, USA.
Ornduff R.. 1969. Reproductive biology in relation to systematics. Taxon 18: 121-133.[CrossRef]
Pigliucci M. Preston K.. 2004. Phenotypic integration: studying the ecology and evolution of complex phenotypes Oxford University Press, New York, New York, USA.
Rice W. R.. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225.[CrossRef][Web of Science]
Roff D. A.. 1996. The evolution of genetic correlations: An analysis of patterns. Evolution 50: 1392-1403.[CrossRef][Web of Science]
Rollins R. C.. 1963. The evolution and systematics of Leavenworthia (Cruciferae). Contributions from the Gray Herbarium of Harvard University 192: 3-98.
Runions C. J. Geber M. A.. 2000. Evolution of the self-pollinating flower in Clarkia xantiana (Onagraceae). I. Size and development of floral organs. American Journal of Botany 87: 1439-1451.
Sokal R. Braumann C.. 1980. Significance tests for coefficient of variation and variability profiles. Systematic Zoology 29: 50-66.
Sokal R. Rohlf F. J.. 1995. Biometry Freeman, New York, New York, USA.
Stanton M. Young H. J.. 1994. Selecting for floral character associations in wild radish, Raphanus sativus L. Journal of Evolutionary Biology 7: 271-285.[CrossRef][Web of Science]
Thompson F. L. Hermanutz L. A. Innes D. J.. 1998. The reproductive ecology of island populations of distylous Menyanthes trifoliata (Menyanthaceae). Canadian Journal of Botany 76: 818-828.[CrossRef][Web of Science]
Ushimaru A. Itagaki T. Ishii H. S.. 2003. Variation in floral organ size depends on function: a test with Commelina communis, an andromonoecious species. Evolutionary Ecology Research 5: 615-622.[Web of Science]
Ushimaru A. Nakata K.. 2001. Evolution of flower allometry and its significance for pollination success in the deceptive orchid Pogonia japonica. International Journal of Plant Sciences 162: 1307-1311.[CrossRef]
Ushimaru A. Nakata K.. 2002. The evolution of flower allometry in selfing species. Evolutionary Ecology Research 4: 1217-1227.[Web of Science]
Waitt D. E. Levin D. A.. 1998. Genetic and phenotypic correlations in plants: a botanical test of Cheverud's conjecture. Heredity 80: 310-319.[CrossRef][Web of Science]
Wyatt R.. 1988. Phylogenetic aspects of the evolution of self-pollination. In L. D. Gottlieb and S. K. Jain [eds.] Plant evolutionary biology 109-131 Chapman and Hall, London, UK.
Yang C. F. Guo Y. H.. 2004. Pollen size–number trade-off and pollen–pistil relationships in Pedicularis (Orobanchaceae). Plant Systematics and Evolution 247: 177-185.[Web of Science]
CiteULike Complore Connotea Del.icio.us Digg Facebook Reddit Technorati Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  J. S. Miller and P. K. Diggle Correlated evolution of fruit size and sexual expression in andromonoecious Solanum sections Acanthophora and Lasiocarpa (Solanaceae) Am. J. Botany, October 1, 2007; 94(10): 1706 - 1715. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Flowering Newsletter bibliography for 2006 J. Exp. Bot., April 20, 2007; (2007) erm028v2. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
